THE EPISTEMOLOGY OF TESTIMONY

This page intentionally left blank

The Epistemology of
Testimony
Edited by
JENNIFER L ACKEY
and
ERNEST SOSA

CLARENDON PRESS · OXFORD

 1
Great Clarendon Street, Oxford ox2 6dp
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide in
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Oxford is a registered trademark of Oxford University Press
in the UK and in certain other countries
Published in the United States
by Oxford University Press Inc., New York
© The several contributors 2006
The moral rights of the authors have been asserted
Database right Oxford University Press (maker)
First published 2006
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
without the prior permission in writing of Oxford University Press,
or as expressly permitted by law, or under terms agreed with the appropriate
reprographics rights organization. Enquiries concerning reproduction
outside the scope of the above should be sent to the Rights Department,
Oxford University Press, at the address above
You must not circulate this book in any other binding or cover
and you must impose this same condition on any acquirer
British Library Cataloguing in Publication Data
Data available
Library of Congress Cataloging in Publication Data
Data available
Typeset by Laserwords Private Limited, Chennai, India
Printed in Great Britain
on acid-free paper by
Biddles Ltd., King’s Lynn, Norfolk

ISBN 0–19–927600–5 978–0–19–927600–4

ISBN 0–19–927601–3 (Pbk.) 978–0–19–927601–1 (Pbk.)

1 3 5 7 9 10 8 6 4 2
 Contents

List of Contributors vii

Introduction 1
Jennifer Lackey

I . T E S T I M O N Y A N D T H E L E G AC Y
OF THOMAS REID
1. Testimony, Credulity, and Veracity 25
Robert Audi
2. Reid on the Credit of Human Testimony 50
James Van Cleve

I I . T E S T I M O N Y A N D I TS P L AC E I N E PI S T E M O LO G Y
3. The Epistemic Role of Testimony: Internalist and Externalist
Perspectives 77
Richard Fumerton
4. Liberal Fundamentalism and Its Rivals 93
Peter Graham
5. Knowledge: Instrumental and Testimonial 116
Ernest Sosa

I I I . R E D U C T I O N I S M A N D N O N - R E D U C T I O N IS M
I N T H E E PI S T E M O LO G Y O F T E S T I M O N Y
6. Reductionism and the Distinctiveness of Testimonial Knowledge 127
Sanford C. Goldberg
7. Testimony and Trustworthiness 145
Keith Lehrer
8. It Takes Two to Tango: Beyond Reductionism
and Non-Reductionism in the Epistemology
of Testimony 160
Jennifer Lackey
vi Contents

I V. T E S T I M O N Y A N D T H E E X T E N T O F O U R
D E PE N D E N C E O N OT H E R S
9. Testimonial Justification and Transindividual Reasons 193
Frederick F. Schmitt
10. Testimony and Epistemic Autonomy 225
Elizabeth Fricker

V. N EW A R E A S A N D N EW D I R E C T I O N S I N T H E
E PI S T E M O LO G Y O F T E S T I M O N Y
11. Pathologies of Testimony 253
C. A. J. Coady
12. Getting Told and Being Believed 272
Richard Moran

Index 307
 List of Contributors

Robert Audi, Professor of Philosophy and David E. Gallo Professor of Business Ethics,
University of Notre Dame
C. A. J. Coady, Professorial Fellow in Applied Philosophy at the Centre for Applied
Philosophy and Public Ethics, University of Melbourne
Elizabeth Fricker, University Lecturer in Philosophy and Fellow, Magdalen College,
Oxford University
Richard Fumerton, F. Wendell Miller Professor of Philosophy, University of Iowa
Sanford C. Goldberg, Associate Professor of Philosophy, University of Kentucky
Peter Graham, Assistant Professor of Philosophy, University of California, Riverside
Jennifer Lackey, Assistant Professor of Philosophy, Northern Illinois University
Keith Lehrer, Regents Professor of Philosophy, Emeritus, University of Arizona
Richard Moran, Professor of Philosophy, Harvard University
Frederick F. Schmitt, Professor of Philosophy, Indiana University
Ernest Sosa, Romeo Elton Professor of Natural Theology and Professor of Philosophy,
Brown University; Distinguished Visiting Professor, Rutgers University
James Van Cleve, Professor of Philosophy, University of Southern California
This page intentionally left blank
 Introduction
Jennifer Lackey

Our dependence on testimony is as deep as it is ubiquitous. We rely on the

reports of others for our beliefs about the food we eat, the medicine we ingest,
the products we buy, the geography of the world, discoveries in science, historical
information, and many other areas that play crucial roles in both our practic-
al and our intellectual lives. Even many of our most important beliefs about
ourselves were learned at an earlier time from our parents and caretakers, such
as the date of our birth, the identity of our parents, our ethnic backgrounds, and
so on. Were we to refrain from accepting the testimony of others, our lives would
be impoverished in startling and debilitating ways.
Despite the vital role that testimony occupies in our epistemic lives, traditional
epistemological theories focused primarily on other sources, such as sense percep-
tion, memory, and reason, with relatively little attention devoted specifically to
testimony. In recent years, however, the epistemic significance of testimony has
been more fully appreciated, and the current literature has benefited from the
publication of a considerable amount of interesting and innovative work in this
area. The Epistemology of Testimony is intended to build on and further devel-
op this work by bringing together new papers by some of the central figures in
the field.
Most of the papers collected for this volume discuss issues in the epistemo-
logy of testimony from a contemporary point of view, though some offer treat-
ments that are historically grounded in the work of Thomas Reid and, to a less-
er extent, David Hume (Audi and Van Cleve). Some examine the similarities
that testimony bears to other kinds of knowledge (Audi, Van Cleve, Sosa, and
Schmitt), while others point to new areas of inquiry in the testimony literature
(Coady) and new approaches to theorizing about these issues (Graham, Lackey,
and Moran). Many of the papers address the debate between non-reductionist
and reductionist theories of testimonial justification and knowledge, either dir-
ectly (Goldberg, Lackey, and Lehrer) or indirectly (Fumerton, Graham, Sosa,
Schmitt, Fricker, and Moran). This debate has rightly assumed a position of

For valuable assistance with this Introduction, I would like to thank Pete Nichols and, especially,
Baron Reed and Ernie Sosa.
2 Introduction

central importance in the literature on testimony. For at issue is a disagreement

over the proper understanding of the relationship between epistemic agents and
the broader social world: are the normative requirements for testimonial justific-
ation and knowledge a burden that falls primarily on individual knowers, or are
they largely satisfied by features of the social interaction that takes place between
speakers and hearers?

1 . T E S T I M O N Y A N D T E S T I M O N I A L LY- B A S E D B E L I E F

One of the main questions in the epistemology of testimony is how we suc-

cessfully acquire justified belief or knowledge on the basis of what other people
tell us. This, rather than what testimony is, is often taken to be the issue of
central import from an epistemological point of view. Because of this, those
who are interested in the epistemic status of testimonial beliefs often embrace
a very broad notion of what it is to testify, one that properly leaves the dis-
tinction between reliable and unreliable (or otherwise epistemically good and
bad) testimony for epistemology to delineate. So, for instance, Elizabeth Frick-
er holds that the domain of testimony that is of epistemological interest is that of
‘tellings generally’ with ‘no restrictions either on subject matter, or on the speak-
er’s epistemic relation to it’ (Fricker 1995: 396–7). Similarly, Robert Audi claims
that in accounting for testimonial knowledge and justification, we must under-
stand testimony as ‘people’s telling us things’ (Audi 1997: 406). And Ernest Sosa
embraces ‘a broad sense of testimony that counts posthumous publications as
examples . . . [it] requires only that it be a statement of someone’s thoughts or
beliefs, which they might direct to the world at large and to no one in particu-
lar’ (Sosa 1991: 219).¹
But not every expression of thought is appropriately regarded as an instance of
testimony. For instance, suppose that we are walking down the street and I say,
‘Ah, it is indeed a beautiful day’. Suppose further that such a statement, though it
expresses my thought that it is indeed a beautiful day, is neither offered nor taken
as conveying information for any proposition; it is simply a conversational filler,
comparable to a sigh of contentedness. Or consider a case in which Ned tells a
joke among a group of our friends and I casually say, ‘He sure has a great sense
of humor’. Again, though I am stating my thought that Ned has a great sense of
humor, the context is such that we all know Ned and we all know that he has a
great sense of humor; thus, my statement is simply a polite response to a friend’s
joke.² Since the concept of testimony is intimately connected with the notion of
conveying information, either in terms of being offered as such or of being taken
as such, it is arguable that both examples should fail to qualify as instances of
testimony.³
An adequate account of testimony, therefore, should recognize the distinction
between entirely non-informational expressions of thought and testimony.
 Jennifer Lackey 3

Moreover, let us use the general term act of communication to include both verbal
and written assertions as well as communicative physical gestures, such as nods,
points, and so on. Given this, a rough characterization of testimony may be
expressed in something like the following way:
T: S testifies that p by making an act of communication a if and only if (in
part) in virtue of a’s communicable content, (1) S reasonably intends to
convey the information that p, or⁴ (2) a is reasonably taken as conveying
the information that p.⁵
The clause ‘(in part) in virtue of a’s communicable content’ is intended to rule
out cases such as the following: suppose that I sing ‘La, la, la’ in a soprano voice
and I offer this as conveying the information that I have a soprano voice in virtue
of its perceptual content.⁶ Such an act of communication should not qualify as
testimony because it was not offered as conveying information even in part in vir-
tue of its communicable content; rather, it was offered as conveying information
entirely in virtue of its perceptual content.⁷ Furthermore, while there need not be
a direct correspondence between the content of a proffered act of communication
and the content of the proposition testified to—my saying that there are umbrel-
las in the closet, for instance, may properly be taken as conveying the information
both that it is raining outside and that there are umbrellas in the closet—a reas-
onable connection must exist between such contents. For instance, my intending
to convey the information that corn on the cob is yellow by virtue of saying that
grass is green fails to satisfy (1) of T because, in the absence of rather unusual
circumstances, there fails to be a reasonable connection between the contents of
these two propositions. T, then, is broad enough to properly leave the distinc-
tion between instances of testimony that are truth-conducive and those that are
not for epistemology to delineate while, at the same time, distinguishing between
entirely non-informational expressions of thought and testimony.⁸
However, even if a speaker’s assertion is properly connected with the convey-
ing of information, not everything we learn from the testimony of others qualifies
as testimonially based knowledge. For instance, suppose that I say that ten people
have spoken in this room today and you, having counted the previous nine, come
to know that ten people have spoken in this room today.⁹ Here, my statement
may certainly be causally relevant with respect to your forming this belief, but
your knowledge is based on your having heard and counted the speakers in the
room today, thereby rendering it perceptual in nature. What is of import for
justification or knowledge that is distinctively testimonial is that a hearer form
a given belief on the basis of the content of a speaker’s testimony. This precludes
cases—such as that above—where a belief is formed, either entirely or primarily,
on the basis of features about the speaker’s testimony from qualifying as instances
of testimonial justification or knowledge.
There are also intermediate cases in which a hearer has relevant background
information and uses it to derive knowledge from the statement of a speaker. For
4 Introduction

example, suppose that you know from past experience that I report that there is
no milk in the refrigerator only when there is some. Now when I report to you
that there is no milk in the refrigerator, you may supplement my testimony with
your background information and hence derive knowledge that there is milk in
the refrigerator. Because the epistemic status of beliefs formed in these types of
cases relies so heavily on memory and inference, the resulting justification and
knowledge are only partially testimonially based. Hence, such beliefs may fall
outside the scope of theories purporting to capture only those beliefs that are
entirely based on testimony.

2 . N O N - R E D U C T I O N IS M A N D R E D U C T I O N I S M

Even if an expression of thought qualifies as testimony and the resulting belief

formed is entirely testimonially based for the hearer, however, there is the further
question of how precisely such a belief successfully counts as justified belief or an
instance of knowledge. Indeed, this is the question at the center of the epistem-
ology of testimony, and the current philosophical literature contains two central
options for answering it: non-reductionism and reductionism.
According to non-reductionists—whose historical roots are standardly traced
back to Reid—testimony is just as basic a source of justification (warrant, enti-
tlement, knowledge, etc.) as sense perception, memory, inference, and the like.
Accordingly, so long as there are no relevant defeaters, hearers can justifiedly
accept the assertions of speakers merely on the basis of a speaker’s testimony.¹⁰
Otherwise put, so long as there is no available evidence against accepting a speak-
er’s report, the hearer has no positive epistemic work to do in order to justifiedly
accept the testimony in question.¹¹
There are two different kinds of defeaters that are typically taken to be here
relevant. First, there are what we might call psychological defeaters. A psycholo-
gical defeater is an experience, doubt, or belief that is had by S, yet indicates
that S’s belief that p is either false or unreliably formed or sustained.¹² Defeat-
ers in this sense function by virtue of being had by S, regardless of their truth
value or justificatory status.¹³ Second, there are what we might call normative
defeaters. A normative defeater is a doubt or belief that S ought to have, yet
indicates that S’s belief that p is either false or unreliably formed or sustained.¹⁴
Defeaters in this sense function by virtue of being doubts or beliefs that S should
have (whether or not S does have them) given the presence of certain available
evidence.¹⁵ The underlying thought here is that certain kinds of experiences,
doubts, and beliefs contribute epistemically unacceptable irrationality to doxastic
systems and, accordingly, justification and knowledge can be defeated or under-
mined by their presence.
Moreover, a defeater may itself be either defeated or undefeated. Suppose,
for instance, that Wendy believes that there is a hawk nesting in her backyard
because she saw it there this afternoon, but Ted tells her, and she thereby comes
 Jennifer Lackey 5

to believe, that the bird is instead a falcon. Now, the justification Wendy had for
believing that there is a hawk in her backyard has been defeated by her belief that
the bird is a falcon. But since psychological defeaters can themselves be beliefs,
they, too, are candidates for defeat. For instance, suppose that Wendy consults
a bird guidebook to check whether the bird in her backyard is in fact a falcon
and she discovers that it is a Cooper’s hawk. In this case, the belief that she
acquires from the bird book provides her with a psychological defeater for the
belief that she acquired via Ted’s testimony, and hence it provides her with a
defeater-defeater for her original belief that there is a hawk nesting in her back-
yard. And, as should be suspected, defeater-defeaters can also be defeated by
further experiences, doubts, and beliefs, which, in turn, can be defeated by fur-
ther experiences, doubts, and beliefs, and so on. Similar considerations involving
reasons, rather than experiences, doubts, and beliefs, apply in the case of normat-
ive defeaters. Now, when one has a defeater D for one’s belief that p that is not
itself defeated, one has what is called an undefeated defeater for one’s belief that p.
It is the presence of undefeated defeaters, not merely defeaters, that is incompat-
ible with testimonial justification and knowledge.
In contrast to non-reductionism, reductionists—whose historical roots are
typically traced back to Hume—maintain that in order to justifiedly accept
the testimony of speakers, more is needed than the mere absence of undefeated
defeaters. In particular, proponents of reductionism argue that hearers must
have sufficiently good positive reasons for accepting a given report, reasons that
are not themselves ineliminably based on the testimony of others. Typically,
these reasons are the result of induction: for instance, we observe a general
conformity between facts and reports and, with the aid of memory and reason,
we inductively infer that certain speakers, contexts, or types of reports are reliable
sources of information. In this way, the justification of testimony is reduced to the
justification we have for sense perception, memory, and inductive inference.¹⁶
There are, however, at least two different answers given to what relata are
involved in the relevant testimonial reductions. The first answer, and the stronger
version of reductionism—a view sometimes called global reductionism—is that
the justification of testimony as a source of belief reduces to the justification of
sense perception, memory, and inductive inference. In particular, global reduc-
tionists maintain that in order to justifiedly accept a speaker’s report, a hearer
must have non-testimonially based positive reasons for believing that testimony is
generally reliable. The second, and weaker, version of reductionism—often called
local reductionism—is that the justification of each particular report or instance of
testimony reduces to the justification of instances of sense perception, memory,
and inductive inference. Specifically, local reductionists claim that in order to jus-
tifiedly accept a speaker’s testimony, a hearer must have non-testimonially based
positive reasons for accepting the particular report in question.¹⁷
Both non-reductionism and reductionism have been subject to various
objections. For instance, because non-reductionists do not require any positive
6 Introduction

epistemic work from recipients of testimony in order to acquire justified

belief or knowledge, such a view is often attacked for sanctioning gullibility,
epistemic irrationality, and intellectual irresponsibility.¹⁸ In addition, while non-
reductionists maintain that testimony is an epistemically basic source, it has been
argued that a source of epistemically basic beliefs must be an a priori source,
and testimony is clearly not an a priori source.¹⁹ On the other hand, because
reductionists hold that testimony is ultimately reducible to other epistemic
sources, such a view is frequently criticized for underestimating or devaluing
the contributions made by testimony to our cognitive lives. Moreover, it is
often claimed that the non-testimonial positive reasons required by reductionists
simply cannot be possessed in all of those cases in which testimonial justification
and knowledge are intuitively present, either because most (or all) of our
knowledge is already indebted to testimony or because we simply lack the
cognitive resources to acquire such reasons.²⁰ Some of these problems have
recently led those working in the epistemology of testimony to develop qualified
or hybrid views of either non-reductionism or reductionism, some of which can
be found in this collection (Graham, Goldberg, Lehrer, and Lackey).²¹

3. TRANSMISSION

Despite the important differences between non-reductionists and reductionists,

there is nonetheless surprising agreement between proponents of these two views
regarding a central thesis about the epistemic status of testimony, which we may
call the Transmission of Epistemic Properties thesis (hereafter, TEP). Roughly,
the thought expressed by TEP is that a testimonial exchange involves a speak-
er’s belief, along with the epistemic properties it possesses, being transmitted to a
hearer. There are two dimensions to TEP; one is a necessity thesis and the other is
a sufficiency thesis. More precisely,
TEP-N: For every speaker, A, and hearer, B, B’s belief that p is warranted
(justified, known) on the basis of A’s testimony that p only if A’s
belief that p is warranted (justified, known).²²
TEP-S: For every speaker, A, and hearer, B, if (1) A’s belief that p is war-
ranted (justified, known), (2) B comes to believe that p on the
basis of the content of A’s testimony that p, and (3) B has no
undefeated defeaters for believing that p, then B’s belief that p is
warranted (justified, known).²³
There is much that is intuitive about both of these theses. For, in many respects,
a testimonial chain seems to be much like a bucket brigade: in order to give you a
full bucket of water, I must have a full bucket of water to pass to you. Moreover,
if I give you a full bucket of water, then—spills aside—the bucket of water you
now possess as a result of our exchange will also be full. Similarly, in order to
transmit to you a warranted belief, I must have a warranted belief to pass to you.
 Jennifer Lackey 7

Moreover, if I transmit to you a warranted belief, then—defeaters aside—the

belief that you now possess as a result of our exchange will also be warranted.
Further support for TEP, particularly for TEP-N, derives from the purported
observation that testimony, unlike, for instance, sense perception and inference,
is not a generative epistemic source. Indeed, in this respect, testimony is said to
bear striking similarities to memory: while the former is capable of only trans-
mitting knowledge from one speaker to another, the latter is capable of only
preserving the epistemic status of beliefs from one time to another. In this way,
just as I cannot know that p on the basis of memory unless I non-memorially
knew that p at an earlier time, the thought underlying this picture of testimonial
knowledge is that a hearer cannot know that p on the basis of testimony unless
the speaker from whom it was acquired herself knows that p.
While this picture of testimony is widely accepted by both non-reductionists
and reductionists (indeed, several papers in this volume include endorsements of
different versions of TEP, such as Audi, Schmitt, and Fricker), various objections
have been raised to both TEP-N and to TEP-S. For instance, it has been argued
against TEP-N that unreliable believers may nonetheless be reliable testifiers —as,
for example, may happen when a teacher who is a devout creationist nevertheless
reliably teaches her students about evolutionary theory—and hence speaker-
warrant or knowledge is not necessary for testimonial warrant or knowledge.²⁴
Against TEP-S, it has been argued that reliable believers may nonetheless be
unreliable testifiers—as, for example, may happen when a speaker who knows
and correctly reports that p is nevertheless such that she would have easily
reported that p even if not-p had been the case—and hence speaker-warrant
or knowledge combined with the absence of defeaters is not sufficient for
testimonial warrant or knowledge.²⁵ These types of cases have led some either
to qualify the relevant theses of TEP or to restrict their scope to certain kinds
of testimonial knowledge only, such as that acquired through the particular
mechanism of trust.²⁶

4 . S U M M A R I E S O F E S S AY S

In ‘Testimony, Credulity, and Veracity’, Robert Audi contrasts testimony-based

belief—which is belief that arises naturally, non-inferentially, and typically unself-
consciously to what someone says—with belief based on the other standard
epistemic sources. Showing the ways in which his view of testimony bears strik-
ing similarities to Thomas Reid’s, Audi argues that testimony is a source of
basic knowledge—that is, knowledge not grounded in knowledge of some other
proposition—though it is not a basic source of knowledge—that is, it cannot
produce knowledge without the cooperation of another epistemic source.
Audi then compares testimonial knowledge with testimonial justification,
arguing that in order for a hearer to acquire testimonial knowledge that p, the
speaker from whom it was acquired must also know that p. In contrast, a hearer
8 Introduction

can acquire testimonially justified belief that p without the speaker from whom
it was acquired herself being justified in believing that p. Moreover, Audi argues
that while testimonial knowledge does not require that a hearer have knowledge
or justification regarding the attester’s credibility, justification concerning the
attester’s credibility is required in order for a hearer to acquire testimonially
justified belief.
Finally, Audi discusses at least five ways in which testimony differs from the
standard basic sources of perception, memory, and introspection. First, while
the reliability of one of the basic sources cannot be tested without relying on
that very source, the reliability of testimony can be tested without relying on
it. Second, unlike the basic sources, there is no domain for which continued
testimony is in principle necessary for a significant increase in knowledge or jus-
tification. Third, testimony-based belief, unlike that based on the basic sources,
passes through the will. Fourth, because of the dependency on understanding an
artificial language, testimony-based knowledge and justification must go through
convention, while knowledge and justification based on the basic sources need
only go through nature. Fifth, although testimony can increase the number of
knowers in the world, it cannot, as the basic sources can, increase the number of
propositions known.
In ‘Reid on the Credit of Human Testimony’, James Van Cleve explores the
analogy between perception and the credit we give to human testimony as it is
discussed in the work of Thomas Reid. He first attempts to properly understand
the two principles Reid identifies as fundamental to the acquisition of testimonial
knowledge: the principles of veracity and credulity. Van Cleve shows that there
are two different ways to understand both the principle of veracity—i.e., as either
V1: (It tends to be the case that) if A says p, p is true, or as V2: (It tends to be the
case that) if A says p, A believes p—and the principle of credulity—i.e., as either
C1: (It tends to be the case that) if A says p, B believes p, or as C2: (It tends to be
the case that) if A says p, B believes that A believes p. Reid’s writings suggest that
the principle of veracity should be understood as V2 and the principle of credu-
lity as C1. But this poses a problem, as Reid also claims that the two principles
are meant to tally with each other. That is, the principles should combine with
each other so as to imply that when a speaker says p, we tend to form true beliefs.
The combinations that successfully achieve this, however, are V1 with C1 and V2
with C2, not Reid’s V2 with C1. Van Cleve suggests that this problem can be
solved, since a good Reidian case can be made for accepting all four of the prin-
ciples. This results in two pairs of principles that tally with each other—V1 with
C1 and V2 with C2.
Van Cleve then turns to Reid’s discussion of the analogy between the testi-
mony of nature given by the senses and the testimony of human beings given in
language. According to Reid, there are four phenomena involved—original per-
ception, natural language, acquired perception, and artificial language—which
give rise to two analogies, one between acquired perception and artificial language
 Jennifer Lackey 9

and another between original perception and natural language. Acquired percep-
tion and artificial language are alike in that the connection between signs and
what they signify is known by experience, but they differ in that the connection
for artificial language, unlike acquired perception, holds by convention rather
than by nature. The analogy between original perception and natural language,
however, is even greater since in both cases nature has established the connection
between sign and thing signified and taught us the interpretation of the sign prior
to experience.
Finally, Van Cleve turns to Reid’s view of the epistemology of testimony. Reid
regards testimony as a source not merely of belief, but also of evident belief or
knowledge. Moreover, this is true even if the hearer in question possesses no reas-
on for accepting a given speaker’s report. Testimony-based beliefs are, thus, epi-
stemically basic on his view, thereby placing Reid clearly in the non-reductionist
or fundamentalist camp in the epistemology of testimony. Van Cleve, however,
finds this view implausible, since he argues that a source of epistemically basic
belief must be an a priori source, and testimony is not an a priori source. Van
Cleve therefore finds himself (uncharacteristically) on the side of Hume rather
than Reid in defending a reductionist view of the epistemic status of testimoni-
al beliefs.
In ‘The Epistemic Role of Testimony: Internalist and Externalist Perspectives’,
Richard Fumerton contrasts the possible ways in which classical internalists and
paradigm externalists might approach the question of how to characterize testi-
monial justification, focusing particularly on whether testimonial inference has
a fundamental or a derivative place in our reasoning. He argues that the plaus-
ibility of countenancing a general and sui generis epistemic principle licensing
testimonial inference is directly proportional to the plausibility of an extern-
alist understanding of probability claims. Moreover, while much recent work
in the epistemology of testimony focuses on whether testimonial inference is
fundamental or derivative, Fumerton argues that the distinction itself is spuri-
ous since arguments that purportedly employ derivative epistemic principles are
better thought of as enthymematic arguments governed by legitimate epistem-
ic principles licensing inferences from premises to conclusion. Thus, derivative
epistemic ‘principles’ aren’t really epistemic principles at all.
In ‘Liberal Fundamentalism and Its Rivals’, Peter Graham argues on behalf
of Liberal Fundamentalism, a view that holds that testimony-based beliefs, like
perception, memory, and introspection-based beliefs, are epistemically direct. In
particular, the Liberal Fundamentalist maintains that it is a priori necessary that
understanding an attester’s presentation-as-true that p confers justification on
the hearer’s belief that p. Such a view, according to Graham, includes two doc-
trines—it combines Liberal Foundationalism about which epistemic principles
are true with Intuitionism about why they are true.
Graham then distinguishes between a strong and a weak version of Liberal
Fundamentalism. The Strong version holds that, in the absence of defeat, the
10 Introduction

event or state of understanding the attester’s presentation-as-true that p confers

on balance justification for the belief that p, whereas the Weak version holds that
the justification provided may fall short of on balance justification even if it is
undefeated. Specifically, the Weak Liberal Fundamentalist—which is Graham’s
preferred view—maintains that no additional support of any kind (reductive
or non-reductive) is required for prima facie pro tanto testimonial justification,
while additional support (whether reductive or non-reductive) is often required
for on balance testimonial justification. According to Graham, then, testimony
is epistemically direct with respect to pro tanto justification, but it is epistemic-
ally inferential with respect to on balance justification. Finally, Graham argues
that his view of the epistemology of testimony has notable advantages over its
rivals, particularly over what he calls Strong Liberal Fundamentalism and Moder-
ate Fundamentalism.
Ernest Sosa argues in ‘Knowledge: Instrumental and Testimonial’ that there
are epistemically important similarities between testimonial knowledge and
knowledge acquired through the use of instruments. In many cases of testimonial
knowledge, for instance, the features that most saliently explain the truth of a
given belief involve, not the individual competence of the subject holding such a
belief but, rather, the testifiers in question and their cognitive accomplishments.
But, according to Sosa, this is not distinctive of testimonial knowledge. For
what carries the explanatory burden in cases of instrumental knowledge is not the
individual competence of the subject possessing the knowledge but, rather, the
safety of the deliverances of the instruments in question. When one knows the
ambient temperature outside through the display of a given thermometer, for
instance, this is explained primarily by the fact that the deliverances of the
thermometer in question are systematically safe—that is, not easily would such
deliverances be false.
Sosa argues, however, that if a responsiveness to a relevant field requires effi-
cient causal input, then this systematic safety of reliable instruments does not
always derive from such a responsiveness. For instance, the facts of arithmetic do
not efficiently cause a calculator to display the correct answer when prompted,
nor are cogito beliefs explained through any efficient causal responsiveness to the
relevant field, yet both sets of deliverances are highly reliable. Moreover, though
Sosa claims that there is a clear sense in which such deliverances are produced
because they are true, this ‘because’ cannot be one of teleology or function, since
an intelligent Swampman caused by lightning serendipitously hitting a swamp
can exhibit epistemically efficient competence. Sosa’s preferred explanation is one
that is compatible with the relevance of evolution or theology to whether and
how we have knowledge.
With respect to exchanges involving testimony, Sosa claims that hearers
must interpret their speakers so as to discern the thoughts behind their oral
or written displays, and hence this interpretative knowledge, in its use of the
instrument of language, is itself a kind of instrumental knowledge. Because
 Jennifer Lackey 11

of this, testimonial knowledge not only shares important epistemic similarities

with knowledge acquired through the use of instruments, it also presupposes
instrumental knowledge.
Finally, while Sosa argues that direct bootstrapping is equally powerless to
explain the justification we have for our senses as it is to explain the justification
we have for trusting ordinary instruments, both our senses and human testimony,
unlike instruments, enjoy a default rational justification. In particular, though
we are default-justified in accepting the deliverances of our senses and of human
testimony, we need a rational basis for justifiedly accepting the deliverances of
instruments.
Sanford C. Goldberg argues in ‘Reductionism and the Distinctiveness of Testi-
monial Knowledge’ that testimonial knowledge is an epistemically unique kind
of knowledge, regardless of whether there are epistemic principles that are unique
to cases of testimony. In particular, only testimonial knowledge is associated with
a characteristic expansion in the sorts of epistemically relevant moves that a sub-
ject can make when identifying the direct epistemic support for a given belief.
According to Goldberg, the specific characteristic expansion in question is that
testimonial knowledge—but no other—gives rise to the hearer’s right to pass the
epistemic buck after her own justificatory resources have been exhausted. Gold-
berg argues that this feature of testimonial knowledge reflects the epistemological
distinctiveness of relying on the epistemic authority of another rational being.
Moreover, Goldberg claims that the arguments on behalf of epistemic buck-
passing do not depend on there being testimony-specific epistemic principles,
and hence that the conclusion that testimonial knowledge is epistemically unique
should be embraced by both reductionists and non-reductionists in the epistemo-
logy of testimony.
Keith Lehrer argues in ‘Testimony and Trustworthiness’ that testimonial
knowledge requires trustworthiness on the part of both the speaker and the
hearer involved in a given exchange. He begins by arguing against causal or
transmission theories of testimonial knowledge, which maintain that knowledge
can be transmitted from a speaker to a hearer in the absence of any evaluation
of trustworthiness. Such theories, according to Lehrer, face what he calls the
opacity objection—a subject may be completely unaware of whether a belief that
she holds is worthy of her trust, that is, the epistemic merits of a belief may be
opaque to the subject holding it, thereby rendering her unable to justify or defend
her own belief. Moreover, because beliefs may be formed in evidentially defective
ways, requiring that the causal relations in question have a reliable or otherwise
appropriate etiology will not provide an adequate response to this objection.
In contrast, Lehrer argues that a speaker’s testimony that p qualifies as evidence
that p only when the speaker is found by the hearer in question to be trustworthy
in what she accepts and what she conveys. In addition, the speaker’s trustwor-
thiness must be successfully truth-connected, and the hearer must be correct in
her evaluation of the speaker’s trustworthiness. The result, according to Lehrer, is
12 Introduction

that testimonial justification is neither reductive, as reductionists would have it,

nor foundational, as non-reductionists would have it, since there is an essential
loop in the justification for accepting the testimony of others. In particular, the
hearer must depend on the testimony of others to evaluate her own trustworthi-
ness, but she must evaluate her own trustworthiness to evaluate the testimony
of others. According to Lehrer, however, there are two reasons why this loop
is a virtuous rather than a vicious one. First, any complete theory of justifica-
tion must explain why the theory itself is justified, and hence an explanatory
loop of this sort should not be regarded as a problem. Second, there is a mutu-
al support between the particular beliefs one has accepted in the past on the
basis of testimony and one’s general trustworthiness in accepting such reports.
Finally, Lehrer argues that what converts the acceptance of testimony into know-
ledge is that acceptance cannot be refuted or defeated by any errors in the evalu-
ation—irrefutable or undefeated justified acceptance of testimony is knowledge.
Jennifer Lackey argues in ‘It Takes Two to Tango: Beyond Reductionism and
Non-Reductionism in the Epistemology of Testimony’, that the current focus in
the epistemology of testimony needs to move beyond the debate between reduc-
tionism and non-reductionism. In particular, while a testimonial exchange typ-
ically involves the transfer of information between two central participants—the
speaker and the hearer—Lackey argues that reductionists and non-reductionists
alike have placed all of the epistemic work on only one or the other of these parti-
cipants and, in so doing, have ignored the positive justificatory contribution that
needs to be made by the other. On the one hand, reductionists focus entirely on
the hearer in a testimonial exchange, thereby failing to account for the fact that
no matter how excellent a hearer’s positive reasons are on behalf of an instance of
testimony, a speaker may still offer a report that is thoroughly unreliable. Non-
reductionists, on the other hand, focus entirely on the speaker, thereby failing to
account for the fact that no matter how reliable a speaker’s testimony is, this can-
not by itself make it rationally acceptable for a hearer to accept her report. In
contrast, Lackey argues on behalf of what she calls dualism in the epistemology of
testimony, which gives proper credence to the dual nature of testimonial justific-
ation by requiring positive epistemic work from both the speaker and the hearer.
Moreover, Lackey claims that this shows that testimonial justification is neither
reducible to nor completely independent from sense perception, memory, and
inductive inference, thereby moving the focus of the current debate beyond the
reductionism and non-reductionism dichotomy.
In ‘Testimonial Justification and Transindividual Reasons’, Frederick F.
Schmitt argues that testimonially justified belief, though not social in a strong
sense, can nonetheless be legitimately regarded as social in a qualified sense. In
support of this claim, he invokes what he calls the Transindividual Thesis, which
consists of the following two parts: first, H’s belief that p is justified on the basis
of testimony only if it is justified on the basis of S’s good reason to believe that p,
unless on the basis of a good reason to believe that p that H possesses herself and,
 Jennifer Lackey 13

second, H’s belief that p can be justified on the basis of testimony even though it
is not justified on the basis of a good reason that H possesses to believe that p.
By way of defending this thesis, Schmitt’s central strategy is to discuss what
he calls the Transtemporal Thesis, a similar though far more intuitively plausible
thesis regarding memorial justification, which consists of the following two parts:
first, a subject, A’s, belief that p is justified on the basis of memory only if it is
justified on the basis of A’s original good reasons to believe that p, unless on the
basis of a current good reason to believe that p and, second, A’s belief that p can
be justified on the basis of memory even though it is not justified on the basis
of a current reason to believe that p that A possesses. Schmitt then argues that
if the intuitively plausible Transtemporal Thesis is accepted, then an analogous
case can be made showing that the Transindividual Thesis is nearly as strong. To
this end, he relies on the intuitively plausible assumption that if A has a justi-
fied belief that p, then it is justified on the basis of a good reason to believe that
p suitably related to A. This assumption, according to Schmitt, equally supports
the first part of both the Transindividual and the Transtemporal Theses. He then
focuses on the second part of these theses. Regarding the Transtemporal Thesis,
he provides examples of memorially justified beliefs that are supported by neither
good current reductive nor good current non-reductive reasons, thereby show-
ing that no good current reasons exist for these justified beliefs. With respect to
the second part of the Transindividual Thesis, Schmitt employs a similar argu-
mentative strategy by providing examples of testimonially justified beliefs that
are supported by neither an ultimately coherentist nor an ultimately foundation-
alist justification, thereby showing that these beliefs are not justified by good
reasons possessed by the cognizer in question. Finally, Schmitt considers and
rejects various attempts to locate an asymmetry between memorial and testimo-
nial justification, including a counterfactual dependency sensitivity requirement,
a constraint on the role of reasons in the process of reasoning, and a responsibil-
ist conception of belief. He then concludes that the case for the Transindividual
Thesis is nearly as strong as that for the Transtemporal Thesis and, accordingly,
that those who accept the latter should similarly endorse the former.
In ‘Testimony and Epistemic Autonomy’, Elizabeth Fricker examines the ideal
of the epistemically autonomous knower—that is, one who believes only where
she herself possesses all of the (non-testimonial) evidence for what is believed,
trusting no one else’s work on any matter—and shows that the subtle and ubi-
quitous dependence each of us has on the testimony of others renders this ideal
practically impossible. She then asks whether there is reason to regret this con-
clusion. By way of answering this question, Fricker focuses on the circumstances
and topics in which one may properly accept and learn from another’s testimony.
She argues on behalf of the following Testimony Deferential Principle: A hearer,
H, properly accepts that P on the basis of trust in a speaker, S’s, testimony that
P if and only if S speaks sincerely, and S is epistemically well enough placed with
respect to P to be in a position to know that P, and S is better epistemically placed
14 Introduction

with respect to P than H, and there is no equally well-qualified contrary testi-

mony regarding P, and H recognizes all these things to be so. She then further
distinguishes between weak deferential acceptance, which occurs when H forms
the belief that P on the basis of trust in S’s testimony that P, when H has no firm
pre-existing belief regarding P, nor would H form a firm belief regarding P were
H to consider the question whether P using only H’s current epistemic resources
(apart from the testimony that P), and strong deferential acceptance, which occurs
when H allows S’s trusted testimony to override H’s own previous firm belief,
or disposition to form a firm belief, regarding P. The former, weaker kind of
deferential acceptance, according to Fricker, is often appropriate when a speak-
er has no superior expertise regarding the topic in question but, rather, is merely
contingently more expert than a hearer, such as when S but not H knows what
happened at a concert because S but not H attended it. In contrast, Fricker claims
that the latter, stronger kind of deferential acceptance is appropriate only when
a speaker has an expertise—that is, an intrinsic epistemic power—superior to a
hearer, such as when S but not H knows what is happening on the stage of a con-
cert because S but not H has excellent distance vision. With respect to both kinds
of expertise, Fricker argues that when H appreciates that S is an expert relative to
H, it is not merely rationally permissible, but rationally mandatory, for S to defer
to H regarding the domain in question.
After examining the various ways in which another person is sometimes better
placed epistemically than oneself to make judgments regarding a certain domain,
Fricker turns to a consideration of the way one’s epistemic self-governances can
be maintained in the face of one’s practically inevitable dependence on the testi-
mony of others. She first shows that our reliance on testimony does result in an
important loss of autonomy, for if H knows that P through being told that P
and trusting S, there is or was someone who knows that P in some other non-
testimonial way. This second-hand nature of testimonial knowledge reveals its
epistemic inferiority, since the hearer does not directly possess the evidence for
the belief in question and, accordingly, is less able to monitor for defeaters rel-
ative to such a belief. Nevertheless, Fricker argues that epistemic self-governance
can be maintained in spite of our dependence on others by critically assessing the
sincerity and competence of those whom we trust and by discriminatingly, rather
than blindly and universally, accepting what we are told.
In ‘Pathologies of Testimony’, C. A. J. Coady discusses three forms of telling
things to others that are either morally or epistemically suspect—gossip, rumor,
and urban myth—and evaluates the extent to which these suspicions are legit-
imate. He begins with gossip, some necessary features of which are that it is
typically conveyed by those who believe it to be true, its subject matter is per-
sonal, there must be something unwelcome to the gossipee about the conveying
of the information in question, and it can involve the transmission of justified
beliefs. He then proceeds to evaluate both the moral and the epistemic status of
gossip. With respect to the former, Coady claims that gossip can both reinforce
 Jennifer Lackey 15

positive existing norms and break down negative social norms, it need not violate
the respect of the gossipee, and it can be quite enjoyable. Regarding the epistem-
ic status of gossip, Coady considers two objections: first, since gossip is inspired,
not by a consideration of the truth but by malice, it is epistemically defective and,
second, it excludes the gossipee’s falsification or correction of the report in ques-
tion, which is an important epistemic source. He responds that gossip need not
be malicious—one may gossip about a person’s upcoming promotion simply out
of interest in the topic—and, even if it is not inspired by the purest or loftiest
of intentions, it need not be disconnected from a concern for the truth. To the
second objection, Coady argues that there is often little reason to suppose that
the gossipee will be a reliable source for confirming or disconfirming the report in
question. Hence, Coady argues that gossip is not in fact a pathology of testimony
at all but, rather, it is a perfectly normal form of testimony whose moral status is
less problematic than is typically thought and whose epistemic status is actually
quite epistemically reasonable.
Coady then turns to rumor and urban myth. Though similar to gossip in some
respects, rumor differs in that it has no strong justificatory base, its subject mat-
ter is not necessarily personal, and it is not restricted in its circulation. Though
rumor may have some epistemic merits, such as prompting the creation of hypo-
theses for further exploration, Coady argues that it is nonetheless a pathology of
testimony. For unlike cases of normal testimony, many cases of rumor involve
neither an original source of the information being conveyed nor a speaker who
has competence with respect to this information. More precisely, Coady argues
that rumor is what J. L. Austin calls a ‘misfire’ rather than an ‘abuse’ of the
speech act of testifying. Whereas misfires simply involve speech acts that have
gone wrong in a way that nullifies them, abuses constitute genuine but irregular
performances of these acts. In this way, the lack of credentials on the part of the
person who spreads rumor renders her assertion void as testimony, thereby res-
ulting in rumor being a pathological form of testimony. Urban myths, however,
differ from rumors in that they have much higher levels of complexity, are presen-
ted in the form of fully fledged stories, have an abiding nature, and are invariably
false and ill-founded, though commonly believed to be true. Coady then argues
that whether urban myths are pathologies of testimony depends on the degree to
which they are presented as testimony or as fiction.
Coady concludes by pointing out two general lessons that should be drawn
from his discussion. First, whether a certain communication is a pathology of
testimony cannot simply be read off from the form or content of its telling and,
second, pathologies of testimony need not be either morally or epistemically
worthless.
In ‘Getting Told and Being Believed’, Richard Moran focuses on the reciproc-
al relations of a speaker’s telling his audience something and his audience’s believ-
ing him, where ‘telling’ is distinct from but includes asserting, and believing
the person is distinct from but includes believing the proposition asserted. He
16 Introduction

approaches this through a contrast between what he calls the Evidential view and
the Assurance view of testimony. The Assurance view is developed through a dis-
cussion of an earlier paper by Angus Ross (1986), where Ross claims that it is
inconsistent with the spirit in which testimony is offered that the speaker present
his utterance as evidence for the truth in question. One advantage Moran sees in
the Assurance view is that it provides a way of understanding how the recogni-
tion of the intentional character of the utterance can contribute positively to the
reason-providing force of the speaker’s words, rather than being seen as evidence
that has been deliberately tampered with. Investigating the nature of the audi-
ence’s dependence on the speaker’s free assurance leads to a discussion of Grice’s
formulation of non-natural meaning in an epistemological light, concentrating
on just how the recognition of the speaker’s self-reflexive intention is supposed to
count for his audience as a reason to believe P. This is understood as the speaker’s
explicitly assuming responsibility for P’s being true, and thereby constituting his
utterance as a reason to believe.
Moran claims that several features of testimony are better understood in terms
of the Assurance view than the Evidential view. These include the fact that for
something a speaker does to be evidence for something else, it doesn’t matter
whether the speaker knows or understands that it is evidence, whereas this is cru-
cial to the reason-giving force of testimony; the contrast between our epistemic
relations to photographs and speech; the fact that something only counts as an
assertion if it is something done freely and consciously; and the speaker’s role in
determining the epistemic import of his utterance, as contrasted with the absence
of any authority he has with respect to the evidential significance of his actions,
including his verbal ones.

REFERENCES
Adler, Jonathan E. (1994), ‘Testimony, Trust, Knowing’, Journal of Philosophy, 91:
264–75.
(2002), Belief ’s Own Ethics (Cambridge, Mass.: MIT Press).
Alston, William P. (1989), Epistemic Justification: Essays in the Theory of Knowledge
(Ithaca, NY: Cornell University Press).
Audi, Robert (1997), ‘The Place of Testimony in the Fabric of Knowledge and Justifica-
tion’, American Philosophical Quarterly, 34: 405–22.
(1998), Epistemology: A Contemporary Introduction to the Theory of Knowledge
(London and New York: Routledge).
Austin, J. L. (1979), ‘Other Minds’, in his Philosophical Papers, 3rd edn. (Oxford: Oxford
University Press).
Bergmann, Michael (1997), ‘Internalism, Externalism and the No-Defeater Condition’,
Synthese, 110: 399–417.
(2004), ‘Epistemic Circularity: Malignant and Benign’, Philosophy and Phenomeno-
logical Research, 69: 709–27.
 Jennifer Lackey 17

BonJour, Laurence (1980), ‘Externalist Theories of Epistemic Justification’, Midwest

Studies in Philosophy, 5: 53–73.
(1985), The Structure of Empirical Knowledge (Cambridge, Mass.: Harvard Univer-
sity Press).
and Sosa, Ernest (2003), Epistemic Justification: Internalism vs. Externalism, Found-
ations vs. Virtues (Oxford: Blackwell Publishing).
Burge, Tyler (1993), ‘Content Preservation’, Philosophical Review, 102: 457–88.
(1997), ‘Interlocution, Perception, and Memory’, Philosophical Studies, 86: 21–47.
Chisholm, Roderick M. (1989), Theory of Knowledge, 3rd edn. (Englewood Cliffs, NJ:
Prentice-Hall).
Coady, C. A. J. (1992), Testimony: A Philosophical Study (Oxford: Clarendon Press).
(1994), ‘Testimony, Observation and ‘‘Autonomous Knowledge’’ ’, in Matilal and
Chakrabarti (1994: 225–50).
Dummett, Michael (1994), ‘Testimony and Memory’, in Matilal and Chakrabarti (1994:
251–72).
Evans, Gareth (1982), The Varieties of Reference (Oxford: Clarendon Press).
Faulkner, Paul (2000), ‘The Social Character of Testimonial Knowledge’, Journal of
Philosophy, 97: 581–601.
(2002), ‘On the Rationality of our Response to Testimony’, Synthese, 131:
353–70.
Foley, Richard (1994), ‘Egoism in Epistemology’, in Schmitt (1994: 53–73).
Fricker, Elizabeth (1987), ‘The Epistemology of Testimony’, Proceedings of the Aristoteli-
an Society, suppl. vol. 61: 57–83.
(1994), ‘Against Gullibility’, in Matilal and Chakrabarti (1994: 125–61).
(1995), ‘Telling and Trusting: Reductionism and Anti-Reductionism in the Epi-
stemology of Testimony’, Mind, 104: 393–411.
(2002), ‘Trusting Others in the Sciences: A priori or Empirical Warrant?’, Studies
in History and Philosophy of Science, 33: 373–83.
(forthcoming), ‘Knowledge from Trust in Testimony is Second-Hand Know-
ledge’, Philosophy and Phenomenological Research.
Goldman, Alvin I. (1986), Epistemology and Cognition (Cambridge, Mass.: Harvard Uni-
versity Press).
(1999), Knowledge in a Social World (Oxford: Clarendon Press).
Graham, Peter J. (1997), ‘What is Testimony?’, Philosophical Quarterly, 47: 227–32.
(2000), ‘Conveying Information’, Synthese, 123: 365–92.
Greco, John and Sosa, Ernest (eds.) (1999), The Blackwell Guide to Epistemology (Oxford:
Blackwell Publishers).
Hardwig, John (1985), ‘Epistemic Dependence’, Journal of Philosophy, 82: 335–49.
(1991), ‘The Role of Trust in Knowledge’, Journal of Philosophy, 88: 693–708.
Hawthorne, John (2004), Knowledge and Lotteries (Oxford: Oxford University Press).
Hume, David (1967), An Enquiry Concerning Human Understanding, in L. A.
Selby-Bigge (ed.), Hume’s Enquiries (Oxford: Oxford University Press).
Insole, Christopher J. (2000), ‘Seeing Off the Local Threat to Irreducible Knowledge by
Testimony’, Philosophical Quarterly, 50: 44–56.
Lackey, Jennifer (1999), ‘Testimonial Knowledge and Transmission’, Philosophical
Quarterly, 49: 471–90.
18 Introduction

Lackey, Jennifer (2003), ‘A Minimal Expression of Non-Reductionism in the

Epistemology of Testimony’, Noûs 37: 706–23.
(2005), ‘Testimony and the Infant/Child Objection’, Philosophical Studies, 126:
163–90.
(forthcoming a), ‘Learning from Words’, Philosophy and Phenomenological Research.
(forthcoming b), ‘The Nature of Testimony’, Pacific Philosophical Quarterly.
(forthcoming c), ‘Knowing from Testimony’, Philosophy Compass.
Lipton, Peter (1998), ‘The Epistemology of Testimony’, Studies in History and Philosophy
of Science, 29: 1–31.
Lyons, Jack (1997), ‘Testimony, Induction and Folk Psychology’, Australasian Journal of
Philosophy, 75: 163–78.
Matilal, Bimal Krishna and Chakrabarti, Arindam (eds.) (1994), Knowing From Words
(Dordrecht: Kluwer Academic Publishers).
McDowell, John (1994), ‘Knowledge By Hearsay’, in Matilal and Chakrabarti (1994:
195–224).
Millgram, Elijah (1997), Practical Induction (Cambridge, Mass: Harvard University
Press).
Nozick, Robert (1981), Philosophical Explanations (Cambridge, Mass: The Belknap
Press).
Owens, David (2000), Reason without Freedom: The Problem of Epistemic Normativity
(London and New York: Routledge).
Plantinga, Alvin (1993), Warrant and Proper Function (Oxford: Oxford University Press).
Pollock, John (1986), Contemporary Theories of Knowledge (Totowa, NJ: Rowman and
Littlefield).
Reed, Baron (forthcoming), ‘Epistemic Circularity Squared? Skepticism about Common
Sense’, Philosophy and Phenomenological Research.
Reid, Thomas (1993), The Works of Thomas Reid, ed. Sir William Hamilton (Charlottes-
ville, Va.: Lincoln-Rembrandt Publishing).
Reynolds, Steven L. (2002), ‘Testimony, Knowledge, and Epistemic Goals’, Philosophical
Studies, 110: 139–61.
Root, Michael (2001), ‘Hume on the Virtues of Testimony’, American Philosophical
Quarterly, 38: 19–35.
Ross, Angus (1986), ‘Why Do We Believe What We Are Told?’, Ratio, 28: 69–88.
Ross, James (1975), ‘Testimonial Evidence’, in Keith Lehrer (ed.), Analysis and Metaphys-
ics: Essays in Honor of R. M. Chisholm (Dordrecht: Reidel).
Rysiew, Patrick (2002), ‘Testimony, Simulation, and the Limits of Inductivism’, Aus-
tralasian Journal of Philosophy, 78: 269–74.
Schmitt, Frederick F. (ed.) (1994), Socializing Epistemology: The Social Dimensions of
Knowledge (Lanham, Md.: Rowman and Littlefield).
(1999), ‘Social Epistemology’, in Greco and Sosa (1999: 354–82).
Sosa, Ernest (1991), Knowledge in Perspective: Selected Essays in Epistemology (Cambridge:
Cambridge University Press).
Stevenson, Leslie (1993), ‘Why Believe What People Say?’, Synthese, 94: 429–51.
Strawson, P. F. (1994), ‘Knowing From Words’, in Matilal and Chakrabarti (1994:
23–7).
 Jennifer Lackey 19

Webb, Mark Owen (1993), ‘Why I Know About As Much As You: A Reply to Hardwig’,
Journal of Philosophy, 90: 260–70.
Weiner, Matthew (2003), ‘Accepting Testimony’, Philosophical Quarterly, 53: 256–64.
Welbourne, Michael (1979), ‘The Transmission of Knowledge’, Philosophical Quarterly,
29: 1–9.
(1981), ‘The Community of Knowledge’, Philosophical Quarterly, 31: 302–14.
(1986), The Community of Knowledge (Aberdeen: Aberdeen University Press).
(1994), ‘Testimony, Knowledge and Belief ’, in Matilal and Chakrabarti (1994:
297–313).
Williams, Michael (1999), Groundless Belief: An Essay on the Possibility of Epistemology,
2nd edn. (Princeton: Princeton University Press).
Williamson, Timothy (1996), ‘Knowing and Asserting’, Philosophical Review, 105:
489–523.
(2000), Knowledge and its Limits (Oxford: Oxford University Press).

N OT E S

1. For views of the nature of testimony with various additional restrictions, see Ross
(1975), Coady (1992), Graham (1997), and Lackey (forthcoming b).
2. Similarly, Peter Graham says, ‘it should be noted that mere statements are not testi-
mony. Saying ‘‘It is a nice day’’ is not usually taken as testimony about the weather
(though it is when said by the weatherman). Repeating what you have already said
over and over does not count as testimony either, unless you have forgotten each
previous utterance’ (1997: 231).
3. Of course, the claim here is not that such conversational fillers and polite responses
should never qualify as instances of testimony but, rather, that they should not
always qualify as testimony. For instance, if I say to my blind companion, ‘It is a
beautiful day today’, such a remark may qualify as testimony in this context since its
function may be to convey information, not to merely fill a gap in the conversation.
4. This, of course, is not an exclusive ‘or’; both (1) and (2) could be satisfied simultan-
eously.
5. Clause (2) of T may need to be modified in something like the following way:
(2*) a is or should be reasonably taken as conveying the information
that p.
This modification would allow for the following type of case to qualify as testi-
mony: Thelma, while engaged in a soliloquy, confesses to murdering her husband.
Although I overhear her make this confession, I do not take her act of commu-
nication as conveying the information for any proposition since I think it is far
too outlandish a possibility. Nevertheless, we may still wish to regard Thelma’s
statement as an instance of testimony because I should have reasonably taken it as
conveying the information that she murdered her husband. If this is correct, then
(2*) can be substituted for (2) in T.
6. This is a modified version of an example found in Audi (1997).
7. The ‘in part’ clause is included since an act of communication can, for instance, be
reasonably offered as conveying information in virtue of both its perceptual and its
20 Introduction

communicable content and yet still qualify as testimony, e.g., I reasonably offer as
conveying the information that I have a soprano voice my saying, in a soprano voice,
that I do.
8. For more on the account of testimony found in T, see Lackey (forthcoming b).
9. This type of example is found in Sosa (1991).
10. Proponents of various versions of non-reductionism include Austin (1979), Wel-
bourne (1979, 1981, 1986, and 1994), Evans (1982), Ross (1986), Hardwig (1985
and 1991), Coady (1992 and 1994), Reid (1993), Burge (1993 and 1997), Plantinga
(1993), Webb (1993), Dummett (1994), Foley (1994), McDowell (1994), Strawson
(1994), Williamson (1996 and 2000), Goldman (1999), Schmitt (1999), Insole
(2000), Owens (2000), Rysiew (2002), Weiner (2003), Sosa (Chapter 5 in this
volume), and Goldberg (Chapter 6 in this volume). Some phrase their view in
terms of knowledge, others in terms of justification or entitlement, still others in
terms of warrant. Audi (1997, 1998, and Chapter 1 in this volume) embraces a
non-reductionist view of testimonial knowledge, but not of testimonial justifica-
tion. Stevenson (1993), Millgram (1997), and Graham (Chapter 4 in this volume)
defend restricted versions of non-reductionism.
11. This is a broad characterization, with subtler versions of non-reductionism not
always clearly subsumed by it. (See, for instance, Graham (Chapter 4 in this volume)
and Goldberg (Chapter 6 in this volume).)
12. To be even more precise, there are two different kinds of psychological defeaters:
rebutting defeaters are those that indicate the target belief is false while undercutting
defeaters are those that indicate the target belief is unreliably formed or sustained.
See Pollock (1986) for further development of the distinction between rebutting and
undercutting defeaters.
13. For various discussions of what I call psychological defeaters see, for
example, BonJour (1980 and 1985), Nozick (1981), Pollock (1986), Goldman
(1986), Plantinga (1993), Lackey (1999, 2003, 2005, and Chapter 8 in this
volume), Bergmann (1997 and 2004), and Reed (forthcoming).
14. Following the distinction in n. 12, there are rebutting and undercutting normative
defeaters. The central difference is that while psychological defeaters are doubts or
beliefs had by the subject, their normative counterparts are doubts or beliefs that the
subject should have. For more on this, see Lackey (1999, 2003, 2005, and Chapter 8
in this volume) and Reed (forthcoming).
15. For discussions involving what I call normative defeaters, approached in a number of
different ways, see BonJour (1980 and 1985), Goldman (1986), Fricker (1987 and
1994), Chisholm (1989), Burge (1993 and 1997), McDowell (1994), Audi (1997
and 1998), Williams (1999), Lackey (1999, 2003, 2005, and Chapter 8 in this
volume), BonJour and Sosa (2003), Hawthorne (2004), and Reed (forthcoming).
What all of these discussions have in common is simply the idea that evidence can
defeat knowledge (warrant, justification) even when the subject does not form any
corresponding beliefs from the evidence in question.
16. Proponents of different versions of reductionism include Hume (1967), Fricker
(1987, 1994, 1995, 2002, and Chapter 10 in this volume), Adler (1994 and 2002),
Lyons (1997), Lipton (1998), and Van Cleve (Chapter 2 in this volume). For a nice
discussion of Hume’s version of reductionism, see Root (2001). Faulkner (2000),
 Jennifer Lackey 21

Lackey (Chapter 8 in this volume), and Lehrer (Chapter 7 in this volume) develop
‘hybrid’ or qualified reductionist/non-reductionist views of testimonial justification
and knowledge.
17. ‘My reliance on a particular piece of testimony reduces locally just if I have adequate
grounds to take my informant to be trustworthy on this occasion independently of
accepting as true her very utterance’ (Fricker 1995: 404).
18. See, for instance, Fricker (1987, 1994, and 1995), Faulkner (2000 and 2002),
and Lackey (2003, 2005, and Chapter 8 in this volume).
19. See Van Cleve (Chapter 2 in this volume).
20. See, for instance, Webb (1993), Foley (1994), Strawson (1994), and Schmitt (1999).
21. See also Faulkner (2000).
22. Proponents of different versions of the necessity thesis (TEP-N) include Welbourne
(1979, 1981, 1986, and 1994), Hardwig (1985 and 1991), Ross (1986), Burge
(1993 and 1997), Plantinga (1993), McDowell (1994), Williamson (1996 and
2000), Audi (1997, 1998, and Chapter 1 in this volume), Owens (2000), Reynolds
(2002), and Schmitt (Chapter 9 in this volume). For slightly weaker versions of
this thesis, see Dummett (1994) and Fricker (forthcoming and Chapter 10 in this
volume).
23. Proponents of different versions of the sufficiency thesis (TEP-S) include Austin
(1979), Evans (1982), Fricker (1987), Coady (1992), and Owens (2000). Burge
(1993), Williamson (1996), and Audi (1997) endorse qualified versions of this thes-
is. For instance, Burge claims that ‘[i]f one has acquired one’s belief from others in a
normal way, and if the others know the proposition, one acquires knowledge’ (1992:
477 n. 16; emphasis added). Timothy Williamson writes that ‘[i]n normal circum-
stances, a speaker who asserts that P thereby puts a hearer in a position to know that
P if (and only if ) the speaker knows that P’ (1996: 520; emphasis added). Similarly,
Audi writes, ‘Concerning knowledge, we might say that at least normally, a belief
that p based on testimony thereby constitutes knowledge . . . provided that the attest-
er knows that p and the believer has no reason to doubt either p or the attester’s
credibility concerning it’ (1997: 412; emphasis added).
24. See Lackey (1999, 2003, and forthcoming a) and Graham (2000).
25. See Lackey (forthcoming a).
26. See Fricker (forthcoming).
This page intentionally left blank
 PART I
T E S T I M O N Y A N D T H E L E G AC Y
OF THOMAS REID
This page intentionally left blank
 1
Testimony, Credulity, and Veracity
Robert Audi

The social aspects of knowledge and justification are an important dimension of

epistemology, and in the history of epistemology they have often been neglected.
Recent work in social epistemology has partially filled this gap,¹ and the literature
of epistemology now contains much discussion of testimony and other aspects of
the communication of knowledge and, if to a lesser extent, of the communicative
sources of justification.² In the history of philosophy, Thomas Reid stands out
among major philosophers both for his extensive treatment of testimony and for
his distinctive conception of it. My main aim here is to extend my own account
of testimony, but the extensions in question are best accomplished in the light
of some of what we learn from Reid. I begin with a partial statement and brief
defense of my account. I then proceed to describe some major elements in Reid’s
view of testimony and to compare the two positions. In that light, I make a num-
ber of points both about how testimony yields knowledge and justification and
about its status as a source of each.

I . T E S T I M O N Y- B A S E D B E L I E F

When philosophers speak of testimony, they usually have in mind not the
formal reporting of the court witness (which is what, in many societies, first
comes to mind when testimony is spoken of), nor even the self-conscious
delivery of information, but virtually any instance of someone’s telling somebody
something, where this is telling that—propositional telling—as opposed to telling
to—imperatival telling.³ Telling that is roughly a matter of saying the kind of
thing from which we learn facts from other people. In this wide sense, ‘testimony’
applies to nearly everything we say to others. To be sure, there is the expository
saying illustrated when, in setting out someone else’s view, we drop ascriptive
expressions like ‘for her’ and express the view from the inside. There is also the
theatrical saying characteristic of acting and the narrative saying exemplified by
reading a story aloud to others. What is said in these cases does not constitute

This paper has benefited much from discussions with William Alston, Elizabeth Fricker,
Christopher Green, Thomas Kelly, Jennifer Lackey, Fritz Warfield, and, especially, Peter Graham.
26 Testimony, Credulity, and Veracity

testimony, at least in the full-blooded sense, as opposed to the kind illustrated by,
say, testimony given in a play. For cases that do constitute testimony, the verb
‘attest’ often serves better than ‘testify’ (as do ‘tell’ and ‘say’, used with ‘that’), and
I will often speak of attesting, attesters, etc., as well as of testimony.
If the philosophically interesting notion of testimony is immensely wide, the
philosophically most important notion of testimony-based belief is much less so.
I take this to be the kind of belief that arises naturally, non-inferentially, and
usually unselfconsciously in response to what someone says to us. I ask you the
time; you tell me it is nine o’clock; and straightaway I believe this on the basis of
your saying it. This basis relation which your testimony bears to my belief has a
causal component, but that is not its only element. Thus, if I mishear you, your
saying, not that it is nine, but that it is noon, could cause me to believe it is nine;
but my belief is not then based on your testimony that it is noon. It is in fact not,
properly speaking, based on your testimony at all, but only on your testimonial
act. I hear and am affected by this, but misunderstand it. To be sure, I might still
be said to believe that it is nine because of that testimony, as I certainly might
believe that you have a cold because of your attesting to this in a nasal voice. The
latter is possible even when I do not understand what you have said and so do not
believe this on the basis of your testimony that you have a cold.
Another case in which belief is produced by testimony, but not based on it, is
this. Suppose I believe something on the basis of premises supporting your testi-
mony, as where the content seems implausible by itself but I judge you to be both
highly competent and unassailably sincere, and for that reason I believe what you
say. Here I may also be said to believe it from your testimony (which is a trigger-
ing causal basis of my forming the belief) and, if I come to know it, to know it
(mainly) through that testimony. But this is not belief or knowledge on the basis
of your testimony. My basis is a combination of your testimony and my beliefs
about you. To be sure, what is believed on the basis of testimony is also believed
from it; but the converse is false.
It is not only in the case of testimony that a mere causal relation between a
source of knowledge and a belief based on that source is not sufficient to render
that belief knowledge of the kind distinctive of the cognitive products of that
source. A perceptual belief, for instance, is more than a belief caused by seeing,
hearing, or otherwise perceiving a perceptible object. It must certainly be more if
it is to constitute perceptual knowledge of the object. A machine that stimulates
our brains in a certain way could cause us to believe (truly) that this machine is in
a certain room; it can do so without providing us with any ground for believing
this and without our perceiving it there or otherwise knowing that it is there.
Here, as in further respects, testimony is like other sources of knowledge and
justification.
Perception provides more than a source of analogies (and disanalogies) with
testimony. To yield a testimony-based belief, an attestation must figure as a per-
ceptual object, for instance as an assertion one hears or as a sentence one sees
 Robert Audi 27

in a letter from a friend. It must be appropriately received, and the appropriate

reception requires that it be a perceptual as well as a semantic object. That is why
the belief that someone has a cold, formed only through noting the nasal tone of
an attestation, is not testimony-based belief, but only a kind of belief from testi-
mony. The content of what is said is irrelevant to its grounding. In a sense, then,
testimony-based belief requires semantic interpretation or, at any rate, seman-
tic construal. This is why misinterpreting an attestation can prevent the proper
functioning of the testimony.
It would be a mistake to think that some conscious activity of interpretation is
generally required for testimony-based knowledge. Typically, we simply under-
stand what is said and believe it. There may be a kind of interpretation process
accomplished by the mind (or perhaps just by the brain)—‘construal’ seems
preferable to ‘interpretation’ here. This certainly seems to occur where, in the
light of context, we avoid misinterpretation or see past an apparent ambiguity
produced by syntax, intonation, multivocality, or some other source of poten-
tial misunderstanding. But for normal adults receiving testimony in their native
language, no conscious activity of, say, parsing is invariably required.
Even where one must think about what is said and laboriously interpret it, as
with a complex message or an utterance by certain non-fluent non-native speak-
ers, it does not follow that one’s belief finally arising from accepting the message
one discerns is inferential. Indeed, testimony-based belief, as I construe it, and as
I think it is normally understood, is never inferential. If, before taking your word
for my student’s having left a paper for me by yesterday’s deadline, I must note
that it could not have gotten in the mail room (which was locked at the end of the
day) any later than the deadline, my belief is at most partly based on your testi-
mony. And if, as a ground for believing what you say, I must infer your credibility
from background information about you, my belief of your attestation, though
acquired through your testimony, may not be said without qualification to be
based on it. In the first case, I may have two or more independent grounds for
p (the proposition in question). It is not easy to explicate such cases. As I con-
ceive this example, both grounds are necessary for my forming the belief. But I
can also have two independent sets of grounds, say one perceptual and the other
testimonial, that are each sufficient (and hence neither by itself necessary) for my
knowing that p. There are other aspects of what we might call mixed grounds, but
we may here set them aside. It is quite enough if we can understand belief based
wholly on testimony, as much belief surely is.
One further point is needed before we can proceed to epistemological aspects
of testimony. To say that testimony-based belief is not inferential is not to imply
that it is ‘‘uncritical’’ or even manifests a degree of credulity incompatible with
a critical habit of mind. This seems to me to hold even for much testimony-
based belief formed where one has no information about the attester beyond what
the context provides, which is often very little. We all have background beliefs
that constrain what we accept, for instance by preventing our accepting stock
28 Testimony, Credulity, and Veracity

market forecasts without special evidence; and some people may have standards
of plausibility that go beyond the constraints provided by these beliefs.
Consider, for instance, how one might be habituated to taking intonation
and facial features into account. These elements are important constraints on
acceptance of much oral testimony, but no specific beliefs need express fully the
way such elements constrain the formation of testimony-based beliefs. At least
for non-skeptics, a critical stance is possible without reasoning from any of its
standards to the acceptability of the testimony, and indeed without inference
at all. Our critical habits and even our critical standards need not all reside in
propositions we believe. Even active monitoring of testimony is possible without
making inferences: if nothing is noticed that requires raising questions or drawing
inferences, no questioning or inference need arise even from attentive monitoring
by a critical listener.

I I . T E S T I M O N Y A S A B A S I S O F K N OW L E D G E A N D
J U S T I F I C AT I O N

If testimony-based belief is non-inferential, then if such belief can constitute

knowledge, it might seem to follow that it constitutes non-inferential knowledge.
This does not obviously follow, however. Perhaps it is possible for a belief’s status
as knowledge to depend on inference in certain ways even when the belief is not
itself inferentially based on any other belief. From the perspective of a certain
kind of epistemological coherentism, this is how it might be: even if I non-
inferentially believe what you tell me, I do not know it except on the basis of
either inferring it (perhaps unconsciously) from something else, or at least on the
basis of having both a set of suitable premises as grounds for my believing it and
an appropriate readiness to infer it from them. I have elsewhere argued against
even the weaker of these inferentialist pictures of knowledge⁴ and will assume the
independently plausible view that we commonly acquire non-inferential know-
ledge from testimony.
I do not deny that there is here (as elsewhere) a negative dependence on
inference: our knowledge can be defeated if we draw certain inferences, for
instance infer from certain telltale signs that the attester is unreliable. But to treat
testimony-based knowledge as inferential on that ground is rather like treating
the absence of poisonous gas as a source of one’s physical well-being because, if
such gas were present, one would be ill. Vulnerability to defeat by inferences that
could occur does not entail dependence in any positive sense. Quashing skeptical
worries might require drawing certain self-protective inferences, but testimony-
based knowledge is possible without the support of precautionary inference.⁵
It is one thing to say that testimony can be a source of non-inferential know-
ledge; it is quite another to say under what conditions it achieves this epistemic
success in producing testimony-based belief. In earlier work I have suggested that
 Robert Audi 29

the attester must know that p and that the recipient of testimony, in addition
to acquiring a belief that is non-inferentially based on the testimony, must not
have certain kinds of grounds for doubting or denying p. A great deal must be
said to explicate all of these conditions, but much of that comes from general
epistemology and can be presupposed here.⁶
One kind of case, however, should be considered now.⁷ Imagine a teacher
(Luke, let us say) who disbelieves the theory of evolution but teaches it conscien-
tiously. He tells his students, on the basis of his correct reading of the theory and
his knowledge of fossil discoveries, that there were homo sapiens in a certain place.
Suppose he is giving his students correct information for which there is adequate
evidence. May we not conclude that testimony-based belief (theirs) can consti-
tute knowledge without Luke’s knowing the proposition in question (which he
disbelieves)? This is an interesting case, since the hearers do have a testimony-
based true belief that seems adequately grounded. But is it adequately grounded,
if the teacher would have taught a false theory in the same disbelieving way, had
this been required by his job? Even if the theory itself is (an item of) ‘‘know-
ledge’’ (as one might say if it is known by someone), he isn’t a reliable link in the
chain from the fossil record through the theory, since he neither knows the theory
nor even believes it, hence does not believe it on the kind of ground that would
protect him from error in the way the (truth-conducive) grounding of knowledge
does. By his lights, in fact, he is deceiving the children—a point important in itself
for the epistemology of testimony. Moreover, it appears that he would have been
as likely to state a false proposition if the school required his teaching a mistaken
theory that seemed to him no more pernicious than this one. Such a person might
well be teaching a false theory or one that is not well evidenced and just happens
to be correct.
The case has more plausibility on the assumption that the school would not
require anyone to teach a theory that is not well evidenced. Let us suppose so.
But do the students perhaps believe or presuppose something to this effect? In
any case, might an essential part of their basis for believing Luke, or for knowing
on the basis of his testimony, be that background belief or presupposition? If
so, either their belief that p would be bolstered by background elements in a
way that implies that it is at most in part testimony-based and hence not what
normally counts as a testimony-based belief or, on the other hand, their belief
would not count as knowledge owing to the falsity of the presupposition in
the case of Luke. In order to believe what (in this kind of case) Luke says—or
at least in order to know its truth—they would presumably have to believe
or presuppose something to the effect that this is what the school is teaching.
If, on the other hand, they simply take his word, they are taking the word of
someone who will deceive when job retention requires it and (let us charitably
assume) there is a plausible rationale for the proposition in question. It is highly
doubtful that this kind of testimonial origin would be an adequate basis of
knowledge.
30 Testimony, Credulity, and Veracity

A related case (readily generalizable to other subject-matters) is more difficult

for the thesis that in order for the recipient of testimony to acquire testimony-
based knowledge that p from an attestation that p, the attester must know that
p. Suppose Luke is fallibilist, but doxastically incontinent. He thus too readily
forms beliefs which he takes to be unjustified on his evidence, particularly where
he wants to believe the proposition in question, say for reasons of an emotional
or religious kind. Again imagine that p is both true and well supported by the
evidence he has. Imagine that in the light of his evidence he believes that p is very
likely to be true (and considerably more likely to be true than any contrary), but
(while realizing the good chance of his being in error), for religious reasons, he
still disbelieves p. We can even assume that in the case at hand he would not (per-
haps for moral reasons) teach that p apart from believing that it is well supported
by the evidence. It may now seem that it is in general unlikely that he would teach
a false proposition.
One question raised by the case is whether the kind of cognitive malfunc-
tion in question makes Luke insufficiently reliable for his (true) attestation to
ground knowledge. For to fail to believe on evidence one takes to be good (to the
degree required by the case) is a kind of malfunction (Luke does not have to have
highly specific beliefs about the evidence but his standard must be high enough
to make it plausible to say that its role enables the students to know that p). A
related problem is this. Given that Luke’s beliefs may go against his evidence,
why should we think that his assessments of the evidence (including probability-
ascribing beliefs) are sound? We have to imagine a cognitive malfunction that
occurs isolated from what is clearly a similar one—both malfunctions are mat-
ters of responsiveness to grounds, whether they are constituted by evidence itself
or by a sense thereof, or by a belief, regarding some source, to the effect that it
is evidence for p. If he fails to believe (or even disbelieves) p when he takes the
evidence to give it strong support, what is to prevent his failing to take evidence
to show what it does, either believing (say) that it does not support p when in
fact it does or at least failing to believe this truth? If his cognitive response to evi-
dence is unreliable, why should his belief that he has it (or sense of it) render his
(unbelieving) testimony reliable?
A further problem is this. Given Luke’s willingness to deceive the students, why
should we view what he says when he has evidence as sufficiently representative
of his testimonial communications to give his present testimony the epistemic
(knowledge-grounding) status it is supposed to have? The problem here is not, as
above, that there might be relevantly similar situations in which his evidence is
inadequate and what he is saying (p) is false. The problem is that, given his willing-
ness to deceive the students regarding p, there are relevantly similar cases in which
he does so without even believing the evidence supports p and where it is false.
Why, given that willingness, should we take his devotion to the existence of evid-
ence to be adequately stable? Here one might point out that virtually anyone will
lie on some occasions; perfect devotion to truth or evidence is not a requirement on
 Robert Audi 31

generating testimony-based knowledge. Granted. But there are what we might call
credibility requirements on testimony that seem pertinent here. Compare a case in
which, at gunpoint, A commands B to tell us whether p. Should we believe B
here? There is a good reason why B might lie, and we cannot in general acquire
testimony-based knowledge regarding p. And what of a case in which we see that
an attester stands to risk bankruptcy depending on what the person says? It may
well be that a kind of sincerity on the part of the attester is a requirement on the
recipient’s gaining testimony-based knowledge and that this partly explains why
in these and certain other cases testimony-based knowledge apparently cannot be
acquired.⁸ But credibility requirements are broader still, and we are surely entitled
to ask whether it might not be largely good fortune that the students are being told
something true rather than something false that Luke thinks is well evidenced and
believes he can express without reprisals from authorities. If reasons of emotion
can lead to his willingly deceiving his students regarding what he believes, we are
invited to wonder whether a similar kind of insincerity might lead Luke to deceive
them regarding what is true. To be sure, it is not clear how strong grounds of
genuine knowledge must be, and more remains to be said for a full assessment of
the kind of case we are imagining. But my tentative conclusion is that it seems
doubtful that cases like Luke’s show that testimony-based knowledge can arise
from a basis with defects of the general kind in question.
If testimony is limited in that an attestation cannot ground testimony-based
knowledge that p if the attester does not know that p, testimony is unlimited in
the diversity of knowledge it can convey and in the ease with which it enables us
to learn things straightaway. Two other limitations should be mentioned. First,
the success of testimony depends on the operation of another source of know-
ledge, namely perception, broadly conceived. Second, what is known through
testimony must apparently be not only known by the attester but, in addition,
known (by the attester or someone earlier in the testimonial chain) at least in part
on the basis of another source, such as perception or reflection. In order to receive
your testimony about the time, I must hear you or otherwise perceive⁹—in some
perhaps very broad sense of ‘perceive’—what you say; and in order for me to
know the time on the basis of your testimony, you or someone else must have
read a clock or come to know the time on some other basis that is at least partly
non-testimonial. If all this is right, testimony is, as a source of knowledge, both
operationally dependent on perception and epistemically dependent on at least one
non-testimonial source of knowledge.
The case with justification is different. I do not need any justification for
believing p in order for my testimony to provide others with ample justification
for believing it. The credibility of testimony may be objectively ill-grounded in
this way without that fact’s providing any reason for the recipient to doubt either
what is said or the attester’s credibility (where objective grounding of justification
for a belief is the kind that counts, as does normal perceptual grounding,
toward the belief’s constituting knowledge). The crucial question is whether the
32 Testimony, Credulity, and Veracity

recipient is justified in believing my testimony (something quite possible without

the recipient’s having any actual belief, as opposed to dispositions to form beliefs,
concerning my credibility).
There is a related contrast: whereas we can acquire testimony-based knowledge
that p from testimony that p without having any knowledge of (or even justifi-
cation for affirming) the credibility of the attester, we apparently cannot acquire
justified testimony-based belief that p without having any justification concern-
ing that credibility. The former, positive point is amply illustrated by the acqui-
sition of such knowledge on the part of tiny children who are too conceptually
unsophisticated to have knowledge (or justification) concerning the credibility of
attesters. The latter, negative point is controversial (and I return to it in the final
section), but it is quite plausible once we note that having justification for a prop-
osition does not entail believing it; hence, having justification for presupposing
the attester’s credibility does not entail the psychologically implausible require-
ment that the recipient must have beliefs about that credibility. The point is still
more plausible if we assume, as I do here, that the degree of justification required
on the recipient’s part is only that degree we normally have regarding ordinary
attesters speaking within a domain concerning which we have no reason to think
they are either misinformed or motivated to deceive us. I take the position (which
I grant sets skepticism aside) that much of what people say to us and around
us meets this condition. I view normal people as very commonly, and in many
kinds of situations quite typically, both veracious and competent in the matters
on which they give testimony.
I certainly do not wish to imply that we acquire testimony-based justification
only if, independently of the testimony, we need justification for p itself. Even
inductivist reductionists can grant that this requirement would undermine our
capacity to acquire much of the justification that, on a non-skeptical view, we do
acquire from what we are told. Of both points—the common veracity and usual
competence of attesters, and the absence of any need for justification regarding
each attested item we are justified in believing—I consider Reid a persuasive pro-
ponent. Let us consider some of his main ideas concerning testimony.

I I I . S O M E M A J O R E L E M E N TS I N R E I D ’ S C O N C E P T I O N
OF TESTIMONY

Reid is well known to have taken humanity to have a ‘‘natural credulity’’: ‘‘I
believed by instinct whatever they [my parents] told me, long before I had the
idea of a lie, or thought of the possibility of their deceiving me. Afterwards, upon
reflection, I found that they had acted like fair and honest people.’’¹⁰ Later, he
includes a principle of credulity with one of veracity. Speaking of God as intend-
ing that ‘‘we should be social creatures’’ and as implanting ‘‘in our natures two
principles that tally with each other,’’ he says:
 Robert Audi 33

The first of these principles is, a propensity to speak truth . . . Truth is always uppermost,
and is the natural issue of the mind . . .
Another original principle implanted in us by the Supreme Being, is a disposition to
confide in the veracity of others, and to believe what they tell us. This is the counterpart
to the former; and, as that may be called the principle of veracity, we shall, for want of a
more proper name, call this the principle of credulity. (pp. 94–5)
In short, Reid sees us as endowed by God with natural tendencies toward both
truthfulness and trustfulness.
Developing the idea of credulity in relation to what he apparently views as
reasonable belief-formation in situations in which one receives testimony, he
continues:
It is evident that, in the matter of testimony, the balance of human judgment is by nature
inclined to the side of belief; and turns to that side of itself when there is nothing put
into the opposite scale. If it was not so, no proposition that is uttered in discourse would
be believed, until it was examined and tried by reason; and most men would be unable
to find reasons for believing the thousandth part of what is told them. Such distrust and
incredulity would . . . place us in a worse condition than that of savages. (p. 95)
If one thinks of credulity as implying naivety, one has the wrong picture of
Reid’s view. As he says in comparing the inductive principle with that of credu-
lity, ‘‘This principle, like that of credulity, is unlimited in infancy, and gradually
restrained and regulated as we grow up’’ (p. 102).
If, however, Reid finds important similarities between testimony and other
sources of knowledge and belief, he also sees differences:
There is, no doubt, an analogy between the evidence of sense and the evidence of testi-
mony. Hence, we find, in all languages, the analogical expressions of the testimony of sense
. . . and the like. But there is a real difference . . . In believing upon testimony, we rely
upon the authority of a person who testifies; but we have no such authority for believing
our senses. (Essays on the Intellectual Powers, p. 203)
It is interesting that in comparing memory with perception Reid does not draw
this contrast:
If we compare the evidence of sense with that of memory, we find a great resemblance,
but still some difference. I remember distinctly to have dined yesterday with such a com-
pany . . . I have a distinct conception and firm belief of this past event; not by reasoning,
not by testimony, but from my constitution. And I give the name of memory to that part
of my constitution by which I have this kind of conviction of past events.
I see a chair on my right hand. What is the meaning of this? It is, that I have, by my
constitution, a distinct conception and firm belief of the present existence of the chair in
such a place . . . (Essays on the Intellectual Powers, pp. 203–4)
Reid’s view here seems to be that whereas our reliance on testimony is instinctual
initially and regulated as we mature, our reliance on memory, like our reliance
on perception, is constitutional. (What is constitutional can also be instinctual,
34 Testimony, Credulity, and Veracity

and Reid’s contrast between testimony and, on the other hand, perception and
memory, is consistent with our having to rely on God’s design in the latter cases,
in the sense that—as Reid seemed to think—the reliability of all our faculties is
dependent on the ‘‘Author of Nature.’’)
Our reliance on perception, moreover, is associated with its feeding directly
into cognition: perception, for Reid, seems at least normally belief-entailing
(a property he apparently does not attribute to memory impressions as such).
Shortly after noting that ‘‘the perception of an object implies both a conception
of its form, and a belief of its present existence’’ and that this belief ‘‘is the
immediate effect of my constitution’’ (An Inquiry, p. 84), he says, ‘‘My belief is
carried along by perception, as irresistibly as my body by the earth’’ (p. 85).
The irresistibility of belief-formation given certain perceptions and perhaps
certain memory impressions does not preclude our regulating our responses to
these two sources, say in how strong a belief we form or in our disposition to
form related higher-order beliefs, such as the belief that we may be deceived in
believing p. But, for what one might call basic perceptions, as opposed to the
‘‘acquired’’ ones that we have when we have learned to see the three-dimensional
roundness of a globe rather than just its circularity, Reid does not emphasize any
regulation.¹¹ He seems, moreover, to treat memory and perception as playing
roles in the development of our knowledge that are quite different from the role
of testimony.
It would be a mistake to think, on the basis of Reid’s contrast between testi-
mony and memory, that he took the importance of testimony as a source of
knowledge to be a secondary matter. He describes as a first principle of contin-
gent truths ‘‘That there is a certain regard due to human testimony in matters of
fact, and even to human authority in matters of opinion’’ (Essays on the Intellectual
Powers, p. 281). It is noteworthy, however, that what immediately follows this
is not (as I read it) mainly an indication of why testimony should be considered
worthy of regard, but something quite different though certainly closely related:
an affirmation of its essential role in developing human knowledge: ‘‘Before we
are capable of reasoning about testimony or authority, there are many things
which it concerns us to know, for which we can have no other evidence’’ (p. 281).
Perhaps he took the unavailability of other evidence here to indicate worthiness
of belief, if only through the good grace of God. But the concepts in ques-
tion—certainly that of knowledge—are connected with the domain of reliability
as well as with worthiness of regard, and in a way that implies the possibility of
justification of testimony-based beliefs by reasons.¹² This possibility is consist-
ent with Reid’s view that we do not normally need reasons to believe testimony.
Not needing does not entail the impossibility—or even the abnormality—of
having.
We might speculate regarding one reason why Reid might think that testi-
mony is worthy of regard. (I am taking it that there can be a reason for this and
indeed that one can have it, even if the principle that testimony is due a ‘‘certain
 Robert Audi 35

regard’’ is self-evident and one rightly thinks it is.¹³) Consider the principle he
introduces just prior to setting forth the principle of testimony just quoted:
Another first principle I take to be, That certain features of the countenance, sounds of the
voice, and gestures of the body, indicate certain thoughts and dispositions of the mind . (Essays
on the Intellectual Powers, p. 279)
It is important to see that these features, sounds, gestures, and presumably other
expressive aspects of human behavior are ‘‘natural signs’’: ‘‘It seems to me incred-
ible, that the notions men have of the expression of features, voice, and gesture,
are entirely the fruit of experience’’ (p. 280). Now, suppose Reid took testimony
to be characteristically accompanied by natural signs of believing what is attested
to (roughly, of conviction); he might have thought of testimony as, if only in ele-
mentary or otherwise special cases, naturally indicating, not as just reporting, or
linguistically expressing, belief. In that case, it is easy to see how he could regard as
so natural the kind of credulity that enables one to form testimony-based beliefs
non-inferentially and without either screening for insincerity or seeking some
premise to support what is attested to. Moreover, the recipient’s ground for tak-
ing the attester to be sincere could be considered to be characteristically more
‘‘direct’’ than it would be if one needed to infer sincerity; there would normally
be no liability to interference by deceitfulness or misspeaking, for instance. Mis-
speaking occurs, of course, but usually it is corrected before the recipient forms
the relevant belief or at least before that becomes in any sense fixed.
We have seen that Reid thinks of belief as an inevitable product of perception
(recall the passage quoted from p. 85 in the Inquiry), and it is plain that he takes
perception as a major and reliable source of true belief. This is clear in connec-
tion with his principle 5, concerning perception; and the same holds for other
major sources of belief as he sees them: he has ‘‘first principles’’ for conscious-
ness and memory, for instance. Thus, even if, as it seems to me, he does not treat
testimony as basic in the way perception, consciousness, and memory are—in
part because we do not in those cases rely on someone else’s authority—he may
think of it as characteristically putting before us not just a verbal expression of
a proposition but the attester’s belief of it. We can sometimes ‘‘see that’’ one
person believes another to be guilty of an offense, ‘‘feel’’ or sense a person’s big-
oted beliefs underlying a proposal, and the like. Granted, we see what a person
believes by hearing what the person says or by some other indication; but we also
see (or otherwise know) persons themselves by seeing (or otherwise perceiving)
their behavior or bodily surfaces. In neither case is inference needed. From the
possibility that appearances can deceive us, we must not infer—and certainly
common-sense philosophers will not infer—that we are never non-inferentially
acquainted with what they manifest.
Even apart from the possibility that belief can be somehow present in testi-
mony, if my belief that p is non-inferentially (and in that sense directly)—and
‘‘reliably’’—produced by your belief that p, is there any reason why my belief
36 Testimony, Credulity, and Veracity

cannot inherit the strength, or at least a good proportion of the strength, of your
own grounds? Those grounds might have to be ultimately at least in part non-
testimonial, but that would not prevent my getting a kind of knowledge that
is both non-inferential and not evidentially dependent on inductive grounds. (I
return to this point in the last section.) My response to your belief is ultimately
a response to your response to your grounds; and if, as seems plausible, the relev-
ant knowledge-sustaining relation of being a response to is transitive, my belief is
also a response to your grounds.¹⁴ Such a chain may be long, either owing to the
number of its essential links or to the length of some of them, say those constituted
by memory connections. But, in relation to transmitting knowledge, if a chain
can be no stronger than its weakest link, it also need not be weaker. (Justification
functions differently here. For one thing, one attester may be only minimally
credible to the recipient in question, who is thereby only minimally justified in
the resulting testimony-based belief that p, while further down the line an attester
is highly credible and gives the last recipient much better justification for p than
that gained by the recipient of the minimally credible testimony that p.)

I V. T E S T I M O N Y A S A N E PI S T E M I C A L LY E S S E N T I A L
S O U RC E

It will be obvious that much of what Reid says about testimony is highly con-
sonant with the sketch of its psychological and epistemological aspects I have
presented on the basis of my own work on the topic. Indeed, I am aware of noth-
ing he says that is clearly inconsistent with my conception of testimony.
It may seem that my view that testimony is not a basic source of knowledge
precludes treating it as an essential one, but this is not so. It need not be basic,
in the sense that it can produce knowledge without the recipient’s relying on
some other source, even if it is epistemically essential, in the sense that what we
think of as ‘‘our knowledge,’’ in an overall sense, would collapse if the contribu-
tion to it made by testimony were eliminated: what remained would be at best
fragmentary. Testimony is in this respect globally essential for human knowledge
(and presumably also for our overall justification, which would also be fragmen-
ted if such elimination occurred). Indeed, I agree with Reid in the view that—in
human life as we know it—one simply could not develop a body of knowledge
at all apart from the instruction one receives in childhood, in which testimony is
central.¹⁵
It is more difficult to see where Reid might stand on the matter of testimony-
based justification. He speaks of reasons for believing testimony as well as of
regard for it; both expressions suggest justification, and it seems quite possible
that he takes testimony to be a source of justified belief much as it is of know-
ledge (though the term ‘justified belief’ is not one he regularly uses). He notes,
however, that most of us cannot give reasons for a ‘‘thousandth part of what
 Robert Audi 37

we are told,’’ and I take it this is meant to indicate that for the vast majority of
propositions we believe, we would also lack what might be called ‘‘independent
evidence’’ (a justification, in one sense). He surely thought that for a huge pro-
portion of these propositions that are objects of our knowledge, testimony, by
contrast with the kinds of inductive grounds Hume believed to be required for
testimony-based knowledge and justification, is a (focally) essential basis of our
knowledge of them as well as of our justification for believing them.
Would Reid think, then, that it is a mistake to maintain that in order to
acquire testimony-based justification one must have some justification either for
taking the attester to be credible or for accepting p? I believe he need not. He
would think that far more often than not we do not have independent suffi-
cient reason for believing p; but that is utterly common in cases where we do
have justification for taking the attester to be credible. Given his emphasis on the
naturalness of veracity and on our utterly pervasive dependence on testimony in
childhood, he at least leaves room for my view that—once we are mature enough
to have justification at all—we do have (prima facie) justification for believing
normal testimony (the kind we typically get from family and friends).
Similarly, there is no reason to think that Reid must deny another point sup-
porting my view that in order to obtain testimony-based justification for p one
needs a measure of justification: apart from perceptual justification for believing
something to the effect that you attested to p, I cannot acquire justification for
believing p on the basis of your testimony. I need perceptually-based justification
for believing you said something and a related (semantic) justification for taking
it to be that p; and again we find a contrast between testimony and an apparently
more basic source. He says, for instance, that
In artificial language [what is now called ‘natural language’] the signs are articulate
sounds, whose connection with the things signified by them, is established by the will of
men; and in learning our mother tongue, we discover this connection by experience . . .
(An Inquiry, p. 91)

He would presumably hold that one could culpably fail to have or exercise
this knowledge of the relevant connection, whereas one would be deficient,
but not necessarily culpable, if one lacked the innate power of perceptual
belief-formation. Misinterpreting the words of one’s language would tend to be
avoidable and blameworthy in a way that being congenitally deaf is not.
In suggesting that Reid’s view of testimony provides for its playing a role in
knowledge (and presumably justification as well) that is less basic than the role
of the standard basic sources, I may appear to undercut the point that testimony
is (for Reid as for me) an essential source of knowledge and indeed of justifica-
tion. But nothing I have said about testimony in comparison with other sources
undermines the point that without testimony we might not even acquire the con-
cepts essential for so much as believing the propositions in question. In this way,
testimony is globally essential in a genetic sense, as well as globally essential in the
38 Testimony, Credulity, and Veracity

epistemic sense already noted: if we were deprived of all the knowledge and jus-
tification we have that arose from testimony, either directly or by inference (and
is now memorially retained), we would be at best reduced to (as Reid put it) a
condition worse than that of savages.
Moreover, everything I have said, and everything I am aware of in Reid’s treat-
ment of testimony, points to another aspect of the epistemological importance
of testimony. It is a source of basic knowledge, i.e., (propositional) knowledge
not grounded in knowledge (or in justified belief ) of some other proposition
(basic knowledge need not be propositional, but my concern here is propositional
rather than objectual knowledge).¹⁶ This is important. It distinguishes testimony
from another non-basic source of knowledge—inference. Inferential knowledge
is premise-dependent as well as source-dependent, in being dependent on a basic
source; testimony-based knowledge has only source-dependence. (Even in the
case of knowledge by virtue of an inferential operation of reason, as with math-
ematical proofs, the conclusion is known or believed on the basis of a premise.
Hence it is not basically known or basically justified, though it may still be a pri-
ori.)
It might seem that testimony cannot be a source of basic knowledge if I am
right in thinking that although one need not believe that the attester is giving
testimony that p, one must have grounds adequate for knowing or justifiedly
believing this. But there is no inconsistency here: having grounds for believing
this does not entail believing it. Even if it did, the belief in question would not
necessarily be needed as a basis for believing the testimony, nor are the grounds a
basis of the belief. Surely we do not normally form such a belief. We may in some
sense presuppose its truth, but that is a quite different matter.
The point that testimony is a source of basic knowledge helps to explain why
it is natural to hold the stronger view that testimony is a basic source of know-
ledge. For it is typical of such sources that they yield non-inferential knowledge,
and it is necessary for a source’s being basic at all that it do this with respect to the
kind of knowledge for which it is a basic source.¹⁷ Moreover, testimony is one of
the distinctive and (in my view) irreducible ways in which we acquire knowledge.
Still, a distinctive and irreducible mode of knowledge acquisition may nonethe-
less be both epistemically and operationally dependent on some other source of
knowledge.

V. T E S T I M O N Y I N C O M PA R I S O N W I T H OT H E R
S O U RC E S O F K N OW L E D G E A N D J U S T I F I C AT I O N

There are at least five further points that distinguish testimony from what I shall
call the standard basic sources: perception, introspection, memory, and reason
(or reflection), with intuition, in one sense, as a special case of the operation of
reason. I take these four basic sources in turn.
 Robert Audi 39

First, we cannot test the reliability of one of these basic sources or even con-
firm a deliverance of it without relying on that very source.¹⁸ I refer to externally
testing for reliability, roughly for reliability in yielding true beliefs in the appro-
priate domain given internal consistency. A source with inconsistent outputs is
not even a candidate for external reliability.¹⁹ One can of course internally test
for reliability by ascertaining whether there is inconsistency or probabilistic inco-
herence among the cognitive outputs of the source (though such testing would
be only negative —aimed at ruling out unreliability, since consistency and indeed
even coherence, would not by themselves imply truth²⁰). With perception, for
instance, quite apart from any question of its overall reliability (which is not
my concern here) one must, in a given case of mistrust, look again (or otherwise
rely on perception). With memory, in order to overcome mistrust of a particular
deliverance, one must try harder to recall or must consult other memories—and
one must remember the original belief being examined, lest the target of confirm-
ation be lost from view. With testimony, one can, in principle, check reliability
using any of the standard basic sources.
The qualification ‘in principle’ is important. Even apart from the point that
in normal human life one cannot acquire concepts without reliance on testi-
mony, there may be certain facets of its reliability that we are in no position to
check without dependence on its deliverances. Consider, for instance, testimony
about technical matters on which, apart from relying on other testimony by col-
leagues or teachers, even experts do not know all one needs to know in order
to understand these matters. It remains true, however, that important aspects of
testimonial reliability can be checked without counting on specific knowledge
grounded essentially in testimony, and the counterpart of this point does not
seem to hold for the basic sources.
The second point about the standard basic sources in comparison with testi-
mony concerns memory in particular and has already been suggested. Memory
is central, in a way testimony is not, for both our retaining and our extending of
knowledge at any given moment—at least if extension of knowledge, as where
inference adds to what we know, is conceived as adding to knowledge we retain,
as opposed to simply increasing the number of propositions we know. We could
speak here of additive extension to distinguish this (standard) case from mere
quantitative extension (‘increase’ might be a better term), the case in which one
simply gets new knowledge that increases the number of propositions one knows.
That could occur where our brains are so affected that, without our memories’
playing any role in the process, new knowledge of wider scope than our present
knowledge (or of a larger number of propositions) supplants the knowledge we
now have. But given the power of memory, even if knowledge could not be
acquired at all without at least the amount and kind of testimony needed to
learn a language (a process in which what parents or others attest to is crucial
for acquiring a vocabulary), once we climb that linguistic ladder, we can discard it
and still retain what we know.
40 Testimony, Credulity, and Veracity

As to significantly extending our knowledge, if we could not remember our

premises after inferring our conclusion, then even if rehearsing them had sufficed
for causing us to believe it, we would either not come to know it at all, or might
be unjustified in (or at least unable to justify) our believing it.
The epistemic role of memory can be better understood in relation to the oth-
er standard basic sources. These differ from memory in a way that adds to their
contrast with testimony. They are basic with respect to knowledge as well as jus-
tification, whereas memory is basic only with respect to justification—assuming
that what we know from memory we must have come to know on the basis of
some other source. Reason in some minimal form is indispensable to possessing
any knowledge (at least in protecting us from pervasive inconsistency), and cer-
tainly to inferential extension of our knowledge, which depends on our being at
least implicitly guided by deductive or inductive logic. Consciousness and per-
ception are essential for development of new knowledge in their domains. There
is, however, no domain (except possibly that of other minds) for which contin-
ued testimony is in principle needed for significant increase of knowledge. Similar
(but not entirely parallel) points hold for justification.
The third point is of a rather different sort. There is a sense in which testimony-
based belief passes through the will—or at least through agency. There are at
least two cases here. Consider the side of the attester. The attester must in some
sense, though not necessarily by conscious choice, select what to attest to, and
in doing so can also lie or, in a certain way, mislead, and in these cases the
testimony-based belief does not constitute knowledge²¹ (and the justification
the recipient may get is, in a certain way we need not pursue here, objectively
deficient in a sense that goes with the notion of misleading evidence). For the
basic sources, there is no comparable analogue of such voluntary representation
of information. Now consider the side of the recipient. The recipient commonly
can withhold belief, if not at will then indirectly, by taking on a highly cautionary
frame of mind (I am taking withholding to be roughly a kind of blocking of belief
formation when a proposition is presented).
In many cases, then, testimony-based beliefs commonly pass through agency
twice over. This passage does not entail an exercise of agency. The point is that
testimony is subject to its power. The attester can falsify it, and the recipient
commonly has both a monitoring system and the voluntary capacity to withhold
belief of a huge range of propositions even if not of just any belief. To be sure,
the contrast with the basic sources can be exaggerated. But we commonly can,
and sometimes do, withhold belief from attested propositions in a way we cannot
withhold belief from propositions strongly supported directly by experience or by
reason.
What, then, of a case of double epistemic support, as where someone attests
to a plainly self-evident proposition one had not thought of but intuitively sees
to be true directly upon hearing it asserted? What is attested may be irresist-
ibly obvious. Here withholding the proposition is not possible. It is usually not
 Robert Audi 41

possible where we already believe what we hear—though some people’s attesting

to certain propositions could even give us reason to disbelieve them! In that case,
withholding may result.
I would also stress that the common possibility of withholding belief of testi-
mony does not imply that in each case we must consider whether to believe
the testimony, though we often do consider this. The point is that, commonly,
whether we withhold belief is ‘‘up to us.’’ For many cases, we may also have
more voluntary control over how we interpret testimony than we have regard-
ing normal deliverances of the standard basic sources. The unreasonable exercise
of this control could lead to our forming a testimony-based belief that is either
not justified or does not constitute knowledge. In our voluntary control over our
interpretation of testimony, as perhaps concerning withholdability of believing
it, I grant that the contrast between testimony and standard basic sources is prob-
ably more one of degree than of kind.
The conditions under which a person can withhold belief are a contingent
matter. Some people may be able to learn to withhold—even if not to
reject—the natural belief that those speaking to them are people as opposed to
robots.²² Moreover, the possibility of withholding belief of what is plausibly and
truly attested to would not show that testimony is not a basic source, but only
that it is not an irresistible one. Still, it is significant that the normal level of
control for the standard sources is different from that applicable to testimony.
Reid seems to have seen something like this. For one thing, he considered
both receptivity to testimony and our use of induction to be ‘‘regulated’’ as
we gain experience. He also contrasted testimony with memory and, especially,
perception, on this score, as in the passage quoted above comparing memory and
perception.
In addition, with respect to testimony, appraisal of credibility may always
involve both the kinds of doubts we may have about basic sources and any we may
have about the attester’s response to them. To be sure, we sometimes speak of
the ‘‘testimony of the senses.’’ But this is metaphor, at least insofar as it suggests
that the senses derive knowledge from another source, as attesters must ultimately
do, since knowledge that p cannot derive from an infinite or circular testimonial
chain in which no person giving testimony that p knows it even in part on a non-
testimonial basis.²³
One might think that the contrast between testimony and the standard basic
sources as regards the will is important in part because testimony is less reliable
than those other sources of justification and knowledge. Whether this is so is
a contingent matter. In any case, the point would have limited significance. It
would at best indicate a difference of degree rather than of kind and might give
misleading support to the idea that, given adequate reliability, testimony, or some
kind of testimony, is a basic source.²⁴ As the case of valid inference shows, how-
ever, the conditional reliability of a source can be readily accounted for without
taking it to be basic. Moreover, for justification as opposed to knowledge, I do
42 Testimony, Credulity, and Veracity

not see that actual reliability is the crucial ground (as I have argued elsewhere²⁵),
but that issue is too large to pursue here.
A fourth point of contrast between testimony and the standard basic sources
has already been suggested. It concerns the need for grounds for the semantic
construal of what is said on the basis of which it is taken to be that p. This is not a
justificatory or epistemic burden intrinsic to the standard basic sources. In Reidi-
an terms, we might say that testimony-based knowledge and justification must
go through convention, since they depend on our understanding an ‘‘artificial’’
language; knowledge and justification grounded in the standard basic sources
need only go through nature. The point is not that no kind of interpretation is
ever needed in order for (say) perception to yield knowledge or justification. It is
that testimony-based knowledge that p arises only when we have both perceptual
grounds for believing that the attester has said something and semantic grounds
for believing that it is the proposition that p which is attested to; and testimony-
based justification that p arises only if we have the corresponding perceptual
and semantic justification. This point illustrates both the operational dependence
of testimony on perception and the epistemic dependence of testimony-based
beliefs on semantic considerations.
We might speculate that with Reidian natural language, such as the ‘‘threat-
ening or angry tone of voice’’ by which ‘‘Children, almost as soon as born, may
be frightened’’ (Essays, Essay 6, ch. 5), what makes the language natural is that
something about its oral expression is a criterion, in a Wittgensteinian sense,
of some element it conveys. Here, however, we have knowledge from testimony
rather than testimony-based knowledge; for just as we can know that people have a
cold from their nasal tone regardless of what they attest to, a child can know that
Father is angry quite apart from what he is saying.²⁶
Granted, much a priori knowledge and justification is acquired through con-
sideration of linguistic expressions of propositions. But perception of such an
expression is not required, nor is there a comparable need (if any) for semant-
ic interpretation. On the most plausible account of the nature and basis of such
knowledge and justification, its object is in principle accessible without reliance
on semantic construal; the ground of justification or knowledge is in any case
a kind of understanding of the proposition in question or, perhaps more dir-
ectly, the concepts figuring in or essential to it. To be sure, one might think
that testimony could somehow convey propositional content directly to anoth-
er mind without using any semantic ‘‘vehicle.’’ If this is possible, it would surely
require at least a symbolic vehicle, such as an image produced in my mind that
I can see to come from some other mind. If someone simply causes me to have
an image that produces a belief that p or, especially, simply causes me, in a dir-
ect way, to believe that p, then there is no testimony-based belief, but only some
kind of cognitive product of action by someone else. Even where testimony is not
verbal, then, if it retains its essential character, the fourth point of contrast with
the basic sources seems unlikely to disappear. The contrast might be thought to
 Robert Audi 43

be somewhat attenuated, but this speculative possibility is not what is normally in

question in the epistemology of testimony, and it would in any case not affect the
first three points of contrast.
It should also be granted that a lack of semantic understanding will normally
restrict the range propositions that are even candidates for one’s a priori know-
ledge or justification, since one’s comprehension of language will (for most of
us, at least) limit the range of propositions we can get before our minds. Seman-
tic misunderstanding —which is of course possible even in people of wide and
deep semantic comprehension—may give us the wrong proposition or range
of propositions. Nonetheless, neither of these defects need affect how good our
grounds are once the right object is before us. To be sure, defeaters of know-
ledge or justification can come from semantically interpreted items and can afflict
beliefs deriving from any of the standard sources; but none of those sources seems
dependent on semantic grounds in the way that testimony is.
The fifth contrast I want to draw is of a quite different sort. It presupposes that
testimony yields testimony-based knowledge in the recipient only if the attest-
er knows that p. This point is controversial (in part owing to the kind of case
illustrated by Luke, discussed in Section II), but the exceptions to it are at best
rare and probably of special kinds that can be described in a way that enables us
to set them to one side. In any case, given this presupposition, we can say that
although testimony can increase the number of knowers in the world, it cannot
increase the number of propositions known. Thus, in one sense it cannot increase
the amount of knowledge in the world. This is an important limitation. It would
be a mistake, however, to think that it detracts from the importance of testimony.
Far from it: the indefinitely extensive testimonial communicability of knowledge
is an incalculably valuable characteristic of testimony, and it is not paralleled
by any of the standard basic sources. Moreover, in this limitation testimony is
like memory: it, too, cannot (by itself) increase the amount of knowledge in the
world.
Testimony-based justification works differently here. Testimony can both
increase the number of justified beliefs in the world and contribute to the
degree of their justification. It can thus add to the amount of justification in
the world. This would not be so unless it could produce new justification. It
can justify people not already justified in holding the relevant propositions and
can enhance the justification of those who are already so justified. Indeed, as
suggested already, there is much that we believe which we would not believe at
all, much less justifiedly believe, apart from our reliance on testimony. Again,
the comparison with memory is significant. For whereas concurring testimony
from many people can increase the justification of a single belief, concurring
memory impressions that p on the part of different people cannot—in the same
way—increase the justification of a single belief. They can do so only indirectly,
say by yielding testimony based on the several people’s memories; but this is quite
different.
44 Testimony, Credulity, and Veracity

These contrasts between testimony and the basic sources are not meant to
impugn the importance of testimony. In addition to being a source of basic
knowledge and an essential source of our overall knowledge, it is apparently an
essential source of much of our justified belief as well. Our overall knowledge of
the world, including even many things we know about ourselves, depends on it
in far-reaching ways, though not perhaps as much as, and certainly not in quite
the same ways as, it depends on memory. The most important thing memory
and testimony have in common may be that they do not generate, as opposed to
(respectively) preserving and transmitting, knowledge (the case with justification
is different, since memory, unlike testimony, is a basic source of that).
In one way, however, testimony is a more far-reaching source than memory:
although they share unlimited epistemic breadth, in the sense that in principle
anything that can be known can be known at least in part on the basis of them,
we can learn from testimony in a way we cannot learn, as opposed to recov-
ering, from memory.²⁷ As to how testimony differs from both perception and
memory, there is far more to say than can be said here. I have already stressed that
the important differences are not a matter of reliability, either of the process by
which these various cognitive sources produce belief or of their cognitive outputs.
No matter how reliable testimony in fact is in either respect, the acquisition of
knowledge or even of justified belief on the basis of testimony depends (as noted
earlier) on the agency of another person. The attester must not lie, or (in certain
ways) seek to deceive, in attesting to p if we are to come to know that p on the
basis of the testimony. By contrast, our responses to the deliverances of the basic
sources are not normally mediated by anyone else’s action. Testimony may be
unreliable—or fail to justify one’s believing that p—both because of natural con-
nections between the state(s) of affairs the testimony concerns and because of the
person’s exercise of agency. This is not normally so for the testimony of the senses
or of memory or of reason. The point is not that the exercise of agency cannot be
a purely ‘‘natural’’ phenomenon—though philosophers who think that freedom
is incompatible with determinism may argue that it cannot—but that the con-
cepts of knowledge and justification apparently presuppose that if it is a natural
phenomenon, it is nonetheless special.²⁸
Whether Reid could accept all these conclusions is not clear to me; but he
does treat testimony differently from the standard basic sources, and I see no
bar to his granting the specific differences I have noted. In any case, it should
be plain that the contrasts I have drawn between, on the one hand, testimony-
based knowledge and testimony-based justification and, on the other, knowledge
and justification deriving from the basic sources, do not imply that testimony is
inessential in a normal human life or that its authority in cognitive matters is
only contingent and empirical. Testimony is both globally and focally essential
in our lives. It is indeed of virtually unlimited breadth in its epistemic power:
virtually anything that can be known firsthand can also be known on the basis of
testimony. Testimony is a source of basic knowledge; it is normally the starting
 Robert Audi 45

point of everyone’s conceptual learning; for all I have said, it may even have a
measure of a priori justificatory authority;²⁹ and it is, as Reid so vividly shows us,
as natural a source of knowledge and justification as any of the others.

REFERENCES

Adler, Jonathan E. (1994), ‘Testimony, Trust, Knowing’, Journal of Philosophy, 91.

(2002), Belief ’s Own Ethics (Cambridge, Mass.: MIT Press).
Alston, William P. (1985), ‘Thomas Reid on Epistemic Principles’, History of Philosophy
Quarterly, 2/4: 435–52.
Audi, Robert (1993a), ‘The Foundationalism-Coherentism Controversy: Hardened Ste-
reotypes and Overlapping Theories’, in Audi (1993b: 117–64).
(1993b), The Structure of Justification (Cambridge and New York: Cambridge Uni-
versity Press).
(1997), ‘The Place of Testimony in the Fabric of Knowledge and Justification’,
American Philosophical Quarterly, 34/4: 404–22.
(1999a), ‘Doxastic Voluntarism and the Ethics of Belief ’, Facta Philosophica, 1:
87–102.
(1999b), ‘Self-Evidence’, Philosophical Perspectives, 13: 205–28.
(2001), The Architecture of Reason (Oxford and New York: Oxford University
Press).
(2003), ‘An Internalist Theory of Normative Grounds’, Philosophical Topics,
29/1–2: 19–46.
(2004), ‘The A Priori Authority of Testimony’, Philosophical Issues, 14: 18–34.
Beanblossom, Ronald, and Lehrer, Keith (eds.) (1983), Thomas Reid’s Inquiry and Essays
(Indianapolis: Hackett).
Coady, C. A. J. (1992), Testimony (Oxford: Clarendon Press).
Corlett, J. Angelo (1996), Analyzing Social Knowledge (Lanham, Md.: Rowman and
Littlefield).
Dretske, Fred (1981), Knowledge and the Flow of Information (Cambridge, Mass.: MIT
Press).
Foley, Richard (2001), Intellectual Trust in Oneself and Others (Cambridge: Cambridge
University Press).
Fricker, Elizabeth (1987), ‘The Epistemology of Testimony’, Proceedings of the Aristoteli-
an Society, suppl. vol. 61.
(1994), ‘Against Gullibility’, in B. K. Matilal and A. Chakrabarti (eds.), Knowing
from Words (Dordrecht: Kluwer), 125–61.
(2003), ‘Trusting Others in the Sciences: A Priori or Empirical Warrant’, Studies in
History and Philosophy of Science, 33.
(2004), ‘Testimony: Knowing Through Being Told’, in I. Niinilouto, M. Sin-
tonen, and J. Wolenski (eds.), Handbook of Epistemology (Dordrecht: Kluwer).
Goldman, Alvin I. (1999), Knowledge and the Social World (Oxford and New York:
Oxford University Press).
Graham, Peter J. (2000a), ‘Conveying Information’, Synthese, 123: 365–92.
(2000b), ‘The Reliability of Testimony’, Philosophy and Phenomenological Research.
46 Testimony, Credulity, and Veracity

Graham, Peter J. (2000c), ‘Transferring Knowledge’, Noûs, 34.

Hardwig, John (1985), ‘Epistemic Dependence’, Journal of Philosophy, 82/7: 355–49.
Lackey, Jennifer (1999), ‘Testimonial Knowledge and Transmission’, Philosophical
Quarterly, 49: 471–90.
Matilal, B. K. and Chakrabarti, A. (eds.) (1994), Knowing from Words (Dordrecht:
Kluwer).
Plantinga, Alvin (1993), Warrant and Proper Function (Oxford: Oxford University Press).
Schmitt, Frederick (ed.) (1994), Socializing Epistemology: The Social Dimensions of Know-
ledge (Lanham, Md.: Rowman and Littlefield).
Sosa, Ernest (1991a), Knowledge in Perspective (Cambridge and New York: Cambridge
University Press).
(1991b), ‘Testimony and Coherence’, in Sosa (1991a).
(forthcoming), Knowledge in Focus, Skepticism Resolved.
Van Cleve, James (1999), ‘Reid on the First Principles of Contingent Truths’, Reid Stud-
ies, 3/1: 3–30.
Webb, Mark Owen (1993), ‘Why I Know about As Much as You’, Journal of Philosophy,
90.
Wolterstorff, Nicholas (2001), Thomas Reid and the Story of Epistemology (Cambridge:
Cambridge University Press).

N OT E S

1. See, e.g., Schmitt (1994); Goldman (1999); and Corlett (1996).

2. Among the recent works on or treating testimony are Coady (1992); Matilal and
Chakrabarti (1994); Fricker (1987); Hardwig, (1985); Sosa (1991a); Plantinga
(1993); Webb (1993) in part a critique of Hardwig (1985); and Adler (1994) in
part a critique of Webb (1993); Lackey (1999); and Graham (2000a, 2000b, and
2000c); Foley (2001); and Adler (2002).
3. Much space could be devoted to refining this idea, e.g. explaining why fictional cases
do not count and ironic declarations can apparently tell us one thing when another
is intended (in which case ‘testimony’ may be the wrong word for either reading, but
certainly for the former).
4. In, e.g., ‘‘The Foundationalism-Coherentism Controversy,’’ ch 4 in Audi (1993b).
5. It might be argued, however, that apart from a certain level of critical reflection on
the grounds of our beliefs, we cannot arrive at something plausibly called reflective
knowledge. This concept, and the requirements it imposes, have been explicated in
detail by Ernest Sosa. See e.g., Sosa (1991b, and forthcoming).
6. For valuable recent discussion, see Lackey (1999) and Graham (2000a). In one
example he brings against the requirement that the attester know that p if the recipi-
ent is to know it (fully) on the basis of testimony, a woman who cannot distinguish
two twins, Judy and Trudy, says, on the basis of seeing her do it, that Judy broke a
vase. She is right, but does not know, since she would have taken Trudy to be Judy
if Trudy had been in the same position. Nonetheless, Bill, who sees Trudy before
him and knows she is Trudy, but does not witness the vase-breaking, knows, from
the testimony given just after the event, that it was indeed Judy. Graham grants that
 Robert Audi 47

Bill makes use of his background knowledge. On my view, however, Bill also does
not have a testimony-based belief: rather, on the basis of what he is told and his
knowledge that Trudy did not break the vase, he believes something to the effect that
the person the attester knows to have broken it and takes to be Judy on the basis
of appearing to be Judy, is indeed Judy. He knows Judy broke it by knowing what
grounded the testimony, not on the basis of the testimony.
7. The case is from Lackey (1999). She presents others, but some of what I believe
should be said about those may be implicit in this paper; further discussion must
await another occasion.
8. I have emphasized the importance of sincerity for the epistemology of testimony
in Audi (1997). The notion deserves more attention than it can be given here, but
I take it as clear that a kind of insincerity is implied in telling someone that p when
one believes it false and expects it to be believed by the recipient. If one simply does
not believe p the case is less clear, but insincerity of a kind still seems implied. Peter
Graham has noted that accepting p, which does not entail believing it, is a special
case here. It is and deserves discussion not possible in this paper.
9. It may be that the sense is very broad indeed, since I might perhaps receive testimony
by a telepathic reception of the attester’s message. Perhaps the relation between testi-
mony and perception is such that as our inclination to treat reception of stated
information as non-perceptual increases, our inclination to regard it as testimonial
decreases.
10. See Thomas Reid, in Beanblossom and Lehrer (1983: 87). Later references to Reid
in this paper are to this edition.
11. Reid discusses acquired perceptions in Essay II, ch. 21 of the Essays, as well as in
ch. 22. For discussion of his conception of such perception in relation to testimony,
see Nicholas Wolterstorff ’s chapter on Reid’s epistemological views on testimony
in Wolterstorff (2001).
12. Justification by reasons does not, of course, exhaust justification by reason; and Reid
may take it that the worthiness in question is self-evident. For extensive critical dis-
cussion of how he conceived his first principles see Alston (1985: 435–52), and Van
Cleve (1999: 3–30), which explores the intriguing thesis that Reid’s main first prin-
ciples are not general epistemic principles such as those quoted, which are supposed
to be self-evident indications of what conditions generate true belief (or knowledge
or perhaps justification and knowledge) but the beliefs (of particular propositions)
licensed by those propositions, say the belief that I just had Van Cleve’s paper in my
hands, grounded in my memory.
13. That the self-evident can be evidenced by something else, even though it does not
stand in need of evidence, I have argued in some detail in Audi (1999b). Here,
however, I depart from what Reid says in at least one place. At one point he speaks
of things which, ‘‘though they admit of illustration yet, being self-evident, do not
admit of proof ’’ (Essays on the Intellectual Powers, p. 153).
14. In Audi (2001), I stress this element of responsiveness to experience in explicating
the nature of belief and treat its formation as a discriminative response because of the
way in which it is ‘‘selective’’ and alters with alteration in the relevant stimuli. Cf.
Dretske’s (1981) information-theoretical account of knowledge and its application
to testimony in the papers by Peter Graham cited above.
48 Testimony, Credulity, and Veracity

15. In Audi (1997), I have described conceptual learning in relation to propositional

learning and emphasized the importance of both in the development of justification
and knowledge.
16. For a detailed account of the epistemology testimony that is in some ways inferen-
tialist and in any event presupposes the formation of more beliefs on the part of the
recipient as a condition for acquiring testimony-based knowledge, see Fricker, esp.
(1994) and for a more recent rounded statement of her position (2004).
17. It may seem that a basic source never yields inferential knowledge; but that is per-
haps not altogether clear. If we take reason as a basic source of knowledge and
consider its application to judging a poem, its intuitive operation might produce
non-inferential knowledge of such things as emotional tones in certain lines and its
inferential operation might (indirectly) produce, on the basis of this knowledge and
knowledge of criteria of interpretation, inferential knowledge constituting an inter-
pretation of the poem. It may be, however, that the relevant intuitive knowledge
should be considered quasi-perceptual—we do, after all, have aesthetic perception
here—and that the case is better conceived as one in which rational perceptions
serve as premises for reasoning, so that we have a combination of sources yielding
inferential knowledge.
18. I do not take this feature of untestability for reliability without dependence on the
source in question to be essential to a basic source; but it does seem to hold for the
basic sources I am comparing with testimony.
19. If a source has inconsistent outputs, how can it be a candidate for reliability at all?
Perhaps it can if the inconsistencies cause no trouble, e.g. because the subject recog-
nizes them and can isolate them for scrutiny and elimination. It is not at all easy
to say under what conditions inconsistencies are innocuous; this is one reason the
preface paradox is difficult to resolve.
20. I argue for this limitation on coherence in Audi (1993a). It is true that coherence
among the belief outputs of a source implies the truth that if the person believes
these outputs to be mutually coherent, the belief is true. But this is trivial; the ques-
tion is reliability of a general kind.
21. If we allow that a testimony-based belief may also be based on another source than
the testimony in question—including other testimony—then lying testimony may
not preclude knowing the proposition. I say ‘may not’ because it is not self-evident
that a person is lying when attesting to a true proposition but believing the prop-
osition false and intending to deceive the recipient in attesting to it. Unless other-
wise indicated, however, I am making the natural assumption that calling a belief
testimony-based implies that it is essentially based on the testimony in question,
where this requires at least that the testimony is a necessary basis of it.
22. I discuss the issue of voluntary control of belief and cite much relevant literature
in Audi (1999a: 87–102).
23. This point is explained and defended in Audi (1997).
24. Support for this point is provided in Peter Graham’s informative paper (Chapter 4
in this volume), in which he argues that the point that testimony (as I would put it)
goes through the will does not favor either the reductionist view of testimony or the
view that it is an irreducibly credible source.
 Robert Audi 49

25. For a detailed defense of a moderate internalism concerning justification, see Audi
(2003: 19–46).
26. It is noteworthy that the most relevant first principle is, ‘‘That certain features of the
countenance, sounds of the voice, and gestures of the body, indicate certain thoughts
and dispositions of the mind’’ (Essays, Essay VI, ch. 5). These natural signs do not
even require making any assertions.
27. There are subtleties here that I cannot go into, but the need for ‘at least in part’
should be explained: for simple self-evident propositions it may be possible to know
them only in part on the basis of testimony because an understanding of them suffi-
cient to come to know them on the basis of testimony at all is such that given even
the minimal kind of entertaining of them required for receiving testimony that they
are true, one must know them on the basis of that understanding. Some detailed
discussion of knowledge on the basis of understanding is given in Audi (1999b).
28. This point may support my view, defended in Audi (1997), that to acquire justi-
fication for p from testimony, one needs some degree of justification for taking the
attester to be credible, though one can of course get it on the occasion of the testi-
mony. (I do not think one needs a counterpart of this in order to acquire prima facie
justification from one of the standard basic sources.)
29. The prospects for testimony’s having a priori epistemic authority, and the difference
between that and a kind of practical authority, are discussed in detail in Audi (2004).
 2
Reid on the Credit of Human Testimony
James Van Cleve

Thomas Reid is perhaps the first philosopher to call attention to ‘‘the analogy
between perception, and the credit we give to human testimony’’—the topic of
chapter 6, section 24 of his Inquiry into the Human Mind (hereinafter abbrevi-
ated as IHM). In this essay, I explore the extent of Reid’s analogy. I begin by
trying to arrive at a proper understanding of the two principles he identifies as
fundamental to our acquiring knowledge from the information of others—the
principles of veracity and credulity. Next, I investigate the similarities Reid finds
between perception and testimony considered as mechanisms of belief formation.
Finally, I consider whether the analogy between perception and testimony can
be extended from psychology into epistemology. In particular, I discuss whether
beliefs based on testimony, no less than beliefs based on sense perception, may be
regarded as epistemically basic or foundational. This is the chief issue that divides
Reid from Hume in the epistemology of testimony.

I . T H E P R I N C I P L E S O F V E R AC I T Y A N D C R E D U L I T Y

Reid introduces the two key principles of his theory of testimony in the Inquiry in
the following passage:
The wise and beneficent Author of Nature, who intended that we should be social
creatures, and that we should receive the greatest and most important part of our know-
ledge by the information of others, hath, for these purposes, implanted in our natures
two principles that tally with each other.
The first of these principles is, a propensity to speak truth, and to use the signs of
language, so as to convey our real sentiments. (IHM, p. 193)
Another original principle implanted in us by the Supreme Being, is a disposition to
confide in the veracity of others, and to believe what they tell us. This is the counter-part
to the former; and as that may be called the principle of veracity, we shall, for want of a
more proper name, call this the principle of credulity. (IHM, p. 194)

Thanks to Albert Chan and Gideon Yaffe for helpful discussion of the issues in this paper.
 James Van Cleve 51

These principles are the key elements in his account of how the words of others
come to be signs conveying to us things we would not have come to know on
our own.
Reid’s formulation of the principle of veracity immediately prompts a vital
question. By ‘‘a propensity to speak truth,’’ does Reid mean a propensity to speak
what is in fact the truth? Or does he mean a propensity to speak what the speaker
believes to be the truth? His unqualified use of the expression ‘‘to speak truth’’ in
the first clause suggests the former, but his use of ‘‘to convey our real sentiments’’
in the second clause suggests the latter.
It is possible that Reid’s ‘and’ connecting the clauses is an ‘and’ of genuine
conjunction, in which case he would mean both things. But it is also possible that
his ‘and’ is an ‘and’ of explication, in which case by ‘speaking the truth’ Reid
would simply mean speaking what you believe to be true.
Our question is whether Reid’s principle of veracity should be understood as
affirming the first, the second, or both of the following:
V1 (It tends to be the case that) if A says p, p is true.
V2 (It tends to be the case that) if A says p, A believes p.
(I use ‘say’ in a sense that includes writing as well as speaking, as was surely
Reid’s intent.)
As we read the ensuing paragraphs in which the principle is developed and
defended, it becomes fairly clear that V2 is what Reid intends. Immediately after
formulating the principle, he goes on to contrast speaking the truth with lying:
This principle has a powerful operation, even in the greatest liars; for where they lie once,
they speak truth a hundred times. Truth is always uppermost, and is the natural issue of
the mind. It requires no art or training, no inducement or temptation, but only that we
yield to a natural impulse. Lying, on the contrary, is doing violence to our nature; and is
never practised, even by the worst men, without some temptation. (IHM, p. 193)
You are not a liar just because you say something false. Lying is saying what you
believe false in an effort to deceive another, and that strongly suggests that the
principle of veracity should be understood as V2: we tend to assert to others only
what we believe to be true.
Further confirmation comes in the next paragraph. Reid defends his view that
the tendency to speak truth is innate by arguing that moral and political consider-
ations are insufficient to account for it. The tendency is present even in young
children before such considerations can have any influence upon them. Well,
moral and political considerations would not even be a candidate explanation for
speaking the truth in the sense of getting things right; they could at best induce
us to speak what we believe to be correct. So the fact that Reid sees the moral-
political explanation as a rival to his own shows again that V2 is the principle he
has in mind.
To clinch the point, I note these two comments that Reid makes on his
principle:
52 Reid on the Credit of Human Testimony

By this instinct, a real connection is formed between our words and our thoughts.
(IHM, p. 194)
If there were not a principle of veracity in the human mind, men’s words would not be
signs of their thoughts. (IHM, p. 197)
If what the principle of veracity brings about is a real connection or sign relation
between our words and our thoughts, the principle must be understood as V2.
Let us turn now to the companion principle, the principle of credulity. Here is
Reid’s formulation of it again: ‘‘Another original principle implanted in us by the
Supreme Being, is a disposition to confide in the veracity of others, and to believe
what they tell us’’ (IHM, p. 194).
As we did with the first principle, we may distinguish two possible versions of
this principle. Letting A be the speaker and B the hearer, we have
C1. (It tends to be the case that) if A says p, B believes p.
C2. (It tends to be the case that) if A says p, B believes that A believes p.
I let it go unstated in the antecedents of C1 and C2 that A’s saying p is direc-
ted at B and that B hears his words. Which version of the principle is intended
this time?
The evidence quickly mounts that C1 is the correct reading. In the very formu-
lation of the principle, Reid speaks of a disposition to believe what others tell us.
What they tell us is typically some fact about the wider world (that the fish are
biting today or that the road through the pass is blocked), not merely an autobi-
ographical fact to the effect that they believe this or that. A few lines later (IHM,
p. 194, ll. 29–30), he says that this principle concerns ‘‘proposition[s] that [are]
uttered in discourse,’’ giving us further occasion to make the same point. Finally,
and most tellingly, Reid observes that the principle of credulity confers enormous
practical benefit on those who are regulated by it:
It is evident, that, in the matter of testimony, the balance of human judgment is by
nature inclined to the side of belief; and turns to that side of itself, when there is nothing
put into the opposite scale. If it was not so, no proposition that is uttered in discourse
would be believed, until it was examined and tried by reason; and most men would
be unable to find reason for believing the thousandth part of what is told them. Such
distrust and incredulity would deprive us of the greatest benefits of society, and place us
in a worse condition than that of savages. (IHM, p. 194)
The pragmatic advantage of credulity flows from believing what my informants
report—for example, that the path to water lies in this direction. It does not flow
simply from believing that they believe it themselves.¹
Our preliminary finding, then, is that the principle of veracity should be
understood as V2 and the principle of credulity as C1. But that finding imme-
diately poses a problem. Reid tells us that the two principles are meant to tally
with each other (IHM, p. 193). What does he mean by that? A plausible guess is
that the principles are supposed to combine with each other to imply that when
 James Van Cleve 53

an informant says p, we tend to form true beliefs. But the combinations that yield
that result are V1 with C1 and V2 with C2—not Reid’s mixed combination of
V2 with C1. So what is going on?
We can get at the problem in another way. One of Reid’s objectives in his
discussion of testimony in Inquiry, 6. 24 is to show how men’s words come to
be signs from which we gain knowledge of what they signify. He has explained
earlier in 6. 21 that there are two requisites of knowledge from signs:

But there are two things necessary to our knowing things by means of signs. First, That
a real connection between the sign and the thing signified be established, either by the
course of nature, or by the will and appointment of men. When they are connected by
the course of nature, it is a natural sign; when by human appointment, it is an artificial
sign. Thus, smoke is a natural sign of fire; certain features are natural signs of anger: but
our words, whether expressed by articulate sounds or by writing, are artificial signs of our
thoughts and purposes.
Another requisite to our knowing things by signs is, that the appearance of the sign to
the mind, be followed by the conception and belief of the thing signified. (IHM, p. 177)

In short, for X to be a sign on the basis of which we have knowledge of Y, (i) X

must be a reliable indicator of Y, and (ii) the apprehension of X must produce in
the mind of the subject a belief in Y. It seems clear that Reid intends the prin-
ciples of veracity and credulity to ensure that these two requisites are satisfied
in the case of human testimony: veracity brings about the reliable connection
between sign and thing signified, and credulity produces our belief in the thing
signified. But if veracity and credulity are to play these roles, veracity will have
to be understood as V1 and credulity as C1 or veracity as V2 and credulity as
C2. With Reid’s V2–C1 mix, the principles do not tally with each other in the
required way.
What are we to do about this? The best thing to say in Reid’s defense, I think,
is that he probably believes all four of the principles we have distinguished. As for
credulity, when we hear someone say p, we normally believe p (as C1 says), but
we also take for granted that the speaker believes p (as C2 says). As for veracity,
we may note that an overarching principle of Reid’s epistemology combines with
V2 to yield V1. The principle I have in mind is Principle 7 in Reid’s list in the
Essays on the Intellectual Powers of Man (EIP) of the first principles of contingent
truths—namely, the principle that our natural faculties are not fallacious (EIP,
p. 480). By this principle Reid means to assert that the things we believe as deliv-
erances of our natural faculties (perception, memory, inference, and so on) tend
to be true. If we add this to V2, we arrive at something very close to V1 by means
of the following argument: what people say, they normally believe (V2); what
they believe is normally true (Principle 7); therefore, what people say is normally
true (V1).²
To conclude this section: although there is an apparent disconnect in Reid’s
exposition of the content and functioning of his two principles, it is correctable
54 Reid on the Credit of Human Testimony

given other materials he puts at our disposal. There is a good Reidian case to be
made for all four of the principles we have discussed, as a result of which there are
two pairs of principles that tally with each other: V1 with C1 and V2 with C2.³

I I . T H E A N A LO G Y B E T W E E N PE RC E P T I O N A N D
TESTIMONY

Reid announces at the beginning of Inquiry, 6. 24 that there is a remarkable ana-

logy between the testimony of nature given by the senses and the testimony of
our fellow human beings given in language—an analogy so great that it is natur-
al to use the word ‘testimony’ in each case. He therefore undertakes to consider
together the principles of the mind that are subservient to testimony of the two
sorts. The analogy he sees is divided into two sets of similarities, as there are two
varieties of perception and two varieties of language to be considered:
We have distinguished our perceptions into original and acquired; and language, into
natural and artificial. Between acquired perception, and artificial language, there is a
great analogy; but still a greater between original perception and natural language.
(IHM, p. 190)
To appreciate the two analogies, we must first say a bit about each of the two
distinctions.
Original versus acquired perception: ‘‘Our perceptions are of two kinds,’’ Reid
tells us; ‘‘some are natural and original, others acquired, and the fruit of experi-
ence’’ (IHM, p. 171). Our original perceptions are the perceptions we have prior
to any learning. When I hold a hard, round ball in my hand, the tactile sensa-
tions I receive trigger an immediate belief in a hard, extended object. No learning
is required; I am innately so constituted that on the appropriate sensory occa-
sions, I conceive of and believe in the ball. Similarly, when a certain pattern of
receptors on my retina is stimulated, I am presented with an object of a certain
two-dimensional shape in a certain region of the space before me: I conceive of
and believe in a square patch there. These are examples of original perception.
Others of our perceptions are acquired. By experience, I learn that a certain
two-dimensional array of polygons presented to my eye will be attended by the
tactile perception of a three-dimensional cube. A butcher learns that a sheep of a
certain visual appearance will have a certain heft when he lifts it and places it on
his scales. What was at first a matter of inductive inference or association based on
accumulated experience becomes a matter of immediate, noninferential, quasi-
perceptual belief: I see a cube in the pattern of lines, and the butcher sees the
weight of the sheep. These are examples of acquired perception.
Natural versus artificial language: Reid first draws the distinction between
natural and artificial language in Inquiry, 4. 2. By artificial language, he means
any system of signs whose meaning is fixed by convention, or as he says, ‘‘by com-
pact and agreement’’. By natural language, he means that small but indispensable
 James Van Cleve 55

body of signs that ‘‘previous to all compact or agreement, have a meaning which
every man understands by the principles of his nature’’ (IHM, p. 51). For exam-
ple, a smile is a natural language sign of approval. Reid argues that unless there
were natural language, artificial language could never be invented, for artificial
language requires compacts, and compacts could not be instituted by creatures
who did not have a language of some sort. There must therefore be natural lan-
guage to get things going. One could argue in similar fashion that artificial lan-
guage, once invented, could not be learned by a novice unless there were natural
language signs of reinforcement and dissent.⁴
We are now in a position to delineate the two analogies—the ‘‘great’’ ana-
logy between acquired perception and artificial language and the ‘‘still greater’’
analogy between original perception and natural language. In all four of the phe-
nomena to be considered—original perception, natural language, acquired per-
ception, and artificial language—there are signs and things signified, and the
mind passes from an apprehension of the sign to a belief in the thing signified.
The various similarities and differences Reid notes all concern the origin of the
relation between sign and thing signified and the means whereby we come to
know of this relation.
Original perception: ‘‘The signs in original perception are sensations’’—for
example, the tactile sensations that trigger in us the conception of and belief
in a hard, round ball in our hands. ‘‘Nature hath established a real connection
between the signs and the things signified; and nature hath also taught us the
interpretation of the signs; so that, previous to experience, the sign suggests the
thing signified, and creates the belief of it’’ (IHM, p. 190). In other words, it is by
an innate or hardwired principle that the mind passes from apprehension of the
sign to belief in the thing signified.
Natural language: ‘‘The signs in natural language are features of the face, ges-
tures of the body, and modulations of the voice’’ (IHM, p. 190). The things
signified are the thoughts and dispositions of another’s mind. As in the case of
original perception, nature has both established the connection between sign and
thing signified and taught us the interpretation of the sign previous to experience.
An infant knows instinctively that a smile is a sign of approval and a frown of
anger. As noted earlier, Reid holds that without a basic repertoire of such instinct-
ively understood signs, artificial language could neither be devised nor learned.
A further point of similarity between original perception and natural language
is this: in both cases, the signs ‘‘have the same signification in all climates and
in all nations’’ (IHM, p. 191). Certain tactile sensations indicate hardness to any
human being, and certain facial expressions indicate approval in all cultures.⁵
Acquired perception: ‘‘In acquired perception, the signs are either sensations,
or things which we perceive by means of sensations’’ (IHM, p. 191). As in the
other cases considered so far, the connection between sign and thing signified is
established by nature. But in this case as not in the others, we must discover the
connection by experience and induction. That a red glow in an iron bar signifies
56 Reid on the Credit of Human Testimony

heat, or that the small size of a man on the beach signifies that he is a long way
off, are things we must learn; they are not written into our constitution. Once
the connection is learned, however, the sign automatically suggests the thing sig-
nified, and it is almost as though we see the heat of the bar or the distance of
the man.
Artificial language: ‘‘In artificial language, the signs are articulate sounds,
whose connection with the things signified by them is established by the
will of men: and in learning our mother tongue, we discover this connection
by experience’’ (IHM, p. 191). Artificial language is therefore like acquired
perception in that the connection between signs and what they signify is known
by experience; but it is unlike acquired perception in so far as the connection
holds by convention rather than nature.
Reid notes a further respect in which acquired perception and artificial lan-
guage are like one another, but different from original perception and natural
language. ‘‘Our original perceptions, as well as the natural language of human
features and gestures, must be resolved into particular principles of the human
constitution’’. The emphasis here is on particular: it is by one particular principle
that sensations of a certain sort signify hardness and by another particular prin-
ciple that frowns express disapproval. By contrast, ‘‘our acquired perceptions, and
the information we receive by means of artificial language, must be resolved into
general principles of the human constitution’’ (IHM, p. 191). It may be objec-
ted that the interpretive principles we learn in the latter cases—for instance,
that a red glow indicates heat, or that the word ‘jaune’ means yellow—are as
particular as any. But Reid’s point is that such particular principles are not pro-
grammed into our minds. The principles that are programmed into our minds
are the general principles whereby we learn the particular principles. One of these
general principles is the principle of induction; whether there are others we shall
see presently.
The foregoing points of similarity and dissimilarity are summarized in Table 2.1.
The table enables us to see at a glance in what respects the analogy between
original perception and natural language is greater than the analogy between
acquired perception and artificial language. In the first two columns, all four rows
are filled in the same way, whereas in the second two columns, only two of the
four rows are filled in the same way. Of course, in so far as the universal versus
variable difference in row three is a corollary of the nature versus convention dif-
ference in row one, one may wish to say that at root there is only one difference in
the right two columns. Nonetheless, the overall analogy is still not as great in the
right columns as in the left.
We saw in Section I that Reid’s formulation of the principle of veracity con-
tains an ambiguity about the thing signified—is it the fact that p, or the speaker’s
belief in p? We must note now another blurred distinction that threatens to
confuse his exposition. This time it is an ambiguity about the sign rather than
the thing signified. What are the signs that figure in knowledge from testimony?
 James Van Cleve 57

Table 2.1. Two Analogies.

Greater analogy Lesser analogy

OP NL AP AL

How is the connection between sign Nature Nature Nature Convention

and thing signified established —by
nature or convention?
How is the connection between sign Innately Innately By experience By experience
and thing signified known—
innately or by experience?
Are the principles of signification Universal Universal Universal Variable
universal or variable across cultures?
Are the principles of our Particular Particular General General
constitution particular principles of
signification or general principles
enabling us to learn the particular
principles?

Normally, they are signs belonging to artificial language,⁶ and these, Reid tells
us, are ‘‘articulate sounds’’ (or written signs, as he adds in another place). Are
the signs to be construed simply as not-yet-interpreted sounds issuing from the
mouths of our fellows? Or are they to be construed as sounds already interpreted
as giving voice to propositions? Let me use in tandem the variables ‘S’ and ‘p’
in the following way: ‘S’ for a sentence, such as ‘it is raining today’ or ‘il pleut
aujourd’hui’, and ‘p’ for a proposition expressed by that sentence, such as the pro-
position it is raining today. Then our question may be put thus: are the signs of
concern to Reid in Inquiry, 6. 24 signs of the sort A utters S or signs of the sort A
says p? I fear that the answer is sometimes one and sometimes the other, and that
Reid does not mark the difference.⁷
Why do I say that Reid glosses over this distinction? In brief, it is because when
he is discussing the principles of veracity and credulity, the signs he is concerned
with must be assertions of propositions—items of the sort A says p. But when he
is discussing the analogy between artificial language and acquired perception, the
signs can only be utterances of words—items of the sort A utters S.
As noted above, Reid thinks there are two requisites for knowledge by
signs—there must be a regular connection between the sign and the thing
signified, and the appearance of the sign must induce a belief in the thing
signified. It is always a matter of interest to Reid to ascertain how, in various cases
of knowledge by signs, these requisites are satisfied. Is the sign conjoined with
the thing signified by nature or by convention? Does the appearance of the sign
produce belief in the thing signified by innate knowledge or by experience? In
Inquiry, 6. 24, he advances the principles of veracity and credulity, two principles
‘‘implanted in our natures,’’ as providing the answers to these questions in the
58 Reid on the Credit of Human Testimony

case of testimony: by virtue of veracity there is a reliable connection between the

sign and the thing signified, and by virtue of credulity we believe in the thing
signified when we apprehend the sign. For these claims to be true, the signs
must be assertions of propositions rather than utterings of words. There are no
principles ‘‘implanted in our natures’’ whereby mere words signify a certain state
of affairs or induce us to believe in it.
But now go back and look at the table delineating the various features Reid
attributes to the sign-thing signified relation in artificial language. The connection
holds by convention; it is known by experience; it varies from culture to culture.
Here, obviously, the signs must be utterances of sentences, not assertions of
propositions.
As far as I can see, the unmarked ambiguity in what counts as a sign causes
only one mistake in what Reid tells us. In Reid’s own order of exposition, the ana-
logy between perception and language precedes his discussion of the principles of
veracity and credulity. From the way in which Reid frames his discussion of these
principles (beginning at lines 12–13 of IHM, p. 192 and closing at lines 35–7
of IHM, p. 195), it is clear that they are the principles he regards as ‘‘the general
principles of the human mind which fit us for receiving information from our
fellow-creatures’’. He then goes on (IHM, pp. 195–7) to consider ‘‘the general
principles which fit us for receiving the information of nature by our acquired
perceptions,’’ which turn out to be the uniformity of nature and the inductive
principle. In other words, he is suggesting that uniformity and induction are the
general principles referred to in row four of our table under the heading ‘acquired
perception’, while veracity and credulity are the general principles referred to
under the heading ‘artificial language’. But that is a mistake. If the signification
of words is what is at issue, then induction is the relevant principle under both
headings. It is induction that teaches me that a red glow signifies hotness, and it is
also induction that teaches me what mama means by ‘milk’.⁸
To summarize, Reid’s exposition in Inquiry, 6. 24 of how signs signify what
they do is clouded by two unmarked ambiguities. One is an ambiguity about the
thing signified—is it a worldly fact or a fact about the speaker’s state of mind?
The other is an ambiguity about the sign—is it the utterance of a string of words
or the assertion of a proposition? Fortunately, the ambiguities do not invalidate
any of Reid’s key contentions. Nearly everything he wants to say can still be said,
but we need to take greater care in stating it. We can present the entire picture as
shown in Fig. 2.1.
If we take A says p as the sign and p as the thing signified, the first requisite
of knowledge by signs (reliable connection between sign and thing signified) is
ensured by veracity principle V1 and the second requisite (belief in the thing sig-
nified upon belief in the sign) is ensured by credulity principle C1. If we take A
says p as the sign and A believes p as the thing signified, the first requisite is ensured
by V2 and the second by C2. If we take A utters S as the sign, things get more
complicated, for there is now in effect an intermediate sign, A says p, signified
 James Van Cleve 59

p is true
A says p
A believes p
A utters S
B believes p
B believes
A says p
B believes A believes p

Figure 2.1. From Signs to Knowledge of what they signify.

by A utters S and signifying in turn either p or A believes p. When we consider A

utters S as sign and A says p as (intermediate) thing signified, the first requisite is
ensured by convention and the second by induction. When we consider A utters
S as sign and p as (ultimate) thing signified, the first requisite is ensured by the
logical product of convention and V1, the second by the product of induction
and C1. Finally, when we consider A utters S as sign and A believes p as (ultimate)
thing signified, the first requisite is ensured by convention and V2, the second by
induction and C2. (The reader may wish to label the arrows in Fig. 2.1. On the
left, the upper arrow should be labeled ‘convention’ and the lower arrow ‘induc-
tion’. On the right, the arrows from top to bottom should be labeled as ‘V1’,
‘V2’, ‘C1’, and ‘C2’.)

I I I . E X T E N D I N G T H E A N A LO G Y: A R E T E S T I M O N I A L
B E L I E F S E PI S T E M I C A L LY B A S I C ?

The preceding sections have dealt mainly with Reid’s views on the psychology of
testimony—with the principles whereby we come to believe what we do when
our fellows utter certain things. I turn now to matters more properly epistemo-
logical. What makes a belief acquired on the basis of testimony justified? What
makes it knowledge?
It is not always clear in Reid where psychology stops and epistemology begins.
When we have certain sensory experiences, Reid tells us, we instinctively and
immediately believe in external objects, without any need of reasoning. Very well,
his readers may ask, but is that just a piece of descriptive psychology (with which
Hume could agree), or is it meant as normative epistemology?
In the case of sense perception, I believe it is meant as both. I shall take for
granted here an interpretation of Reid’s perceptual epistemology I have defended
elsewhere (Van Cleve 1999). According to that interpretation, beliefs in physic-
al objects prompted by sensory experiences are not only psychologically immediate
(that is, triggered directly by the experiences without any reasoning or reliance
on background information), but also epistemically basic (that is, justified without
depending for their justification on any other justified beliefs). Beliefs about the
60 Reid on the Credit of Human Testimony

observable features of things in our environment function as first principles —we

appeal to them in justification of other things, but they are justified themselves
without appeal to anything further. ‘‘First principles’’ may sound like too grand
a name for such humble deliverances of perception as there is a tree over there,
but Reid makes it clear that such propositions play the same role in our empirical
knowledge as do axioms in mathematics (IHM, p. 172).
Can we extend Reid’s analogy between the testimony of our fellows and the
testimony of our senses one step further, maintaining that beliefs based on human
testimony are like beliefs based on sense perception in being epistemically basic?
In other words, can the mere fact that someone tells you p make you prima facie
justified in believing p and (if p is true and there are no defeaters for your justi-
fication) thereby give you knowledge of p? On this question, Hume says no, but
Reid says yes.
That Reid accords positive epistemic status, and perhaps even basic status, to
testimonial beliefs can be argued on the basis of his discussion of testimony in the
Essays on the Intellectual Powers of Man. His chapter on first principles includes
the following principle relating to testimony: ‘‘Another first principle appears
to me to be, That there is a certain regard due to human testimony in mat-
ters of fact, and even to human authority in matters of opinion’’ (EIP: 487). I
reproduce here several of the sentences he offers in explanation of this principle,
accompanying each with my own gloss.
‘‘Before we are capable of reasoning about testimony or authority, there are
many things which it concerns us to know, for which we can have no other
evidence.’’ ‘‘No other evidence:’’ so testimony is a source of evidence. This is
confirmed in the chapter on probability (EIP: 557–58), where Reid discusses
testimony as a species of probable evidence.
‘‘The wise Author of nature hath planted in the human mind a propensity to
rely upon this evidence before we can give a reason for doing so.’’ That someone
has usually told us the truth in the past would be a reason for believing what
he says, but his say-so confers evidence even before I am capable of having this
reason. Testimonial beliefs are evident even in the absence of reasons.
‘‘If children were so framed, as to pay no regard to testimony or to authority,
they must, in the literal sense, perish for lack of knowledge.’’ This sentence reaf-
firms the immense practical advantage of credulity Reid had emphasized in the
Inquiry. It also carries the further implication that what testimony bequeaths to
children is knowledge.
When we put these three points together, it emerges that Reid regards
testimony as a source not merely of belief, but also of evident belief or knowledge.
Moreover, it plays this role even if the believer has no reason for his belief.
Testimony-based beliefs therefore qualify for Reid as epistemically basic.⁹
Standing opposed to Reid on this question is Hume. Here is a well-known
passage from his essay on miracles (section X of the Enquiry Concerning Human
Understanding):
 James Van Cleve 61

There is no species of reasoning more common, more useful, and even necessary to
human life, than that which is derived from the testimony of men, and the reports of
eye-witnesses and spectators. This species of reasoning, perhaps, one may deny to be
founded on the relation of cause and effect. I shall not dispute about a word. It will be
sufficient to observe, that our assurance in any argument of this kind is derived from
no other principle than our observation of the veracity of human testimony, and of the
usual conformity of facts to the reports of witnesses. It being a general maxim, that no
objects have any discoverable connexion together, and that all the inferences, which we
can draw from one to another, are founded merely on our experience of their constant
and regular conjunction; it is evident, that we ought not to make an exception to this
maxim in favour of human testimony, whose connexion with any event seems, in itself, as
little necessary as any other. (Hume 1977: 74)
The reason, why we place any credit in witnesses and historians, is not derived from any
connexion, which we perceive a priori, between testimony and reality, but because we are
accustomed to find a conformity between them. (Hume 1977: 75)

Hume is nowadays regarded as the prototypical reductionist about the epistem-

ological status of testimony. Reductionists hold that when B believes p because
A says p, B’s belief in p is justified only if B is justified in believing two further
things: (i) that A did say p, and (ii) that most of the things said by A (or by the
individuals in some wider relevant class of which A is a member) are true. Item
(i) may be factored into the information that A produced certain signs and that
these signs should be interpreted in a certain way. Item (ii) is what Hume refers
to as the ‘‘usual conformity of facts to the reports of witnesses.’’ It is clear that
the reductionist position denies that beliefs based on testimony are epistemically
basic—although they may be psychologically noninferential or immediate, they
owe their justification to other justified beliefs B has about what A has said on this
occasion and about his track record (or that of his cohort) in the past.¹⁰
Who is right—Reid or Hume? To answer this question, I shall set out the
most formidable argument I know of against testimonial fundamentalism (as I
shall call the anti-reductionist position, following Coady)¹¹ and then consider
whether any good response to the argument is available to Reid.
The argument I have in mind consists of two premises and a conclusion:
Premise A: If a source S is a source of epistemically basic beliefs, then it must be an a
priori matter that deliverances of S are prima facie warranted or likely to be true.¹²
Premise B: It is not an a priori matter that testimonial beliefs are prima facie warranted or
likely to be true.
Conclusion: Testimony is not a source of epistemically basic beliefs.

Let me henceforth abbreviate ‘it is a priori that deliverances of S are prima facie
warranted or likely to be true’ to ‘S is an a priori source’ and ‘S is a source of
epistemically basic beliefs’ to ‘S is a foundational source’. Then the argument can
be put briefly as follows: a foundational source must be an a priori source, and
testimony is not an a priori source.¹³ But why should we accept the premises?
62 Reid on the Credit of Human Testimony

For premise A, there is a rationale that many traditional epistemologists would

find compelling. This rationale is nicely laid out in the opening chapter of Pol-
lock’s Knowledge and Justification (Pollock 1974). Pollock begins by noting that
basic beliefs about a given subject matter typically have a distinctive source. Beliefs
about physical objects have perception as their source; beliefs about the past have
memory as their source; beliefs about the future or the unobserved have induction
as their source, and so on. It is part of Pollock’s idea of a source that it is a way of
knowing about its subject matter on which all other ways of knowing about that
subject matter depend (Pollock 1974: 6). We can have knowledge of the physical
world by means other than perception, such as memories and photographs, but
if we couldn’t have knowledge about the physical world through perception, we
couldn’t have it in these other ways, either. We might say that for Pollock, a
source of basic beliefs is both autonomous (it gives knowledge or justification in
its sphere without the help of other sources) and ultimate (no other sources give
knowledge or justification about that subject matter without the help of it).
With this much assumed about sources, Pollock’s argument may be set out
(with some compression) as follows:
1. A putative source of knowledge is not a genuine source unless we can
establish a connection between the source and the facts it is supposed to
deliver.
2. Such a connection can be established only if (a) it can be established
inductively, by showing that there is a reliable correlation between a
belief’s being delivered by the source and its being true, or (b) the con-
nection holds a priori.
3. Since foundational sources are ultimate, alternative (a) is ruled out. We
could establish the reliability of a foundational source only by relying on
that very source and thus reasoning in a circle.
4. Therefore, a foundational source of knowledge must be an a priori
source: if its deliverances are epistemically basic, it must be an a priori
matter that its deliverances are warranted or likely to be true.
That, as I said, is a defense of premise A that many traditional epistemologists
would find compelling.
Let us turn now to premise B, which tells us that there is no a priori connection
between testimonial reports (or beliefs based thereon) and their truth or war-
rant. Why believe that? One reason would be that a priori connections must hold
necessarily,¹⁴ yet there seems to be no discernible necessary connection between a
testimonial report and the truth or warrant of what is reported. That, indeed, was
part of Hume’s case for his stance on testimony as quoted above: he notes that the
connection between human testimony and the event attested to ‘‘seems, in itself,
as little necessary as any other.’’
In Pollock’s scheme of things, there is an additional reason for denying an a
priori connection between testimony and its deliverances. Pollock holds that an
 James Van Cleve 63

a priori connection ‘‘must arise from the meanings of the concepts or statements
involved in the knowledge claims.’’ Traditionally, this demand led to reductive
analyses: for example, phenomenalist analyses of the truth conditions of external
world statements in terms of the patterns of experience that serve as evidence for
them. Pollock offers an alternative to the reductive tradition in which the mean-
ing of a statement is given not by its truth conditions, but by its justification
conditions. On his account, it is part of the meaning of ‘x is red’ that x’s looking a
certain way is a justification condition for it: one’s understanding of what it is for
something to be red is constituted in part by knowing that something’s looking
that way justifies you in believing it to be red. If we applied this strategy in the
case of testimonial belief, we would say that one understands the meaning of a
statement ‘p’ partly by knowing that ‘A says p’ is a justification condition for it.
But ‘A says p’ has ‘p’ embedded within it! It is therefore out of the question that
we understand ‘p’ in terms of ‘A says p’, since we can understand the latter only
if we already understand the former. It is no accident that Pollock’s book, which
devotes a chapter to each of the traditional sources of justification, contains no
chapter on testimony. His program of exhibiting the justification conditions of a
statement as constitutive of its meaning cannot be carried out when the justifica-
tion condition is a testimonial report.
Such, then, is the case against taking testimony as a foundational source: a
source of basic beliefs must be an a priori source, and testimony is not an a priori
source. How might a testimonial fundamentalist defend his position against this
two-pronged attack? I shall discuss two strategies. One denies the first premise,
invoking a reliability theory of justification to argue that a source of basic beliefs
need not be an a priori source. The other denies the second premise, mobilizing
some of the ideas in Wittgenstein’s Private Language Argument to maintain that
testimony is, after all, an a priori source. I discuss the second strategy first.
That testimony should qualify as an a priori source may seem surprising, but
precisely that is an implication of some versions of the Private Language Argu-
ment. I focus on a version that was prevalent before considerations about rule-
following took center stage. (I am indebted in what follows to Saunders and
Henze 1967, especially ch. II.)
For present purposes, let a private language be a language whose terms refer to
‘‘private’’ data—to tickles and twinges and other inner states, regarded as having
no conceptual ties to any manifestations in outward behavior. The argument to
be reviewed maintains that a private language in this sense is impossible, because
no one could know that he was employing its terms consistently and correctly.
Suppose I take myself to be using a term T to refer to the same type of private
state that I have applied it to in the past. Since the referent of T is private, no
one else can be in a position to correct or corroborate my use of T. I will have
nothing but my own memory impressions to vouch for the fact that I am using
T in the same way as before. But I am entitled to trust my memory impressions,
it is insisted, only if there is some way other than memory of checking up on
64 Reid on the Credit of Human Testimony

them. (Corroborating one memory by reference to another is disallowed on the

alleged ground that it would be like buying two copies of the morning paper
to verify that what it says is true [Wittgenstein 1953: para. 265].) The required
independent check is available to speakers of a public language, for they may rely
upon the testimony of their fellows that they are using T in the same way as
before. The speaker of a private language, by contrast, cannot avail himself of any
independent checks. His companions, having no outer criteria of his inner states,
can provide no correction or corroboration. He can only be under the impression
that he is following some semantic rule in the same way as before, but to think
that you are obeying a rule is not the same thing as obeying it. You cannot obey a
rule privately (Wittgenstein 1953: para. 202).
To the Private Language Argument in this incarnation, the traditionalist has
a ready reply. He can say that his opponent’s requirements would keep anyone
from knowing he is using his terms correctly. For what entitles the speaker of a
public language to rely upon the testimony of his fellows in corroboration of his
memory impressions? In order for one to be justified in accepting something on
someone else’s say-so, one must know that the other person’s testimony has been
reliable in the past, and the only way one could know that is through reliance on
one’s own memory impressions. If the Wittgensteinian insists that memory can
be relied upon only if backed up by testimony, he lands us in a vicious circle,
because testimony in its turn can be relied upon only if backed up by memory.
It should be manifest that what the traditionalist’s reply invokes is precisely a
Humean reductionist position on testimony. In the next step in the dialectic—
the crucial one for our purposes—the Wittgensteinian repudiates Hume. He says
that it is a conceptual truth, knowable a priori, that an utterance by one of my
fellows of a past-tense sentence that I would use to say p confers initial probability
on p.¹⁵ Thus there is no need to underwrite the deliverances of testimony by
induction and memory, for such deliverances have their warrant a priori. QED.
The argument just given is, in effect, a transcendental argument for testimony
as an a priori source. It proceeds from the premise that we know we are using the
terms in our language correctly and argues it to be a necessary condition of this
knowledge that the corroborative reports of our fellows have their evidential force
a priori.
What verdict shall we pass on the argument? In my opinion, the argument can
succeed in establishing the a priori credentials of one source of knowledge (testi-
mony) only at the cost of destroying or downgrading the credentials of another
(memory). If we allow that memory is an autonomous source of knowledge, cap-
able of delivering knowledge and justified belief without corroboration by other
sources, we can sidestep the argument entirely. I believe that memory is such an
autonomous source, and I believe Reid would agree. For Reid, memory takes its
place alongside perception as a source of first principles or basic beliefs. When
we seem to perceive that there is a tree over there, we automatically believe there
is a tree over there and are prima facie justified in so believing. Likewise, when
 James Van Cleve 65

we seem to remember some past event, we automatically believe that the event
occurred and are prima facie justified in so believing (see EIP, pp. 253–5 and
474). One can gather what Reid’s attitude would have been toward the thesis that
memory stands in need of independent corroboration from the following remark:
‘‘When I remember a thing distinctly, I disdain equally to hear reasons for or
against it’’ (EIP, p. 476). For better or for worse, then, Reid keeps the Private
Language Argument from getting off the ground. The anti-fundamentalist prem-
ise that testimony is not an a priori source still stands.¹⁶
Let us turn, then, to the other strategy for avoiding the anti-fundamentalist’s
one-two punch. In this strategy we deny that a source of basic beliefs must be
an a priori source. One basis for doing so is provided by the reliability theory of
justification, which has come to prominence only in recent years, but which is
sometimes discerned in the writings of Reid himself.
The tenets of the reliability theory may be set down in the following way,
which is due to Goldman (1979). First, we distinguish two kinds of belief-
forming processes: belief-independent processes, such as perception, which
do not take other beliefs as inputs, and belief-dependent processes, such as
reasoning, which do take other beliefs as inputs. Next, we say that a process of
the first sort is reliable iff it tends to produce only true beliefs as outputs (this
is ‘‘unconditional reliability’’), and we say that a process of the second sort is
reliable iff it tends to produce only true beliefs as outputs when it is given true
beliefs as inputs (this is ‘‘conditional reliability’’). Finally, we offer a recursive
account of justification as follows: a belief is justified iff either (i) it results from an
unconditionally reliable belief-independent process or (ii) it results from justified
beliefs by way of a conditionally reliable belief-dependent process.
A reliability theory along these lines lets us counter the one-two punch in two
ways. First, it undermines the case for premise A. The argument for that premise
relied crucially on the following assumption: a putative source of knowledge is
not a genuine source unless we can establish a connection between the source
and the facts it is supposed to deliver. ‘‘Establishing’’ a connection in this context
means knowing or verifying that it obtains. That is precisely what is not necessary
according to the reliability theory: if the theory is correct, there need only be a
connection, whether anyone knows it obtains or not. For the reliability theory,
the mere fact that a belief has been formed by a reliable process is sufficient to
make that belief justified. That implies that no knowledge by the subject or any-
one else that the process is reliable is necessary. So the key assumption is false: a
source is a source just so long as there is an appropriate connection between the
source and its deliverances, whether anyone can establish it or not.
Second, the reliability theory not only undermines the case for premise A, but
it also enables us to see directly that that premise is false. Suppose that when we
hear others attest to some fact p, we automatically believe p. There is no drawing
of inferences or weighing of reasons—we simply believe p. That, of course, is
what Reid’s principle C1 says. It implies that beliefs generated by the testimony
66 Reid on the Credit of Human Testimony

of others are psychologically immediate, not based on reasons.¹⁷ Suppose next

that it is a contingent and a posteriori fact, but a fact nonetheless, that the things
others attest tend to be true. That, of course, is what Reid’s principle V1 says.
It implies that believing what others say is a reliable method of belief formation,
and thus that beliefs formed by this method are (by reliabilist standards) justified.
Putting this all together, testimonial beliefs are immediate (by C1) and justified
(by V1 and the reliability theory); therefore, they are justified immediate beliefs,
which is another way of saying they are epistemically basic. So testimony is a
source of epistemically basic beliefs, even though the fact that such beliefs are
generally true is knowable only a posteriori.¹⁸
Could Reid avail himself of this way of avoiding the anti-fundamentalist argu-
ment? That depends on whether he is a reliabilist. The best case I know of for
interpreting Reid as a reliabilist has been presented in a book by Philip de Bary
(2002), whose main features I now sketch.
According to de Bary, when Reid draws up his list of first principles, he is in the
first instance simply formulating psychological laws about human belief forma-
tion—laws specifying what sorts of things people instinctively believe in various
circumstances. When I perceive a tree or a star, that is enough to make me believe
one is there; when I remember walking on the beach yesterday, I believe that I did
so; and when a friend or stranger tells me a tree has blocked the road, I believe
forthwith that the road is indeed blocked. If Reid went no further than this, his
list of first principles would ‘‘lack any epistemological bite’’ (de Bary 2002: 65).
But Reid does go further (according to de Bary), embedding his principles in
a framework of reliabilism. Reid believes that ‘‘the instinctive beliefs of healthy
people . . . tend towards the truth’’ (de Bary 2002: 83) and further, that such a
tendency towards truth (given a reliabilist view like Goldman’s) is sufficient (in
the absence of special reasons for doubting them) for their being justified and
amounting to knowledge when true. So the various classes of belief marked out
in Reid’s list of first principles are not only psychologically immediate, but epi-
stemically basic. When true, they are items of basic knowledge, and that goes
for beliefs based on testimony just as much as for beliefs based on perception
and memory.
I am not convinced that de Bary’s interpretation of Reid is right. It is plausible,
I admit, but I think there are alternatives at least equally plausible. I have sketched
one such in ‘‘Reid on the First Principles of Contingent Truths’’ (Van Cleve
1999). According to my alternative, when Reid enunciates his first principles, he
is giving generalizations that are epistemological principles as they stand, not just
psychological laws that acquire epistemological significance only when supple-
mented with facts about reliability and the tenets of the reliability theory. They
are principles according to which the deliverances of introspection, perception,
memory, and credulity (if I may so name the ‘‘faculty’’ through which we believe
testimony) are prima facie justified, regardless of whether they are true or false on
a given occasion and possibly even regardless of whether they generally tend to be
 James Van Cleve 67

true.¹⁹ On this interpretation, Reid’s epistemology is more like Chisholm’s than

Goldman’s. Chisholm provides a list of epistemic principles according to which
the deliverances of introspection, perception, and memory have one or another
positive epistemic status just in virtue of having those sources, not in virtue of any
connection with truth (Chisholm 1977: ch. 4).
If Reid is taken in this way, he could perhaps still hold that testimony-based
beliefs are epistemically basic. However, he would lack de Bary’s reason for so
regarding them: they would not be justified simply in virtue of their reliability,
no matter how great that reliability is. They would be justified instead in virtue of
. . . what?
Here I think Reid would have no good answer. I do not think it entirely
out of the question that perception and memory are a priori sources of justi-
fied belief,²⁰ but testimony just does not look to me like an a priori source, any
more than the readings of a barometer do. If testimonial beliefs are not epistem-
ically basic because they are deliverances of an a priori source, nor because they
are deliverances of a reliable belief-independent source, then how they can they
be epistemically basic at all? This question pushes me strongly in the direction
of testimonial reductionism. I shall therefore conclude this essay by defending
reductionism against what I take to be the most formidable objection to it. I shall
also offer a sketch of what Reid’s epistemology of testimony would look like if he
were to join the reductionist camp.
The objection I have in mind is at least intimated if not explicitly articulated
by several writers, including Anscombe (1979), BonJour (2002), Coady (1992),
and Wolterstorff (2001). The key premise in it is that the vast majority (or perhaps
even the totality) of what passes for corroboration of testimony itself relies on other
testimony. ‘‘The guidebook was right,’’ I say, ‘‘there is such a place as Piccadilly
Circus’’—but in so saying I rely on the street signs that someone has posted at
the site. I cannot construct an inductive argument for the reliability of testimony
because ‘‘the cases that I can investigate firsthand [without relying on further
testimony to corroborate them] amount to only a vanishingly small proportion
of either the persons and other sources that provide testimony or the subject
matters to which such testimony pertains’’ (BonJour 2002: 172–3). An implaus-
ibly strong version of this argument would say that there is in principle no case
in which I can corroborate testimony without relying on further testimony—in
other words, that testimony is an ultimate source. A more plausible version would
say that the proportion of cases I can corroborate firsthand is ‘‘vanishingly small,’’
so that any inductive argument based on them must be ‘‘extremely weak’’ (Bon-
Jour 2002: 173). The first conclusion to be drawn from these premises would be
that if reductionism is correct, I know little or nothing on the basis of testimony.
The second conclusion would be that since I do know a great deal on the basis of
testimony, reductionism is untenable—another transcendental argument.
In my opinion, this argument is far from decisive. To begin with, there is no
plausibility at all in the strong version’s contention that any corroboration of
68 Reid on the Credit of Human Testimony

testimony must rely on further testimony. For many years of my life, I believed
there were such wonders as the Grand Canyon and the Taj Mahal solely on the
basis of books and postcards. Now I have verified the existence of those things
with my own eyes. (More accurately: I have verified with my own eyes the exist-
ence of such things, that is, structures matching a certain general description, for
I admittedly relied on the testimony of the locals to know that building is the
Taj Mahal.) To these dramatic episodes of confirmation may be added thou-
sands of more quotidian occurrences of finding beer in the fridge or a restroom
down the hall on the right after being told where to look. To be sure, these myri-
ad instances in which I have been able to check on the veracity of testimony
firsthand are (as the weak version rightly points out) only a minuscule fraction
of all the instances in which I have believed things on the basis of testimony.
But does it follow that any inductive justification I have for believing testimony
must be ‘‘extremely weak’’? Not at all, for what matters is not the proportion of
testimonial beliefs I have checked, but the proportion of checks undertaken that
have had positive results. I have seen only a tiny fraction of the world’s crows, but
the ones I have seen have been overwhelmingly black, and that is enough to sup-
port my belief that nearly all crows are black. (Of course, the ratio of testimonies
checked to testimonies that have proved true varies with different classes of testi-
mony; I have found geography textbooks to be more reliable than presidential
press conferences.²¹)
How much of what Reid has to say about testimony can we still accept if
we move into the reductionist camp? We can accept nearly everything he says
about the psychology of testimonial belief, especially as regards the principle of
credulity. We may also agree with what he says about the immense practical
advantage of credulity. A viable reductionism had better not take the form of
saying: believe no one whose track record you have not checked out for yourself.
Children (fortunately for them and fortunately for all who were once children
ourselves) go through a credulous phase during which they believe without reas-
on nearly everything they are told. As reductionists, however, we must hold that
these beliefs are justified only in a pragmatic sense, not in an epistemic sense. If
they qualify as knowledge, it must be a kind of knowledge that does not require
justification, but only a reliable connection with the truth, as in what Sosa calls
‘‘animal knowledge’’ (Sosa 1997; compare the view ascribed to Audi in n. 13).
As children grow into adulthood, their credulity diminishes and their ability to
give inductive reasons for what they accept from others grows. As Reid notes,
[Credulity] will be strongest in childhood, and limited and restrained by experience. . . .
When brought to maturity by proper culture, [reason] begins to feel her own strength,
and leans less upon the reason of others; she learns to suspect testimony in some cases,
and to disbelieve it in others; and sets bound to that authority to which she was at first
entirely subject. But still, to the end of life, she finds a necessity of borrowing light from
testimony, where she has none within herself, and of leaning in some degree upon the
reason of others. . . . (IHM, p. 195)
 James Van Cleve 69

In another passage, he tells us that adults are in a position to believe on the basis
of experience and reflection things they originally believed only on instinct:
I believed by instinct whatever [my parents and tutors] told me, long before I had the
idea of a lie, or thought of the possibility of their deceiving me. Afterwards, upon reflec-
tion, I found that they had acted like fair and honest people who wished me well. I found,
that if I had not believed what they told me, before I could give a reason of my belief,
I had to this day been little better than a changeling. And although this natural credu-
lity hath sometimes occasioned my being imposed upon by deceivers, yet it hath been of
infinite advantage to me upon the whole; therefore I consider it as another good gift of
Nature. And I continue to give that credit, from reflection, to those of whose integrity
and veracity I have had experience, which before I gave from instinct. (IHM, p. 170–1)
Things we accepted originally as a gift of nature we can give reasons for as we
grow older. In this way, animal knowledge is replaced by reflective knowledge,
and beliefs that were formerly justified only in a pragmatic or external sense
become justified reflectively.²²
My view, in conclusion, is that testimony gives us justified belief and reflective
knowledge not because it shines by its own light, but because it has often enough
been revealed true by our other lights. On this point, I find myself uncharacterist-
ically on the side of Hume rather than Reid.

REFERENCES
Anscombe, Elizabeth (1979), ‘What Is It to Believe Someone?’, in C. F. Delaney
(ed.), Rationality and Religious Belief (Notre Dame: University of Notre Dame Press),
141–51.
Audi, Robert (1997), ‘The Place of Testimony in the Fabric of Knowledge and Justifica-
tion’, American Philosophical Quarterly, 34: 405–22.
BonJour, Laurence (2002), Epistemology: Classic Problems and Contemporary Responses
(Lanham, Md: Rowman and Littlefield).
Burge, Tyler (1993), ‘Content Preservation’, Philosophical Review, 102: 457–88.
Chisholm, Roderick M. (1977), Theory of Knowledge (Englewood Cliffs, NJ: Prentice-
Hall, 2nd edn.).
Coady, C. A. J. (1992), Testimony: A Philosophical Study (Oxford: Clarendon Press).
de Bary, Philip (2002), Thomas Reid and Scepticism: His Reliabilist Response (London:
Routledge).
Ekman, Paul, Sorenson, E. Richard, and Friesen, Wallace V. (1969), ‘Pan-Cultural Ele-
ments in Facial Displays of Emotion’, Science, ns 164: 86–8.
Gilbert, Daniel T. (1993), ‘The Assent of Man: Mental Representation and the Control
of Belief ’, in Daniel M. Wegner and James W. Pennebaker (eds.), Handbook of Mental
Control (Englewood Cliffs, NJ: Prentice-Hall), 57–87.
Goldman, Alvin (1979), ‘What is Justified Belief?’, in George S. Pappas (ed.), Justification
and Knowledge (Dordrecht, Holland: D. Reidel), 1–23.
Hume, David (1977), An Enquiry Concerning Human Understanding, ed. Eric Steinberg
(Indianapolis: Hackett) (first published in 1748).
70 Reid on the Credit of Human Testimony

Plantinga, Alvin (1993), Warrant and Proper Function (New York: Oxford University
Press).
Pollock, John L. (1974), Knowledge and Justification (Princeton: Princeton University
Press).
Reid, Thomas (1997), An Inquiry into the Human Mind on the Principles of Common
Sense, ed. Derek R. Brookes (Edinburgh: Edinburgh University Press) (first published
in 1764). It is abbreviated in the text as IHM, with ‘6. 24’ indicating chapter 6,
section 24.
(2002), Essays on the Intellectual Powers of Man, ed. Derek R. Brookes (University
Park, Pa: Pennsylvania State University Press) (first published in 1785). It is abbrevi-
ated in the text as EIP.
Saunders, John Turk, and Henze, Donald F. (1967), The Private Language Problem: A
Philosophical Dialogue (New York: Random House).
Shoemaker, Sydney (1963), Self-Knowledge and Self-Identity (Ithaca, NY: Cornell
University Press).
Sosa, Ernest (1997), ‘Reflective Knowledge in the Best Circles’, Journal of Philosophy, 94:
410–30.
Van Cleve, James (1999), ‘Reid on the First Principles of Contingent Truths’, Reid
Studies, 3: 3–30.
(2004), ‘Reid’s Theory of Perception’, in Rene van Woudenberg and Terence Cuneo
(eds.), The Cambridge Companion to Reid (Cambridge: Cambridge University Press),
101–33.
Wittgenstein, Ludwig (1953), Philosophical Investigations, trans. G. E. M. Anscombe
(Oxford: Blackwell).
Wolterstorff, Nicholas (2001), Thomas Reid and the Story of Epistemology (Cambridge:
Cambridge University Press).

N OT E S
1. Interpreted as C1, the principle of credulity finds confirmation in the work of some
contemporary cognitive psychologists. See Gilbert (1993) for descriptions of experi-
ments in which the mere hearing or reading of a proposition tends to induce belief
in it.
2. There is also a route leading from C2 to C1. Suppose B hears A say p; in accordance
with C2, B thereupon believes that A believes p; drawing on Principle 7, he then
draws the inference that p is probably true; in light of that, he commences believing
p, just as C1 says. But this account would falsify what Reid regards as the psychology
of the situation. C1 is meant to describe a mechanism whereby we forms beliefs
immediately, not as the result of any reasoning
3. I am indebted in this section to Nicholas Wolterstorff ’s discussion of Reid’s account
of testimony in Wolterstorff (2001). Wolterstorff notes that Reid blurs the distinc-
tion between the consequents of the principles I have labeled C1 and C2 (2001:
172). He also observes that what Reid really needs by way of a veracity principle
is V1—the disposition to assert only what one believes must be coupled with a
tendency to get things right (2001: 176).
 James Van Cleve 71

4. It may be helpful to relate Reid’s discussion of natural language in IHM 4. 2 to his

discussion of natural signs in IHM 5. 3. In the latter section, he distinguishes three
classes of natural signs. The first class comprises signs ‘‘whose connection with the
thing signified is established by nature, but discovered only by experience’’ (IHM,
p. 59). For example, smoke is a class I natural sign of fire. The second class ‘‘is that
wherein the connection between the sign and thing signified, is not only established
by nature, but discovered to us by a natural principle, without reasoning or experi-
ence’’ (IHM, p. 60). For example, a smile is a class II natural sign of approval. The
third class comprises those signs ‘‘which, though we never before had any notion or
conception of the things signified, do suggest it, or conjure it up, as it were, by a
natural kind of magic, and at once give us a conception, and create a belief of it’’
(IHM, p. 60). For example, certain tactile sensations are class III natural signs of
hardness in bodies. The key difference between class II and class III natural signs is
that although in both cases the connection between our apprehension of the sign and
our conception of the thing signified is hardwired or innate, in the third class alone is
our conception of the thing signified also innate.
What Reid calls ‘‘natural language’’ comprises natural signs of the second class.
What he calls ‘‘artificial language’’ comprises signs whose connection with the thing
signified is established by convention rather than nature and which therefore belong
to none of the three classes of natural signs. His usage is obviously different from
our own, as he would classify English as an artificial language rather than a natural
language.
5. A smile’s a smile the world around. That certain facial expressions are universal
in their signification has been confirmed by contemporary research such as that
of Ekman et al. (1969).
6. There are exceptions. Reid tells us that ‘‘two savages who have no common artificial
language’’ (IHM, p. 52) could communicate with each other exclusively in the nat-
ural language of gesture, facial expression, and tone of voice. Presumably, they could
exchange testimony.
7. I gloss over a further distinction myself. A speaker may utter a sentence S that means
p without thereby saying p in the sense of asserting it (as contrasted with carrying
out a speech act with some other illocutionary force). By ‘saying p’ I mean asserting
p. For A’s uttering S to amount to his saying p in this sense, it is required not only
that S means p, but also that in uttering something that means p, A is asserting
p (rather than reciting a line in a play, for instance). I ignore this second aspect
of the difference between uttering S and saying p. It is discussed in Burge (1993)
and Wolterstorff (2001).
8. Of course, once the language of one’s elders has been learned, induction is no longer
necessary; one simply hears S as saying p, attending to the sense rather than the
sounds. Here is another similarity that Reid might have remarked between artificial
language and acquired perception.
9. Here I am in agreement with Coady (1992: 23 and 123). It may also be the view
of Plantinga (1993) that testimonial beliefs are epistemically basic for Reid.
10. Reductionism is so-called because it holds that testimonial knowledge can be
‘‘reduced to’’ or accounted for in terms of knowledge exclusively from other sources:
B knows p because he knows by perception that A has said p, by memory that A
72 Reid on the Credit of Human Testimony

has said such-and-such other things in the past, by memory and a variety of other
apposite sources that these things were true, and by induction based on the foregoing
that A is probably right in what he says on the present occasion. I dislike the name
‘reductionism’ because of its misleading associations with ‘‘reductive’’ doctrines such
as phenomenalism and behaviorism, in which there is a reduction not of one way of
knowing to others, but of the subject matter known to facts belonging to some other
class. Nonetheless, the label has become well entrenched, so I shall stick with it here.
11. Coady proposes ‘fundamentalism’ as a name for the view that testimony is a source
of basic or foundational beliefs on a par with perception and memory (Coady
1992: 23).
Reductionism and the fundamentalism I ascribe to Reid do not exhaust the
options, as Peter Graham makes clear (Chapter 4 in this volume). One of the other
options is the position that Graham defends under the label ‘weak fundamentalism’.
In this view, a belief based on testimony is not thereby prima facie justified, that
is, justified enough to qualify as knowledge if the belief is true and there are no
defeaters. Instead, it is only pro tanto justified, that is, possessed of a modicum of
justification, but not necessarily enough to qualify as knowledge. Coherence with
other justified beliefs may be required to bring the justification of the belief up to the
level required for knowledge. Graham’s view about testimonial knowledge is thus
analogous to C. I. Lewis’s view about memory knowledge: ostensible memories have
an initial level of justification or credibility just in virtue of being memory reports;
this initial level can be brought up to the level required for knowledge through
the coherence of the reports with one another and with other beliefs. I believe
weak fundamentalism is a more plausible position than the strong fundamentalism I
attribute to Reid. Nonetheless, I think the argument I am about to present, if cogent
at all, would apply to weak fundamentalism as well as to strong. That would be so,
at any rate, if pro tanto justification (no less than prima facie justification) must flow
from an a priori source.
12. This is not to say, of course, that the deliverances themselves are a priori. What is a
priori is the conditional: if a belief is delivered by S, then the belief is warranted or
likely to be true.
13. I might have been tempted to abbreviate ‘S is a source of epistemically basic beliefs’
to ‘S is a basic source’, except that to do so would have been to ride roughshod over
the nice set of distinctions developed by Robert Audi (1997). Audi distinguishes
between ‘source of basic X’ and ‘basic source of X’, where X can be any of the com-
modities belief, knowledge, or justification. Roughly, S is a basic source of {belief,
knowledge, justification} iff S can produce {belief, knowledge, justification} without
the cooperation of another source of the relevant commodity. S is a source of basic
{belief, knowledge, justification} iff S is a source of {beliefs not based on other
beliefs, knowledge not derived from other knowledge, justification not dependent
on other justified propositions}. In Audi’s view, testimony is not a basic source of any
of the three commodities. It is, however, a source of basic belief and a source of basic
knowledge, though not a source of basic justification.
How is it possible for testimony to be a source of basic knowledge without being a
source of basic justification? The answer is that Audi does not define basic knowledge
 James Van Cleve 73

as true belief with basic justification; in fact, he does not define knowledge as requir-
ing justification at all. Reliably formed belief can count as knowledge for him, but
reliability does not suffice for justification.
14. Kripkean reasons for believing that some propositions are both a priori and contin-
gent are plainly not in play here.
15. In the exposition of the argument by Saunders and Henze, the thesis I have just
mentioned is derived from two other theses, the Utterance Thesis and the Testi-
mony Thesis. According to the Utterance Thesis, it is a priori that an utterance by
A of a past-tense statement ‘p’ confers initial likelihood on the proposition that A is
making a memory claim that p. According to the Testimony Thesis, it is a priori that
A’s testifying as to what he remembers (and thus his making the memory claim that
p) confers initial probability on p. Saunders and Henze attribute versions of both
theses to Shoemaker (1963).
Shoemaker maintains (1963: 249–50) that we do not make an inductive infer-
ence from A uttered ‘p’ to A said p. I suspect that in holding this, he is primarily
concerned with the difference between uttering and the illocutionary act of saying.
He seems to have lost sight of the need for induction that is surely involved in learn-
ing what ‘p’ means in the language of one’s society.
16. I should acknowledge that Wittgensteinian worries about private languages are not
the only reasons for regarding testimony as an a priori source. Tyler Burge advances
the following as an a priori principle: ‘‘A person is entitled to accept as true some-
thing that is presented [by another person] as true and that is intelligible to him,
unless there are stronger reasons not to do so.’’ Burge argues for this principle by
deriving it from two subsidiary principles, each of which he considers to be a priori:
there is prima facie reason to regard a message one finds intelligible as coming from a
rational source, and there is prima facie reason to regard a rational source as a source
of truth. I lack the space to discuss Burge’s case for the subsidiary principles here, but
the interested reader may consult Burge (1993).
17. The following issue needs to be discussed: is believing testimony really a belief-
independent process? When I believe a piece of testimony, I come to the belief
only because I have heard (or otherwise perceived) someone say something. Audi
points out that this indicates a dependency of testimonial belief on perception, but
not necessarily a dependence on other beliefs, since I need not form the belief that
A said p. For Reid, however, things may be otherwise. He sometimes (though not
invariably) seems to define perception as involving belief, so that hearing A say p
would include as an ingredient believing that A said p. (For discussion, see Van
Cleve 2004: sect. III.) If that is so, believing what you hear others say would turn out
to be a belief-dependent process rather than a belief-independent process. But I am
going to assume for the sake of discussion here that testimonial beliefs for Reid are
psychologically immediate, i.e., dependent on no other beliefs.
18. In Goldman’s terms, any belief formed by a process that is both belief-independent
and reliable is an epistemically basic belief, and testimonial beliefs are so formed.
19. It is these deliverances—particular propositions believed on the basis of perception,
memory, and the like—that are the first principles on my interpretation, not the
generalizations that single them out.
74 Reid on the Credit of Human Testimony

20. I have characterized an a priori source as a source S such that it is a priori that the
deliverances of S are prima facie warranted or likely to be true. ‘Likely to be true’
is a statistical notion, implying that most of the deliverances of S are true. I cannot
see that it is a priori that most deliverances of perception are likely to be true. How-
ever, ‘prima facie warranted’ is a normative notion, not implying any statistics. Just
conceivably, it is a priori that deliverances of perception are prima facie warranted.
21. Though skeptical about an inductive justification of testimony, BonJour notes the
possibility of a coherentist justification of testimony, appealing to the agreement
among one another of various authorities on matters I am unable to check on
firsthand (2002: 173–7). I think such coherence can indeed boost the epistemic
standing of testimony-based belief, but only provided there is some initial reason to
believe the authorities. I also think such initial reason would have to derive from
induction over some sample of authorities that I have been able to check on for
myself. I do not, therefore, see a coherentist justification of testimony as an inde-
pendent alternative to an inductivist justification.
22. As P. D. Magnus has pointed out to me, the view I am now recommending for Reid
has the consequence that children’s perceptual beliefs qualify as genuine knowledge,
whereas their testimonial beliefs qualify only as animal knowledge. I am not alto-
gether happy with this invidious distinction, but I think it may be pressed upon us if
perception is an a priori source and testimony is not.
 PART I I
T E S T I M O N Y A N D I TS P L AC E I N
E PI S T E M O LO G Y
This page intentionally left blank
 3
The Epistemic Role of Testimony: Internalist
and Externalist Perspectives
Richard Fumerton

Setting aside radical skeptical concerns, it seems almost a truism that much of
what we believe is based on the testimony of others. Beliefs about the distant past
are based on the writings of historians. Beliefs about the microworld are based on
the word of physicists. Beliefs about the names, ages, histories, habits, likes, and
dislikes of friends are largely based on information those friends provide. There
are important distinctions one can make between kinds of testimony.¹ Through-
out this paper, however, I will be relying on a very broad understanding of the
term. Any genuine assertion one person makes for the consumption of another
will count, for these purposes, as that person’s testifying to some putative fact.
The assertion can be oral or written, formal and under oath, or casual in some
familiar context of conversation. Again, understood this way, the road to much of
what we believe travels through the testimony of other people.
Any plausible epistemology must distinguish questions about the genesis of
belief from its epistemic justification.² If it is relatively uncontroversial from the
perspective of commonsense that we very often rely on information provided
by other people, it is far less clear how to construe the nature of the evidential
path we need to travel in getting justified belief through reliance on testimony.
The traditional view of testimony and the way in which it contributes to justi-
fied belief makes the epistemic road long and winding. We hear sounds or see
marks. We then must reasonably interpret those sounds and marks as meaningful
assertions. Critically, we must have some reason to believe that the assertions in
question are likely to be true. Only then are we in a position to reach a rational
conclusion that takes into account what other people say. More recently, how-
ever, philosophers have begun to challenge the idea that the epistemic contri-
butions of testimony are so complex. Some have even seemed to suggest that
reliance on testimony should be viewed as just as fundamental as reliance on
inductive reasoning, memory, or perception.³
The attempt to determine how, if at all, testimony contributes to knowledge
is made more difficult these days as the debate takes place in the shadow of the
internalism/externalism controversy that dominates contemporary epistemology.
78 The Epistemic Role of Testimony

In this paper I want to contrast the way in which classical internalists and para-
digm externalists might approach the question of how to construe the justifica-
tion (if any) provided by testimony. In particular, I am interested in the question
of whether testimonial inference has a fundamental or a derivative place in our
reasoning. In the course of answering this question we will have occasion to
examine more closely this alleged distinction between fundamental and deriv-
ative principles of reasoning.

A C L A S S I C A L I N T E R N A L I S T F O U N D AT I O N A L IS M A N D
A T R A D I T I O N A L A P P ROAC H TO T E S T I M O N Y

On classical internalist foundationalist models of justification, the epistemic tasks

one must complete in order to justify belief based on testimony are intimidating
to say the least. There are, of course, radically different versions of internalism. I
have argued elsewhere (Fumerton 1985 and 1996) that one of the most demand-
ing, but also one of the most plausible, takes S’s foundationally justified beliefs
to be those justified by S’s direct acquaintance with a correspondence between a
belief/thought and the fact that makes it true. That version of foundationalism
need not be a version of internal state internalism, at least if the internal states
of a subject are defined as those states that include as constituents only the self,
its nonrelational properties, and those relations it bears to itself and its nonrela-
tional properties. According to Russell (1959), for example, we can be directly
acquainted with universals and relations that hold between them. Universals are
hardly mind-dependent entities. Also, direct realism as a theory of perception
is making a bit of a comeback. At least some philosophers hold that in veridic-
al perception we are directly acquainted with mind-independent constituents of
physical objects.⁴
On a direct acquaintance theory, one’s stock of foundationally justified beliefs
is a function of the class of facts with which one can be directly acquainted. The
radical empiricists, of course, were convinced that the only contingent truths one
knows noninferentially are descriptions of the current contents of one’s mind.
They were convinced that one is directly acquainted only with one’s own subject-
ive states and the truth makers for necessary truths. That version of the acquaint-
ance theory would take one’s justification for believing contingent truths to be
constituted by one’s internal states.
However the classical foundationalist construes foundational knowledge, that
foundationalist will also need an account of how one moves from foundational
knowledge to the rest of what one justifiably believes. This is another issue which
separates paradigm internalists and externalists. On the view that I call inferential
internalism, in order for S to be justified in believing P on the basis of some other
proposition justifiably believed E, S must be justified in believing that there is a
probabilistic connection between E and P (where entailment can be viewed as the
upper limit of making probable).⁵
 Richard Fumerton 79

Inferential internalism can be made at least initially attractive if we think

about some commonplace epistemic criticisms of inferentially formed beliefs.
The astrologer who forms beliefs about the future of human beings based
on information he possesses concerning the positions of planets and stars has
unjustified beliefs if he lacks reason to believe that there is some sort of
probabilistic connection between the position of celestial bodies and the affairs
of human beings. One who infers that a person will have a long life from the
observation that the person has a long ‘‘life line’’ on the palm of her hand, forms
an irrational belief in the absence of possessing good reason to think that a line on
one’s palm has some connection to the length of one’s life.
One must be cautious, however, before one relies too heavily on the above
intuitions to defend inferential internalism.⁶ The problem is that we are promis-
cuous in our characterization of the evidence from which we infer conclusions.
In casual conversation we often identify as our evidence for believing some pro-
position only a particularly interesting part of the entire body of knowledge upon
which we rely in reaching our conclusion. I call the police and report that I have
been robbed. If asked why I think that is so, I might cite as my evidence that
my window has been broken and my valuables are missing. But it seems fairly
clear on reflection that there is a vast array of unstated background information
upon which I rely in reaching my conclusion. I know that I live in a culture
in which it is not acceptable for friends and relatives to borrow my belongings
with or without my permission and to use force if necessary to enter my home in
order to achieve that end. I also know that glass does not typically break spontan-
eously causing valuables to disappear in nihilo. I also know (or at least think that I
know) that I am not a psychotic prone to staging robberies which I subsequently
can’t remember having staged. To be sure, I probably don’t consciously bring all
these beliefs to mind in reaching the conclusion that I’ve been robbed, but it is
perfectly plausible to suppose that their presence as dispositional beliefs plays a
crucial role in my willingess to reach the conclusion that I’ve been robbed.
Once we realize that many, if not most, accounts of our reasoning are
enthymematic, it is far less obvious what lesson we should learn from the fact
that we reject astrological ‘‘reasoning’’ as illegitimate in the absence of justified
belief concerning connections between celestial events and human affairs. It is
more than likely that we, and for that matter the astrologers, view their reasoning
as enthymematic. No one, not even an astrologer, is crazy enough to suppose
that one can somehow legitimately infer that Susan is going to meet the love
of her life from the fact that she was born in December and Jupiter is aligned
with Mars. Everyone simply assumes that the information concerning birth
dates and planets is only a part of a more complex array of premises which
constitute the complete story of the alleged evidence justifying the astrological
prediction. Furthermore, it is natural to assume that the unstated premises
make assertions about correlations between celestial facts and human affairs. We
reject the astrologer’s conclusions as unjustified not because the astrologer lacks
80 The Epistemic Role of Testimony

justification for believing that the premises of his argument make probable the
conclusion, but because we are convinced that the astrologer lacks justification
for believing a critical but unstated premise upon which he was relying.
Having said all this, I still think that inferential internalism is a plausible
view. Deductively valid arguments are surely the paradigm of arguments whose
premises bear an appropriate relation to their conclusion. And it still seems
obvious to me that if someone infers a conclusion C from some known premise E
when E entails C, it doesn’t follow that the person has a justified belief in C based
on E. If the person in question fails to ‘‘see’’ the connection—doesn’t realize that
the entailment holds—then the person lacks inferential knowledge.

T H E S TAT U S O F T E S T I M O N I A L “ I N F E R E N C E ” O N
T H E “ T R A D I T I O N A L” V I EW

My main concern here is not with the plausibility of inferential internalism.

Rather, I want to see what lessons we can learn from the above discussion con-
cerning the character of inferences that rely on testimony. And the first conclu-
sion we might reach is that the whole idea of relying on testimony as a kind
of inference is potentially misleading. Inferences from testimony might be like
astrological ‘‘inferences’’. We do not, in fact, ever infer that P from the fact that
some person tells us that P. To be sure we talk that way. We’ll identify as our
evidence for believing some proposition the testimony of another person. But a
full and perspicuous representation of the reasoning will make explicit unstated
premises upon which we critically rely in reaching our conclusion that P. The
perspicuous representation of the argument we accept might be something like
the following:
1) Jones said that P in conditions C (where C includes a description of
Jones, his qualifications as an authority, and the circumstances under
which he made the assertion).
2) People who make assertions like P in conditions C are usually saying
something true.
Therefore,
3) P
That something like the above is the more perspicuous representation of the
reasoning involved in reliance on testimony is also suggested by careful character-
izations of the supposed fallacy of appealing to authority. I began this paper by
observing that we rely on authority for much of what we believe. It would indeed
be a bit unfortunate if this reliance involved some sort of fallacy. The writers of
texts on informal fallacies understand clearly that they had better be careful in
characterizing the fallacy in question. As a result, they typically describe it as the
fallacy (better, mistake) of relying on the testimony of someone who is not in
 Richard Fumerton 81

fact a reliable authority concerning the subject of his testimony. An epistemically

more sensitive characterization of the mistake would presumably make reference
instead to the epistemic position of the person relying on the testimony. An epi-
stemic mistake is committed only if one relies on an authority when one has no
reason to believe that the person in question is reliable. But if that’s the mistake
that constitutes the fallacy of appeal to authority, the clear implication is that
when one is not making the mistake one does have some good reason to believe
that the person is reliably testifying—that is, one has some good reason to believe
our premises 1) and 2) above.
Earlier, I suggested that some recent work on testimony appears to focus on
whether testimonial inference is fundamental or derivative. I’m now suggesting
that this way of putting the controversy is needlessly confusing. In general, there
may be no useful distinction between fundamental and derivative principles of
reasoning. It’s harmless enough to suggest that there is an epistemic rule per-
mitting an inference from the litmus paper turning red in the solution to the
conclusion that the solution is acidic, a rule which had to be learned from exper-
ience, and is in that sense derivative. But it is surely more perspicuous to suggest
that there is no rule of inference at all sanctioning an inference from the propos-
ition that the paper is red to the conclusion that the solution is acidic. Rather,
the representation of the reasoning in question is enthymematic. It is only in
conjunction with certain background information that the premise describing
the litmus paper allows us to draw the conclusion. The most obvious unstated
premise is one that describes a correlation between the change in the litmus
paper and the character of the solution. The reasoning, when described fully, is
either deductively valid (if the premise takes the form of a universal proposition
describing the correlation) or inductive (if the premise describes merely observed
or statistical correlations). Either way there is no need to recognize ‘‘secondary’’
epistemic principles sanctioning ‘‘litmus reasoning.’’ We need only keep in mind
that our descriptions of our reasoning are often (indeed, in ordinary discourse,
almost always) enthymematic.
So the most perspicuous characterization of the traditional internalist’s ap-
proach to understanding the role of testimony in acquiring justified belief is
probably that strictly speaking there is no testimonial inference at all. When
we make explicit critical unstated premises, we find that the reasoning that takes
account of testimony is just some other familiar sort of deductive or nondeduct-
ive reasoning that employs at least one premise describing what other people say.
As we noted earlier, on the classical internalist/foundationalist model, the
way in which information about what others say can legitimately be taken into
account in reaching conclusions is complicated indeed. On a radical empiricism
we first need to reach a justified belief that there are real mind-independent
sounds and marks (based on what we seem to hear and see). I am not about
to discuss the problem of perception here, so we’ll just suppose for the sake of
argument that we don’t have any difficulty getting to knowledge of an external,
82 The Epistemic Role of Testimony

mind-independent reality. But our work has just begun. We need some reason to
believe that those sounds and marks are meaningful symbols.
What is involved in rationally believing that sounds or marks are representa-
tions of reality? That also is a question that would take us far afield. To answer
it we’d need a general account of representation and intentionality. On classical
views (which I think are almost obviously correct), we need to draw a distinction
between signs that represent only by convention and signs that are in some sense
‘‘natural.’’⁷ ‘‘Cat’’ represents a certain kind of animal but only because human
beings assigned the mark or sound a certain task. If we collectively decided that we
wanted that symbol to represent something else, we would need only to reach an
alternative agreement. It used to seem obvious to almost all philosophers that not
all symbols could represent by convention. Indeed, it seems plausible to suppose
that unless we could independently think of both the symbol and what we use it
to stand for, conventional representation would be impossible. But if that’s right
and thought itself represents only conventionally, we face a vicious regress. To
assign thought its representational role we’d have to think of the thought and that
for which it stands. To end the regress we need to recognize that there is a way of
representing the world that does not rely on convention. On the traditional view,
conventional representation presupposes a ‘‘language’’ of thought that represents
naturally.
If the above is right, we now need an analysis of ‘‘natural’’ representation.
In virtue of what does something X represent naturally something else Y? The
internalism/externalism debate in epistemology is paralleled by a similar debate
in philosophy of mind. Painting with a very broad stroke, most externalists are
naturalists who attempt to understand representation employing causal analyses.
X represents Y in virtue of the fact that X is nomologically tied to Y in certain
ways. The devil is, of course, in the details. Information theoretic accounts have
labored long and hard to tell us how we single out from among the vastly complex
chain of causes and effects that produce some brain state the one that is represen-
ted by the brain state. Depending on just how the account goes, it may be possible
to get language back on the side of natural representation. The import of much
of Putnam’s work (1975, 1978, 1988) (and before him Sellars (1954) ) is to deny
that one needs a radically different account of how thought represents from the
account one gives of how language represents. The use of word tokens can stand
at the end of causal chains just as surely as can images in the mind, or neurons fir-
ing in the brain, and if they occupy the right place in the right causal chains they
can represent in precisely the same way that images or brain states can represent.
Waiting in the wings for the collapse of naturalistic accounts are ‘‘magical’’
theories—theories that maintain that certain states of mind (and only states of
mind) have intrinsic and sui generis content.⁸ Intentional states are unlike any-
thing else simply in virtue of having the capacity to correspond to reality. That
capacity to correspond defies any sort of reductive analysis. The view is derided
as ‘‘magical’’ just because the critic is convinced that these peculiar states are
 Richard Fumerton 83

dragged into the picture as a kind of deus ex machina to solve fundamental prob-
lems concerning intentionality. Of course, almost every philosopher who engages
in analysis will admit that analysis must begin somewhere. There must be con-
ceptual building blocks if we are to understand anything, and proponents of the
magical theory should not apologize for the fact that something as mysterious as
thought cannot be assimilated to any other natural phenomenon.
Again, we cannot expect to resolve the most fundamental questions concern-
ing intentionality here. Our only concern is to point out that the nature of the
epistemic task we need to complete in construing sounds or marks as testimony
seems to depend directly on the account of intentionality one puts forth. The
naturalists will suppose that the task of discovering that certain sounds or marks
have the content they do is the task of discovering complex facts about their
causal origin. The proponent of the magical theory who thinks that only thought
represents naturally will be convinced that the key to correctly interpreting appar-
ent language is to come up with the right hypothesis about what states of mind
occur in the person (or people) who produce the relevant sounds or marks.
Suppose, for the sake of argument, that the internalist is right. The meanings
of sounds or marks are to be found in the head, or better in the minds, of those
who produce those sounds and marks. To interpret reasonably those sounds or
marks we now need to solve the problem of other minds. We need to find some
reason to suppose that those marks have the meaning we take them to have. That
will minimally involve figuring out what conscious states were involved in their
production. When the language we are interpreting is our own—without beg-
ging the question, when the sounds or marks appear to be of the same type as
those we ourselves use to express thoughts—one might suppose that our best
hope is to rely on some version of an argument from analogy or an argument
to the best explanation. And indeed, I think that’s probably right. The symbols
‘‘There’s a mountain nearby’’ as used by me express a thought with which I am
introspectively acquainted. Suppose I’m on the first expeditionary trip to Mars
and upon setting foot on the planet immediately notice clearly etched on a rock
face the symbols ‘‘There is a mountain nearby.’’ While amazed and bewildered, I
have no doubt that I would be irresistibly inclined to think that this was indeed
a message, and that the message had the same content as the message in my lan-
guage. I’d probably also think that there was a decent chance that the mountain
was nearby. And notice that I’d probably think all this even if I had absolutely no
further explanation of how the marks appeared. My thought would probably be
that it would be such a bizarre coincidence that marks with that form and syntax
appeared without having some connection to my own symbols with the meanings
they have, that any other hypothesis is simply unbelievable. Note well that I don’t
have to be very clear at all about the connection in question. It could be that what
I take to be our respective symbols acquired the meanings they have due to some
unknown common cause. It could be that one of us somehow caused the other
to adopt the relevant conventions. But the existence of some connection or other
84 The Epistemic Role of Testimony

just seems more plausible than the detailed convergence of form and syntax that
would otherwise be mysterious.
The above is not intended to be a very convincing argument. At the risk of
setting aside all of the really difficult epistemological problems, I’ll again simply
beg off solving the difficult questions concerning arguments by analogy and argu-
ments to the best explanation that would need to be explored in depth as part
of any serious attempt to uncover the epistemic justification we possess for inter-
preting the familiar symbols we encounter. I will note in passing that the ap-
proach sketched above is committed to rejecting Davidsonian (1984) arguments
(also embraced by Coady) for the idea that we must inevitably presuppose the
truth of what people say (and also that what they say has a certain content) if we
are to arrive at translations. If what is said above is correct, there may be a much
cruder way of arriving at conclusions about meaning that are at least prima facie
plausible.
On the traditional approach, reaching a rational conclusion about what sym-
bols mean and that they are used as assertions is, of course, not the last step in
the implicit reasoning involved in reliance on testimony. We need some reason to
believe that the assertions are likely to be true. Given a radical empiricism, that
reason will need to be traced again back to what we know about ourselves. Setting
aside again some more extreme skeptical concerns,⁹ we find ourselves in epistemic
situations in which there are certain truths that it is pretty easy for us to come to
know. When we have reason to believe that there is another person in a similar
epistemic situation, we have reason to believe that that person would have similar
access to those kinds of truths. For example, when conscious I know when I have
a headache and when I don’t. I expect that if you are remotely like me you are
similarly positioned to know whether or not you have a headache. I also know
of myself that I don’t usually lie about such matters, and in the absence of any
other information, therefore, I will, based on analogy, take what you say about
your headache as a pretty good indicator of truth.
I deliberately started with testimony about the simplest of truths. Things get
more complicated the more controversial the background assumptions about epi-
stemic position get, and the more controversial background assumptions about
motives to mislead are. All this mirrors precisely the controversies we encounter
concerning whether or not to rely on testimony. Hume was quite right in sug-
gesting that there are all kinds of live hypotheses as to why people might testify
falsely concerning the occurrence of miracles. Alibis provided by mothers and
lovers of the accused don’t carry nearly as much weight as alibis provided by
people the accused doesn’t even know. The testimony of other philosophers con-
cerning the truth of their philosophical views carries almost no epistemic weight
at all for philosophers when it comes to evaluating those views.
None of this is very original. While I do think that the traditional approach
to understanding the evidential role of testimony is quite right—indeed almost
obviously right—my primary concern in this paper is to see what alternatives are
 Richard Fumerton 85

available to both internalist and externalists. More specifically how might one
try to find room within one’s epistemology for genuine fundamental testimonial
inference? The search for sui generis fundamental epistemic principles that sanc-
tion moves from hearing testimony to forming beliefs might be motivated partly
by phenomenology. Classical foundationalists have often been accused of radic-
ally over-intellectualizing the processes by which we form beliefs. The view that
we reach conclusions about the objective external world based on truths we dis-
cover about subjective and fleeting experience has often been criticized on the
grounds that we rarely even pay attention to subjective appearance. There may be
an appearance/reality distinction, but it takes a certain skill—the kind of skill
acquired by painters, for example—to even notice the many and subtle ways
in which appearance is constantly shifting. The person who actually wanders
around consciously inferring truths about his physical environment from truths
about appearance is probably not destined to stay that long in this world. If you
can’t react instinctively to sensory stimulation with the quick realization that the
bus is bearing down on you, you’ve had it.
Just as it seems implausible to suppose that our beliefs about objects directly in
front of us are produced through inference from truths we notice about appear-
ance, so also it might seem implausible to suppose that in relying on testimony we
travel anything like the long and winding road postulated by the radical empiri-
cist. When I’m on the golf course and hear someone yell ‘‘Fore,’’ I’d better duck.
If I stand there trying to complete the steps of a rational reconstruction of my
ultimate reliance on the ‘‘testimony’’ provided by that golfer’s warning, I’m in
serious danger.
It’s not clear that the traditional approach need worry much about this alleged
phenomenological data. Earlier, I suggested that there is no difficulty in sup-
posing that dispositional beliefs can play a critical causal role in both producing
and sustaining belief. I need not consciously rely on background information in
forming some conclusion for my background beliefs to be playing the critical
causal role. Just as I have background beliefs about truths that can serve as impli-
cit premises, so also I may have background beliefs about inferential connec-
tions. Furthermore, those dispositional beliefs may concern particular inferences
rather than general epistemic rules. If anything is obvious, it is that the mind
is extraordinarily complex and it would surely not be surprising if much of the
inference that takes place does not take place at the conscious level.
Still, one might worry about the fact that the traditional foundationalist’s
reconstruction of reliance on testimony requires so many problematic steps.
Traditional foundationalism generally is fertile ground for skepticism. It might
be a relief if we could understand the way in which testimony contributes to
rational belief in a more straightforward way. But the world doesn’t always
cooperate to make life easy, and we need to figure out whether it is at all plausible
to suppose that there is a more straightforward epistemic route from hearing
testimony to forming rational beliefs. In what follows, I want to emphasize that
86 The Epistemic Role of Testimony

the prospects for finding that more straightforward route are directly related to
one’s position on the internalism/externalism debate and one’s corresponding
position on the content and modal status of epistemic principles.

T H E M O D A L S TAT U S O F E PI S T E M I C P R I N C I P L E S
A N D I N T E R N A L I S T A N D E X T E R N A L I S T P RO S PE C TS
F O R R E C O G N I Z I N G I N D E PE N D E N T T E S T I M O N I A L
REASONING

We haven’t said much about epistemic principles. I’ve hinted already that the
distinction between fundamental and derivative epistemic principles might itself
be spurious. Arguments that employ so-called derivative epistemic principles are
probably better thought of as enthymematic arguments governed by legitimate
epistemic principles that license the inference from premises to conclusion. Strictly
speaking derivative epistemic ‘‘principles’’ aren’t epistemic principles at all.
Epistemic principles can be thought of and described in a number of different
ways. Consider the following:
1) If S has property X then S is justified in believing P.
2) S’s having property X makes prima facie probable P for S.
Property X can be understood as broadly as you like. I don’t want to prejudice
any questions concerning what can justify a belief. So property X can be the
property of having other beliefs, having other justified beliefs, being in a cer-
tain conscious state, having a brain state with a certain causal origin, or what
have you. Principles governing inferential justification presumably license infer-
ence from believing one proposition justifiably to believing another. Principles
governing noninferential justification license belief when one is in certain non-
doxastic states. 1) and 2) might be just alternative ways of saying the same thing,
but it is nevertheless important to be clear about which epistemic concept one
takes to be conceptually fundamental. Chisholm and his followers, for example,
clearly take as primitive certain epistemic properties of belief—specifically the
comparative property of being more reasonable to believe than. Keynes (1921)
and his followers took the most fundamental concept in epistemology to be the
logical concept of probability. On Keynes’s view there are relations of making
probable holding between propositions that are directly analogous to relations of
entailment holding between propositions.¹⁰ When we make a reasonable infer-
ence, the rationality of belief in our conclusion is in part a function of our ability
to ‘‘see’’ the relation of making probable holding between our premises and our
conclusion. On Keynes’s view, propositions asserting probability relations are
necessary truths knowable a priori. Such a view would be a godsend to inferen-
tial internalists who are convinced that inferential justification requires awareness
of probabilistic connections between premises and conclusions. The inferential
 Richard Fumerton 87

internalists must obviously terminate a potential regress when it comes to getting

justification for believing that the relevant inferential connections obtain. If pro-
positions asserting those connections are necessary truths knowable a priori, it is
plausible to suppose that knowledge of such connections can be foundational.
There are, of course, alternatives to construing the probability appealed to in
2) as Keynesean logical probability. One could hold that the relevant probability
has something to do with frequency. Roughly, the idea is that we’d have to assign
exemplifying the pair of properties, being X and believing P to a pair kind. We
could then understand the probability claim as asserting that usually when the
first member of the pair kind is instantiated by a subject, the second member (the
belief ) is true. The attempt to construe the probability appealed to in 2) in terms
of frequency (or propensity) in effect makes 2) as a statement of an epistemic
principle a version of reliabilism.¹¹ Any view that takes epistemic principles to be
assertions of probability and then understands the relevant probability in terms
of frequency will inevitably render epistemic principles contingent truths know-
able only a posteriori. With such a view one would do well to eschew inferential
internalism for there will be no plausible way to terminate a regress of justifica-
tion that arises in connection with the possibility of justifying belief in inferential
connections.
With the above as background, let us return to testimony. As we saw, on one
rather natural understanding of the role of testimony in justifying belief there is
no need to recognize epistemic principles taking the form of 1) or 2) that gov-
ern specifically testimony. Reasoning from testimony can easily be construed as a
species of some other familiar sort of reasoning (inductive reasoning or reasoning
to the best explanation), a form of reasoning which includes among its premises
information about what other people assert. Is there an alternative to that view?
The answer, of course, will depend in part on what one takes an epistemic prin-
ciple to assert and on what one takes its modal status to be. Let’s consider some
candidates for an epistemic principle governing testimony.
T) When R hears a sentence ‘‘S’’ (e.g. the words ‘‘There is a dog outside’’)
in the sort of conditions that characterize a context of genuine assertion,
that makes probable for R that S (e.g. that there is a dog outside.)
T) certainly seems an unlikely candidate for the kind of synthetic necessary truth
that a Keynesean would take a genuine epistemic principle to assert. Clearly there
are all sorts of possible situations in which hearing those words would not make
even prima facie probable (in any sense relevant to epistemology) the truth of
the proposition that there is a dog outside. For one thing, there are all sorts of
situations in which those words have an entirely different meaning.
Don’t confuse the above with a quite different and clearly bad argument.
One might suppose that no epistemic principle asserting a probability connection
could be a necessary truth. After all, the fact that E only makes probable P
suggests that E can be true while P is false. But if that is so then it can hardly
88 The Epistemic Role of Testimony

be a necessary truth that E makes probable P. But that is to confuse the modal
status of the conditional (If P then Q) with the modal status of the claim that P
makes (prima facie) probable Q. The fact that I vividly seem to remember having
a headache earlier might make it probable that I did have the headache. Can we
imagine a world in which my seeming to remember that experience doesn’t make
it prima facie likely that I had the experience? Well, I can certainly imagine a
world in which I seemed to remember having the experience when I didn’t have
it. I can probably even imagine a world in which the conjunction of my seeming
to remember having the headache together with some other proposition (e.g. that
my memory is hopelessly bad and I’m prone to ‘‘hallucinatory’’ memory states)
doesn’t make probable that I had the headache. But all that is perfectly consistent
with its being true—indeed, necessarily true—that my seeming to remember
having the headache makes it prima facie likely that I had the headache. Again,
it is not the least bit plausible to suppose it is a necessary truth that my hearing
the sounds ‘‘there is a dog outside’’ makes it likely for me that there is a dog
outside and if genuine epistemic principles are necessary truths, T) isn’t a genuine
epistemic principle.
The situation is much more complicated if we adopt the position that the
relevant epistemic principle simply asserts a statistical correlation of some sort
between the processing of certain kinds of input and the truth of output beliefs.
On most versions of reliabilism there is no a priori restriction on what can count
as an unconditionally or a conditionally reliable belief-forming process. Plantinga
(2000) points out, quite correctly, that there might be a Holy Spirit who is
causally responsible for one’s acquiring true belief in the existence of God. Should
such a being exist and be causally active in producing true beliefs about God’s
existence, the resulting belief would be a prime candidate for a noninferentially
justified belief (noninferentially, because the input, by hypothesis, involves no
justified beliefs).
Have speakers evolved in such a way that when they take as input auditory
experiences of certain symbols they immediately and unreflectively believe what
they take the symbols to assert? And critically, does this process typically res-
ult in true beliefs? It is not perhaps wildly implausible to suppose that there
may be such ways of forming beliefs that are generally reliable. A great deal (as
always) depends on how the frequentist/reliabilist addresses the generality prob-
lem—how they specify in detail the relevant input–output mechanism. But if
one includes enough mundane situations—situations in which people give you
relatively unproblematic information about the time of day, the weather, their
names, their ages, and the like—it may be plausible to suppose that something
like T) is true when the probability is understood in terms of frequency. It is, of
course, an empirical question. In fact, as I suggested earlier, I suspect that there
are all sorts of background beliefs playing a critical causal role in the resulting
‘‘output’’ beliefs. I doubt therefore that T) accurately describes an actual process
of forming beliefs that we employ. Even if T) were true, if we are never actually
 Richard Fumerton 89

induced to believe a proposition based on the satisfaction of the conditions de-

scribed in T)’s antecedent, we are not getting justified beliefs by implicitly ‘‘fol-
lowing’’ T).¹² But I’m a philosopher not a psychologist/sociologist/cognitive
scientist. One can at least imagine our evolving in such a way that we are now
so constituted that we skip whatever intermediate premises upon which the tradi-
tional foundationalist thought we needed to rely.
From the perspective of one who takes epistemic principles to state neces-
sary truths, I said that T) above is a rather pathetic candidate for an epistemic
principle concerning testimony. But there are, of course, others. One could, for
example, modify T) as follows:
T*) When S hears someone say ‘‘There is a dog outside’’ and rationally
takes that sentence to be a sincere assertion that there is a dog out-
side, then it is prima facie reasonable for S to believe that there is a dog
outside.
On one way of thinking about it, T*) isn’t much more implausible than
doxastic conservatism—the view that a proposition acquires a certain prima
facie probability for S from the mere fact that S believes it. On T*) one takes
the fact that someone believes P to make it prima facie probable that P is
true. One must emphasize again that the proposition that someone’s sincerely
assenting to P makes probable P does not imply that P is probable relative to
everything we know. As we saw earlier, one must take account of all sorts of
other relevant truths in calculating the probability of P relative to one’s total
body of evidence. Controversial assertions automatically lose whatever probability
is conferred on them from the fact that some person sincerely assents to the
proposition asserted. What makes an assertion controversial, trivially, is that
there isn’t much agreement on its truth—some people assert the proposition
while others assert its negation. So P might be made probable by the fact that S
sincerely asserts it while not-P is made probable by the fact R asserts it. Relative to
the fact that roughly equal numbers of people assent as dissent from P, and those
facts alone, P is presumably no more likely than not-P. Relative to important
information I might possess, P’s probability might be 1 regardless of how many
people believe not-P (if for example P is a proposition describing some obvious
fact about my current conscious states).
Is T*) a candidate for the sort of fundamental epistemic principle to which
a Keynesean would be committed? Well I suggested that it is not much more
implausible than principles of epistemic conservatism. But then I’ve never found
principles of epistemic conservatism very plausible. It has never seemed to be
very plausible to suppose that the mere fact that someone believes a proposition
(even if I am that person) confers the least likelihood on the proposition’s being
true (leaving aside those trivial cases in which the having of the belief entails its
truth—e.g. the belief that there are beliefs). So I’m not inclined to think that
the Keynesean inferential internalists should recognize T*) as a true epistemic
90 The Epistemic Role of Testimony

probability principle having as much credibility as a principle of induction or

a principle of memory. But the issue really now hinges on deep and difficult
methodological issues in epistemology. In his famous discussion of the problem
of the criterion Chisholm (1966) suggests that we simply have to decide whether
or not we are going to take skepticism seriously. If we do not—if we take the
fact that a philosophical view leads to skepticism as a reductio of that view—then
we should adjust our epistemic principles until they allow us to achieve our non-
skeptical conclusions. If one thinks that the traditional story of how to trust the
testimony of others is fatally compromised by the need to rely on inductive argu-
ments that proceed from the limited sample of one’s own case, then one may
simply need to supplement the epistemic principles that entitle us to form beliefs
by adding to them a principle like T*). But we can’t be Keyneseans (or Chishol-
mians, for that matter) and ignore the supposed modal status of the principles
we add to our stock of epistemic principles in order to achieve desired epistemic
ends. Though Chisholm is more coy than is Keynes, both will in the final analysis
insist that epistemic principles state necessary truths knowable a priori. And it is
an odd justification of commitment to a necessary truth that we ‘‘need’’ it to get
where we want to go. The principle really should strike you as being necessary in
precisely the same way that other synthetic necessary truths (What is red all over
is not blue all over) strike you as necessary.
Once again T*) is a perfectly plausible candidate for a fundamental epistemic
principle on an externalist/frequency understanding of the reference to probab-
ility. I’m not really interested in acquiring the empirical evidence that would be
required in order to discover that people actually do process data to form beliefs
where the processing would accord with such a principle. That’s a task better left
to cognitive scientists, psychologists, or perhaps evolutionary theorists.

C O N C LU S I O N

The plausibility of recognizing a general and sui generis epistemic principle sanc-
tioning testimonial inference is directly proportional to the plausibility of an
externalist understanding of probability claims. Let reliabilism be a paradigm
of externalism. Just as reliabilism places no a priori restrictions on what kinds
of beliefs might be noninferentially justified (because there is no end of possible
belief-independent unconditionally reliable belief-forming processes), so also reli-
abilism places no a priori restrictions on what interestingly different kinds of
inferentially justified beliefs there are (because there is no end of possible belief-
dependent reliable belief-producing processes). I have argued elsewhere (1995)
that the very ease with which noninferential and inferential justification prolifer-
ates on most externalist views might give one pause. If one insists that inferential
justification requires awareness of inferential connections, then the prospects for
finding epistemic principles sanctioning sui generis testimony inferences are slim.
 Richard Fumerton 91

REFERENCES
Addis, Laird (1989), Natural Signs (Philadelphia: Temple University Press).
Bergmann, Gustav (1964), Logic and Reality (Madisons University of Wisconsin Press).
Brewer, Bill (1999), Perception and Reason (Oxford: Oxford University Press).
Coady, C. A. J. (1992), Testimony (Oxford: Oxford University Press).
Chisholm, Roderick (1966), Theory of Knowledge (Englewood Cliffs, NJ: Prentice-Hall).
Davidson, Donald (1984), Inquiries into Truth and Interpretation (Oxford: Clarendon
Press).
Fumerton, Richard (1985), Metaphysical and Epistemological Problems of Perception (Lin-
coln, Nebr.: University of Nebraska Press).
(1995), Metaepistemology and Skepticism (Lanham, Md.: Rowman and Littlefield).
(2002), Realism and the Correspondence Theory of Truth (Lanham, Md.: Rowman
and Littlefield).
(2004a), ‘Inferential Internalism and the Presuppositions of Skeptical Argument’,
in Richard Schantz (ed.), The Externalist Challenge (De Gruyter).
(2004b), ‘Epistemic Probability’, Philosophical Issues, 14 (2004).
Goldman, Alvin (1979), ‘What Is Justified Belief?’, in George S. Pappas (ed.), Justifica-
tion and Knowledge (Dordrech): D. Reidel Pub. Co.)
(1986), Epistemology and Cognition (Cambridge, Mass.: Harvard University Press).
(1988), ‘Strong and Weak Justification’, in James Tomberlin (ed.), Philosophical
Perspectives 2: Epistemology (Ridgeview Publishing Co.).
Huemer, Mike (2002), ‘Fumerton’s Principle of Inferential Justification’, Journal of
Philosophical Research, 27: 329–40.
Keynes, John (1921), A Treatise on Probability (New York: MacMillan).
Plantinga, Alvin (2000), Warranted Christian Belief (Oxford: Oxford University Press).
Russell, Bertrand (1959), The Problems of Philosophy (Oxford: Oxford University Press).
Sellars, Wilfred (1957), ‘Intentionality and the Mental’, in Herbert Feigl,
Michael Scriver, and Grover Maxwell (eds.), Minnesota Studies in the Philosophy of
Science, vol. ii (University of Minnesota Press).

N OT E S

1. See Coady (1992) for an extended discussion of different sorts of testimony and a
definition (1992: 32, 42) of what testimony involves. Coady’s definition is problem-
atic in that he seems to argue that S’s statement constitutes testimony only when that
S has the competence, authority, or credentials to state truly that P. In order to eval-
uate the epistemic worth of testimony we surely need a way of characterizing it that
leaves open the competency of the person who puts forth the testimony.
2. That the two are distinct doesn’t mean that they aren’t related in various ways. On
most views a justified belief must be based on adequate justification, and the basing
relation is often construed as causal.
3. See again Coady (1992). I’m not sure what precisely Coady’s final view is. He some-
times seems to suggest that one can know a priori that testimony is generally reliable
92 The Epistemic Role of Testimony

or at least prima facie credible (1992: 96). In other places he seems only to suggest
that testimony is a fundamental source of evidence on a par with perception and
memory (1992: 145). I’ll have more to say about the implicit distinction between
fundamental and derivative sources of evidence in what follows.
4. See e.g. Brewer (1999).
5. See Fumerton (1995, 2004a, 2004b) for an extended discussion of inferential intern-
alism.
6. I have profited enormously and been influenced heavily by Mike Huemer’s (2002)
thoughts on these matters.
7. The locution ‘‘natural sign’’ is used by Addis (1989). A version of the view defen-
ded by Addis was defended by Bergmann (1964). And variations of it are defended
by Fumerton (1985, 1995, 2002).
8. The expression ‘‘magical’’ theory was coined by Putnam.
9. In characterizing the skepticism as extreme, I do not mean to diminish its threat.
10. With the emphasis on analogous. There are, of course, important differences between
the quasi-logical relation of making probable that Keynes took to hold between pro-
positions. From the fact that P entails Q it follows that the conjunction of P and
any other proposition entails Q. From the fact that P makes probable Q it does not
follow that the conjunction of P with any other proposition makes probable Q.
11. By far the most sophisticated versions of reliabilism were put forth by Goldman
(1979, 1986, 1988). Note the discussion of justification rules in Goldman (1986)
and the similarity between that view and the view that takes epistemic probability to
be defined statistically.
12. Put another way, philosophers typically insist that for a belief that P to be justified
by a belief that E, the belief that E must be based on the belief that E. If basing is
to be understood, at least partially, in causal terms, then if that actual cause of my
belief that P involves far more than my belief that E it is misleading to suggest that
my belief that P is based on my belief that E.
 4
Liberal Fundamentalism and Its Rivals
Peter Graham

Many hold that perception is a source of epistemically basic (direct) belief: for
justification, perceptual beliefs do not need positive inferential support from oth-
er justified beliefs, especially from beliefs about one’s current sensory episodes.
Perceptual beliefs can, however, be defeated or undermined by other things one
believes, and so to be justified in the end there must be no undefeated undermin-
ing grounds. Similarly for memory and introspection.¹
Testimony-based beliefs are as indispensable as perception, memory, and intro-
spection-based beliefs.² Many of our testimony-based beliefs are epistemically
justified. Indeed, most of what we justifiably believe we believe, at least in part,
on the basis of comprehending and accepting the word of others.
The testimony debate is largely over whether testimony-based beliefs are
epistemically inferential or, like perception, memory, and introspection-based
beliefs, epistemically direct. One side holds that a testimony-based belief is
justified just in case the hearer has no reason to believe that the speaker is either
insincere or unreliable. The other holds that a testimony-based belief is justified
only if the hearer does possess positive reasons to think that the speaker is either
sincere or reliable or both.³ Advocates of the direct view include Burge (1993,
1997, 1999), Coady (1973, 1992), Dummett (1993), Goldberg (Chapter 6 in
this volume), McDowell (1998), Quinton (1973), Ross (1986), Rysiew (2000),
Stevenson (1993), Strawson (1994), and Weiner (2003) among others. It goes
back to Reid. Those who reject the direct include Adler (2002), Audi (1997,
2002, 2004, Chapter 1 in this volume), Kusch (2002), Lackey (2003, Chapter
8 in this volume), Lehrer (1994), Lyons (1997), Faulkner (2000), Fricker (1987,
1994, 1995, 2002, Chapter 10 in this volume), and Root (1998, 2001), among
others. It goes back to Hume.

An earlier version of this paper circulated under the title ‘‘Reductionism and Anti-Reductionism
about Testimony’’. Another version circulated under the title ‘‘Fundamentalism and its Rivals’’. I
have significantly altered the terminology from previous versions. I hope I have also made a number
of improvements. For comments that led to changes, I am grateful to Jonathan Adler, Jennifer
Lackey, Paul Hurley, Ted Hinchman, Brian Keely, Peter Thielke, Duncan Pritchard, and especially
Peter Kung and Robert Audi. The two referees for this volume indirectly prompted substantial
changes. I’m grateful to both.
94 Liberal Fundamentalism and Its Rivals

In this essay I articulate and defend a version of the direct view. I shall call
it ‘Liberal Fundamentalism’. The Liberal Fundamentalist holds (to be qualified
below) that it is a priori necessary that comprehending an attester’s presentation-
as-true that P confers justification on the recipient’s belief that P. There is a
Strong and a Weak version. The Strong version holds that (absent defeat) the
event or state of comprehending the attester’s presentation-as-true that P provides
on balance justification for the belief that P, whereas the Weak version holds that
the justification provided or conferred may fall short of on balance justification
(even if undefeated).
This paper has two parts. The first articulates Liberal Fundamentalism and
some of its central rivals. The second articulates and defends the Weak version.
The theme of the first part is that what one says about the testimony debate is
driven in large part by one’s overall theoretical orientation on the nature of epi-
stemic justification, including one’s epistemology of epistemology. The theme of
the second is that one particular version of the direct view (Weak Liberal Fun-
damentalism) is more plausible than two of its immediate rivals (Strong Liberal
Fundamentalism and Moderate Fundamentalism).
I use three new ideas. The first is a list of different versions of foundationalism.
The second is a new taxonomy of theories of epistemic justification. The third is
the distinction between pro tanto and on balance justification.

L I B E R A L F U N D A M E N TA L I S M

Four Versions of Foundationalism

Liberal Fundamentalism is a combination of two doctrines. The first is about
which epistemic principles (given below) are true, and the second is about why
they are true. The first doctrine I call ‘Liberal foundationalism’ and the second I
call ‘Intuitionism’. In the rest of this section I articulate Liberal foundationalism.
In the next I explain Intuitionism.
Different versions of foundationalism are defined by which of the following
epistemic principles they accept as true:
(AP) If it seems to S upon understanding P that P is self-evident or
necessary, then the belief that P is prima facie pro tanto justified.
(INT) If it introspectively seems to S as if S is occurrently having a sens-
ory, perceptual or otherwise conscious experience such and such,
and this causes or sustains in the normal way the belief that S
is experiencing such and such, then that confers justification on
S’s belief.
(DED) If S believes P and believes (P entails Q) and believes Q on the
basis of inferring Q from P and (P entails Q), then S’s belief that
Q is conditionally justified.
 Peter Graham 95

(MEM) If S seems to remember that P and this causes or sustains in the

normal way S’s belief that P, then that confers justification on S’s
belief that P.
(EIND) If S possesses a sufficiently large and representative (nonbiased)
inductive base where all (most) Fs are Gs, then were S to infer that
all (most) Fs are Gs on that basis, then S’s belief that all (most) Fs
are Gs would be conditionally justified by the inference.
(IBE) If S possesses one explanation that better explains S’s evidence
than any other available alternative explanation, then S is justified
in believing that explanation on the basis of the evidence.
(PER) If S’s perceptual system represents an object x as F (where F is
a perceptible property), and this causes or sustains in the normal
way S’s belief of x that it is F, then that confers justification on S’s
belief that x is F.
(TEST) If a subject S (seemingly) comprehends a (seeming) presentation-
as-true by a (seeming) speaker that P, and if that causes or sustains
in the normal way S’s belief that P, then that confers justification
on S’s belief that P.

One can find four versions of foundationalism embraced in the literature:

Reactionary, Conservative, Moderate, and Liberal foundationalism (the labels are
new, the positions are familiar). The Reactionary accepts the first three but no
more, the Conservative the first five, the Moderate the first six, and the Liberal
all seven.

Reactionary: AP, INT, DED

Conservative: AP, INT, DED, MEM, EIND, IBE
Moderate: AP, INT, DED, MEM, EIND, IBE, PER
Liberal: AP, INT, DED, MEM, EIND, IBE, PER, TEST

The Liberal foundationalist thus has three foundationalist rivals. The pure
coherentist is a rival to all foundationalist views. The pure coherentist rejects the
direct/inferential distinction altogether, and so rejects all of the principles listed
above. The pure coherentist embraces only COH:

(COH) If the belief that P is a member of S’s coherent set of beliefs R, then
S’s belief that P is justified to the degree that R is coherent.

I set aside coherentism about testimony in what follows. I intend to treat it else-
where.
Liberal Fundamentalists are Liberal foundationalists that give one of four pos-
sible answers to why the epistemic principles they embrace are true. I characterize
the four possible answers next.
96 Liberal Fundamentalism and Its Rivals

Four Theoretical Perspectives

Elsewhere I have developed and defended a new taxonomy of theories of epistem-
ic justification (forthcoming). On my new taxonomy, there are four different the-
ories on the nature of epistemic justification. Each theory or perspective provides
an answer to why any epistemic principle is true. I shall describe and make use of
that taxonomy here.
The standard taxonomy relies upon two distinctions: foundationalism vs.
coherentism and internalism vs. externalism. Though the standard taxonomy is
useful and important, my new taxonomy relies on two different distinctions.
The first concerns the relationship between justification and truth. The second
concerns the epistemic (a priori or empirical) and modal (necessary or continent)
status of the epistemic principles.
The first distinction involves the relation between justification and truth.
Everyone agrees that epistemic justification is connected to truth (Audi 1988;
BonJour 1999, 2002; Burge 2004). Disagreement emerges when one asks how it
is connected to truth. There are two possible answers: either justification makes
a belief objectively more likely to be true, or justification properly aims belief at
the truth (Audi 1988). An ‘‘Actual Result’’ theorist holds the former, a ‘‘Proper
Aim’’ theorist holds the latter. Actual Result theorists are like consequential-
ists about moral rightness where an act is right provided that it has good con-
sequences, and Proper Aim theorists are like non-consequentialists. One way
to get a grip on the distinction is to reflect on the demon-worlds objection to
reliabilism. If you think a subject fooled by an evil demon may still enjoy justi-
fied perceptual, memorial, inductive, abductive, and other beliefs (even though
they are not de facto reliably held), then you are more likely to be a Proper
Aim theorist. If, on the contrary, you think those beliefs cannot be justified
because not likely to be true, then you are more likely to be an Actual Res-
ult theorist. To telegraph, the Liberal Fundamentalist is a Proper Aim
theorist.
The second distinction involves the epistemic and modal status of the epi-
stemic principles. The ‘‘Fundamentalist’’ holds that the true epistemic principles
are conceptually necessary, a priori knowable truths. The ‘‘Non-Fundamentalist’’
rejects this; the true epistemic principles are contingent, only empirically know-
able truths. For example, if PER is true, the Fundamentalist thinks it is an a priori
necessary, conceptual truth, whereas the Non-Fundamentalist thinks it is only
a contingent, empirical truth. If PER is known to be true, the Fundamentalist
thinks this is a piece of philosophical knowledge, whereas the Non-Fundamenta-
list thinks this is a piece of empirical knowledge, part of the subject-matter of the
natural or social sciences.
The two distinctions are orthogonal; they determine four possible (and famili-
ar) theories of justification: Cartesianism, Reliabilism, Intuitionism, and Pragmat-
ism (see Fig. 4.1).
 Peter Graham 97

Actual-Result Proper-Aim
Fundamentalism Cartesianism Intuitionism
Non-Fundamentalism Reliabilism Pragmatism

Figure 4.1 Four Theoretical Perspectives on Justification.

Using our definitions of the two distinctions, we can define the four positions:
Cartesianism: a belief is justified only if held in a way which is a priori
known or knowable to either necessarily make the belief true
or make the belief true more likely than not in all worlds. The
way held confers justification only if it is a priori knowable
that it is either every-instance reliable or all-worlds reliable.
Reliabilism: a belief is justified only if held in a way that de facto makes
the belief more likely than not to be true in the actual circum-
stances of use. The way held confers justification only if de
facto reliable.
Intuitionism: a belief is justified only if held in a way that is a priori known
or knowable to constitute properly aiming belief at truth,
where ‘‘properly aiming belief at truth’’ means conformity
to a priori necessary epistemic principles (listed above), and
does not require de facto or all-worlds reliability.⁴
Pragmatism: a belief is justified only if held in a way that de facto con-
stitutes properly aiming the belief at truth, where ‘‘properly
aiming belief at truth’’ means conformity to our deepest held
norms of proper belief formation (where ‘‘our’’ can mean
the subject, the discipline, the community, the tradition, or
the species).
Although each perspective, as stated, only places a necessary condition on justi-
fication, I shall, for the sake of illumination, treat each perspective as placing both
a necessary and sufficient condition on justification.
Each perspective places conditions on what it takes for a belief held in a certain
way to enjoy justification. Each perspective explains why an epistemic principle
is true if true. It will also explain why a principle is false if false. It is easiest to
see this in the case of the Cartesian. The Cartesian will only accept, at best, AP,
INT, and DED, for only (some) a priori insight, introspection of one’s current
sensory episodes, and deductive reasoning, are likely candidates for ways of form-
ing and holding beliefs that pass the Cartesian test; they are the only three ways of
holding belief likely to be reliable in all worlds. The Cartesian will reject the other
principles as false.
98 Liberal Fundamentalism and Its Rivals

Which principles the other three theoretical perspectives would accept is much
harder to determine. The Intuitionist accepts only those principles that are a
priori, conceptually necessary truths, but it is not obvious right at the start which
ones pass that test and which ones do not. One aim of this essay is to contribute
to sorting out just which considerations are relevant and which ones are not to
determining whether TEST, for example, is a priori necessary.⁵
The Reliabilist accepts only those principles that govern de facto reliable meth-
ods of belief acquisition and retention. It is not the job of the philosopher to
figure out which ones are reliable, but rather the job of the cognitive scientist.
Which ones will show up on the Reliabilist’s list is an open question until the
empirical inquiry is complete. It is the job of the philosopher to analyze justi-
fication and reliability; it is the job of the scientist to discover which processes
are reliable.
The Pragmatist accepts only those principles that govern methods of belief
acquisition and retention that are individually or socially embraced as the right
methods. Which ones will show up on her list, is to be decided by the individual,
the sociologist or the anthropologist.⁶

Liberal Fundamentalism and its Rivals

The two pieces are now in place to characterize both the Liberal Fundamentalist
and certain rivals. The Liberal Fundamentalist is a Liberal foundationalist about
which epistemic principles are true and an Intuitionist about why they are true;
all seven epistemic principles are, for the Liberal Fundamentalist, a priori neces-
sary truths.
The Liberal Fundamentalist has a number of rivals. Within the Intuitionist
camp, the Reactionary, the Conservative, and the Moderate Fundamentalist are
all rivals. A Coherentist that accepts Intuitionism (and so AP as well as COH) is
also a rival. So too is the Pure Coherentist. Any non-Intuitionist view is, by defin-
ition, a rival; Cartesians, Reliabilists, and Pragmatists are ipso facto rivals, even
if they are sympathetic to, or even embrace, TEST. Any Non-Fundamentalist
view is, by definition, a rival; if you don’t believe epistemic principles are a pri-
ori knowable conceptual truths, then a fortiori you don’t believe TEST is a priori
necessary. If you are a Coherentist (Adler), a Reactionary (Fumerton), a Conser-
vative (BonJour, Feldman), a Moderate (Pollock, Huemer), a Cartesian (Fumer-
ton), a Reliabilist (Goldman), or a Pragmatist (Foley, Kusch), the Liberal Funda-
mentalist is one of your rivals.⁷
An interesting rival is the Moderate Fundamentalist. This is for three reasons.
Firstly, there are many Moderate Fundamentalists; it is a live position. Secondly,
the ‘‘testimony debate’’ (described at the opening) receives a sharp formulation
when characterized as the debate between the Moderate and the Liberal Fun-
damentalist; the Liberal thinks (many) testimony-based beliefs are direct, the
Moderate thinks they are all (necessarily) inferential. Lastly, the Moderate is a
 Peter Graham 99

close rival to the Liberal. If there is a deep and convincing reason to think the
Liberal cannot be right, it is a reason the Moderate should be able to articulate
consistent with her position. Moderate Fundamentalism is thus a live position
that is a clear occupant of one side of the testimony debate, and if there is a reason
not to be a Liberal, the Moderate should be able to advance it.
In the rest of this section I say more about the Moderate-Liberal debate. In
the next I explain why the Moderate is not entitled to make four particular argu-
ments against the Liberal. This shows that whether something is a good reason
for (or against) an epistemic principle is largely a function of which theoretic-
al perspective is true. If Intuitionism is true, some considerations are relevant,
others are not.
To better understand the Moderate-Liberal debate, consider first the parallel
disagreement between the Conservative and the Moderate. The Conservative
rejects PER; the Moderate embraces it. The Conservative thinks perception is
epistemically neutral: a perceptual representation is, in itself, no reason or ground
to believe anything at all about the external environment. For the Conservative,
a perceptual belief is justified only if it can be inferentially supported by other,
non-perceptual beliefs. Traditionally this means the subject must be able to infer
that the way things seem to her in perceptual consciousness is best explained by
the real world hypothesis. She cannot essentially rely upon any perceptual beliefs
as premises. She needs to be able to infer that how things introspectively seem
to her corresponds to the way they are in the world. If she can, she will have
epistemically ‘‘reduced’’ perceptual beliefs to beliefs based on introspection and
reason. Perceptual beliefs are, for the Conservative, epistemically inferential. The
Moderate, on the other hand, is not so demanding. Perceptual beliefs are, for
the Moderate, epistemically direct. The Conservative is a ‘‘reductionist’’ about
perceptual beliefs; the Moderate is an ‘‘anti-reductionist’’.
The Moderate-Liberal disagreement is analogous. The Moderate holds that
testimony-based beliefs, if justified, are justified inferentially on the basis of non-
testimony-based beliefs; comprehending the presentation-as-true of another is,
in itself, epistemically neutral. It is, as such, no reason or ground to believe that
what the speaker said is true (Pritchard 2004: 328–30). The subject must be
able to infer from non-testimony-based beliefs that testimony-based beliefs are,
for the most part, reliable or justified in order for her testimony-based beliefs to
be justified. The qualification ‘‘non-testimonial’’ is essential. The hearer cannot
appeal to testimony-based beliefs about the reliability of testimony in an inelim-
inable way for that would presuppose that (at least some) testimony-based beliefs
are justified without inferential support. If A’s say-so that P is, in itself, no reas-
on to believe P, then B’s say-so that A is trustworthy is, in itself, no reason to
believe A.
The natural way to ‘‘reduce’’ (inferentially support in the required way)
testimony would be for the hearer to appeal to his own first-hand experience of
the reliability of the particular speaker, or speakers of that kind, or of testimony
100 Liberal Fundamentalism and Its Rivals

in general.⁸ The hearer would have to (be able to) ‘‘reduce’’ his testimony-
based beliefs to beliefs purged of testimonial reliance, using either enumerative
induction or inference to the best explanation. If she could do it, her testimony-
based beliefs would thereby epistemically ‘‘reduce’’ to inductively based
(reasoned) beliefs, beliefs inferred from or based on a non-testimonial induction
base; justified testimonial beliefs just are beliefs ‘‘reductively’’ justifiable.
The Moderate is more demanding than the Liberal, just as the Conservative
is more demanding than the Moderate. The Liberal does not require the subject
to ‘‘reduce’’ testimony-based beliefs to non-testimony-based beliefs; the Liberal is
an ‘‘anti-reductionist’’ about testimony while the Moderate is a ‘‘reductionist.’’
There are two standard objections to reductionism about testimony. They
parallel two standard objections to reductionism about perception. The first is
the ‘‘paucity of evidence argument’’. The argument is that the reduction is not
possible, for actual agents do not possess enough first-hand evidence to carry
it out. Hence if ordinary testimony-based beliefs are, by and large, justified,
then ‘‘reductionism’’ (Moderate foundationalism, Fundamentalist or not) is
false (Coady 1992; cp. Fricker 1994, 1995; Lipton 1998; Lyons 1997). The
second is that even if the reduction is possible, requiring it is overly demanding;
the requirement to reduce hyper-intellectualizes testimonial justification (Burge
1993; Strawson 1994; cp. Adler 2002). These two objections parallel objections
to reductionism about perception. The first is that the ‘‘reduction’’ cannot
succeed; subjects cannot derive the justification for perceptual beliefs from non-
perceptual beliefs. The second is that the ‘‘reduction’’, even if it is possible, is too
demanding on ordinary subjects.
So far I have introduced two new ideas: the four versions of foundational-
ism defined in terms of the epistemic principles, and the four theories of the
nature of epistemic justification. I used those new ideas to describe the Liberal
Fundamentalist and her rivals: Cartesians, Reliabilists, Pragmatists, coherentists
and other foundationalist Intuitionists. I then compared the Liberal to a close
rival, the Moderate Fundamentalist. In the next section I show why four possible
arguments against the Liberal are ineffective on the assumption that at least the
first six principles (the ones the Moderate embraces) are a priori necessary truths;
i.e. on the assumption that Intutionism is correct. This will show that what one
thinks about which principles are true is largely a function of which of the four
theoretical perspectives one employs. The next section concludes the first main
part of the paper.

Four Arguments Against Liberal Fundamentalism

Let us assume that Intuitionism is correct, and also that at least the first six
epistemic principles are all a priori necessary truths, but that it is still an open
question whether TEST is also a priori necessary. In this section I explain why
four arguments against TEST are ineffective on that assumption.
 Peter Graham 101

The first argument goes like this. Testimony is not a necessarily reliable
process; error and deceit might outnumber truth and sincerity (Graham 2000a).
Hence it cannot be a priori necessary that comprehending the presentation-as-
true of another confers justification on belief in what the speaker presented-
as-true (cp. Adler 2002; BonJour 2002; Fricker 1994; Faulkner 2000). This
argument is ineffective on the assumption that Intuitionism is correct, for
necessary reliability is not a necessary condition upon justification (either
direct or inferential justification). The Cartesian places this condition upon
justification; the Intuitionist (and so the Moderate Fundamentalist) does not. If
Intuitionism is true (and so Cartesianism is false), the mere fact that testimony is
not necessarily reliable is neither here nor there.
The second argument has two versions. The first goes as follows. Perceptual
beliefs are de facto more reliable than testimony-based beliefs, hence perceptu-
al justification is direct and testimonial justification is inferential. The second
goes like this. Testimony-based beliefs inferentially backed by non-testimony-
based beliefs are more reliable than testimony-based beliefs without such backing.
Hence testimony-based beliefs without inferential backing from non-testimony-
based beliefs are not justified but beliefs with such backing are. Hence testimonial
justification is inferential and not direct. Both versions rely upon the principle
that differences in degree of reliability determine differences in epistemic kind
(inferential vs. direct) (cf. Goldman 1979, 1992; Pritchard 2004: 343–4). But
if the Moderate is right this principle is false. This is because introspection may be
more reliable than perception, and perception may be more reliable than memory
(with or without backing), but introspection, perception, and memory are
all, according to the Moderate, epistemically direct. The Moderate does not reas-
on from differences in degree of reliability to differences in epistemic kind. A
fortiori, if perception supported by other beliefs is more reliable than percep-
tion without such support, it does not follow that perceptual-beliefs without
such support enjoy no justification (that they are epistemically neutral). And
so, if testimony-based beliefs epistemically supported by other beliefs are more
reliable than beliefs without such support, it again does not follow that beliefs
without such support enjoy no justification as such, that they are epistemically
neutral.
The third argument goes like this. TEST is true only if testimony without
inferential backing is a de facto reliable belief-forming process. However, it is not.
Hence TEST is false. But the Moderate Fundamentalist cannot argue this way,
for the Moderate is not a Reliabilist. At best he can offer defeaters by appeal to
de facto reliability considerations. If Intuitionism is correct and Reliabilism is not,
considerations of de facto reliability do not in themselves determine what neces-
sary a priori epistemic principles are true. Reliability considerations must play a
part in a complete treatment of testimony (especially for testimonial knowledge),
but if Intuitionism is correct, de facto reliability (itself ) does not enter into wheth-
er a source does or does not confer justification.⁹
102 Liberal Fundamentalism and Its Rivals

The fourth argument goes as follows. Justified perceptual beliefs sometimes

defeat testimony-based beliefs. Hence perceptual beliefs are ‘‘epistemically prior’’
to testimony-based beliefs. Hence testimony-based beliefs depend upon positive
epistemic support from perceptual beliefs for justification; testimony is inferential
and not direct. The principle this argument relies upon is that if a belief from
source N sometimes defeats a belief from source M, source M cannot be direct
but must be inferential. The Moderate Fundamentalist, however, must reject this
principle. That is because the Moderate accepts that perception, memory, and
introspection are all sources of direct justification, but also must accept that per-
ceptual beliefs sometimes defeat memory beliefs, that memory beliefs sometimes
defeat perceptual beliefs, that introspective beliefs sometimes defeat perceptual
beliefs, that perceptual beliefs sometimes defeat introspective beliefs, that intro-
spective beliefs sometimes defeat memory beliefs, and that memory beliefs some-
times defeat introspective beliefs. The Moderate accepts INT, MEM, and PER,
and so holds that these sources produce justified beliefs. If they do, they also pro-
duce defeaters for other beliefs, sometimes beliefs from those very sources. That a
source is a source of defeaters for beliefs from another source, or even from itself,
does not show that the other source depends for justification on inferential sup-
port from another source, or even from itself. Applied to testimony, this means
that the argument does not show that testimony depends upon non-testimonial
sources for justification. The fact that my perception defeats your testimony does
not show testimony is inferential and not direct. Indeed, the fact that testimony-
based beliefs sometimes defeat perceptual beliefs does not show that testimony is
prior to perception. For example, you say you see a VW coming over the horizon
and I correct you. Then my testimony would defeat your perceptual belief. But
that possibility, though familiar, clearly does not show that you must inferentially
support (derive) all of your perceptual beliefs from testimony-based beliefs.
These four arguments all fall short if we assume that the first six epistemic
principles are a priori necessary, that at least Moderate Fundamentalism is cor-
rect. They all also fall short even if we assume Conservative Fundamentalism, for
the Conservative does not rely upon necessary reliability, differences in degree of
reliability, de facto reliability, or relations of defeat to determine epistemic kinds.
What if we were to assume Cartesianism (and so assume Reactionary Funda-
mentalism, for instance) instead of Intuitionism? That would show that TEST is
false, for testimony is not necessarily reliable (BonJour 2002). But it would also
show that PER and other sources of direct and inferential justification are false. It
would not give us a reason to think that TEST itself is suspect.
What if we were to assume Reliabilism or Pragmatism? Then, possibly, some
of the arguments just listed would show that there are (contingent) differences in
epistemic kind; they could show that while perception, memory, and introspec-
tion (or certain kinds of perceptual belief forming processes and methods, and
so on for memory and introspection) are direct, testimony is inferential. If testi-
mony is not de facto reliable but perception is, or perception is considerably more
 Peter Graham 103

reliable than testimony, then perception confers justification without inferential

backing (according to the Reliabilist) but testimony does not. Or if testimony-
based beliefs without inferential backing are held by the relevant community to
fall short of justification but beliefs with such backing are held to pass the test,
then (according to the Pragmatist), perception is direct but testimony is inferen-
tial (Goldman 1979, 1992).
What one thinks about the epistemic status of testimony is thus largely a func-
tion of what theoretical perspective one holds about the nature of epistemic justi-
fication generally and further what one thinks about the relevant facts. If you are
an Intuitionist, certain familiar considerations, I have just argued, do not under-
mine Liberal Fundamentalism. But if you are not an Intuitionist, then you think
Liberal Fundamentalism is ipso facto false and TEST is not a priori necessary.
TEST may still be an empirical contingent truth, but that depends upon which
rival perspective is correct and how the facts turn out. What you think about
testimony is driven in large part by what theoretical perspective on epistemic jus-
tification you think is correct. It will largely determine what considerations lead
you to embrace or reject TEST. If you are not an Intuitionist, certain consider-
ations are relevant. But if you are, those very considerations may be neither here
nor there.
I have discussed only four arguments against the Liberal here. There are a num-
ber of other possible arguments. I shall discuss one such argument in the next
section. I have discussed some other arguments elsewhere.¹⁰
I have now concluded the first main part of this essay. I have described Lib-
eral Fundamentalism and (some of) its rivals and shown that certain arguments
against it are ineffective on the assumption that Intuitionism is correct. I have
not, however, argued that Intuitionism is correct. That is a very large topic best
treated at length on another occasion. Nor have I argued that the Liberal is cor-
rect. That too is a large topic for another time. However, I do say a few things in
defense of the Liberal in the next part of the paper.

W E A K L I B E R A L F U N D A M E N TA L I S M

In this part I make use of the third new idea, the distinction between pro tanto
and on balance justification. I use it to distinguish Strong and Weak readings of
the epistemic principles, and then two versions each (Strong and Weak) of both
Moderate and Liberal Fundamentalism. I compare Weak Liberal Fundamental-
ism with its Strong Liberal and Moderate rivals. I then respond to an objection
against the Liberal. The objection does not require the falsity of Intuitionism for
its force; it is an objection the Moderate is free to lodge against her Liberal rival.
The Weak version prima facie avoids the objection while the Strong version does
not. Weak Liberal Fundamentalism thus emerges as the more plausible variant.
104 Liberal Fundamentalism and Its Rivals

Pro Tanto vs. On Balance

In this section I distinguish between pro tanto and on balance justification. I
should first say a few words about prima facie justification.
It is customary to qualify principles governing justification with the phrase
‘‘prima facie’’. Each principle given above thus should include ‘‘prima facie’’
in between the words ‘‘confers’’ and ‘‘justification’’. Prima facie justification is
defeasible justification. Further information may undermine or override the jus-
tification one holds for a belief. So if I seem to see a red apple on the picnic
table, then the experience, many would hold, confers perceptual justification on
the belief that there is a red apple before me. But if I am told or have reason to
believe that it is a fake, or if I remember taking mind-altering drugs just before
coming to the picnic, then my justification has been defeated. In order to regain
the justification from the perceptual experience I would need to defeat the defeat-
ers with even further information. I would have to have a reason to believe that
you are fooling, or that it is probably is not a fake, or a reason to believe that
the mind-altering drugs have no effect on how well I see things. If there are no
defeaters in the first place, or if the defeaters present have been defeated in turn
by further information, then the prima facie justification I enjoy is not defeated
(defeat is absent).
Is lack of defeat sufficient to convert prima facie justification into on balance
justification? The standard view is that it is. Witness Alston and Huemer:
I have proposed an account of the prima facie epistemic justification of beliefs. . . . The
justification will be ultima facie provided there are not sufficient overriders from within
the subject’s knowledge and justified beliefs. (Alston 1988: 227)
When a belief is said to be prima facie justified . . . the belief ’s justification can be defeated
by countervailing evidence. . . . The appearances are presumed true, unless proven false.
This means that when it seems as if P and no evidence emerges contravening P, it is
reasonable to accept P. (Huemer 2001: 100)
Though commonplace, I believe this is incorrect, at least for both perceptual
and testimonial justification. Undefeated prima facie justification is not a priori
equivalent to on balance justification. I distinguish pro tanto from prima facie jus-
tification. I then distinguish Strong from Weak versions of both Moderate and
Liberal Fundamentalism. The Strong view holds that undefeated prima facie jus-
tification necessarily converts to on balance justification. The Weak view does
not. In the rest of this section I shall focus on perception, turning to testimony in
the sections following. The Weak view for perceptual justification is supported by
the examples that mark the distinction.
‘‘Pro tanto’’, as I understand it, means ‘‘as far as it goes’’ or ‘‘to that extent’’.
I contrast it with ‘‘on balance’’. A pro tanto justification is a consideration in
favor of a certain belief. If a certain belief is pro tanto justified by a perceptual
 Peter Graham 105

experience it does not follow that the belief is justified on balance. A pro tanto
justification may only justify the belief to a certain degree. Epistemologists have
shown sensitivity to this distinction (e.g. Audi 2001; Pritchard 2004). Those
sensitive to the distinction, however, have not marked it. Its significance should
not be overlooked.¹¹
The distinction between some evidence and enough is obvious in the case of
inductive reasoning. The distinction between some and enough also applies to
cases of psychologically non-inferential, prima facie justified, perceptual beliefs.
I will here show by a series of steps that the justification one has for an empir-
ical perceptual belief comes in degrees. This is, of course, widely believed. But
it implies that there is a point at which the justification converts from pro tanto
to on balance justification, for it is also widely believed that many perceptual
beliefs—beliefs held roughly as strongly as each other—are on balance justified,
but not that all of them are (Goldman 1979).
Imagine looking through a narrow steel pipe and seeming to see only the out-
lines of a red apple. Here your visual experience is focused in on a single object
and you have no other information about any other objects and relations before
you. The rest is, as it were, all dark; you can’t see anything else because the pipe
is too narrow. Here it seems all you have justification for is that there is probably
a red apple out there, and little justification for beliefs about its particular size or
distance, whether it is sitting on something or being held up, and so forth, from
the experience itself. You normally use other information that is presently lack-
ing to help figure those things out. Suppose the experience causes you to believe
(automatically and non-inferentially in the normal way) that there is a red apple
of ordinary size a certain distance away. The experience confers prima facie pro
tanto justification on the belief, but surely not on balance justification.
Now imagine removing the pipe and picking up more information. You can
now see much more of the scene; your experience represents a good deal more
than before. The visual experience of the red apple continues to confer some jus-
tification on the belief. The other parts of the enlarged experience confer more
justification, and do so in two ways. First, they confer justification on related
beliefs, such as that there is a table below the apple, a tree to the left, and so
forth. All of these beliefs in turn confer some justification on the belief that there
is an apple on the table by integration. And second, other parts of your visual
field indirectly confer justification on the belief. The more the rest of the visual
field makes intelligible the existence of a red apple before you, the more justified
your belief is. All of your visual (and other sensory) experiences and beliefs fit
together.
Consider third moving around the table. Then you will have a number of addi-
tional and distinct experiences of, and beliefs about, the apple. These will confer
additional justification. Consider fourth the experiences retained in memory.
They will also contribute. Consider also possible interactions with other people
106 Liberal Fundamentalism and Its Rivals

who also seem to see the apple. If they act as if all is normal, or talk about
the apple, or pick it up and eat it, all of this confers additional justification on
your belief.
All of this ‘‘evidence’’ is evidence ‘‘in the foreground’’. But consider also all
of your background knowledge about apples, picnic tables, medium sized-dry
goods, the nature of vision and light, and so forth. These beliefs also contribute
to the justification of your belief. These beliefs make up your evidence ‘‘in the
background’’ (Adler 2002).
Your first visual experience through the pipe confers some justification on your
belief that an apple is there. It looks like an apple. Additional experiences and
beliefs then confer additional justification. Moving around the apple furthers this
process. Touching it, eating it, talking about it, and so forth, makes a difference.
It feels like an apple, tastes like an apple, and everyone agrees that it is an apple.
At some point your pro tanto justification converts to on balance justification.
Indeed, you go from some (less than enough) to more than enough.
Although the example involves an apple, the example is a standard one of per-
ceptual belief. However, being an apple might not be a perceptible property. If
not, the example can be substituted with a red solid sphere at a certain distance. A
first quick glance from one eye with few surrounding distance cues of a partially
occluded sphere may automatically and normally cause a belief that there is a
red sphere at a certain distance. Binocular vision would confer more information
and justification. Walking up to the sphere, walking around it (seeing all sides),
touching it, and so on, all contribute justification. The first quick glance, though
sufficient for belief, is not on its own sufficient for on balance justification.
It is clear that ‘‘coherence’’ (very broadly understood) plays an important
role in converting pro tanto to on balance justification. Integration is essential.
Many of the various visual and other perceptual experiences and perceptual
beliefs present the same ‘‘content’’: that there is an apple. It is conceivable that
on balance justification for a belief is only conferred on a belief that comes in
a ‘‘cluster’’ of other beliefs, or comes along with a cluster of sensory or other
relevant experiences (cf. Sosa 2002). A belief not inferentially based on another
may receive epistemic support from other beliefs, or from other ‘‘experiences’’,
from one or more sources, where the beliefs and experiences are all appropriate to
the content of the target belief. Perhaps no on balance justified belief is an island.
This view of perceptual justification, though perhaps not entirely novel
(cf. Haack 1993), is motivated by what moves both the foundationalist and, I
think, the coherentist, without embracing the idea that justification somehow
emerges from relations of mutual inferential support (where P justifies Q and
Q justifies P), or that only a belief can confer justification on another belief.
True, justification starts with experiences as the foundationalist supposes. True,
such justification is often not enough as the coherentist supposes. False, I claim,
that experience alone or inferential support from other beliefs alone is typically
sufficient for on balance perceptual justification. Clusters of experience and
 Peter Graham 107

belief convert pro tanto to on balance justification, at least for many ordinary
perceptual beliefs.
Psychologically non-inferential beliefs are (or at least can be), from the epi-
stemic point of view, both direct and indirect. They do not depend upon other
beliefs for prima facie pro tanto justification, and so in that sense are immediately
justified: direct. But they positively, and not just negatively, depend upon other
beliefs for on balance or sufficient justification, and so in that sense are mediately
justified: indirect. Thus the Moderate Fundamentalist should be an epistemic
inferentialist about on balance justification, while remaining a non-inferentialist
about prima facie pro tanto justification.
One might deny, or at least try to deflate the significance of, the distinction
between pro tanto and on balance justification by distinguishing between either
all-out-belief from degrees of belief, or belief that definitely P from belief that
probably P, and then claiming that all pro tanto justification for a belief amounts
to is just on balance justification for a belief that P held to a certain degree, or
for a belief that P is likely to a certain degree. But this move would confuse the
psychological with the epistemic. A wise man may proportion his belief to the
evidence, but a tentative endorsement that P may still be strongly justified, and
a resounding endorsement of P may only be weakly justified. We need terms to
mark these facts.
The logical distinction between prima facie justification and pro tanto justifica-
tion is this. Prima facie justification is necessarily defeasible. Pro tanto justification
is not; indefeasible pro tanto justification is a conceptual possibility. There is a
further contrast. Undefeated prima facie justification does not necessarily imply a
possible falling short of on balance justification. Undefeated prima facie justifica-
tion could, for some source of justification, ipso facto count as on balance justifica-
tion. Pro tanto justification, however, just means justification that may fall short.
‘Prima facie’ means defeasible justification, ‘pro tanto’ means some justification.
I will not here try to say what exactly converts some justification into enough, or
when some is enough. This is a complicated issue. I hope to discuss it elsewhere.
We can now state Strong and Weak readings of PER.
(PERs ) If S’s perceptual system represents x as F (where F is a perceptible
property), and this causes or sustains in the normal way S’s belief of
x that it is F, then that confers prima facie justification on S’s belief.
(PERw ) If S’s perceptual system represents x as F (where F is a perceptible
property), and this causes or sustains in the normal way S’s belief of
x that it is F, then that confers prima facie pro tanto justification on
S’s belief.
The Weak Moderate foundationalist embraces PERw without commenting on
its modal status, and the Intuitionist version, the Weak Moderate Fundamental-
ist, embraces PERw as a priori necessary. With the distinction between pro tanto
and on balance justification in hand, and the corresponding distinction between
108 Liberal Fundamentalism and Its Rivals

Strong and Weak readings of the epistemic principles, I am now in a position to

state and defend Weak Liberal Fundamentalism in the next section.

Weak Liberal Fundamentalism

Just as there are two possible versions of PER, so too there are two possible
versions of TEST:
(TESTs ) If a subject S (seemingly) comprehends a (seeming) presentation-
as-true by a (seeming) speaker that P, and if that causes or sus-
tains in the normal way S’s belief that P, then that confers prima
facie justification on S’s belief that P.
(TESTw ) If a subject S (seemingly) comprehends a (seeming) presentation-
as-true by a (seeming) speaker that P, and if that causes or sus-
tains in the normal way S’s belief that P, then that confers prima
facie pro tanto justification on S’s belief that P.
I favor the weaker version. I say why below. Weak Liberal Fundamentalism, how-
ever, is not the standard version of Liberal Fundamentalism. The distinction
between Weak and Strong versions has not been marked (Burge 1993: 467–8,
1997: 21, 22, 45 n. 4; Dummett 1993: 423; McDowell 1998: 435; Weiner
2003: 257; cp. Fricker 2002: 379; Lackey 1999: 474).¹²
Weak Liberal Fundamentalism is a position intermediate between Strong Lib-
eral Fundamentalism and Moderate Fundamentalism (either Weak or Strong).
The Strong Liberal Fundamentalist holds that a hearer’s belief that P based on
comprehending a presenter’s presentation-as-true that P enjoys on balance justi-
fication absent defeat. The Strong position holds that on balance justification is
(absent defeat) direct. The Moderate Fundamentalist, on the other hand, denies
that testimony as such confers any justification on belief. Testimony is, accord-
ing to the Moderate Fundamentalist, evidentially neutral. If any testimony-based
belief enjoys any justification, its epistemic status is entirely inferential. Not only
that, the inferential support must be of a certain sort; the reasons supporting
the belief must not be ineliminably and wholly testimonial; the reasons must
be genuinely ‘‘reductive’’ reasons. The Strong Liberal Fundamentalist denies that
testimony-based beliefs must be supported by reasons, a fortiori by reductive reas-
ons. And the Moderate Fundamentalist denies that any justification, pro tanto or
on balance, for testimony-based beliefs is direct; a fortiori it denies that testimony
enjoys direct pro tanto justification.
The Weak Liberal stands in between these two positions. She agrees with the
Strong Liberal that reductive reasons are not necessary for on balance justifica-
tion. She agrees with the Moderate that testimony-based beliefs do not as such
enjoy on balance justification (absent defeat). But she disagrees with the Strong
Liberal that additional epistemic support is not often needed for on balance justi-
fication for testimonial beliefs, and it disagrees with the Moderate that genuinely
reductive reasons are always required for any justification for testimony-based
 Peter Graham 109

beliefs. The Weak Liberal holds that no additional support of any kind (reductive
or non-reductive) is required for prima facie pro tanto justification, but also that
additional support (whether reductive or non-reductive) is often required for on
balance justification. When it comes to pro tanto justification, the Weak Liber-
al holds that testimony is epistemically direct. But when it comes to on balance
justification, it holds that testimony is epistemically inferential. Weak Liberal
Fundamentalism is weaker (it claims less about on balance justification) than
Strong Liberal Fundamentalism but stronger than Moderate Fundamentalism.
An analogy may be helpful. Suppose an on balance justified belief that P costs
a dollar. The Strong view holds that comprehending a presentation-as-true that P
provides the hearer with an entire dollar. Absent other fees (defeaters), the hearer
can buy the on balance justified belief. The Weak view holds that comprehending
a presentation-as-true that P often fails to provide the hearer with an entire dollar,
but for all that it provides the hearer with some money. With a little more money
to spend, the hearer can buy the on balance justified belief. The Moderate Funda-
mentalist, on the other hand, thinks comprehending a presentation-as-true that P
is like a check. In itself it is worthless. Unless there is money in the bank backing
up the check, you can’t buy anything with it. And the money in the bank, accord-
ing to the Moderate Fundamentalist, can’t come from testimony either (just as
you can’t pay your credit card bill with that very same credit card). The Strong
Liberal view holds that comprehending the presentation-as-true of another that
P as such provides enough justification (absent defeaters). The Weak Liberal view
holds that it provides some. The Moderate view holds that it provides none.

Is Liberal Fundamentalism Too Permissive?

I now discuss an objection to Liberal Fundamentalism. It is an objection the
Moderate Fundamentalist (among others) is entitled to make, unlike the four
objections discussed above. The objection is that Liberal Fundamentalism is too
strong or too permissive, that it would entail that many testimony-based beliefs
without inferential support would be justified where, intuitively, they are not.
Merely comprehending another’s presentation-as-true is not, the objection goes,
sufficient for justified belief in the content of what was presented, at least in very
many cases, even absent defeat. Additional support is necessary for on balance
justification.¹³
This objection clearly targets Strong Liberal Fundamentalism. Strong Liberal
Fundamentalism is also the standard version of the view in print, as noted above.
But it is not clear that this complaint applies to Weak Liberal Fundamentalism.
The complaint is that acceptance without additional support is too permissive.
The Strong view disagrees: undefeated acceptance without support is fine. The
Weak view leaves that question open. It is entirely consistent with Weak Liberal
Fundamentalism that additional support is necessary for on balance justification.
Hence the objection seems to have no, or at least considerably less, force against
110 Liberal Fundamentalism and Its Rivals

the Weak position. Since I think (though have not argued here) that there are
good prima facie reasons in favor of the Liberal view generally, and since the
Weak view is more defensible than the Strong, I prefer Weak Liberal Fundament-
alism. It seems defensible against a standard complaint.

Perception and Testimony Compared

I now want to make a point about the comparison between perception and testi-
mony. I believe the right account for both is the Weak account, PERw and
TESTw . Both principles are supported by the data. Just as the justification from
testimony as such often fails to confer on balance justification for a testimony-
based belief, so too (I argued above) that justification from perception as such
often fails to confer on balance justification for a perceptual belief. The right
accounts of perception and testimony, I believe, parallel one another.
There is, however, a tendency in the literature to compare perception and testi-
mony unfavorably. Just having a perceptual experience of something being so and
so is supposed to be enough for on balance justification in the perceptual case,
but not enough in the testimonial case. Hence it is inferred that there must be a
fundamental difference in kind between perception and testimony (e.g. Pritchard
2004: 343–5).
This inference, I believe, is based on a mistake. I think some are led into think-
ing there is such a contrast because they fail to see that in the perception case
there are a large number of interlocking experiences and beliefs that individu-
ally confer pro tanto justification and collectively confer on balance justification,
but in the testimony case only one piece of pro tanto justification is in focus, the
one presentation-as-true whether P. For instance, compare looking at an apple
on a table and being told that there is an apple on a table in the park over yon-
der hill. In the visual case, over a relatively small portion of time, one will have a
number of distinct but interlocking experiences of, and beliefs about, the apple.
One’s belief that there is an apple will enjoy plenty of justification; its justifica-
tion on balance will, most likely, be over-determined (Conee and Feldman 1985:
29–30). Now compare the hearer who is told by a speaker (a speaker that the
hearer doesn’t know first-hand and knows very little about otherwise) that there
is an apple on the table yonder. Here all the hearer may have in favor of the belief
is this one presentation-as-true. There may be no, or only a very little, additional
support. There is thus a difference between the perceptual case and the testimony
case. The perceiver (with many experiences and other related beliefs to go on)
will enjoy on balance justification for that same belief, but the hearer (with only
one presentation-as-true to go on) will only enjoy pro tanto justification for the
belief that there is an apple on the table. I grant this (de facto) difference. But
it is no reason to compare perception and testimony unfavorably. All this differ-
ence shows is that typically in the perceptual case enough justification is present
for on balance justification, but that in many testimonial cases only pro tanto
 Peter Graham 111

justification is present.¹⁴ If we were to compare just one perceptual experience

with one presentation-as-true that might help bring the analogy between the epi-
stemologies of perception and testimony into sharp relief. One report may really
be like just one perceptual experience.¹⁵ Both are some, but often not enough,
for justification on balance. Comparing memory with testimony reinforces this
point. Remembering something may be just as good, all else being equal, as
being told.

Additional Supporting Reasons

If the Weak Liberal is right, on balance justification for testimony-based beliefs
will often require additional support. No on balance justified testimony-based
belief, the slogan would go, is an island. Additional support, obviously, can come
from a number of sources. You tell me it is an apple, and I take a look myself. But
additional support can also come from other testimony-based beliefs. There are
four ways this might happen.
First, additional presentations-as-true may fit together with the original
presentation, though they are about different things. Suppose you meet someone
on a plane and you start a conversation. She tells you that she is a highly trained
mathematician working for Xerox in Palo Alto engaged in pure research. You
may find this an unlikely occupation, but you do not disbelieve her. During
the conversation she starts talking about Palo Alto, about other researchers at
Xerox, and where she studied mathematics, and engages you in a discussion
about the nature of numbers and sets. You may only have been pro tanto justified
in believing that she is a mathematician doing pure research at Xerox at the start
of the conversation. Indeed, if that was all she said and then she turned away,
you might only have been justified on balance in believing that she said she was a
mathematician. But at the end of the actual conversation it seems that you are on
balance justified in believing that she is a Xerox employed mathematician.¹⁶
The second, and perhaps more obvious way additional presentations may help,
is by going directly to the fact that the original presentation-as-true is about. A
witness may tell you that he saw the killer drive away in a red van. Another, inde-
pendent witness, may tell you the same thing. And so on. The more independent
witnesses, the Liberal holds, the better. At some point, presumably, the justifica-
tion the hearer’s belief enjoys converts from pro tanto to on balance justification.
The third and certainly very common way additional presentations often help
is by going to the trustworthiness of the target interlocutor. That is, if you are
talking to A, being told by B that A is trustworthy will confer some additional
warrant on trusting A.
A fourth way is by answering defeaters. Suppose I have reason to think C is
not trustworthy. D can tell me that she is generally trustworthy, or that she
has an adequate reason in this case to be sincere. Or suppose I have reason to
think E, though obviously sincere, does not know what he is talking about. F
112 Liberal Fundamentalism and Its Rivals

can tell me that, appearances notwithstanding, E really does know what he is

talking about.
Justified beliefs are often supported by experiences and beliefs from a number
of different sources. They all work together. Testimony-based beliefs support
perceptual beliefs; perceptual beliefs support memory beliefs; memory beliefs
support testimony-based beliefs, and so on. According to the Liberal, perception,
memory, introspection, testimony, and reason all as such confer justification on
the beliefs they normally cause and sustain.
My overall goal in this essay is to bring more clarity to the testimony debate.
Although I favor the Liberal position, I have not argued for it directly here.
Hopefully the clarity speaks in its favor, or at least removes some sources of
opposition. Given one’s overall point of view in epistemology, the Liberal posi-
tion may be a non-starter, or it may be a rather plausible extension of what one
already believes. Making explicit one’s overall point of view should reveal why
one might reject it off the bat, or reveal why one might find it a plausible account-
ing for our justified reliance upon the word of others.

REFERENCES
Adler, Jonathan (2002), Belief ’s Own Ethics (Cambridge, Mass.: MIT Press).
Alston, William (1988), ‘An Internalist Externalism’, Synthese, 74: 265–83; repr. in his
Epistemic Justification: Essays in the Theory of Knowledge (Ithaca, NY: Cornell Univer-
sity Press, 1989).
Audi, Robert (1988), ‘Justification, Truth, and Reliability’, Philosophy and Phenomenolo-
gical Research, 49: 1–29.
(1997), ‘The Place of Testimony in the Fabric of Knowledge and Justification’,
American Philosophical Quarterly, 34: 405–22.
(2001), The Architecture of Reason (Oxford: Oxford University Press).
(2002), ‘The Sources of Belief ’, in Paul Moser (ed.), Oxford Handbook of Epistemo-
logy (Oxford: Oxford University Press).
(2004), ‘The A Priori Authority of Testimony’, Philosophical Issues, 14.
BonJour, Laurence (1999), In Defense of Pure Reason (Cambridge: Cambridge University
Press).
(2002), Epistemology: Classic Problems and Contemporary Responses (Lanham, Md:
Rowman and Littlefield).
Burge, Tyler (1993), ‘Content Preservation’, Philosophical Review, 102: 457–88.
(1997), ‘Interlocution, Perception, Memory’, Philosophical Studies, 86: 21–47.
(1999), ‘Comprehension and Interpretation’, in L. Hahn (ed.), The Philosophy of
Donald Davidson (La Salle, IU.: Open Court).
(2003), ‘Psychology and the Environment: Reply to Chomsky’, in M. Hahn and
B. Ramberg (eds.), Reflections and Replies (Cambridge, Mass.: MIT Press).
(2004), ‘Perceptual Entitlement’, Philosophy and Phenomenological Research, 67:
503–548.
 Peter Graham 113

Coady, C. A. J. (1973), ‘Testimony and Observation’, American Philosophical Quarterly,

10: 149–55.
(1992), Testimony: A Philosophical Study (Oxford: Clarendon Press).
Conee, Earl, and Feldman, Richard (1985) ‘Evidentialism’, Philosophical Studies, 48:
15–34.
Dummett, Michael (1993), ‘Testimony and Memory’, in his The Seas of Language
(Oxford: Oxford University Press).
Faulkner, Paul (2000), ‘The Social Character of Testimonial Knowledge’, Journal of
Philosophy, 97: 581–601.
Fricker, Elizabeth (1987), ‘The Epistemology of Testimony’, Proceedings of the Aristoteli-
an Society.
(1994), ‘Against Gullibility’, in B. K. Matilal and A. Chakrabarti (eds.), Knowing
from Words, (Dordrecht: Kluwer Academic Press).
(1995), ‘Telling and Trusting: Reductionism and Anti-Reductionism in the Epi-
stemology of Testimony, A Critical Notice of Coady 1992’, Mind.
(2002), ‘Trusting Others in the Sciences: a priori or Empirical Warrant?’, Studies in
History and Philosophy of Science, 33: 373–83.
Goldman, Alvin (1979), ‘What Is Justified Belief ?’, in G. Pappas and M. Swain (eds.),
Knowledge and Justification (Dordrecht: Reidel).
(1992), ‘Epistemic Folkways and Scientific Epistemology’, in his Liaisons: Philo-
sophy Meets the Cognitive and Social Sciences (Cambridge, Mass.: MIT Press).
Graham, Peter J. (1997), ‘What is Testimony?’, Philosophical Quarterly, 47: 227–32.
(2000a), ‘The Reliability of Testimony’, Philosophy and Phenomenological Research,
61: 695–708.
(2000b), ‘Conveying Information’, Synthese, 123: 365–92.
(2000c), ‘Transferring Knowledge’, Noûs, 34: 131–52.
(2004), ‘Metaphysical Libertarianism and the Epistemology of Testimony’, Americ-
an Philosophical Quarterly, 41: 37–50.
(forthcoming), ‘Theorizing Justification’, in M. O’Rourke and J. Campbell (eds.),
Knowledge and Skepticism: Contemporary Topics in Philosophy, vol. v (Cambridge,
Mass.: MIT Press).
Haack, Susan (1993), Evidence and Inquiry (Oxford: Blackwell).
Huemer, Michael (2001), Skepticism and the Veil of Perception (Lanham, Md.: Rowman
and Littlefield Publishers).
Kusch, Martin (2002), Knowledge by Agreement (Oxford: Oxford University Press).
Lackey, Jennifer (1999), ‘Testimonial Knowledge and Transmission’, Philosophical
Quarterly, 49: 471–90.
(2003), ‘A Minimal Expression of Non-Reductionism in the Epistemology of Testi-
mony’, Noûs 37: 706–23.
Lehrer, Keith (1994), ‘Testimony and Coherence’, in B. K. Matilal and A. Chakrabarti
(eds.), Knowing from Words (Dordrecht: Kluwer Academic Publishers).
Lipton, Peter (1998), ‘The Epistemology of Testimony’, British Journal for the History
and Philosophy of Science, 29: 1–31.
Lyons, Jack (1997), ‘Testimony, Induction, and Folk Psychology’, Australasian Journal of
Philosophy, 75: 163–78.
114 Liberal Fundamentalism and Its Rivals

McDowell, John (1998), ‘Knowledge by Hearsay’, in his Meaning, Knowledge, and

Reality (Cambridge, Mass.: Harvard University Press).
Pritchard, Duncan (2004), ‘The Epistemology of Testimony’, Philosophical Issues, 14:
326–48.
Quinton, Anthony (1973), ‘Autonomy and Authority in Knowledge’, in his Thoughts
and Thinkers (London: Duckworth).
Root, Michael (1998), ‘How to Teach a Wise Man’, in Kenneth Westphal (ed.),
Pragmatism, Reason, and Norms (New York: Fordham University Press).
(2001), ‘Hume on the Virtues of Testimony’, American Philosophical Quarterly, 38:
19–35.
Ross, Angus (1986), ‘Why do We Believe what We are Told?’, Ratio, 28: 69–88.
Rysiew, Patrick (2000), ‘Testimony, Simulation, and the Limits of Inductivism’, Aus-
tralasian Journal of Philosophy, 78: 269–74.
Sosa, Ernest (2002), Epistemic Justification, with Laurence BonJour (Oxford: Blackwell
Publishers).
Stevenson, Leslie (1993), ‘Why Believe What People Say?’, Synthese, 94: 429–51.
Strawson, P. F. (1994), ‘Knowing from Words’, in B. K. Matilal and A. Chakrabarti
(eds.), Knowing from Words (Dordrecht: Kluwer Academic Publishers).
Weiner, Matthew (2003), ‘Accepting Testimony’, Philosophical Quarterly, 53: 256–64.

N OT E S
1. Robert Audi distinguishes between (a) basic or direct belief (belief not inferentially
based upon another belief ), (b) sources of basic belief, and (c) basic sources, sources
of belief that do not operationally depend upon other sources of belief. He holds
that testimony-based beliefs are direct, that testimony is a source of basic beliefs,
but denies that it is a basic source, for it operationally depends upon perception.
See Audi (1997, 2002, 2004, and Chapter 1 in this volume). As I am using the terms
‘‘direct’’ and ‘‘basic’’, I mean to imply epistemic independence, that a belief does not
depend upon another for prima facie pro tanto justification. I mean what he means
by the ‘‘a priori authority’’ of a source.
2. I have characterized testimony-based beliefs elsewhere (Graham 2000a).
3. The disjunctive phrase ‘‘either sincere or reliable or both’’ is meant to accommodate
‘‘hybrid’’ views like Faulkner’s (2000) that require only positive inferential support
for sincerity, but not for reliability. Fricker sometimes suggests she is OK with sin-
cerity, but not with reliability (1994).
4. Though it is consistent with the requirement that the way held is per se reliable: reli-
able in normal conditions when functioning normally (cf. Burge 2004 on ‘‘reliably
veridical’’).
5. I hope to discuss the Intuitionist on PER on another occasion.
6. The principles the Reliabilist or the Pragmatist accepts may be (some of ) the prin-
ciples listed above, close analogues, distant cousins, or wholly new. It is unlikely,
however, that some of the principles, or close analogues, would not show up on
either the Reliabilist’s or the Pragmatist’s list at all.
 Peter Graham 115

7. Elizabeth Fricker’s position, elaborated in a number of papers (1994, 1995, 2002), is

complex; it defies easy categorization. I hope to discuss it at some length on another
occasion. For some discussion, see Weiner (2003).
8. Another way one might ‘‘reduce’’ testimony is to demand that the hearer be able to
show a priori that the speaker is trustworthy. I will not discuss this route in what
follows.
9. De facto reliability surely matters for knowledge (as opposed to justification). And,
again, as above, per se reliability may be required for justification (I remain neutral
here), and this is consistent with Intuitionism.
10. One commonly made argument against the Liberal is that there is a fundamental
epistemic difference in kind between perception and testimony because perception is
a natural process operating according to natural laws but testimony goes through the
will of the speaker, i.e. the speaker may always choose to lie or mislead. This fact in
turn generates a number of different arguments against the Liberal when combined
with other premises. I have discussed one of these arguments in Graham (2004).
11. Burge uses the phrase ‘prima facie pro tanto’ as well (1993: 467–8; 2003: 463).
However, he seems to use ‘prima facie’ and ‘pro tanto’ interchangeably. I do not.
12. Pritchard (2004) may be an exception. He distinguishes a ‘‘modest’’ version of
‘‘credulism’’ (the Liberal view) from a ‘‘bare’’ version. From what he says, the modest
version may parallel the Weak Liberal view as here defined. He does not defend such
a view. In fact, he argues against it. See n. 15.
13. I have heard this objection many times. Something like it seems to drive Frick-
er’s rejection to the Liberal view in her ‘‘Against Gullibity’’ (1994) and elsewhere.
Pritchard states it explicitly (2004: 328–30). It seems to be widely shared.
14. However, if one were to look to the background for support in the testimonial case,
there would be a good deal there as well (Adler 2002; Fricker 2002). The difference
seems to be that in the perceptual case in the foreground lots of additional support is
obvious, but that is not as clearly so in the testimonial case.
15. Pritchard disagrees. He says if A sees his car in the driveway, but had no reason to
believe it would or would not be there, then A’s belief is (on balance) justified. But if
A is told that his car is there, and has nothing else to go on, then his belief is not (on
balance) justified. ‘‘In the perceptual case, it does seem entirely plausible to suppose
the agent is justified in forming this belief. . . . The situation is very different, how-
ever, when it comes to the [testimony-based belief that the car is in the driveway]. . . .
Intuitively, without independent . . . grounds this . . . belief is not justified’’ (2004:
342–3). Pritchard does not consider the additional support the perceptual belief
enjoys from the entire visual field over time, as I do in my case of the apple, and he
does not consider comparing the two cases for equivalence in pro tanto justification.
Perhaps he might change his mind after taking these facts into consideration. He
appeals to differences in the extent and nature of the reliability of perception versus
testimony to support his conclusion that perception and testimony differ (2004:
343–4). I have replied to this kind of argument above. It does not, I think, support
his conclusion.
16. This example is due to Audi (1997).
 5
Knowledge: Instrumental and Testimonial
Ernest Sosa

If a belief held on authority turns out to be correct, what most saliently explains
this fact must surely involve the discovery and transmission of the relevant
information. Relatively little of the credit belongs to the ultimate believer, by
comparison, if all he did was to trust the authoritative source without question.
In order to constitute knowledge, a testimony-derived belief must be accurate
because competent, which should not be thought to require that the most salient
explanation of its being right must involve the individual competence manifest
by the subject in holding that belief. The explanatorily salient factors will prob-
ably lie elsewhere; what mainly accounts for the belief ’s correctness will likely
involve others and their cognitive accomplishments.
That insightful point must be properly appreciated and accommodated.¹
Testimonial knowledge is a collaborative accomplishment involving one’s
informational sources across time. Consider what is required: the gathering,
retaining, transmitting, and receiving of information, with pertinent controls
applied each step of the way. Consider the aptitude, competence, or intellectual
virtue required for any full account of how the ultimate belief outcome amounts
to knowledge when true because competent. Many people might be involved,
acting mostly individually, while unaware of the others. Think of the documents
consulted by a historian, of those responsible for their production, for their
preservation and transmission unaltered, and so on; think of those who help with
the production of a text, and of those who collaborate to produce a book; think of
the copies of the book preserved, relevantly unaltered, by librarians and others; all
of which eventuates in one’s reading of the text and acquiring certain information
about something far away and long ago.
Accordingly, there is a large external element in the knowledge of the members
of a civilization advanced enough to exploit testimony as extensively as we do.
Our knowledge will depend deeply and extensively on factors beyond the scope
of anyone’s reflective perspective. That is not, however, distinctive of knowledge
through testimony, as may be seen if we compare a closely related ‘‘instrumental’’
sort of knowledge.
 Ernest Sosa 117

A deliverance of a proposition by an instrument is epistemically reliable only

if that proposition belongs to a field, and that instrument is so constituted and
situated, that not easily would it then deliver any falsehood in that field.
The deliverances of an instrument are answers to questions. By punching cer-
tain keys we pose to a calculator questions of the form ‘What is the sum of x
and y?’ By placing a thermometer at a certain location and time, we can ask it
a question of the form ‘What is the ambient temperature there and then?’ The
deliverances of an instrument are its answers to such questions that might be
posed to it. An instrument is reliable insofar as it would tend to answer them
correctly.
The reliability of an instrument varies with its situation. What makes it reli-
able for a given situation is that so situated it would not easily answer relevant
questions incorrectly.
It is the thermometer that is a reliable instrument, not just its screen. What is
the difference that makes this difference? True, the screen needs the aid of the
attached thermometer. But so does the thermometer need to be properly situ-
ated. It cannot be insulated, for example, nor can the temperature in the relevant
space be too heterogeneous. If the thermometer is to tell the ambient temperat-
ure reliably, it must be appropriately situated in certain contingent ways, ways in
which it might not have been situated, perhaps very easily might not have been
situated.
Whether physically unified or dispersed, however, it is the larger instrumental
system that is deeply operative. The guidance device in your car, for example, is
not just a screen for the broader GPS system. It is more interestingly a seat of its
relevant functions than is the thermometer screen. The thermometer screen is not
a similarly important seat of its temperature-indicating functions. Nevertheless,
the device in your car is not like the whole thermometer either.
When we trust the instrument readings on which we increasingly rely, such
as displays on screens, we presuppose a field of propositions, and a situation,
such that we take an instrument so situated to be reliable for that field: not
easily would it then deliver a proposition in that field unless that proposition
were true.
The man in the street needs no deep understanding of the instruments on
which he relies.² He relies on his GPS devices, cellular telephones, atomic
watches, and computer terminals with little or no awareness of how they depend
on relations to other devices that more importantly seat the relevant functions.
Let us use the term ‘quasi-instruments’ for devices that come more fully within
our purview than do the fuller instruments involved.
Take the gauges that we face as driver of a late-model car. Most of us have
a paltry conception of them as little more than screens, displays, that keep us
informed about the amount of fuel in our tank, our speed, the rpm of our motor,
etc. We take the display to be part of a fuller instrument that reliably delivers its
118 Knowledge: Instrumental and Testimonial

deliverances. But who knows how the display on our dashboard reliably connects
with its relevant subject matter? Our conception hardly extends beyond the
distinctive screen or display.
When we thus rely epistemically on a quasi-instrument, then, even in the
near-limiting case of the screen, we presuppose reliability. In thus relying we
make manifest our assumption of reliability. We take the deliverance, even when
understood as just the display on the screen, as more than accidentally connected
with its truth. We take it to be at least safe, in that not easily would that screen
display a false deliverance. This trust could be properly acquired in any of several
ways. I might have it simply because I then think as I am told. I might acquire
it, alternatively, through inductive generalization, even through trial and error.
Some such trust is required, but it is worthless if just arbitrary.
What, more explicitly, might be the content of the required trust, however
acquired? Whether our trust derives from testimony or from our own inductive
generalization, we trust that a proposition in the relevant field would be true, or
would tend to be true, if delivered by this instrument.
That is only a first approximation, however, since we still need to distinguish
between outright and dependent safety, which yields the following improved
condition:
S knows that p on the basis of an indication I(p) only if either (a) I(p)
indicates the truth outright and S accepts that indication at face value, or
(b) for some condition C, I(p) indicates the truth dependently on C and
either (i) S accepts that indication as such guided by C (so that S accepts
the indication as such on the basis of C), or else (ii) C is constitutive of the
normal conditions for the operation of that source.
Such reliability is required in the instruments and quasi-instruments on which
we rely, if our reliance upon them is to be epistemically effective. It might
be thought that instrumental knowledge can be reduced to non-instrumental
knowledge, including testimonial knowledge, but this is dubious. Our access to
the minds of others is after all mediated by various instruments, and we must trust
such media at least implicitly in accessing the testimony all around us. So there is
a kind of instrumental knowledge prior to and essentially involved in testimonial
knowledge.

Many of our epistemic instruments are reliable because they are responsive to
their environment. This seems true of thermometers, speedometers, fuel gauges,
and many other instruments, whose deliverances are safe because they are thus
responsive. A thermometer is reliable, for example, because its deliverances
are safe; not easily would they be false. This is because it senses the ambient
temperature, being so constituted and so related to its surroundings that the
ambient temperature will cause it to read accordingly. Given such responsiveness,
no wonder the thermometer’s deliverances are systematically safe: not easily
 Ernest Sosa 119

would it issue an incorrect deliverance. (That is, it might easily issue deliverances,
if it is consulted, but not easily false ones.)
Not all reliable instruments are reliable through a systematic safety of their
deliverances that derives from responsiveness to their proper field, not if this
requires efficient causal input. Consider, for example, a calculator, about as reli-
able an instrument as any. If you give it a question, your calculator returns a
correct answer with extremely high safety and reliability. But the reliability of a
calculator, and the associated systematic safety of its deliverances, do not derive
from responsiveness to its field through efficient causation. The facts of arithmet-
ic do not efficiently cause the calculator to display the right answer on its screen
when you consult it.
Yet in some sense the calculator gives its answer because it is the true answer.
Indeed, we can predict its answers with extreme power and reliability if we pre-
dict that it will answer with truth. But in what sense does it answer as it does
because it is the true answer? What sort of causation could be at work if not effi-
cient causation? Here the explanation will presumably run via the fact that the
artefact is designed to be accurate by an efficient and intelligent designer. But this
shifts the question to that of how that designer gets to be himself so reliable a
calculator, given the lack of efficient causation from the facts of arithmetic to the
contents of his mind.
We have no need of efficient causation to explain the reliability and systematic
safety of cogito beliefs such as the belief that one thinks and the belief that one
exists. Consider, indeed, any propositional content whose conditions of under-
standing or truth preclude its being entertained while untrue. No such content
could possibly gain our assent without being true, which makes our competence
in such assents infallibly reliable.
One might even extend the scope of cogito assent far beyond the cogito itself,
to demonstrative introspective thought. One is bound to be right in assenting to
<This is thus>, if the conditions of reference for the relevant uses of ‘this’ and
‘thus’ are constituted essentially through the episode attended to and its content.
If it succeeds at all in attributing something to anything, <This is thus> is bound
to be true if the property attributed by ‘‘is thus’’ is selected by the attribution
from among phenomenal properties of the item picked out as ‘this’.
Cogito propositional contents hence extend beyond cogito claims in a strict
sense restricted just to cogito itself and perhaps sum. Take any propositional
content whose truth is introspectively accessible, and whose conditions of
understanding and truth guarantee that it must be true if entertained with
understanding. Any such propositional content can now count as a cogito content
in a broader sense. Assenting to cogito contents thus broadly conceived manifests
an epistemic competence, one highly reliable in its safe deliverances. But its
reliability and safety are not to be explained through any efficient responsiveness
to its field.
120 Knowledge: Instrumental and Testimonial

The same goes for the reliability of calculators and the systematic safety of their
deliverances. Indeed, no more than that of a mechanical calculator is our own
calculating competence to be explained through our efficient responsiveness to
the facts of arithmetic, these being incapable of causal efficacy.
A wide variety of competencies are reliable, therefore, and their deliverances
systematically safe, independently of any responsiveness to causal efficacy. Yet
such deliverances are still somehow delivered ‘‘because’’ they are true. What sort
of ‘‘because’’ might this be, if not that of efficient causation?
Is it a ‘‘because’’ of teleology or function, involving either a calculator con-
sciously designed to give true answers, or the design of a Creator, or of fitness-
selecting Mother Nature? No, epistemically efficient competence need not derive
from any design, whether intentional or unintentional, divine or evolutionary.
Swampman cases show this clearly enough. Lightning in a swamp might
serendipitously cause molecules to come together into the form and substance
of an intelligent Swampman. It may be doubted that Swampfolk are really
physically possible; but this does not show them to be metaphysically impossible
a priori. Insofar as we aim to understand what it would be to enjoy epistemically
worthy beliefs, ones that are epistemically justified or even amount to knowledge,
insofar as we are trying to understand what is involved with a priori necessity
in such justification and knowledge, to that extent will Swampfolk be relevant
even if they are physically impossible. For all we know, BIVs, Rings of Gyges,
Twin Earth, and transplanted split brain hemispheres may all be physically
impossible, but that would not make them irrelevant to philosophical inquiry
aimed at understanding the fundamental nature of morality, personal identity, or
reference and mental content.
Our stance is quite compatible with the relevance of theology or evolution
to whether we know and how we know. After all, the better we know about
our competencies and how they yield our beliefs, the better we understand their
sources. The more justifiedly we can attribute our beliefs to them, moreover, the
better also is the epistemic quality of these beliefs. Some limit is hence built into
the epistemic prospects of Swampfolk. For, at an important juncture they are
denied any further possibility of explanation and deeper assurance about their
epistemic competencies.
Some of our justification for trusting the instruments that we rely on does
plausibly have an inductive basis. Once we trust a particular device repeatedly
with good results, we gain inductive support for its reliability. Justification for
trusting our instruments derives also, of course, from testimony. But our aware-
ness of testimony as testimony itself relies on instrumental knowledge. We must
interpret our interlocutors, so as to discern the thoughts or statements behind
their linguistic displays. From oral or written displays we can tell what someone is
saying, and thinking.
 Ernest Sosa 121

Interpretative knowledge, I am suggesting, is a kind of instrumental know-

ledge. You ask a question of someone. Assuming sincerity and linguistic compet-
ence, what they utter reveals what they think (and on similar assumptions reveals
also what they say). This means that we can tell what they think (or say) based
on a deliverance conveyed by their utterance. Interpretative knowledge of what a
speaker thinks (says) is thus instrumental knowledge that uses the instrument of
language. Language is a double-sided instrument serving both speaker and audi-
ence. Hearers rely on the systematic safety of the relevant deliverances. Not easily
would the speaker’s utterance deliver that the speaker thinks (says) that such and
such without the speaker’s indeed thinking (saying) that such and such.
Speakers do not speak just about what they think. On the default assumption,
however, which must be that of sincerity, as is known to both speaker and audi-
ence, the speaker’s utterance does give to understand what the speaker thinks.
So, the utterance carries a deliverance as to the speaker’s mind, as well as any
deliverance it may deliver as to its more direct subject matter.
If we are to know a speaker’s mind through his utterances, the speaker must
have a reliable competence to state his mind. He must be able through his utter-
ances to deliver safe deliverances about what he thinks on the topic at hand.
These must be deliverances that would not be delivered unless their content (con-
cerning what the speaker thinks) were true.
If any of this is put in serious enough doubt, the supposed instance of testi-
mony will be disqualified as a source of knowledge about its direct content, for
that audience at that time. Testimonial knowledge thus presupposes instrument-
al knowledge, and it is out of the question to reduce all instrumental knowledge
to testimonial knowledge.
It might be countered that the linguistic instrumental knowledge highlighted
here itself counts as a kind of testimonial knowledge, for it is the testifier who
through his utterance gives to understand that he believes such and such. Since
‘‘giving to understand’’ is a kind of communication, we might well consider it
a kind of testimony. This terminological option would distinguish two kinds
of testimony: the assertive and the non-assertive. Orthodox terminology would
restrict testimony to assertive testimony, however, so that non-assertive testi-
mony might properly be grouped with deliverances of instruments more gener-
ally. Labels matter less, in any case, than a proper delineation of the phenomena,
with appreciation of what they interestingly share.

One way in which we could not hope to attain adequate justification for believing
an instrument to be reliable is through simple bootstrapping, whereby we accept
its deliverances on the sole basis of their being so delivered, and base a belief that
its deliverances would tend to be correct, and safe, simply on the inductive base
thus formed. That could not be how you gain all your justification for thinking
122 Knowledge: Instrumental and Testimonial

an instrument reliable; indeed, it could not be how you gain any, if you previ-
ously had none.
Just as one is a calculator, though not as accurate or powerful as an arte-
factual calculator, so one is a thermometer, though not as good as artefactual
thermometers. What applies to us as temperature-sensors, moreover, applies to
us as sensors more generally, as perceivers. Indeed, the instruments on which we
depend most extensively and fundamentally are the perceptual modules included
in our native endowment.
Much perceptual knowledge can thus be seen as instrumental. If our modules
are reliable, we gain knowledge and epistemic justification by accepting their safe
deliverances at face value.
We could hardly gain a justified belief by accepting a deliverance of an
instrument trusted arbitrarily. Our trust in the instruments that we use to pry
information off our environment cannot be just arbitrary. If we assume them
to be reliable, and assume their deliverances to be systematically safe, these
assumptions cannot be just intellectual whims. But nor can they all be justified
exclusively on the basis of testimony. Nor can they possibly just lean on each
other, with no support under them. How then can we ever be justified in
thinking an instrument reliable, and its deliverances systematically safe?
That seems especially troubling when we see that among such instruments are
to be found our perceptual modules. How could we come to know the reliability
of these instruments? We could not do so through testimony in general, nor of
course through direct bootstrapping. Could we perhaps manage it through some
more indirect form of coherence-involving bootstrapping, where we do rely on
perceptual input at some fundamental level without doing so in the ludicrous
way of direct bootstrapping?
Epistemically justified trust in our sensory sources is a gift of natural evolution,
which provides us with perceptual modules that encapsulate sensory content and
reliability in a single package. We accept their deliverances at face value as a
default stance, properly so. This is because the content delivered requires the
reliability of the delivery in normal circumstances. What gives these introspect-
able sensory states the content that they have is substantially the fact that they
normally respond to the truth of their content. They are thus apt for normally
mediating between the relevant environmental facts suitable for such sensory
uptake and the beliefs they tend to prompt.
Our senses are thus distinguished epistemically from ordinary instruments.
We can have reasons for trusting our senses as we do, a trust justifiably based on
these reasons. What is distinctive of our senses as epistemic instruments is that we
do not need, and cannot have, sufficient reason for trusting each, with absolutely
no reliance, either now or earlier, on any of the others.
Compare an ordinary instrument: a fuel gauge, thermometer, or speedometer,
etc. As we go through an ordinary day, we find ourselves trusting such devices
implicitly at many turns. We accept their deliverances at face value. What justifies
 Ernest Sosa 123

such implicit trust? It could not possibly be direct bootstrapping. That would
involve a vicious circle. The data required for our simple bootstrapping cannot be
acquired with justification unless our implicit trust in the instrument is already
justified.
Direct bootstrapping is of course as powerless to explain our justification for
trusting our senses as it is to explain our justification for trusting our ordinary
instruments. So, this is not what distinguishes our senses on one side from the
instruments on the other. The difference is rather that our senses enjoy a kind
of default rational justification denied to (ordinary) instruments. That is to say,
we are default-justified in accepting the deliverances of our senses, but we need a
rational basis for accepting the deliverances of our instruments.
We no longer need a current rational basis once enculturated as a competent
instrument user in a technological society. The difference between instruments
on one hand, and senses on the other, emerges only through memory. At some
point we need a rational basis for trusting our instruments (unlike our senses),
though epistemic justification can then be preserved through sheer memory even
if we are later unable to recall the rational basis. But this is the way of epi-
stemic justification generally and nothing peculiar to instrumental knowledge in
particular.
Human testimony stands with the senses in providing default rational justific-
ation. And the same goes for the instrumental knowledge that gives us access to
testimony through the instrument of language. The instrument of one’s language
is among those that we master through sub-personal means involving animal
processes below the level of any kind of reasoning. And yet the instrumental
knowledge that it enables is an essential link in the chain whereby we come to
know much of what we know, whereby we attain our knowledge at its best, and at
its most rational.

N OT E S
1. The point is made by Jennifer Lackey in her review of Michael DePaul and Linda
Zagzebski (eds.), Intellectual Virtue: Perspectives from Ethics and Epistemology, in Notre
Dame Philosophical Reviews (Aug. 2004).
2. Nor does the woman; but I trust that context will continue to make it clear enough
when my terms are gender-free.
This page intentionally left blank
 PART I I I
REDUCTIONISM AND
NON-REDUCTIONISM IN THE
E PI S T E M O LO G Y O F T E S T I M O N Y
This page intentionally left blank
 6
Reductionism and the Distinctiveness
of Testimonial Knowledge
Sanford C. Goldberg

abstract This paper aims to correct what I will argue are two common and
related misconceptions in discussions of the epistemology of testimony. The first
misconception is that testimonial knowledge is an epistemically distinct kind of
knowledge only if there are testimony-specific epistemic principles implicated
in the justification of beliefs formed through testimony. The second misconcep-
tion is that anyone who endorses a reductionist position regarding the epistemic
status of testimony, and so who denies the existence of testimony-specific epi-
stemic principles, ipso facto ought to be hostile to the hypothesis that testimonial
knowledge is epistemically distinctive. In this paper I argue against both miscon-
ceptions; I will do so by arguing for the distinctiveness hypothesis in a way that
involves no premise any reductionist should want to deny.

1 T H E C AT E G O RY O F T E S T I M O N I A L K N OW L E D G E

The core thesis of this paper is that even those who deny a certain uniqueness
claim regarding the epistemic principles involved in testimonial justification, to
the effect that
(1) The justification of beliefs through testimony implicates epistemic prin-
ciples unique to testimony cases
should acknowledge a certain uniqueness claim regarding the distinctiveness of
testimonial knowledge, to the effect that
(2) Testimonial knowledge is an epistemically unique kind of knowledge.
The significance of my core thesis lies in the fact that, to the extent that the mat-
ter of (2) is taken up, those who deny (2) typically do so on the basis of their
rejection of (1); if my core thesis is correct, then such an argument is misguided in
principle and even those who deny (1) should endorse (2).

I would like to thank Lizzie Fricker, Jennifer Lackey, and Ted Poston for very helpful comments
on earlier versions of this paper.
128 Reductionism and the Distinctiveness of Testimonial Knowledge

Given my core thesis, it is very important to begin with a clear characterization

of the subject-matter, which I am designating as ‘testimonial knowledge’. Intu-
itively, testimonial knowledge is knowledge acquired on the (epistemic) authority
of another speaker. Since we will want to focus on cases that bring out what, if
anything, is distinctive of testimonial knowledge, we will restrict our attention to
cases answering to the schema that is characteristic of such cases. Fricker (1994:
136–7) aptly characterizes this schema as ‘‘coming to know that something is so,
through knowing that a certain speaker has asserted it to be so.’’¹
But we can move beyond this intuitive characterization to a more explicit one,
as follows. A has testimonial knowledge that p if and only if
(A) A knows that p;
(B) There is a speaker S whom A observed to offer testimony on occasion
O, such that the proposition that p was understood by A to be
presented-as-true in S’s testimony on O; and
(C) A’s knowledge that p depends for its status as knowledge on both
(i) the reliability of S’s testimony on O,² as well as (ii) A’s epistemic
right to rely on that testimony.
Condition (A) is straightforward: having testimonial knowledge that p entails
knowing that p. Condition (B) is perhaps slightly less straightforward, but the
basic idea is clear enough: having testimonial knowledge that p presupposes
observing and understanding testimony in which that p is presented-as-true.³
The need for the roundabout formulation on which the proposition that p is
described as ‘‘presented-as-true’’ in the testimony, as opposed to a formulation
which would describe the testimony simply as ‘‘testimony that p’’, is needed if we
are to acknowledge that what the hearer can know on the basis of the testimony
need not correspond exactly to what was asserted in the testimony itself. (That
we should acknowledge this possibility is explicitly argued in (Goldberg 2001)
and is implicit in a remark in (Burge 1993: 482 n. 20).) I should add that if
the proposition that p is not presented-as-true in the testimony A observed, then
A’s acquisition of the belief that p from the testimony, even if it does attain the
status of knowledge, would not be knowledge acquired on the epistemic authority
of the speaker. This brings me to condition (C), which expresses the epistemic
dimension characteristic of testimony cases: qua knowledge, one’s testimonial
knowledge that p implicates a testifier’s reliability regarding the truth of p (on
the occasion of testimony to that effect), as well as one’s own epistemic right to
rely on that testimony. As we might put it: (C)(i) ensures that one’s knowledge
is acquired from what in fact was reliable testimony, and (C)(ii) ensures that one
was epistemically entitled to accept that testimony. Although it is controversial
to say what is involved with being epistemically entitled in this way—as we will
see, this is an issue that separates reductionists from anti-reductionists regarding
testimony—the need for the addition of (ii) is indicated once we recognize that
it is one thing to be confronted with what in fact is reliable testimony, and quite
 Sanford C. Goldberg 129

another to be epistemically entitled to believe that the testimony was reliable. (I

will return to this below.)
I submit that conditions (A)–(C) state individually necessary and jointly suf-
ficient conditions on a subject A’s having testimonial knowledge that p. I also
believe that conditions (A)–(C) are neutral with respect to (1) and (2). How-
ever, before I can confirm this, I want to identify the two main positions within
the epistemology of testimony, reductionism and anti-reductionism, since it is
in terms of a theorist’s commitment to one of these positions that she typic-
ally will reason to her verdict regarding (1) and (2). After briefly presenting the
reductionist and anti-reductionist positions, I will argue that it begs no question
against either side to regard (A)–(C) as stating individually necessary and jointly
sufficient conditions on A’s having testimonial knowledge that p. Simply put,
accepting this characterization of testimonial knowledge is consistent with any
position one takes regarding (1) and (2) respectively.

2 REDUCTIONISM AND ANTI-REDUCTIONISM

There are two traditional positions that have been formulated in connection with
the epistemology of testimony. In characterizing these positions I am going to be
following Fricker (1994). My reasons for doing so are strategic. First, Fricker’s
characterization of these positions links them directly to their respective views
regarding (1), the thesis regarding epistemic principles unique to testimony. For
this reason her characterization of these positions enables a quick summary of
how one might think to argue from a denial of (1) to a denial of (2)—precisely
the sort of argument I aim to be criticizing. Second and relatedly, Fricker herself
appears to endorse the sort of argument I aim to be criticizing. Consequently, my
use of her characterization of the positions will inoculate my argument against
the charge that I have somehow begged the question e.g. by accepting a loaded
characterization of the debate in the epistemology of testimony. Finally I might
add, as a third reason to follow Fricker’s characterization, that I find it to be a
good characterization anyway.
According to Fricker, the core question confronting the epistemologist in con-
nection with testimony pertains to the justification of beliefs through testimony.
She characterizes the ‘‘problem of justifying belief through testimony’’ as ‘‘the
problem of showing how it can be the case that a hearer on a particular occa-
sion has the epistemic right to believe what she is told—to believe a particular
speaker’s assertion’’ (1994: 128). She then identifies two possible ‘solutions’ to
this problem:
The solution can take either of two routes. It may be shown that the required step—from
‘S asserted that P’ to ‘P’—can be made as a piece of inference involving only familiar
deductive and inductive principles, applied to empirically established premisses. Alternat-
ively, it may be argued that the step is legitimised as the exercise of a special presumptive
epistemic right to trust, not dependent on evidence. (1994: 128)
130 Reductionism and the Distinctiveness of Testimonial Knowledge

The first route is the reductionist route, the latter is the anti-reductionist route.
From this it is clear that anti-reductionists accept, while reductionists deny,
(1) (the thesis of epistemic principles unique to testimony cases).
Return now to the contention that the characterization of testimonial know-
ledge in Section 1 is neutral, both on the issue between the reductionist and the
anti-reductionist, and on the truth-value of both (1) and (2).
Regarding the issue between the reductionist and the anti-reductionist, I sub-
mit that this matter is aptly cast as a debate over what it takes to satisfy (C)(ii).⁴
Regarding (C)(ii), the reductionist will think that a hearer A has an epistemic
right to rely on the say-so of a speaker S on a given occasion only if A has reasons
(including those of a standard inductive sort) to think that S’s assertion that p is
a reliable indication of the truth of p. Such reasons will presumably involve A’s
beliefs regarding S’s credibility in general, S’s position vis-à-vis the subject-matter
of her testimony, the prima facie plausibility of what S said (as determined by A’s
background beliefs), the absence of the trappings of guile or insincerity, and so
on. (Of course, A’s reasons for accepting the testimony need not involve all of
these, though they may.) The important point for the reductionist is simply that
the justification of A’s belief in what another speaker S says does not appeal to
any epistemic principle unique to testimony cases. Contrast the anti-reductionist
position. For her part, the anti-reductionist will think that A has an epistemic
right to rely on S’s say-so as long as A has no (doxastic, normative, or factual)
defeaters for the claim that S’s say-so was reliable.⁵ This sort of position goes
hand-in-hand with another characteristic thesis endorsed by anti-reductionists,
to the effect that there is a defeasible but presumptive right to trust other speak-
ers. (This is the sort of testimony-specific epistemic principle whose existence the
reductionist is anxious to deny.) It should be clear that (C)(ii) does not settle this
matter, but instead only indicates that it is a matter to be settled (as part of a
complete account of testimonial knowledge).
Regarding the truth-value of (1) and (2), I submit that the above character-
ization of testimonial knowledge entails nothing in this regard. Given the point
of the previous paragraph, we have already seen this in connection with (1): the
truth or falsity of (1) does not affect the adequacy of (A)–(C) as a characteriza-
tion of testimonial knowledge, but rather affects how we should conceive of the
satisfaction of condition (C)(ii). To see that the truth-value of (2) is independent
of the adequacy of that characterization, it suffices to show that the supposition
of either truth-value is consistent with that characterization. Suppose that (2) is
true, that testimonial knowledge is a distinctive sort of knowledge. Then presum-
ably something will make it distinctive, epistemologically speaking. What makes
it distinctive may be that it is knowledge whose status as knowledge depends on
epistemic principles unique to testimony; but perhaps what makes it distinctive is
something else. (Indeed, I will be arguing for the latter case.) The point is simply
that the supposition that testimonial knowledge is distinctive is consistent with
treating (A)–(C) as individually necessary and jointly sufficient conditions on
 Sanford C. Goldberg 131

the possession of the sort of knowledge we are presently supposing is distinctive.

Suppose then that (2) is false, that testimonial knowledge is not a distinctive sort
of knowledge. Then presumably (1) is false; but nothing regarding our charac-
terization of testimonial knowledge would follow. What would follow, rather, is
that the sort of knowledge picked out by (A)–(C) is not an interesting epistemic
kind. Under such circumstances, we might still want to retain the category of
‘testimonial knowledge’ so understood; but if we did we should be aware that it
would have the same status as the category of, for example, ‘microscope know-
ledge’, where this is knowledge acquired by reliance on microscopes.⁶ The point
would then simply be that though we might have pragmatic reasons to continue
to use the category, it is a bit of a grab-bag category that should not be taken to
mark off a distinctive epistemic kind.

3 R E D U C T I O N I S T A N T I PAT H Y F O R T H E
D I S T I N C T I V E - K I N D H Y P OT H E S I S

My aim in this paper is to argue for (2), the epistemic distinctiveness of testimo-
nial knowledge, in a way that does not depend on (1), the hypothesis of epistemic
principles unique to testimony cases. To be sure, theorists who accept (1) have a
reason to accept (2): if there are epistemic principles unique to testimony cases,
then there is some reason to think that the sort of knowledge acquired through
testimony is unique, insofar as a complete account of that knowledge will pre-
sumably invoke those principles unique to testimony (the uniqueness of those
principles carrying over to render the knowledge unique as well). What I deny is
the converse, that (2) is true only if (1) is. On the contrary, I want to argue that
testimonial knowledge is epistemologically distinctive whether or not there are
epistemic principles unique to testimony cases.
Before presenting the case for such a claim, I want to point out why I think it
is worthwhile making this claim. To do so I want to cite reasoning from the liter-
ature in which (by all appearances) the issue of (2), the distinctness of testimonial
knowledge, is taken up from the perspective of the assumption that (2) is true
only if (1) (the claim asserting testimony-specific epistemic principles) is. The
passage I cite is from Fricker; though I suspect that she is not the only philosopher
to approach (2) from the perspective of the offending assumption.
Starting off a section of her 1994 paper with the question, ‘‘Is knowledge
through testimony a distinctive category of knowledge at all?’’ (1994: 136), Frick-
er goes on to defend a qualified negative answer to this question. An examina-
tion of her reasoning suggests that the offending assumption, that an affirmative
answer to her question turns on the existence of a testimony-specific principle of
justification, may be playing a significant role in her reflections on the distinctive-
ness (or not) of testimonial knowledge. I quote her at length:
Testimony . . . does indeed constitute a distinctive kind of epistemic link. There is a
distinctive type of connection, characteristic of testimony, between a state of affairs, and
132 Reductionism and the Distinctiveness of Testimonial Knowledge

a hearer’s coming to believe in its obtaining. This connection runs through another per-
son, a speaker—her own original acquisition of the same belief, her other mental states,
her subsequent linguistic act, which transmits that belief to the hearer. There being this
distinctive type of link between a hearer, and what she comes to believe, in testimony,
means that there is a distinctive type of justification associated with testimony . . .: we
can identify a characteristic justificatory schema S [involving the move from ‘Speaker
S asserted that p’ to ‘p’]. A hearer has knowledge through testimony just when she has
knowledge whose content is given by appropriate instances of the elements of S, and
can cite such knowledge, or evidence for it, in defense of her belief. But what there is
not . . . is any new principle of inference or other normative epistemic principle involved,
which is special to testimony. . . . This makes the ‘problem of justifying testimony’ unlike
the ‘problem of induction’. In the latter, the task is to show the legitimacy of a general
principle of inference, one which is broadly comparable to the principles of deductive
inference in the way in which it validates particular inferences of the form in question.
It is therefore appropriate to approach the ‘problem of induction’ at a completely gen-
eral level. The task is to show that an arbitrary inductive inference is valid, by showing
that the principle of inference involved in any such inference is a valid one. . . . Now the
anti-reductionist may mistakenly suppose that the task of justifying testimony must be
approached by looking for some highly general premise or principle which would serve to
justify an arbitrary testimony belief. . . . My local-reductionist approach avoids the initial
mistake. . . . If what were in question were a special normative epistemic principle, con-
cerning testimony as a distinctive and unitary category of knowledge, then it would indeed
apply indifferently to an arbitrary piece of testimony, and the task of justifying it would
need to be conducted at an abstract general level. . . . But if there is no special epistem-
ic principle in question, and what is common to all and only instances of knowledge
through testimony is just a characteristic kind of belief-producing causal process, then
there is no reason why what justifies belief in particular instances of testimony must be
some proposition or principle applying to testimony in general. Instead, what justifies a
hearer’s belief in a particular assertion may be her knowledge of relevant facts about that
assertion and speaker, which warrant her in trusting him. (Fricker 1994: 136–7; some
italics added)
Fricker goes on to write that
Looking for generalizations about the reliability or otherwise of testimony, in the inclus-
ive sense of serious assertions aimed at communication of belief, as a homogenous whole,
will not be an enlightening project. Illuminating generalizations, if there are any, will be
about particular types of testimony, differentiated according to subject matter, or type of
speaker, or both. True, there is a belief-producing process characteristic of testimony, and
consequently a generic type of justification, as captured in S. But when it comes to the
probability of accuracy of speakers’ assertions, and what sorts of factors warrant a hearer
in trusting a speaker, testimony is not a unitary category. . . . One aspect of this disunity
is . . . that while there are certain limited epistemic rights to trust involved in particu-
lar types of testimony, there is no blanket [presumption that one is entitled] to believe
what is asserted without needing evidence of trustworthiness, applicable to serious asser-
tions aimed at communication as a whole, regardless of subject matter and circumstance.
(Fricker 1994: 137–8; italics mine)
 Sanford C. Goldberg 133

Now it is not entirely clear that Fricker here is assuming that (2) only if (1),
that is, that testimonial knowledge is a distinctive epistemic kind only if there
are testimony-specific epistemic principles. But various things about these pas-
sages suggest that Fricker is making this assumption. First, having started off the
section with the unqualified question whether ‘‘knowledge through testimony’’
is a ‘‘distinctive category of knowledge’’, Fricker moves without comment to the
question whether there are any epistemic principles unique to testimony. (The
lack of comments regarding such a transition would be explained if she is making
the offending assumption.) Second, by the end of the section the only manner in
which she addresses the unqualified question is by noting that testimony is not a
‘‘unitary category’’ insofar as ‘‘the probability of accuracy of speakers’ assertions,
and what sorts of factors warrant a hearer in trusting a speaker’’ are concerned.
Admittedly, this leaves open the issue whether testimonial knowledge is a unit-
ary category in ways that go beyond the fact (noted by Fricker) that there is ‘‘a
belief-producing process characteristic of testimony’’. But nowhere does Fricker
herself note this possibility, let alone endorse it. This silence is curious if she really
does allow for the possibility envisaged: one would be forgiven for thinking that
if Fricker does want to allow for the possibility that testimony is a distinctive cat-
egory of knowledge, in ways that go beyond the fact that such knowledge involves
a characteristic belief-producing process, she would have said so.⁷
These considerations do not vindicate the charge that Fricker is making the
offending assumption that (2) is true only if (1) is. But they do suggest that some-
thing like this assumption may be playing a guiding, if implicit and perhaps
unacknowledged, role in her thinking about the distinctiveness of testimonial
knowledge.⁸ In what follows I argue that such an assumption ought not to figure
into our thinking about the distinctiveness of testimonial knowledge—whether
we are reductionists (like Fricker) or not.

4 I N S U P P O RT O F T H E D I S T I N C T I V E - K I N D
H Y P OT H E S I S

In Section 2 I defended the contention that nothing in the category of testimo-

nial knowledge, understood in terms of (A)–(C), begs any relevant questions. I
now want to draw on that characterization to argue that testimonial knowledge
so understood is a distinctive epistemic kind. The main interest of the argument
to follow is that it assumes nothing that a reductionist should want to deny. If
successful, it shows that a case can be made for (2), the epistemic distinctive-
ness of testimonial knowledge, even if we do not assume (1), the hypothesis of
epistemic principles unique to testimony cases.
I want to begin with a programmatic statement of the position I will defend.
My main contention is that testimonial knowledge is a distinctive kind of
134 Reductionism and the Distinctiveness of Testimonial Knowledge

knowledge in that this sort of knowledge, but no other, is associated with a char-
acteristic expansion in the sorts of epistemically relevant moves that can be made by
the subject in her attempt to identify the direct epistemic support enjoyed by her belief.
As I will try to bring out, this feature of testimonial knowledge reflects the fact
that there is something epistemologically distinctive about relying on the epi-
stemic authority of another rational being: it is because of what is distinctive in
relying on the epistemic authority of another rational being, that there is a char-
acteristic expansion in the sort of moves that can be made in defense of a belief
acquired on such authority. The characteristic expansion, I suggest, is that testi-
monial knowledge gives rise to the hearer’s right to pass the epistemic buck (in a
sense to be characterized below) after her own justificatory resources have been
exhausted. After arguing for this, I will argue further that the phenomenon of
epistemic buck-passing should be recognized as a core part of the phenomenon
that is testimonial knowledge, no matter one’s attitude towards the claim that
there are testimony-specific epistemic principles. This is because what generates
the hearer’s right to pass the epistemic buck is not any testimony-specific epi-
stemic principle. That right is generated rather by the key features at play in cases
in which one rational being relies on the authority of another rational being, in
the course of shaping its beliefs.⁹
I want to begin with a simple case. Smith wants to know the temperature. She
consults a thermometer visible from her kitchen window. The thermometer reads
44◦ F. Smith has strong inductive reasons to believe that the thermometer is reli-
able. So she forms the belief that it is 44◦ F. If challenged to defend her belief that
it is 44◦ F, she will cite her reading of the thermometer, together with all of her
evidence for thinking that the thermometer was reliable. (Some of this evidence
will involve her inductive reasons for thinking that the thermometer is reliable;
but presumably other evidence will come in the form of the coherence of the
present deliverance of the thermometer with Smith’s background beliefs regarding
such matters as average temperatures for this time of year, what she can deduce
regarding the range of possible temperatures from the day’s visual appearances,
etc.) Once she has cited all of this, she has exhausted the responses available to her
in the face of a challenge to defend (or to identify the epistemic support enjoyed
by) her belief that it is 44◦ F.
Now contrast this case with the following. Smith, wanting to know the tem-
perature, asks Jones, who replies by telling her that it is 44◦ F. Smith knows Jones
to be a most reliable source—someone who never asserts something for which
he lacks substantial evidence. Consequently, on the strength of Jones’s word,
Smith forms the belief that it is 44◦ F. If challenged to defend her belief that it
is 44◦ F, Smith will cite her observation of Jones’s testimony, together with all of
her (Smith’s) evidence for thinking that Jones is a reliable interlocutor. (Again,
some of this evidence will involve Smith’s inductive reasons for thinking Jones to
be a reliable interlocutor; but presumably other evidence will come in the form of
the coherence of Jones’s present testimony with Smith’s background beliefs about
 Sanford C. Goldberg 135

average temperatures for this time of year etc.) But—and this is the key—even
after Smith cites all of this, there remains a further response available to her in the
face of a challenge to defend (or to identify the epistemic support enjoyed by) her
belief that it is 44◦ F. She can simply pass the buck to Jones, as follows:
I’ve told you all the reasons I have for trusting Jones’s testimony; so if what you want is
more in the way of a defense of the claim that it is 44◦ F (as opposed to more in the way of
a defense of the claim that Jones’s say-so to that effect was trustworthy), you can simply
ask Jones himself. He’ll be able to provide considerations that more directly support the
claim that it is 44◦ F (or else, failing that, will be able to direct you to the person from
whom he acquired the belief ).

Several points are worth noting about this sort of response.

First, the sort of challenge that might prompt this reply is a sort of chal-
lenge that can, and often does, arise in connection with beliefs formed through
testimony. It is very important to Dien to know who won last night’s baseball
game. (No media sources are presently accessible to him.) Anita tells him that the
Yankees won. Dien asks Anita how she knows this (this was not the reply Dien
wanted!), and she tells him that Brad told her so. Because the outcome of the
game is extraordinarily important to Dien, he wants more in the way of a defense
of the claim that the Yankees won, than Anita’s reasons for thinking Brad’s say-
so to that effect was reliable. (As is common in connection with testimony cases,
Anita’s reasons for trusting Brad’s testimony fall short of establishing the truth
of what Brad said.) Seeing this, Anita replies with a move of last resort: ‘‘OK,
then, go ask Brad. Since he’s the one who told me so, he’ll be in a better position
to address directly your concern regarding whether the Yankees won last night.’’
There can be little doubt that these sorts of situations arise with some regularity.
Second, this ‘‘move of last resort’’ is an epistemically appropriate move for
Anita to make at this point in the dialectic. Having exhausted her own justi-
ficatory resources in giving Dien her reasons for trusting Brad’s testimony, and
seeing that more epistemic support for what was attested is wanted, Anita points
Dien in the direction where further epistemic support can be found: to Brad,
the source himself. And it is clear that Anita is justified in believing that it is to
Brad that Dien ought to be looking, if further epistemic support (for the claim
that the Yankees won last night) is what Dien wants. Indeed, Anita’s reasons for
thinking Brad trustworthy on the present occasion are ipso facto reasons for think-
ing that further support for the claim at issue can be found in connection with
him—either in the form of reasons Brad himself would be able to offer in sup-
port of the claim that the Yankees won last night, or else in the form of warrants
generated by the process through which Brad formed the belief that the Yankees
won last night (warrants that we might be able to discern, even if Brad himself
cannot, by reflecting on those processes¹⁰).
Third and relatedly, nothing in the story told so far should offend any reduc-
tionist sensibilities. This is for the simple reason that the move of last resort is a
136 Reductionism and the Distinctiveness of Testimonial Knowledge

move that is epistemically appropriate only given that the recipient of the testi-
mony (in this case, Anita) has already justified her reliance on that testimony. The
reductionist may place as stringent conditions as she likes on the justified accept-
ance of testimony. The point remains that, even by the lights of a substantive
reductionist requirement on testimonial justification, not all testimony a hearer is
(by the lights of the reductionist conception) justified in accepting is true.¹¹ For
this reason the question can still be raised whether the testimony, which the hear-
er justifiably accepted, was true. To address this question, the hearer’s appeal to
the source interlocutor is perfectly epistemically appropriate; indeed, it is unclear
whether anything other than such an appeal can address this question at all.¹²
Fourth, the take-home point here concerns the epistemic significance of rely-
ing on the presentation-as-true of another rational source. Return to the case
above in which we contrasted a subject’s reliance on a thermometer’s ‘testimony’
with reliance on another rational source’s testimony to the effect that the tem-
perature is 44◦ F. The point I wish to make is that reliance on the deliverances of
another rational being has an epistemological significance absent from the case of
reliance on mere instruments (e.g. thermometers); and that this difference reflects
the different metaphysical status that rational beings have (relative to mere instru-
ments). The difference in metaphysical status is reflected in the fact that rational
beings themselves rely on (and shape and express their views in accordance with)
evidence of the sort that reliable instruments merely offer. A rational being engages
in the project of shaping its beliefs to fit the evidence it has. Because this project
is to some degree under the being’s own rational control, this shaping process
can be done in better and worse ways (epistemologically speaking), in ways that
are epistemically sanctioned, and in ways that are not. Consequently, the notion
of epistemic responsibility finds a home here. The result is that, in relying on a
rational being’s testimony, one is relying on that being to have lived up to her
relevant epistemic responsibilities.¹³ A merely reliable instrument, by contrast,
operates according to the laws of nature.¹⁴ Because there is no ‘rational control’
to speak of, the notion of epistemic responsibility has no home here. In relying
on the ‘testimony’ of a merely reliable instrument, one is relying on the laws of
nature and the systematic and predictable effects these have in connection with
the instrument’s construction (when no noise-introducing factors intervene). It
makes no sense to say that one is relying on the instrument to have lived up to its
relevant epistemic responsibilities; the instrument has none.
The upshot of these reflections is that the reliance on the authority of
another rational being is itself epistemologically significant—something that
should be accommodated by any adequate epistemological account of testimony-
invoking belief and knowledge. The proposal on offer is that the epistemological
significance of such reliance is seen in the characteristic expansion of
epistemically appropriate moves that can be made in response to challenges
to defend (or to identify the epistemic support enjoyed by) one’s belief. To a
first approximation—though a slight modification will be introduced below, in
 Sanford C. Goldberg 137

Section 5¹⁵—the point is this: in cases in which the knowledge was not acquired
on the basis of testimony, the only epistemically appropriate moves that can be
made in response to a ‘‘How do you know?’’ question are those in which one
produces one’s justifications for the belief in question—roughly, the reasons
one has for believing it.¹⁶ If the belief was acquired on the basis of another’s
testimony, however, a subject has not exhausted the epistemically appropriate
moves available to her, once she has exhausted her justification for the belief.
On the contrary, when the knowledge is testimonial knowledge, even after the
subject has exhausted her justification for the belief, there remains the ‘move of
last resort’, whereby she passes the epistemic buck onto her interlocutor. Since
this move is epistemically appropriate—the subject is epistemically entitled to
make this move—only if she has already vindicated her epistemic right to rely
on the testimony in question, the hypothesis of epistemic buck-passing begs no
question against even the most rabid reductionist regarding testimony.

5 O N T H E D I S T I N C T I V E N AT U R E O F T E S T I M O N I A L
S U P P O RT

My claim so far has been that testimonial knowledge is a distinctive epistemic

kind, not for implicating some testimony-specific rule of inference, but rather
for being associated with the phenomenon of epistemic buck-passing. But it
might be wondered whether this case for the distinctiveness of testimonial know-
ledge succeeds. One worry on this score concerns whether what I have been
calling the phenomenon of epistemic buck-passing is involved in cases of non-
testimonial knowledge.¹⁷ If it is, then this feature of testimonial knowledge does
not demarcate it as a distinctive epistemic kind. Another issue concerns whether
the phenomenon of ‘epistemic buck-passing’ should be understood as an epi-
stemic phenomenon in the first place. Perhaps this phenomenon merely registers
an interesting psychological fact about testimonial knowledge, that when speakers
have such knowledge, they are disposed to pass the buck in the manner described
above. If so, then this feature of testimonial knowledge does not demarcate it as a
distinctive epistemic kind. The comments that follow attempt to address both of
these worries; I address the latter worry first.
Various things can be said in defense of the contention that the sort of buck-
passing described above is to be understood in the first instance in epistemological
terms.
The first point to be made builds on something implicit in the discussion in
Section 4. Implicit in the discussion of the phenomenon of buck-passing is that
we should distinguish between (a) the justifications that a hearer has to accept
for what was attested, on the one hand, and (b) the total epistemic support that a
hearer has for her belief that p, when this belief is a belief in what was attested.
Take a case in which hearer A is justified in taking S’s testimony to be a reliable
138 Reductionism and the Distinctiveness of Testimonial Knowledge

indication of the truth of what is being attested. Since A is justified in taking S’s
testimony to be a reliable indication of the truth of what is being attested, A is
justified in thinking that the testimony is properly based to count as knowledge.
As a result, A is prima facie justified in thinking that whatever it is that provides
the proper base for S’s testimony also provides a base, albeit inaccessible to A her-
self, for A’s own belief in what was attested.¹⁸ This reflects the difference between
(a) and (b): A is at least prima facie justified in believing that the epistemic sup-
port that her testimony-based belief enjoys outstrips what she herself can provide
by way of a justification for that belief.
It is worth characterizing the nature of the sort of support at issue in (b). Our
question is this: what sort of support, beyond the support provided by A’s justi-
fication for accepting S’s testimony, is provided by that testimony itself? I will
call this sort of support distinctly testimonial (epistemic) support. Two immediate
things should be noted about distinctly testimonial support: first, it is a sort of
support that will typically be inaccessible to the hearer A herself ; and second, nev-
ertheless it is a sort of support that bears more directly on A’s testimony-based belief
than A’s own justification does.
First, distinctly testimonial support is a sort of support that will typically be
inaccessible to the hearer herself. This merely reiterates the point that speaker
S’s basis for believing what she expresses in her testimony is itself not typically
accessible, and certainly not accessible ‘from the armchair’, to hearer A. But at
the same time, distinctly testimonial support bears more directly on A’s belief
than A’s own justification does. As we noted above, A’s justification for believ-
ing that p is grounded in A’s belief that S’s testimony was reliable (or at least
that it was not unreliable), which in turn is supported e.g. by A’s beliefs that S
is generally trustworthy, that S is competent with respect to the subject matter
of her present testimony, that what S said coheres with A’s relevant background
beliefs, etc. We might say: these beliefs of A’s provide direct epistemic support
for A’s belief that S’s testimony was reliable, and hence (given that S’s testimony
was to the effect that p) provide indirect epistemic support to A’s belief that p.¹⁹
This distinction between direct and indirect support is a key to understanding
the possibility (noted above) of someone challenging A’s belief that p, even after
recognizing that belief as justified. Confronted with such a challenge, nothing
that A says to establish her right to accept the testimony will have a direct bearing;
what will have such a bearing is the basis A’s source (= S) had for the belief she (S)
expressed in the testimony A accepted.²⁰
This distinction between direct and indirect epistemic support enables us
to respond to the first worry mentioned at the outset of this section, to the
effect that testimonial knowledge is not the only sort of knowledge associated
with the phenomenon of epistemic buck-passing. Consider the following case.
McSorley reports that there was a red car in Richardson’s driveway a half-hour
ago. Richardson (who was not home at the time) asks him how he knows that.
McSorley responds that he distinctly remembers having seen a red car there. A lay
 Sanford C. Goldberg 139

epistemologist by nature, McSorley adds that memory and perception are reliable
belief-forming processes. But when Richardson presses McSorley regarding how
he knows that, McSorley simply shrugs: ‘‘I don’t know; go ask the experts
working on perception and memory.’’ If this is a case of epistemic buck-passing,
it challenges the idea that the phenomenon of epistemic buck-passing is what
singles testimonial knowledge out as a distinctive epistemic kind.
Two things can be said in response. First, it is not clear that the foregoing is
a case of epistemically acceptable buck-passing; some epistemic internalists will
want to deny this. But even if we waive this first point, the challenge posed by this
case can be met by appeal to the distinction between direct and indirect epistem-
ic support. Testimonial knowledge is unique in that the epistemic buck-passing
appropriate to it is an epistemically appropriate move whose point is to identify
a direct source of epistemic support for the content believed. When Anita passes
the buck to Brad, she does so in a context in which it is already settled that she
has an epistemic right to rely on his testimony; what is at issue concerns the dir-
ect epistemic support for the content at issue (namely, that the Yankees won last
night). Her passing the buck to Brad speaks precisely to this issue of direct epi-
stemic support for the content believed. The contrast with the case above is clear.
McSorley passes the buck to the relevant experts on perception, not to identify
a source of further direct epistemic support for his claim that there was a red car
in Richardson’s driveway a half-hour ago, but rather to identify a source of direct
epistemic support for the claim that perception and memory are reliable belief-
forming processes. The result is that, even if he is entitled to pass the buck in this
way, his doing so is a move that is epistemically distinct from the buck-passing
move appropriate to testimony.
The contrast, then, is this. Testimony cases are those in which the subject’s
own justification for her belief provides only indirect support for what she
believes—the point of buck-passing in such cases being to manifest the subject’s
awareness of further, direct epistemic support (albeit introspectively inaccessible
to her) for what she believes. The case of McSorley’s report of a red car in
Richardson’s driveway is one in which the subject’s own justification for his
belief already provides direct support for what he believes—the point of buck-
passing in such cases being to provide further direct support for a hypothesis
(the reliability of perception and memory) that itself provides further indirect
support for what he believes. Distinctly testimonial support for a hearer’s
belief that p, then, is epistemic support that (in virtue of the hearer’s reliance
on her interlocutor’s authority) both directly bears on the truth of p, and is
introspectively inaccessible to the hearer herself.
It is worth noting that other authors have appreciated something very much
in the spirit of my characterization of distinctly testimonial support. Thus Burge
(1993: 485–6) distinguishes between a speaker’s ‘proprietary justification’ and
her ‘extended body of justification’, in a way that corresponds to my talk above
of (a) the justifications that a hearer has to accept what was attested, on the one
140 Reductionism and the Distinctiveness of Testimonial Knowledge

hand, and (b) the total epistemic support that a hearer has for her belief that p. And
similarly, Faulkner’s (2000) ‘hybrid’ theory of testimony distinguishes between
the matter of being justified in accepting testimony, and the epistemic support a
recipient would have once she has accepted a piece of testimony. Faulkner desig-
nates the former sort of support as falling under the ‘principle of assent’ and the
latter as falling under the ‘principle of warrant’; and he goes on to note that the
‘extended warrant’ of a proposition presented by testimony is
the conjunction of the warrant possessed by the original speaker for believing this pro-
position together with any further warrant provided by the chain(s) of communication.
(2000: 591)²¹
My present point is a version of the same claim. Only the significance that I am
attaching to this claim is not restricted to the nature of testimonial support, but
rather is being used to suggest the sense in which testimonial knowledge is a dis-
tinctive epistemic kind. My claim has been that this distinctiveness is seen in
what we all should recognize as the familiar ‘move of last resort’ appropriate to
testimony cases; and that even reductionists about testimony ought to accept the
hypothesis of distinctly testimonial knowledge, so defended.

6 C O N C LU S I O N

The main ambitions of this paper have been (first) to defend the hypothesis
that testimonial knowledge is a distinctive epistemic kind, and (second) to offer
a defense that should not offend any party to the debate regarding the proper
account of testimonial justification. If successful, my argument should put to
rest any thought of denying the distinctiveness of testimonial knowledge on the
grounds that there are no testimony-specific epistemic principles. More posit-
ively, it should convince all parties that (and in what sense) testimonial know-
ledge is an epistemically distinctive kind after all.

REFERENCES
Audi, R. (1997), ‘The Place of Testimony in the Fabric of Knowledge and Justification’,
American Philosophical Quarterly, 34/4: 405–22.
Burge, T. (1993), ‘Content Preservation’, Philosophical Review, 102/4: 457–88.
Faulkner, P. (2000), ‘The Social Character of Testimonial Knowledge’, Journal of Philo-
sophy, 97/11: 581–601.
Fricker, E. (1987), ‘The Epistemology of Testimony’, Proceedings of the Aristotelian Soci-
ety, suppl. vol. 61: 57–83.
(1994), ‘Against Gullibility’, in B. K. Matilal and A. Chakrabarti (eds.), Knowing
from Words (Amsterdam: Kluwer Academic Publishers), 125–61.
Goldberg, S. (2001), ‘Testimonially Based Knowledge from False Testimony’, Philosoph-
ical Quarterly, 51/205: 512–26.
 Sanford C. Goldberg 141

Graham, P. (2000), ‘Transferring Knowledge’, Noûs, 34/1: 131–52.

Grice, H. P. (1989), Studies in the Way of Words (Cambridge, Mass.: Harvard University
Press).
Jack, J. (1994), ‘The Role of Comprehension’, in B. K. Matilal and A. Chakrabarti (eds.),
Knowing from Words (Amsterdam: Kluwer Academic Publishers), 163–93.
Lackey, J. (1999), ‘Testimonial Knowledge and Transmission’, Philosophical Quarterly,
49/197: 471–90.
(2003), ‘A Minimal Expression of Non-Reductionism in the Epistemology of
Testimony’, Noûs, 37/4: 706–23.

N OT E S

1. As Fricker herself remarks (1994: 130), it is appropriate to restrict one’s attention

to these cases, since it is in these cases that one can get the clearest handle on what,
if anything, is distinctive of the sort of epistemic support testimony provides. See
also Jack (1994: 163), who speaks of such cases as ‘radical communication’, where ‘‘a
hearer comes to believe what another says and there is no basis for his belief which is
independent of the testimony itself.’’
2. See Fricker (1994: 147) and Graham (2000: 142) for a characterization of the rel-
evant notion of reliability. (Here I aim to be neutral on how best to understand this
notion.) And see also Lackey (2003) for a defense of the claim that among the neces-
sary conditions on testimonial knowledge will be a condition on the environment in
which testimony is received. (If Lackey is correct, this will affect the relevant notion
of reliability.)
3. I borrow the notion of a presentation-as-true, as applied to testimony, from Burge
(1993).
4. It might be thought that the issues between the reductionist and the anti-reductionist
carry over to (C)(i) as well. That is, it might be thought that reductionists and anti-
reductionists will disagree over the sort of epistemic support provided by the de facto
reliability of a piece of testimony: the reductionist will maintain, while the anti-
reductionist will deny, that the epistemic support provided by a piece of reliable
testimony can be subsumed by the factors that entitle the hearer to accept that testi-
mony. (Call this the ‘Factors Generating Entitlement Exhaust Epistemic Support’
thesis, or FGEEES for short.) However, one of the conclusions of this paper will be
that the reductionist should not endorse the FGEEES thesis. On the contrary, reduc-
tionism is best construed (à la Fricker) as a position regarding the conditions on a
hearer’s having the epistemic right to accept what she is told. Construed in this way,
reductionism is consistent with denying the FGEEES thesis; and in Sections 4 and 5
I present arguments to the effect that the reductionist ought to reject FGEEES. See
also Lackey (Chapter 8 in this volume) for an argument against FGEEES. Although
Lackey (who appears to assume that reductionism is committed to FGEEES) takes
her argument against FGEEES to be an argument against reductionism itself, I note
that reductionists have the option of taking her argument to be a reason not to be
committed to FGEEES.
5. I borrow the notion of doxastic, normative, and factual defeaters from Lackey (1999:
474–6).
142 Reductionism and the Distinctiveness of Testimonial Knowledge

6. This way of putting the point was suggested to me by an anonymous referee for
another paper of mine.
7. One final consideration might be worth noting. Fricker’s point about ‘no illuminat-
ing generalizations’, which is meant to support the verdict that testimonial knowledge
is not a ‘unitary category’, settles that matter only if the assumption of illuminating
generalizations is required by the hypothesis that testimonial knowledge is a unitary
category. So far as I can tell, the only reason for thinking that the assumption of
illuminating generalizations is required by the hypothesis that testimonial know-
ledge is a unitary category, is if one thinks that the unitary category claim rests
on the existence of testimony-specific epistemic principles (which the illuminating
generalizations might then provide, or at least justify).
8. Although I should say that Fricker herself has told me (in private correspondence)
that she did not intend to be making the offending assumption.
9. Some readers may recognize the appeal to the distinctiveness of relying on a rational
being from Burge (1993). Unlike the argument from Burge (1993), however, here
I aim to appeal to what is distinctively involved in relying on a rational being, not
to warrant an anti-reductionist thesis, but rather to warrant a claim regarding the
distinctiveness of testimonial knowledge. I share a premise with Burge, but I want to
establish a different conclusion.
10. It might be thought: insofar as we are relying on the warrants (as opposed to the
justifications) Brad had for claiming that the Yankees won last night, we are treating
him as a (merely) reliable mechanism—thereby undermining the ‘contrast’ I am try-
ing to make between, for example, the case of relying on a thermometer and the case
of relying on a person’s testimony regarding the temperature. But the contrast is not
undermined even in this sort of scenario. First, the manner in which ‘testimony’ is
‘produced’ differs according to whether the source is a thermometer or a person. To
wit: if a person asserts something for which she lacks adequate grounds, we blame
her for not living up to her epistemic responsibilities. There is no analogue of this in
the case of a thermometer: even if the thermometer is not functioning properly, and
so in this sense ‘provides’ a false temperature reading, we would not blame the ther-
mometer as failing to live up to its epistemic responsibilities—for it has none. And
second and relatedly, unlike a thermometer, a person is, or at least is regarded as hav-
ing the epistemic responsibility to be, rationally sensitive to conditions that would
defeat her warrant for a given claim. So even if we are relying on Brad’s warrant for
the claim that the Yankees won last night, still, the fact that we are relying on Brad
qua rational agent has epistemic significance. The proof of this is that, had some-
thing gone wrong in the process by which Brad arrived at his baseball belief—what
he was watching on TV was not the Yankees game itself but rather a bunch of act-
ors pretending (lamely!) to be playing a baseball game—we might well have blamed
Brad for not having been alive to that. (He should have known as much: although
the game was supposed to be Yankees vs. Red Sox, neither of the two teams wore the
Yankees’ tell-tale pinstripes.) No such blame has any place in connection with our
reliance on a merely reliable instrument. I should say that, in appealing to a notion of
epistemic responsibility and epistemic blame, I am not assuming that these notions
exhaust the notion of epistemic justification; I regard that as an open question.
 Sanford C. Goldberg 143

11. There may be epistemic externalists who would deny that there are justified but
false beliefs (whether testimonial or otherwise). To those, the claim I am making
could be reformulated by replacing ‘justified’ with one of the internalist or subjective
justification notions they recognize. (If they recognize none, then the point could be
case in terms of ‘seeming subjective justification’ or some such.)
12. It should be clear from this that, in passing the epistemic buck in the way appropriate
to testimonial knowledge, the subject is not shirking her epistemic responsibilities.
Rather she is pointing to a source of further (and, I will argue below, direct) epi-
stemic support for the content believed—support which the recipient of testimony
herself is not typically in a position to provide herself (at least not on the basis of
what she can introspect).
13. It is worth noting in this connection that the relevant epistemic responsibilities
include those attending the performance of the speech act of assertion (and any oth-
er speech acts, if any, apt for the transmission of knowledge). Consider for example
Grice’s norm of Quality, which he took to include the maxim, ‘‘Do not say that for
which you lack adequate evidence’’ (Grice 1989: 27).
14. It might be thought that this contrast between merely reliable instruments and
rational agents is weakened if the behaviour of the latter is governed by psycholo-
gical laws. There may be something to this point. At the same time, and without
putting too fine a point on it, we can say that even if rational agents are governed
by psychological laws, it would seem that our knowledge of these laws, and our abil-
ity to apply these to particular cases to predict what a speaker will say, is (and, as
a matter of practical fact, is likely to remain) deeply impoverished. This being so,
there will remain an epistemic asymmetry between our reliance on merely reliable
instruments and our reliance on rational agents. When the workings of the former
are sufficiently well understood (take the case of the thermometer), we can vindicate
our belief in their reliability by appeal to laws whose predictions in the present case
we can discern; not so in the case of rational agents. I would hope that more can be
said in defense of this asymmetry than the present epistemological point; but doing
so will have to be the aim of another paper. (I thank Jennifer Lackey for indicating
the need to address this matter.)
15. The modification will involve distinguishing between direct and indirect types of
epistemic support. In Section 5 I discuss the need for this distinction.
16. I am not supposing, however, that what a subject does in response to a ‘‘How
do you know?’’ question—namely, offering one’s justification—is producing the
sort of thing that renders one’s true belief knowledge. Some writers (e.g. Plantinga)
have argued that knowledge does not require justification at all; still others (e.g.
Goldman), that the sort of justification required by knowledge is not the sort of jus-
tification that needs to be accessible to the subject herself. My present point does not
require entering into these controversies.
17. I thank Ted Poston and Jennifer Lackey, both of whom (independently) made this
suggestion to me. (The illustration to follow is Poston’s.)
18. A’s justification, for thinking that whatever it is that provides the proper base for
S’s testimony also provides a base for A’s own belief in what was attested, is prima
facie: perhaps A has other beliefs which would defeat the epistemic support that S’s
144 Reductionism and the Distinctiveness of Testimonial Knowledge

testimony (and the evidence and warrants S herself had for what she asserted) would
otherwise provide to A.
19. I admit that I am not entirely certain how best to explicate this distinction between
direct and indirect epistemic support; but it has sufficient intuitive appeal, and it
is sufficiently clear at least in outline, that it can be appealed to prior to having a
complete analysis in hand.
20. Of course S’s own ground for her testimony that p may not settle all questions
regarding the truth of p. But my present point is only that S’s ground can settle some
questions which A’s justification leaves open—there is an epistemic asymmetry here.
It goes without saying that complications arise when S, in turn, is merely a link in
a still-further chain originating prior to her. But then S will be in a position to pass
the buck to her interlocutor, and so on, until such time as the original interlocutor is
reached.
21. I note that a view such as Faulkner’s, which allows for the transmission of warrants
through chains of testimony, need not be seen as committed to the transmission
view of testimonial knowledge, according to which in order for A to know or be
justified in believing that p on the basis of S’s testimony, S (or someone else in the
chain preceding S) must know or be justified in believing that p.
 7
Testimony and Trustworthiness
Keith Lehrer

This paper is concerned with the relationship between testimony, trustworthi-

ness, and knowledge. Testimony leads to the acquisition of knowledge under
some circumstances but not all circumstances. The problem is to analyze when
testimony suffices for the acquisition of knowledge. A solution to the problem
depends on the analysis of knowledge and testimony. I shall argue that when
the knowledge acquired is discursive knowledge (Lehrer 1997, 2000), which is
knowledge that involves justification and defensibility, the trustworthiness of self
and others is a condition of knowledge. The condition of trustworthiness of oth-
ers appears to pose a problem, however. I must evaluate the trustworthiness of
others to evaluate their testimony. This leads to the question of whether I am
trustworthy in my evaluation of their trustworthiness. That leads to the further
question of how I am to evaluate my own trustworthiness without appeal to the
intersubjective constraint based on the testimony of others. If, however, I must
evaluate the trustworthiness of others before accepting their testimony, and I
must evaluate my own trustworthiness by appeal to the testimony of others, then
I am involved in a circle of evaluation from which there is no exit. The solution to
the paradox is to be found in a virtuous explanatory loop of trustworthiness.
There is a simple theory about the role of testimony in knowledge that might
be called a causal theory of knowledge or a transmission theory of knowledge.
Such theories bypass the issue of evaluation as unnecessary. The model is bottom
up and escapes from the circle of evaluation by ignoring it or adding it as a refine-
ment. So the idea is that some fact causes a first person to form a belief. Then,
as the result of communication, the first person causes a belief in another, and
the causal sequence effected by communication continues from person to person.
Assuming that the form of causation suffices for knowledge in the original belief
and in the transmission of belief from person to person, causation suffices for the
acquisition of knowledge from testimony.
The simple theory has a certain plausibility without any question. Reid (Lehr-
er and Beanblossom 1975; Smith and Lehrer 1985) noticed that a young child
seems to be gullible by nature beyond the restraint of reason, partly because it
does not understand the distinction between veracity and deception. Moreover,
all of us retain this initial inclination to believe what we are told, according to
146 Testimony and Trustworthiness

Reid, though we become constrained by our experience of error and deception.

The causal theory has a plausibility derived from the causal influence of the
speech acts of others. The acknowledged problem is that there are causal rela-
tionships between facts and beliefs and between the beliefs of one person and
those of another that do not suffice to yield knowledge in the first instance or in
the transmission. For example, suppose the subject of the belief is totally ignor-
ant of the causal relation and has no idea why he or she believes what she does.
This problem is met by complicating the simple theory, arriving at the modi-
fied simple theory. The modified simple theory says only that there is some kind
of causal relationship that suffices to yield knowledge in the original belief, and
some kind of causal transmission of the belief by testimony that suffices to transfer
knowledge. Philosophers defending the modified simple theory have undertaken
to explain what kind of causal relationship is required in both cases.
There is, however, a basic objection to such theories for obtaining discursive
knowledge, namely, that the subject of the belief, whether of the original belief
or the transmitted belief, might have no idea whether the belief is worthy of their
trust. I have called this (Lehrer 2000) the opacity objection. The merits of the
belief may be opaque to the subject. Therefore, the subject may not be in a posi-
tion to justify or defend the belief. There is a reply to this objection that will take
us to the heart of the issue. It is that some kinds of causal relations may have the
result that the belief is formed in a special way, a reliable way, for example, or in
conformity to some norm or rule that insures that the belief is justified. The reas-
on that this takes us to the fundamental issue in the epistemology of testimony is
that it assumes that there is some way that a belief is formed that is necessary and
sufficient for knowledge. There are reasons to doubt that it is necessary and ones
to doubt that it is sufficient. Let us consider them briefly.
A belief may be formed in some way that is evidentially defective. I may believe
what someone tells me when I have reason to distrust their testimony. As I eval-
uate the belief, however, perhaps as the result of obtaining other evidence for
the belief after forming it, I may decide that the person was trustworthy in this
instance in spite of my original reasons for mistrusting the person. I acquire
knowledge as the result of my subsequent evaluation. Does that mean that I had
knowledge when I formed the original belief based on testimony that I had reason
to distrust? I do not think I knew at the time that I formed the original belief.
There was at that stage an objection I was in no position to meet, namely, that I
had reasons to distrust the testimony on which I based my belief. It was not the
formation of the belief but the subsequent evaluation of the belief on the basis of
evidence obtained later that resulted in knowledge. If the later evaluation of the
belief in terms of evidence was necessary, then the earlier formation of the belief
on the basis of testimony was not sufficient. Moreover, if the evidence for the
belief was sufficient for obtaining knowledge later, then the testimony might not
have been necessary either.
 Keith Lehrer 147

There have been enough examples offered in support of the claim that caus-
al theories about the formation and transmission of belief are neither necessary
nor sufficient for knowledge so that the addition of further examples is not likely
to influence conviction any more than the reiteration of such examples. Why
would someone refuse to acknowledge the relevance of examples of people form-
ing beliefs on the basis of epistemologically defective evidence of testimony and
later acquiring knowledge on the basis of evidence obtained in the interval? One
reason is that the simple theory can be modified in an ad hoc manner to require
that the person have no reasons to distrust the testimony. It is natural to proceed
in this way to refine a causal or reliablist theory of knowledge, but the refine-
ment expands the account to include background considerations of the subject,
namely, what reasons the person has for not accepting the testimony.
Once reasons for not accepting testimony are acknowledged to be necessary
for the acquisition of knowledge, however, why should we deny that the reasons
that we have for accepting testimony are relevant to the acquisition of knowledge?
The negative route of requiring that we not have reasons for not accepting testi-
mony is an attempt to conceal the importance of the positive route of acknow-
ledging that we have reasons for accepting some testimony as being worthy of
our trust and other testimony as unworthy of our trust. These reasons are relev-
ant determinants of knowledge. When one seeks to justify or defend a claim to
knowledge on the basis of testimony, the reasons one has for trusting or distrust-
ing testimony become salient. Discursive knowledge of testimony depends upon
such reasons.
So let us consider the role of testimony in the acquisition of discursive know-
ledge. Suppose someone tells me that p, and I accept what the person tells me,
that p, as a result. Under what conditions do I know that p? Suppose that I ask
myself whether I know that p and consider objections that might be raised against
the claim that I know that p as the result of accepting that p on testimony. It
might be objected that the person who has told me that p, let us call her the
informant, does not know that p. Should we conclude that if the informant does
not know that p, then we cannot know that p as a result of accepting that p
because the informant tells us that p? Is it necessary for us to know that p that the
informant knows that p?
It is surely an objection to the claim to know on the basis of testimony that the
informant whose testimony one accepts does not know whereof she speaks, that
is, does not know that p. The question is whether it is decisive, and the answer
is negative. The reason is that the informant might have come to believe that p
as a result of being told that p by another, a man she has reasons to think might
deceive her. She does not know that p because there is an objection to accepting
that p on the testimony of the man, namely, that he is not worthy of her trust,
and she cannot meet the objection because of her reasons for thinking the man is
untrustworthy.
148 Testimony and Trustworthiness

However, suppose I know that the man knows whereof he speaks, contrary to
reasons the informant has for thinking he is untrustworthy. I know that she is
misled about his being untrustworthy because, whatever he tells her about the
truth of p, he has obtained and conveyed the truth when he tells her what he does.
So, when she tells me that p, having herself come to believe that p on the basis of
the testimony of the man, despite her reasons for thinking that man is untrust-
worthy, I might be in a better position on the basis of my background system,
my evaluation system, for evaluating the claim of the informant than she was in
evaluating what the man told her. I am in a position to evaluate the testimony of
the informant as having originated with a trustworthy source while she is not in a
position to make that evaluation.
The foregoing example of the informant who herself believes and conveys
something when she has reasons to consider her source untrustworthy is inter-
esting in that she is herself untrustworthy both in believing what she is told and in
the way she conveys what she is told to another. I acquire knowledge from what
she has told me, not because she is trustworthy, but because I know that she is an
accurate conduit of truth from a source who has knowledge. She has successfully
obtained and transmitted truth, however untrustworthy she has been in forming
and conveying her beliefs. There is a remaining question about the case, however.
It is whether it is her testimony that justifies me in what I accept and converts
to knowledge. The issue is subtle, but the answer is negative. She is a conduit of
truth, and I may know that, but her testimony is not the evidence or justifica-
tion that converts to knowledge, for she is untrustworthy. The justification that
converts to knowledge is that she is conveying what the man said, and he is trust-
worthy. It is his trustworthiness, not her testimony, that is the evidence for the
truth of what she says.
This reflection enables us to make the question concerning the evidence or jus-
tification of testimony more precise. Sometimes the testimony of a person that p
is evidence that p, and sometimes not. It depends, though not only, on whether
the person is herself a trustworthy informant in the sense that she has acquired
and conveyed the information in a trustworthy way. Whether the informant’s
testimony is a source of evidence and justification also depends on the evaluation
of the informant by the person who receives the information. The informant may
be trustworthy when the person receiving the testimony thinks the informant is
not trustworthy. Or, on the contrary, the person receiving the testimony, and
even believing it, may think, with justification, that the informant is trustworthy,
when he is not trustworthy. In the first case, the person receiving the information
is not justified in accepting that the testimony is trustworthy. In the second case,
the person is justified in accepting that the testimony is trustworthy, but the jus-
tification is defeated or refuted by some mistake the person has made, however
unavoidable it may have been, in justifying acceptance of the testimony of the
informant.
 Keith Lehrer 149

The testimony is in itself a source of evidence when the informant is trust-

worthy in the testimony. The testimony in itself does not constitute evidence
otherwise. A person may have evidence, as in an example considered, that what
the informant says is true which is based on knowledge about the etiology of the
testimony when the informant is herself untrustworthy. That does not mean that
the testimony in itself is a source of evidence. Knowledge of the etiology reveals
that what the informant says is true in spite of the fact that the informant is her-
self untrustworthy. She is not worthy of trust even though we have evidence that
what she says is true.
The preceding will not convince everyone that the testimony of the person
is not a source of evidence by itself for the truth of what she says when she is
untrustworthy. It may be modified to strengthen the conviction that the testi-
mony is not by itself a source of evidence by supposing that the informant intends
to lie by saying the opposite of what someone tells her. Suppose we add that the
person who tells her what he does knows she will lie to us. As a result, he tells
her the opposite of what he knows to be the case. If we know this as well, we
will accept what she says. We will reason that the person who told her what he
did told her the opposite of what he accepted and knew to be true. She then told
us the opposite of what he said. So, we conclude what she says is true, since it
corresponds to what the person knew who told her the opposite.
Her testimony, that is, her testifying to the truth of what she says, is no evid-
ence by itself of the truth of what she says, for we know she is lying. However,
given our knowledge of the history of it all, we conclude that what she said is true
because it corresponds to what the man knew who told her the opposite of what
she told us. We may know that what she says is true even though her testifying
to the truth of what she says is no evidence by itself of the truth of her testimony.
Suppose we only knew about her and her intentions. We would conclude that
what she says is probably false because of her intention to lie.
This line of thought tests two divergent theories about testimony discussed by
Coady (1992, 2002). One is reductive and says that the evidence of testimony is
derived from induction and attempts to reduce the evidence of testimony to oth-
er evidence that the person is telling the truth (Hume 1978). On this account,
when, and only when, we have inductive reasons, whether direct or indirect, for
supposing that what a person says is true, does what the person says constitute
evidence for the truth of what the person says. The other is nonreductive and
foundational and supposes that testimony is an original source of evidence that
cannot be reduced to another. On the reductive account, both examples involve
an informant whose testimony is evidence. On the foundational account, the
testimony of the informant is also evidence, though it may be defeated. The cor-
rect view, it seems to me, is neither of these. The informant in the two examples
tells us something, and we have evidence for the truth of what we are told in both
cases. However, the informant testifying to the truth of what she says, contrary
150 Testimony and Trustworthiness

to both accounts, is no evidence by itself, defeated or not, for the truth of what
she says. The evidence for the truth of what she says lies in our knowledge of the
history and etiology of what she says rather than in her testifying to the truth of
it. The latter, considered in terms of our knowledge of her alone, would lead us to
doubt what she says.
So when and how does testimony yield evidence? Part of the answer is that the
person testifying to the truth of what she says must be trustworthy in what she
accepts and what she conveys. Of course, the person does not need to be infal-
lible to be trustworthy. A person can be fallible and worthy of trust. We may
long for perfection, but we recognize the trustworthiness of those who sometimes
err including ourselves. You do not have to be perfect to be trustworthy. Notice,
however, that to be in a position to meet objections to accepting something on
the basis of testimony, we must be in a position to meet the objection that the
person is not trustworthy in testifying to the truth of her testimony. Moreover,
since she may fail to be trustworthy either because of the way in which she accepts
what she does, or in the way in which she conveys what she does, we must be
able to meet the objection that she is not trustworthy in either of these ways.
Furthermore, a person can be trustworthy in what she accepts but fail to be suc-
cessful in obtaining truth. Since we are fallible and capable in principle of being
invincibly deceived, trustworthiness is only contingently connected with success
in obtaining truth. Thus, reliability considered as a high frequency of success in
obtaining truth in what one accepts is not entailed by trustworthiness. This is
contrary to Schmitt (1994, 2002).
It is clear that the trustworthiness of the informant is not sufficient for
the conversion of our accepting what she tells us to knowledge. That she is
untrustworthy is an objection that is met by her being trustworthy. But, as we
have noted, trustworthiness is only fallibly connected with truth. So another
objection is that her trustworthiness is not successfully truth-connected. To
meet the objection we must accept that her trustworthiness is successfully truth-
connected. We must accept not only that her trustworthiness is in general reliably
truth-connected but, also, that in her current situation the trustworthiness of her
testimony is successfully truth-connected. Thus, evidence and justification that
converts testimony to knowledge assumes acceptance of her trustworthiness and
the successful truth-connection of it. Is that sufficient? It is not.
What else is required for the evidence and justification of testimony to convert
to knowledge other than our acceptance of her trustworthiness in her testimony
and our acceptance of the successful truth-connectedness of her trustworthiness?
The conversion requires that we be correct and not in error in our acceptance
of the trustworthiness of her testimony and the successful truth-connectedness
of it. There is a kind of subjective or, as I prefer to say, personal evidence and
justification (Lehrer 2000) resulting from our background system of evaluation,
of acceptance, preference over acceptance and reasoning about acceptance. The
 Keith Lehrer 151

conversion to knowledge, however, requires that such personal justification not

be refuted or defeated by errors in the background system.
We noted that the justification of testimony is neither reductive nor founda-
tional. So what is the source of the justification? It is the background system to
which we have referred and which we use to meet objections to the testimony
we accept. Why are the reductive and foundational explanations of justification
unsatisfactory? The answer becomes apparent when we press further to answer
objections to accepting testimony. No sooner do we evaluate the trustworthi-
ness of our informant, thus meeting the objection that the informant is not
trustworthy, than a deeper question arises. Are we trustworthy in the way in
which we evaluate the trustworthiness of the informant? We are trustworthy in
the way in which we evaluate the trustworthiness of others, in part, as the res-
ult of experience. When we evaluate others as trustworthy on occasions when
they are not, we learn of our mistake. However, one primary way in which we
learn of this is through the testimony of others. False or deceptive testimony is
sometimes revealed directly through experience, but it is very frequently revealed
through the testimony of others. Moreover, intersubjective agreement with oth-
ers, with their testimony, is a primary test of our own trustworthiness. When we
are untrustworthy in what we accept or how we evaluate claims, including the
testimony of others, this is often revealed to us by the testimony of others.
The foregoing reflections contain the explanation for why no foundational
theory, whether reductive or nonreductive, of the evidence of testimony will
prove adequate. There is an interactive loop between the trustworthiness of self
and others in the evaluation and justification of testimony. To see this, consider
the sort of justification that is necessary for discursive knowledge, namely, justi-
fication involving the defensibility of what is accepted in terms of a background
evaluation system. One objection to the acceptance of testimony of another is
that the other is not trustworthy in what he testifies to as the truth. The defense
of accepting his testimony depends on evaluating his trustworthiness and on the
reasonableness of the defense that he is trustworthy. But then one faces the objec-
tion that one is oneself not trustworthy in the evaluation of others. How is one to
defend accepting that one is trustworthy in the evaluation of others?
There are two routes of defense. One is that one is in general trustworthy in
what one accepts. That reply has the advantage, an important advantage, that it
ends the regress of defense with a loop. Suppose one considers the objection that
one is not trustworthy in accepting that one is trustworthy. One has an answer,
for one accepts, contrary to the objection, that one is trustworthy in what one
accepts. So it follows from what one accepts that one is trustworthy in accept-
ing that one is trustworthy in what one accepts. The circularity of the defense is
obvious and Schmitt (2002) has objected to my account on this ground.
An advantage of the circle is that it is explanatory. One’s trustworthiness
in what one accepts explains why one is trustworthy in accepting that one is
152 Testimony and Trustworthiness

trustworthy. So, one committed to maximizing explanation may justify accepting

that one is trustworthy by the explanatory advantage of so doing. However, it
is clear one’s justification for accepting one’s trustworthiness in evaluating the
trustworthiness of oneself and others depends on other information about one’s
successes and failures in evaluating the trustworthiness of oneself and others. It
also depends on information about how one corrects one’s errors. Finally, then,
this information about successes and failures in evaluating such trustworthiness
is obtained, in part, from the testimony of others and from the constraint of
intersubjective agreement.
My conclusion is that the foundational account of the evidence and justi-
fication of testimony fails because there is an essential loop in the justification
for accepting the testimony of others. The briefest account of the loop is that
the justification for accepting the testimony of others depends on evaluation of
the trustworthiness of oneself and others and the consequences thereof. Let me
put the matter in the first person. My justification for accepting my evaluations
of others depends on my trustworthiness in evaluating others. My justification
for accepting my trustworthiness in evaluating others depends on my accept-
ing the testimony of others concerning my trustworthiness in evaluating myself
and others. And there is the loop. My justification for accepting the testimony of
myself and others depends on the trustworthiness of my evaluation of the testi-
mony of others. This, in turn, depends on my accepting the testimony of others
in a trustworthy way concerning my evaluations of the trustworthiness of myself
and others.
The argument might be viewed as paradoxical because the circle might be
viewed as vicious as Schmitt (2002) suggests. I shall argue, as I have before,
that we are involved in a virtuous loop rather than a vicious circle. The reason
is that foundationalism and externalism fail to provide a maximally explanatory
account. Foundations and externalities leave us with the unexplained justification
of first premises resulting from their foundational character or the unexplained
justification of premises resulting from the opacity of the external conditions. By
contrast, an account of the justification of testimony based on the evaluations of
the trustworthiness of others constrained by intersubjective agreement in testi-
mony does not leave us with unexplained justification. We accept that we are
trustworthy in testimony with some qualifications. Either we are trustworthy
or we are not. If we are trustworthy, our agreement about our trustworthiness
explains why we are trustworthy in accepting our trustworthiness. If we are not
trustworthy, our lack of trustworthiness should reveal itself in our experience and
our lack of agreement with others about our trustworthiness.
Whatever trustworthiness we possess must be qualified to reflect the cognitive
defects and duplicities we share. On some occasions, the testimony of others will
not be worthy of our trust, and, if the truth were told, on some occasions our
testimony will not be worthy of the trust of others. Often the proper response to
the testimony of others is doubt and refusal to accept what they tell us. Moreover,
 Keith Lehrer 153

there will also be disagreement about trustworthiness of various individuals as

individuals decide how much weight to give to the testimony of others. With all
the cognitive defects, duplicities, and disagreements, how can we make sense of
the idea that our evaluations of the trustworthiness of ourselves and others are or
should be constrained by intersubjective agreement?
First of all, it is important to remember the frequent robust agreement, in
perceptual cases, for example. Of course, even in some simple perceptual cases
our fallibility will dominate and lead us to be untrustworthy. In other cases,
agreement will be conspiratorial and prove untrustworthy as well. However, our
fallibility is compatible with our trustworthiness. We do not have to be infallible
to be trustworthy. Moreover, the disagreement in evaluations of the trustworthi-
ness of an individual may conceal a fixed point of convergence. I have argued
(Lehrer 1997 and elsewhere) that the weights that individuals assign to others,
and this applies to the weights assigned to the trustworthiness of others as well,
may be aggregated to find convergence. Convergence results if the individuals
giving some positive weight to others are connected by a vector in which each
individual assigns positive weight to the next member of a full sequence of mem-
bers as well as to himself or herself. This is a mathematical idealization. Short of
a mathematical representation of evaluations and the aggregation thereof, there
is sometimes qualitative convergence concerning when the testimony of others is
trustworthy and when one is in a position to be trustworthy in evaluating their
trustworthiness. That is something I assume, and I admit it cannot be proven
to the satisfaction of a sceptic, but our acquisition of knowledge from testimony
depends, in part, on the trustworthiness of others and of ourselves in evaluating
their trustworthiness and our own.
Let me turn again to the virtues of the loop in the justification of accepting
testimony and the role of evaluating the trustworthiness of testimony. The basic
argument for the virtue of the loop is that the trustworthy evaluation of others
depends on the assumption that I am trustworthy in the evaluation of myself as
an evaluator of others. Let me put this in terms of meeting objections. Suppose
1. I accept that p on the testimony of T.
If I can meet all objections, then I can meet the objection
1OT. The testimony of T is not trustworthy.
Now suppose that I hold that it is more reasonable for me to accept that the testi-
mony of T is trustworthy than to accept that it is not, that is, it is more reasonable
for me to accept the denial of 1OT than to accept 1OT. That is a reply to the
objection, at least from my own point of view. However, that reply depends on
my trustworthiness in evaluating the trustworthiness of people because I must
also meet the objection
2OT. You are not trustworthy in evaluating the trustworthiness of people,
in particular, T.
154 Testimony and Trustworthiness

Suppose that I hold (which means it is a consequences of my evaluation system)

that it is more reasonable for me to accept that I am trustworthy in evaluating
the trustworthiness of people, and, in particular, T, than to accept that I am
not. Thus, it is more reasonable for me to accept the denial of 2OT than to
accept 2OT.
So far so good. But now consider the objection:
3OT. You are not trustworthy in evaluating the trustworthiness of people,
yourself included, because you do not accept the corrections of
others concerning your evaluations of the trustworthiness of people,
again, yourself included.
In reply, I must claim that it is more reasonable for me to accept the denial of
3OT than 3OT, where the denial is the reply
3OTR. I am trustworthy in evaluating the trustworthiness of people,
myself included, because I do accept the corrections of others
concerning my evaluations of the trustworthiness of people, myself
included.
This reply will lead us into a loop of justification, for it is clear that corrections
of others I accept are based in part on their testimony. To answer objections to
my accepting the testimony of others, I should accept that I am trustworthy in
evaluating the trustworthiness of others. However, I am trustworthy in accept-
ing this claim about my trustworthiness only if I am trustworthy in accepting the
corrections of others concerning my evaluations, that is, if I am trustworthy in
accepting their testimony about the trustworthiness of my evaluations. So, I need
to be trustworthy in accepting the testimony of others to be justified in accept-
ing what they claim in testimony. I am trustworthy in accepting the testimony
of others only if I am trustworthy in evaluating others. I am trustworthy in eval-
uating others, however, only if I am trustworthy in evaluating the testimony of
others concerning the correctness of my evaluations. The solution to the problem
this raises depends on recognizing that I must accept that I am trustworthy in my
evaluations of people, myself included. Finally, my trustworthiness in accepting
that I am trustworthy in accepting that I am trustworthy in this way depends on
my accepting I am trustworthy in accepting the testimony of others. And that
closes the loop. But, it is a virtuous loop.
To understand the virtue of the loop, consider two possibilities. The first is
that I accept the testimony of others concerning my trustworthiness in evaluat-
ing people, myself included, at least when I evaluate them as worthy of my trust.
The second is that I do not accept the testimony of others concerning my trust-
worthiness in evaluating people, myself included. In the first case, my view that I
am trustworthy in evaluating people is confirmed and corrected by the testimony
of others. In the second case, it is not. Though others might mislead me, their
confirmation and correction, especially when I evaluate them as trustworthy,
is a source of improvement of my own trustworthiness. So I should accept the
 Keith Lehrer 155

testimony of others concerning my trustworthiness in evaluating people, at least

when I evaluate them as trustworthy.
Now, suppose someone objects as follows. (This has similarity to an argument
of Schmitt (2002) though I do not impute it to him.) I evaluate the trustwor-
thiness of others when I accept the testimony that corrects and confirms my
trustworthiness in evaluating people. Now, if I am not trustworthy in evaluat-
ing them, then my acceptance of their testimony, which corrects and confirms
my trustworthiness, is useless. I cannot tell whether their corrections and con-
firmations are trustworthy because I cannot tell whether they are trustworthy in
providing them. On the other hand, if I am trustworthy in evaluating them, then
their corrections and confirmations are otiose. So my accepting their corrections
and confirmations is either useless or otiose.
Notice, in reply, that there is some defect in this argument because my trust-
worthiness in evaluating others surely does depend in an important and useful
way on the corrections and confirmations of others. Without the benefit of the
corrections and confirmations of others, my judgements of the trustworthiness
of people, myself included, would be egocentric and idiosyncratic, rather than
socially and intersubjectively corrected and constrained. Extreme individualists
may, of course, applaud independence from social and intersubjective constraint.
But that position, though it may lead to truth on some occasions, is equivalent to
a preference for ignoring information rather than evaluating it.
The circularity revolves around the following principle of trustworthiness of
evaluation:
TE. I am trustworthy in evaluating the trustworthiness of people, myself
included, in what they accept and testify.
Notice that I might reason to this as a conclusion by the following path:
T. I am trustworthy in what I accept.
Therefore,
TT. I am trustworthy in what I accept about the trustworthiness of
people.
ATE. I accept that I am trustworthy in evaluating the trustworthiness of
people, myself included, in what they accept and testify.
Therefore,
TE. I am trustworthy in evaluating the trustworthiness of people, myself
included, in what they accept and testify.
In this argument, T is the key premise, TT is inferred, though not deduced,
from T, and ATE is a premise. TE is inferred, though not deduced, from TT
and ATE. The reason that the inferences are not deductive is that trustworthiness
is a general capacity to be trustworthy in more specific cases, and the inferences
from a general capacity to trustworthiness in more specific cases are not deduct-
ive. Schmitt (2002) attributes to me the view that T is a universal statement, but
156 Testimony and Trustworthiness

that is contrary to the interpretation I proposed in Lehrer (1997) which he cites.

The general capacity to be trustworthy is fallible in application to particular cases
as are all general capacities. It is like the capacity to lift one hundred pounds that
must allow for failure to succeed in lifting that weight in all cases even though
the inference from the general capacity to success in particular cases is warran-
ted inductively. In short, such inferences, though warranted, are defeasible under
some conditions.
The argument begins with T, I am trustworthy in what I accept. Suppose one
raises the question of whether I am reasonable to accept that premise. As I have
argued elsewhere, I am reasonable to accept it if I am trustworthy in accepting it.
But am I trustworthy in accepting T? Principle T has the important explanatory
feature that if it is true, then it explains why I am trustworthy in accepting T by
an inference of the sort that we have considered. My trustworthiness in accepting
T is explained by my general trustworthiness in what I accept. Notice, however,
that a further and more illuminating explanation of the reasonableness and trust-
worthiness of accepting T is explained by other things that I accept about my
trustworthiness in the past and my success in obtaining truth and avoiding error
thereby. Of course, part of what I accept about my past success will be based on
the confirmations of others, that is, on testimony of others that I accept.
So there are two explanation of my trustworthiness in what I accept. There is
a narrow explanation of why I am trustworthy in accepting T, and by inference,
TE, based on the inference from my general trustworthiness. I am trustworthy in
what I accept about my trustworthiness and the trustworthiness of others. This is
a small but virtuous loop of explanation. It depends on the truth of T, however,
not just the acceptance of it, even though I reason from my acceptance of it to the
reasonableness of accepting it.
There is also a broader explanation based on inference from what I accept
about my past successes in being trustworthy and succeeding in obtaining truth.
Such explanation is more illuminating, but the loop in such explanation is equally
apparent. What I accept about the past can only explain my general trustworthi-
ness in what I accept if I am trustworthy in accepting those claims about the past.
Thus, my general trustworthiness, which narrowly explains why I am trustworthy
in accepting what I do, is broadly explained by what I accept about past successes
in being trustworthy. The latter succeeds as a result of my being trustworthy in
accepting what I do about my past successes, however. The broader explanation
presupposes the narrower explanation as my trustworthiness and my acceptance
of it loops back onto itself.
We have now reached the final question. Why is the loop virtuous rather than
vicious in explanation? There are two answers. First, any complete theory of justi-
fication or trustworthiness will have to explain why we are justified or trustworthy
in accepting the theory itself. So the theory must apply to itself to explain why we
are justified or trustworthy in accepting it. Secondly, and equally important, our
trustworthiness at any given time must result from what we have accepted in the
 Keith Lehrer 157

past, including what we have accepted from the testimony of others. The result
is that there is a kind of mutual support between the particular things we have
accepted and our general trustworthiness in what we accept, including, of course,
the particular things we have accepted. It is the mutual support among the things
that we accept that results in the trustworthiness of what we accept.
The foregoing is an oft-told story of mine. What I add here is the extension
of the argument to testimony. My trustworthiness in evaluating the trustworthi-
ness of others depends on their trustworthiness. Their trustworthiness depends
on both their honesty, on whether they accept what they say, and on whether
they proceed in the right way for the purposes of obtaining truth and avoiding
error. However, as examples of invincible deception reveal our fallibility, we note
that the trustworthiness of ourselves and others may fail to be successfully truth-
connected. We may, after all, proceed in the right way to accept what is true and
fail because we are invincibly deceived.
Suppose, however, that our evaluation of the trustworthiness of others is
optimal. We are trustworthy in accepting that they are trustworthy in their testi-
mony. They are trustworthy in their testimony. Moreover, suppose their
trustworthiness is successfully truth-connected. Then our acceptance of their
testimony will be successfully truth-connected because they are trustworthy,
as we accept them to be, and their trustworthiness is successfully truth-
connected. Suppose we accept their testimony about our trustworthiness in
evaluating the trustworthiness of people. We do so because we evaluate their
testimony as trustworthy. They are trustworthy, and their trustworthiness is
successfully truth-connected. Therefore, our acceptance of their testimony about
our trustworthiness in evaluating the trustworthiness of ourselves and others is
successfully truth-connected because their trustworthiness, which we accept, is
successfully truth-connected. The truth-connectedness of their trustworthiness
explains why we are trustworthy in accepting their testimony about our
trustworthiness in evaluating their trustworthiness. They are trustworthy in
telling us what they do about our trustworthiness. Their trustworthiness explains
our trustworthiness when we accept their testimony.
Am I assuming a principle of parity that Schmitt (2002) suggests I have in
common with Foley (1993, 1994) and Gibbard (1990)? The principle of parity
says, in effect, that we are all trustworthy or none of us are. Each person might
be inclined to accept that he or she is trustworthy. But we are not committed
to assuming that others are trustworthy as a result. The reason is simple. If I
assume that I am trustworthy in the evaluation of others, I might evaluate some
as trustworthy and others as untrustworthy for good reasons contained in my
background evaluation system. They might, of course, return the insult. How-
ever, there is a truth of the matter about trustworthiness. Whether I am justified
in accepting that others are trustworthy or not in a way that converts to know-
ledge will depend both on my evaluations of trustworthiness and the truth of the
matter about these evaluations.
158 Testimony and Trustworthiness

Nevertheless, each person will begin by accepting that they are trustworthy in
what they accept. As I run the argument concerning trustworthiness, I express
my trustworthiness in the first person, recognizing, of course, that the reader may
run the same argument. Each of us, I propose, may employ the trustworthiness
argument beginning with the premise—I am trustworthy in what I accept. As I
allow my readers that premise of their own trustworthiness, am I not committed,
as Schmitt assumes I am, to the premise that all of us are trustworthy in what
we accept? My readers may, with my blessings, trust themselves and accept that
they are worthy of their trust in what they accept. However, not all will be as
worthy of their trust as they accept themselves to be. I concede a need of initially
accepting that one is trustworthy in what one accepts, but to accept something is
not the same as to obtain truth in what one accepts, even concerning one’s own
trustworthiness. Some will accept that they are trustworthy when they are not,
the demented and incompetent, for example. Though they accept that they are
trustworthy in what they accept, and I may commiserate with their situation, I
need not agree that they are trustworthy in what they accept. They are in error in
accepting that they are trustworthy in what they accept.
The final question that arises concerning our acceptance of testimony is this.
What converts our acceptance of the testimony of others into knowledge? The
first part of the answer is that we must be trustworthy in our evaluations of the
trustworthiness of others, and we must accept that this is so. Moreover, our
trustworthiness must be successfully truth-connected, that is, the others must,
in fact, be trustworthy and their trustworthiness must itself be successfully truth-
connected. We must also accept that this is so. In short, our acceptance of their
testimony must be justified in a way that is not refuted or defeated by any errors
that we make in evaluating them and their testimony. Undefeated or irrefutable
justified acceptance of the testimony of others is knowledge.

REFERENCES

Coady, C. A. J. (1992), Testimony: A Philosophical Study (Oxford: Oxford University

Press).
(2002), ‘Testimony and Intellectual Autonomy’, Studies in the History and Philo-
sophy of Science, 33/2: 355–72.
Foley, R. (1993), Working without a Net: A Study of Egocentric Epistemology (Oxford:
Oxford University Press).
(1994), ‘Egoism in epistemology’, in F. Schmitt (ed.), Socializing Epistemology: The
Social Dimensions of Knowledge, (Lanham, Md.: Rowman & Littlefield), 53–73.
Gibbard, A. (1990), Wise Choice, Apt Feeling: A Theory of Normative Judgment
(Cambridge, Mass.: Harvard University Press).
Hume, D. (1978), A Treatise of Human Nature (2nd edn.), ed. L. A. Selby-Bigge, rev.
P. H. Nidditch (Oxford: Oxford University Press) (first published 1739).
Lehrer, K. (1997), Self-Trust: A Study of Reason, Knowledge, and Autonomy (Oxford:
Oxford University Press).
 Keith Lehrer 159

(2000), Theory of Knowledge (2nd edn., Boulder, Colo: Westview Press).

and Beanblossom, R. (1975), Reid’s Inquiry and Essays (Indianapolis: Bobbs Merrill,
Indianapolis) (new edn., Indianapolis Hackett Publishing Co.: 1984).
Schmitt, F. (1994), ‘Socializing Epistemology: An Introduction through Two Sample
Issues’, in F. Schmitt (ed.), Socializing Epistemology: The Social Dimensions of Know-
ledge (Lanham, Md.: Rowman & Littlefield), 1–27.
‘Social Epistemology’, in J. Greco and E. Sosa (eds.), Blackwell Guide to Epistemo-
logy (Oxford: Basil Blackwell), 354–82.
‘Testimonial Justification: The Parity Argument’, Studies in the History and Philo-
sophy of Science, 33/2: 385–406.
Smith, J. C., and Lehrer, K. (1985), ‘Reid on Testimony and Perception’, Canadian
Journal of Philosophy, suppl. vol. 11: 21–38.
 8
It Takes Two to Tango: Beyond
Reductionism and Non-Reductionism
in the Epistemology of Testimony
Jennifer Lackey

How precisely do we successfully acquire justified belief from either the spoken
or written word of others?¹ This question is at the center of the epistemology
of testimony, and the current philosophical literature contains only two general
options for answering it: reductionism and non-reductionism. While reduction-
ists argue that testimonial justification is reducible to sense perception, memory,
and inductive inference, non-reductionists maintain that testimony is just as basic
epistemically as these other sources. The aim of this paper is to challenge the
current terms of the debate by, first, showing that there are serious problems
afflicting both reductionism and non-reductionism and by, second, suggesting an
alternate view of testimonial justification that goes beyond the reductionist/non-
reductionist dichotomy.

1. REDUCTIONISM

There are two central components to reductionism. The first is what we may
call the Positive-Reasons Component: justification is conferred on testimonial
beliefs by the presence of appropriate positive reasons on the part of hearers.
Since these reasons cannot themselves be ultimately testimonially grounded,²
they must depend on resources provided by other epistemic sources—typically,
sense perception, memory, and inductive inference. This gives rise to the
second component—what we may call the Reduction Component. Because
the justification of testimonial beliefs is provided by these non-testimonially

For very helpful comments on earlier versions of this paper, I am grateful to Robert Audi, Marian
David, Sandy Goldberg, Peter Graham, Jason Kawall, Matt McGrath, Joe Shieber, Ernie Sosa, and
audience members at the 2004 Pacific Division of the APA in Pasadena, CA. Most of all, I am, as
always, indebted to Baron Reed in ways too numerous to list. With respect to this paper alone, if
I inserted an endnote at every place in which his comments substantially altered my thinking, the
pages of notes would outnumber those of the paper itself.
 Jennifer Lackey 161

grounded positive reasons, testimonial justification is said to reduce to the

justification of sense perception, memory, and inductive inference.³
But what, one might ask, is being reduced to what? Otherwise put, what are
the relata of testimonial reduction? Two different answers are given to this ques-
tion. The first answer—a view sometimes called global reductionism—is that the
justification of testimony as a source of belief reduces to the justification of sense
perception, memory, and inductive inference. In particular, global reductionists
maintain that in order to justifiedly accept a speaker’s report, a hearer must have
non-testimonially based positive reasons for believing that testimony is generally
reliable.⁴
There are, however, at least three problems with global reductionism that
render it ultimately an untenable view of testimonial justification. The first is
that, before accepting any testimony at all, including that of their parents and
teachers, very young children would have to wait until they have checked the
accuracy of enough different kinds of reports from enough different speakers to
conclude that testimony is generally reliable. Not only is it wildly implausible to
suppose that most young children—or even adults—are capable of engaging in
such a process, it also becomes mysterious how they would be able to acquire the
conceptual and linguistic tools needed for an induction to the general reliability
of testimony without accepting some testimony in the first place.⁵ Thus, if global
reductionism is true, the very cognitive tools needed to acquire testimonial
justification would be inaccessible to epistemic agents, thereby leading ultimately
to skepticism about testimonial knowledge.
The second problem is that in order to have non-testimonially based posit-
ive reasons that testimony is generally reliable, one would have to be exposed
not only to a non-random, wide-ranging sample of reports, but also to a non-
random, wide-ranging sample of the corresponding facts. Both are problematic.
With respect to the reports, most of us have been exposed only to a very limited
range of reports from speakers in our native language in a handful of communit-
ies in our native country. This limited sample of reports provides only a fraction
of what would be required to legitimately conclude that testimony is generally
reliable. With respect to the corresponding facts, a similar problem arises: the
observational base of ordinary epistemic agents is simply far too small to allow the
requisite induction about the reliability of testimony. As C. A. J. Coady says:
it seems absurd to suggest that, individually, we have done anything like the amount
of field-work that [reductionism] requires . . . many of us have never seen a baby born,
nor have most of us examined the circulation of the blood nor the actual geography
of the world nor any fair sample of the laws of the land, nor have we made the obser-
vations that lie behind our knowledge that the lights in the sky are heavenly bodies
immensely distant nor a vast number of other observations that [reductionism] would
seem to require. (Coady 1992: 82)
Moreover, with many reports, such as those involving complex scientific, eco-
nomic, or mathematical theories, most of us simply lack the conceptual machinery
162 It Takes Two to Tango

needed to properly check the reports against the facts. Once again, then, global
reductionism leads to skepticism about testimonial knowledge, at least for most
epistemic agents.
The previous two points focused on our ability to know or determine whether
testimony is a generally reliable epistemic source. A third problem with global
reductionism—one that has not been properly appreciated but is, to my mind,
the most debilitating objection—is that it is questionable whether there even
is an epistemically significant fact of the matter here.⁶ To see this, consider, for
instance, the following epistemically heterogeneous list of types of reports, all
of which are subsumed under ‘testimony in general’: reports about the time of
day, what one had for breakfast, the achievements of one’s children, whether
one’s loved one looks attractive in a certain outfit, the character of one’s polit-
ical opponents, one’s age and weight, one’s criminal record, and so on. Some of
these types of reports may be generally highly reliable (e.g., about the time of
day and what one had for breakfast), generally highly unreliable (e.g., about the
achievements of one’s children, the looks of one’s loved ones, and the character
of one’s political opponents), and generally very epistemically mixed, depending
on the speaker (e.g., about one’s age, weight, and criminal record). Because of this
epistemic heterogeneity, it is doubtful, not only whether ‘‘testimony’’ picks out
an epistemically interesting or unified kind, but also whether it even makes sense
to talk about testimony being a generally reliable source.
Otherwise put, even if it turned out that the majority of testimonial reports
are, as a matter of fact, both true and properly formed, this information would
not have much epistemic significance. For concealed in this percentage are all sorts
of epistemically salient facts: some people offer mostly false reports, some kinds of
reports are mostly false, many true reports are about very mundane facts, and so on.
Because of this, the mere fact that testimony is generally reliable has very little epi-
stemic bearing on any particular instance of testimony. For instance, suppose that
I came to learn that 70 per cent of all reports are both true and properly formed.
What relevance would this information have to whether a particular instance of
testimony is epistemically acceptable? Very little. For this information is so broad
and conceals so many epistemically important differences that it would have vir-
tually no straightforward epistemic application. Thus, even if global reductionism
were entirely successful and it could be shown that testimony is generally reliable,
this conclusion would have very little epistemic significance in itself.
The second version of reductionism—often called local reductionism—is that
the justification of each particular report or instance of testimony reduces to the
justification of instances of sense perception, memory, and inductive inference.
Specifically, local reductionists claim that in order to justifiedly accept a speak-
er’s testimony, a hearer must have non-testimonially based positive reasons for
accepting the particular report in question.⁷
There are, however, two importantly different ways of understanding the local
reductionist’s Positive-Reasons Component. The first is:
 Jennifer Lackey 163

PR-N: Appropriate positive reasons are necessary for testimonial justifica-

tion.
The second, stronger, interpretation is:
PR-N&S: Appropriate positive reasons are necessary and sufficient for testi-
monial justification.
But notice: in order for testimonial justification to be reducible to sense percep-
tion, memory, and inductive inference, the positive reasons in question must
be fully sufficient for justifying the relevant testimonial belief. Otherwise, there
would be an asymmetry between the justificatory status of the testimonial belief
being reduced and the positive reasons doing the reducing, thereby preventing the
possibility of just such a reduction. Because of this, the Reduction Component
of the local reductionist’s view depends on reading the Positive-Reasons Com-
ponent as the stronger PR-N&S. In what follows, however, I shall argue that
such a requirement is false. This conclusion has particular significance. For, since
local reductionism represents the weakest version of reductionism, an argument
against it will thereby show that reductionism in the epistemology of testimony,
in general, is false.⁸
To begin, consider the following:
nested speaker: Fred has known Helen for five years and, during this
time, he has acquired excellent epistemic reasons for believing her to be
a highly reliable source of information on a wide range of topics. For
instance, each time she has made a personal or professional recommend-
ation to Fred, her assessment has proven to be accurate; each time she has
reported an incident to Fred, her version of the story has been independ-
ently confirmed; each time she has recounted historical information, all
of the major historical texts and figures have fully supported her account,
and so on. Yesterday, Helen told Fred that Pauline, a close friend of hers,
is a highly trustworthy person, especially when it comes to information
regarding wild birds. Because of this, Fred unhesitatingly believed Pauline
earlier today when she told him that albatrosses, not the widely believed
condors, have the largest wingspan among wild birds. It turns out that
while Helen is an epistemically excellent source of information, she was
incorrect on this particular occasion: Pauline is, in fact, a highly incompet-
ent and insincere speaker, especially on the topic of wild birds. Moreover,
though Pauline is correct in her report about albatrosses, she came to hold
this belief merely on the basis of wishful thinking (in order to make her
reading of The Rime of the Ancient Mariner more compelling).⁹
Now, does Fred justifiedly believe that albatrosses have the largest wingspan
among wild birds on the basis of Pauline’s report? Intuitively, no. For even though
Helen’s testimony provides Fred with excellent positive reasons for accepting
the report in question, Pauline is not only a generally unreliable speaker, she
is also reporting a belief which, though true, fails to be reliably produced or
164 It Takes Two to Tango

appropriately truth-conducive. Because of this, the testimony that Pauline offers

to Fred also fails to be reliably produced or appropriately truth-conducive,
thereby preventing it from leading to justified belief for Fred.¹⁰
What nested speaker reveals is that the possession of good positive reasons
by a hearer is not sufficient for justifiedly accepting a speaker’s testimony. Why?
Because the possession of positive reasons on behalf of a speaker’s report, even
objectively excellent ones, does not necessarily put one in contact with testimony
that is reliable. There is, then, a further necessary condition for testimonial jus-
tification, one that requires that a speaker’s testimony be reliable or otherwise
truth-conducive. This additional ‘speaker condition’ can be fleshed out in differ-
ent ways. For instance, most non-reductionists require that a speaker herself com-
petently believe the proposition to which she is sincerely testifying in order for it to
qualify as justified belief for her hearers.¹¹ Alternatively, it may be required only
that the speaker’s statement be reliably produced or truth tracking.¹² But however
the speaker condition is fleshed out, the point that is of import here is that it is a
condition that cannot be subsumed merely by requiring the possession of appro-
priate positive reasons on the part of the hearer. The presence of such reasons
is, therefore, not sufficient for testimonial justification; accordingly, the stronger
reading of the Positive-Reasons Component—i.e., PR-N&S—is false.¹³
What, then, are the consequences of this for the Reduction Component of
reductionism? If the Reduction Component is correct, then there cannot be any
difference between the justificatory status of the testimonial belief being reduced
and the positive reasons doing the reducing. That is, to the extent that the positive
reasons are justified, so, too, should the testimony in question be justified. This
is just what it means for testimonial justification to be reduced to the justifica-
tion of sense perception, memory, and inductive inference. But notice: in nested
speaker, we have an example in which there is precisely such a justificatory differ-
ence. In particular, Fred’s positive reasons are fully epistemically justified and yet
the testimonial belief in question is not. This shows that the Reduction Compon-
ent of reductionism is false.
There are, however, at least two central objections that a reductionist may raise
to this conclusion. First, it may be argued that nested speaker simply describes a
testimonial Gettier-type case, in which a hearer’s belief is justified but true merely
by accident. In particular, one may claim that, because of the excellent reasons
he has for trusting Pauline, the true belief that Fred acquires from her testimony
is justified. Our intuition that something is epistemically defective in such a case,
however, can be explained by the fact that Fred’s belief nevertheless falls short of
knowledge.
By way of response to this objection, it will be helpful to compare nested
speaker with the following:
unnested speaker: Max has known Holly for ten years and, over the
course of these years, he has acquired excellent epistemic reasons for believ-
ing her to be a highly reliable source of information on a wide range of
 Jennifer Lackey 165

topics. Indeed, during this time, she has never offered, to Max or anyone
else, a report that has been either insincere or improperly formed. Cur-
rently, however, Holly is in the midst of a personal crisis, which she effect-
ively conceals from those around her, and her emotional state of mind
leads her to report to Max that her purse has been stolen, despite having
absolutely no evidence for thinking this to be the case. Max, detecting
nothing amiss, readily accepts Holly’s testimony. Now, it turns out that
Holly’s purse was in fact stolen: while she was at Caribou Coffee earlier
today, a young man slipped it off her chair and into his backpack.

Notice first that in unnested speaker, Max not only has excellent epistemic
reasons for accepting Holly’s testimony, Holly is also a generally reliable testi-
fier. The problem is simply that, on this particular occasion, Holly acts com-
pletely out of epistemic character and offers a report for which she lacks adequate
evidence. What’s more, she gets veritically (though not financially!) lucky: her
report turns out to be true. Because of this, one may plausibly regard unnes-
ted speaker as a Gettier-type case, i.e., as a case of justified true belief that falls
short of knowledge. In particular, one may argue that Max’s excellent positive
reasons for Holly’s testimony combined with her general reliability as a testifi-
er render his true belief about the stolen purse justified, though not an instance
of knowledge.¹⁴ Accordingly, unnested speaker would fail to show that there
can be a difference between the justificatory status of the testimonial belief being
reduced and the positive reasons doing the reducing.
In nested speaker, however, while Fred has excellent positive reasons to
accept Pauline’s testimony, she is not a reliable testifier in any sense of the word.
For not only does Pauline hold her belief about the wingspan of albatrosses on
the basis of wishful thinking, she is also in general a highly incompetent and
insincere testifier. In other words, with respect to most topics most of the time,
Pauline believes to be true what is in fact false, reports what she herself does
not believe, or both. Thus, positive reasons can come apart from even general
reliability.¹⁵ Moreover, given the degree and depth of Pauline’s unreliability,
there is simply no plausible sense in which the belief that Fred forms on the basis
of her testimony could be justified. Unlike unnested speaker, then, nested
speaker simply cannot be plausibly regarded as a Gettier-type case.¹⁶
One may pursue a second line of resistance to nested speaker, however, by
arguing that Fred’s ‘excellent epistemic reasons’ on behalf of Pauline’s testimony
are not ‘appropriate’ in the relevant sense. For, one may say that the only positive
reasons that are appropriate are ones that render it likely in an objective sense
that the testimony in question is true. Since Helen is incorrect in her assessment
of Pauline, it may be thought that such an assessment fails to provide Fred with
positive reasons that satisfy this criterion.
But as nested speaker is described, Fred does have reasons that, by one meas-
ure, render it objectively likely that Pauline’s testimony is true. In particular,
166 It Takes Two to Tango

Fred’s positive reasons place those beliefs from Pauline’s testimony in a category
that contains beliefs that are or would be mostly true; namely, those beliefs that
are supported by Helen’s testimony. For instance, were Fred to decide between
accepting the reports of two different speakers, one of whom has the support
of Helen’s testimony and another who lacks this support, most of the time Fred
would do well to accept the reports of the former. That is, most of the time,
forming beliefs from sources supported by Helen’s testimony would lead to the
truth. In this sense, then, the positive reasons that Fred possesses for Pauline’s
testimony do render it likely in an objective sense that her testimony is true. The
problem is that, by other measures of objective likelihood, Fred’s positive reas-
ons do not render it likely that Pauline’s testimony is true. Fred’s belief about
the wingspan of albatrosses also belongs to a category that contains beliefs that
are or would be mostly false; namely, those beliefs that are supported by Pauline’s
testimony. Moreover, because Pauline is the direct source of the belief, it is clear
that her unreliability is not offset by the excellence of Fred’s reasons for believing
her. So, although Fred does have excellent positive reasons for believing Pauline’s
testimony, the belief in question is not justified.¹⁷
It is, therefore, not enough for testimonial justification that a hearer have even
epistemically excellent positive reasons for accepting a speaker’s testimony—the
speaker must also do her part in the testimonial exchange by offering testimony
that is reliable or otherwise truth-conducive. Thus, PR-N&S is false and, accord-
ingly, reductionism in the epistemology of testimony is false.

2 . N O N - R E D U C T I O N I SM

In contrast to reductionists, non-reductionists altogether reject both versions of

the Positive-Reasons Component and the Reduction Component as well, that
is, they maintain that positive reasons are neither necessary nor sufficient for
testimonial justification and, accordingly, that testimonial justification is an irre-
ducible epistemic source. What non-reductionists hold is that hearers are justified
in accepting the testimony of speakers so long as two conditions are satisfied:
(i) the report in question has to be reliably produced, typically through the speak-
er being both a competent believer and a sincere testifier, and, (ii) the hearer
in question cannot have any relevant defeaters—that is, any counterbeliefs or
counterevidence—for such a report.¹⁸ In this way, so long as there is no evidence
against accepting a speaker’s report, the hearer has no positive epistemic work
to do in order to justifiedly accept the testimony in question.¹⁹ In what follows,
however, I argue that this is false—hearers do have some positive work to do
for testimonial justification. Specifically, I show that although PR-N&S is false,
PR-N is nevertheless true. Thus, I show that non-reductionism is false.
To begin, notice that in denying even the weaker reading of the Positive-
Reasons Component of reductionism—i.e., PR-N—non-reductionists commit
themselves to saying that testimonial justification can be acquired in the complete
 Jennifer Lackey 167

absence of any positive reasons on the part of the hearer. Let us, therefore,
consider such a case.
alien: Sam, an average human being, is taking a walk through the forest
one sunny morning and, in the distance, he sees someone drop a book.
Although the individual’s physical appearance enables Sam to identify her
as an alien from another planet, he does not know anything either about
this kind of alien or the planet from which she comes. Now, Sam even-
tually loses sight of the alien, but he is able to recover the book that she
dropped. Upon opening it, he immediately notices that it appears to be
written in English and looks like what we on Earth would call a diary.
Moreover, after reading the first sentence of the book, Sam forms the
corresponding belief that tigers have eaten some of the inhabitants of the
author’s planet. It turns out that the book is a diary, the alien does commu-
nicate in English, and it is both true and reliably written in the diary that
tigers have eaten some of the inhabitants of the planet in question.
Now since the book in question is written by an alien, Sam truly has no epi-
stemically relevant positive reasons: he has no commonsense psychological alien
theory, he has no beliefs about the general reliability of aliens as testifiers, he has
no beliefs about the reliability of the author of this book, he has no beliefs about
how ‘diaries’ function in this alien society, and so on. Moreover, if Sam attends
to the narrative voice of the author in the hope of trying to assess her competence
and sincerity, he would be engaged in a fruitless activity since there is no reason
to believe that signs of competence and sincerity on the planet in question cor-
respond to these signs on Earth. Sam cannot even compare the content of the
reports in this diary to his background beliefs since he does not know that the
words in this book are used in the same way that we on Earth use them. So, here
is a case in which a hearer truly fails to have any positive reasons on behalf of a
speaker’s testimony. Let us suppose, further, that there is nothing about the diary
that provides Sam with relevant counterbeliefs or counterevidence. The crucial
question we now need to ask is whether Sam is justified in believing that tigers
have eaten some of the inhabitants of the planet in question on the basis of the
alien’s diary.
Here the answer should clearly be no. Despite the fact that the alien’s report
is both true and reliable, it seems plainly irrational epistemically for Sam to form
the belief in question on the basis of the alien’s testimony. For, it may very well be
accepted practice in alien society to be insincere and deceptive when testifying to
others. Or, normal alien psychology may be what we Earthlings would consider
psychosis. Or, the language that the aliens use, though superficially indistinguish-
able from English, may really be Twenglish, where Twenglish uses the ‘negation’
sign for affirming a proposition. Or, ‘diaries’ in the alien society may be what
we on Earth regard as science fiction, and so on. For all Sam knows when he
reads the book, each of these scenarios is just as likely as the possibility that
168 It Takes Two to Tango

these aliens are reliable testifiers who speak English. But, in the absence of any
way to discriminate among these possibilities, it seems clear that the appropriate
epistemic response is to withhold belief.²⁰
It is of further interest to note that the general diagnosis offered of alien
appeals only to features to which non-reductionists are already committed. To
see this, recall that the second condition of non-reductionism requires that the
hearer in question not possess any relevant defeaters for accepting a speaker’s
report. For instance, if I believe that you frequently lie but nevertheless come
to believe that owls are raptors on the basis of your testimony, then, according
to non-reductionism, my testimonial belief fails to qualify as justified or known
even if it is in fact true and reliably formed. Why? Because even non-reductionists
agree that testimonial justification is incompatible with at least certain kinds of
epistemic irrationality.²¹ What alien reveals, however, is that accepting a speak-
er’s report in the complete absence of positive reasons can be just as epistemically
irrational as accepting such a report in the presence of a defeater—indeed, per-
haps even more so. If I, for example, have a defeater by virtue of believing that
you only occasionally lie, would it be more epistemically irrational for me to trust
your testimony than it would be for Sam to trust the alien’s in the absence of pos-
itive reasons? Not at all. For while Sam knows absolutely nothing about the alien
in question, I have all sorts of beliefs, both about humans in general and about
you in particular, that are relevant to my acceptance of your testimony—for
instance, I believe that humans often speak sincerely, that reports on Earth are
usually offered to communicate information, that you do not exhibit any clear
signs of being deceptive, and so on. Against the background of all of this incred-
ibly rich positive information, my belief that you merely occasionally lie seems
rather epistemically insignificant when compared with the fact that Sam doesn’t
even know whether aliens actually speak English. Accepting testimony in the
absence of positive reasons can, then, be even more irrational than accepting testi-
mony in the presence of defeaters. Thus, by showing that epistemic irrationality
is involved in accepting a speaker’s report in the complete absence of positive
reasons—even more than in some cases in which defeaters are present—alien
poses a challenge to non-reductionists on their own terms.²²
Now, one way the non-reductionist may respond to alien is to deny that Sam
satisfies the relevant conditions in his acceptance of the contents of the ‘diary.’
Specifically, it may be argued that Sam fails the no-defeater condition because a
context like that envisaged above, in which there is absolutely no epistemically
relevant information about the speaker, report, or context, provides the hearer in
question with evidence against the testimony in question.
However, this response is unacceptable since, ex hypothesi, there is nothing
about the diary that suggests that its contents are false or that its author is unreli-
able. Hence, Sam has no relevant defeaters. Any residual discomfort that one may
have about granting justified belief in this sort of case simply reveals one’s intu-
itions that positive reasons are necessary for testimonial justification. For the only
 Jennifer Lackey 169

negative reason it is appropriate to say that Sam has with respect to the alien’s
‘diary’ is the absence of positive reasons. Since the fundamental difference between
non-reductionism and reductionism is precisely over the need for positive reas-
ons, this reply is simply not available to the non-reductionist.
A second strategy for denying the force of alien is to argue that the non-
reductionist’s principle applies only to humans because only those who are
members of our institution of testimony fall under it. The aliens may very
well have their own institution of testimony on their planet, and their practices
and epistemic principles may be quite similar to ours. But we cannot assume
this similarity. Non-reductionism would thus be limited in its applicability to
members of our species only and, accordingly, Sam would fail to be justified in
accepting the alien’s testimony.
An obvious response to this objection is simply to modify the counterexample
so that the testifier in question is in fact a member of our institution of testimony.
For instance, suppose that Sally has been in a coma for the past two months and,
upon waking, discovers that she has lost all of her previous knowledge except
for her competence with the English language. Upon leaving the hospital, she
stumbles upon a diary of an unknown author and begins reading it. Now, ex
hypothesi, Sally no longer has commonsense beliefs about human psychology, she
no longer has beliefs about the general reliability of humans as testifiers, she no
longer has beliefs about how diaries function in our society, and so on. Is Sally
justified in accepting the contents of the diary? Since this case is similar to the alien
example in all epistemically relevant respects, the answer must be no. So, restricting
the scope of non-reductionism to humans will not avoid this objection.
Furthermore, it seems that the primary explanation for why different epistemic
standards would be invoked, depending on whether the speaker in question is a
human or an alien, is precisely that we have all sorts of epistemically relevant beliefs
about our institution of testimony, and fail to have them in the case of the aliens. For
consider: why aren’t we entitled to assume that the aliens are like us in all relevant
respects? The natural answer seems to be that we do not have any reason to believe
that this is the case. Thus, the very criterion for saying who is or is not a member
of our institution of testimony is simply whether we have positive reasons for
their testimony—which begs the question.
We have seen, then, that the hearer must also do her part in a testimonial
exchange by having at least some epistemically relevant positive reasons for
accepting the report in question. Thus, PR-N is true and, accordingly, non-
reductionism with respect to testimonial justification is false.

3 . D UA L I S M

In a testimonial exchange, information is typically transmitted between two cent-

ral participants: the speaker and the hearer. My diagnosis of what has gone wrong
in the epistemology of testimony is this: reductionists and non-reductionists alike
170 It Takes Two to Tango

have attempted to place all of the epistemic work on only one or the other of these
participants and, in so doing, have ignored the positive justificatory contribution
that needs to be made by the other.
Reductionists, on the one hand, focus entirely on the hearer in a testimonial
exchange. For in order for testimonial justification to be reduced to the justifica-
tion of perception, memory, and inference, all of the justificatory work needs to
be shouldered by the hearer since it is precisely her positive reasons that are sup-
posed to provide the reductive base. Reductionists, then, are committed to say-
ing, first, that the reasons possessed by a hearer wholly determine the justificatory
status of a given testimonial belief, and, second, that nothing about the speak-
er (apart from what may already be captured by the hearer’s positive reasons)
has epistemic relevance to the justification of a hearer’s testimonial belief. But, as
we have seen, both of these theses are false. For, as nested speaker showed, no
matter how excellent a hearer’s positive reasons are on behalf of an instance of
testimony, a speaker may still offer a report that is thoroughly unreliable. Because
of this, an adequate account of testimonial justification must include a condition
requiring that the testimony in question be reliable or otherwise truth-conducive.
Non-reductionists, on the other hand, capture the work that needs to be done
by the speaker in a testimonial exchange, but neglect the positive contribution
that a hearer needs to make. Specifically, they correctly require the reliability of
the speaker’s testimony, but then mistakenly assume that the hearer merely has
to satisfy the no-defeater condition.²³ However, as alien showed, no matter how
reliable a speaker’s testimony is, this cannot by itself make it rationally accept-
able for a hearer to accept her report. For this, the hearer needs to have some
epistemically relevant positive reasons on behalf of the testimony in question.
The upshot of these considerations is that it takes two to tango: the justific-
atory work of testimonial beliefs can be shouldered exclusively neither by the
hearer nor by the speaker.²⁴ To put it somewhat crudely, the speaker condition
ensures reliability while the hearer condition ensures rationality for testimonial
justification.²⁵ Thus, we need to look toward a view of testimonial justification
that gives proper credence to its dual nature, one that includes the need for the
reliability of the speaker (from non-reductionism) and the necessity of positive
reasons (PR-N from reductionism). Accordingly, an adequate view of testimonial
justification needs to recognize that the justification of a hearer’s belief has dual
sources, being grounded in both the reliability of the speaker and the rationality
of the hearer’s reasons for belief. More precisely, we should accept what I shall call
dualism in the epistemology of testimony, which includes at least the following:

Dualism:²⁶ For every speaker A and hearer B, B justifiedly believes that p on

the basis of A’s testimony that p only if: (1) B believes that p on
the basis of the content of A’s testimony that p, (2) A’s testimony
that p is reliable or otherwise truth conducive, and (3) B has
appropriate positive reasons for accepting A’s testimony that p.
 Jennifer Lackey 171

Because dualism specifies only necessary conditions, there may be other condi-
tions that need to be added for a complete account of testimonial justification.²⁷
What is of import, here, however, is that testimonial justification requires pos-
itive epistemic contributions from both the speaker and the hearer. Though this
point is not acknowledged by any of the standard views in the current literat-
ure, it should be obvious: acquiring testimonial justification involves an exchange
between two parties. And in order for such an exchange to properly result in justi-
fication, both parties need to do their epistemic work.²⁸
Let us now take a closer look at the specific conditions expressed in dualism.
Regarding (1), since we are here interested in testimonial justification, this con-
dition specifies that the hearer must form the belief in question on the basis of
the content of the speaker’s testimony. This is to preclude cases where a belief is
formed, either entirely or primarily, on the basis of features about the speaker’s
testimony. For instance, if you say, in a soprano voice, that you have a soprano
voice and I come to believe this either entirely or primarily on the basis of hearing
your soprano voice, then my resulting justification is either or partially perceptual
in nature.²⁹ Condition (1), therefore, is included to prevent cases of this sort from
qualifying as instances of testimonial justification.³⁰
With respect to (2), the details of the reliability of the speaker’s testimony can
be fleshed out in several different ways. The most common strategy is to require
that the speaker in question be both a competent believer and a sincere testifi-
er—the speaker must form her own belief in an epistemically acceptable fashion,
and then report to others what she herself believes.³¹ Alternatively, it may be
required that the speaker’s statement, rather than her belief, be somehow reliable,
perhaps by being either sensitive (à la Nozick)³²—the speaker would not state
that p if p were false—or safe (à la Sosa)³³—the speaker would not state that p
without it being so that p. Either way, while more could certainly be said about
the details of (2), there is no reason to question the general tenability of such
a condition.
In contrast, serious doubts over the plausibility of a condition such as (3) have
been expressed repeatedly in the literature. Indeed, all of the standard arguments
against reductionism focus specifically on the Positive-Reasons Component. For
instance, Mark Owen Webb claims that, ‘[t]he cause of the trouble [with reduc-
tionism] seems to be the requirement that our beliefs based on the testimony of
others be based on beliefs in us about the reliability of testimony. This higher-
level requirement . . . places too great a burden on the believing subject, since
it requires of him all kinds of knowledge about people, their areas of expert-
ise, and their psychological propensities, which knowledge most subjects simply
lack’ (Webb 1993: 263). Following this, Richard Foley maintains that the prob-
lem with reductionism ‘is that it threatens to cut us off from expertise and inform-
ation that others have and we lack . . . After all, many people with expertise
and information that we lack are people about whom we know little. Hence,
there may be little or no basis for us to grant them derivative authority’ (Foley
172 It Takes Two to Tango

1994: 57–8).³⁴ Moreover, even when we do have adequate positive reasons for
accepting a speaker’s testimony, it is often argued, as P. F. Strawson does, that
‘the checking process . . . consists in nothing other than seeking confirmation
from other sources of testimony’ (Strawson 1994: 25).³⁵
These quotations express two slightly different concerns: (a) ordinary epistem-
ic agents simply do not have enough information to acquire positive reasons
strong enough to justify accepting most of the testimony that is intuitively jus-
tified, and (b) even when agents do have enough information to justify accept-
ing particular instances of testimony, the positive reasons themselves often are
indebted to testimony. Now, since the Reduction Component was the focus of
my arguments in Section 1, my reasons for rejecting reductionism obviously dif-
fer in significant ways from those expressed in (a) and (b). But, more importantly,
by arguing on behalf of the Positive-Reasons Component (i.e., PR-N), my view
is also targeted by these very objections. My purpose in what follows, then, is to
defend condition (3) of dualism from (a) and (b). To this end, I shall make three
central points.
First, one of the primary reasons that (a) is frequently raised as a problem for
reductionism is that reductionists are committed to PR-N&S rather than to PR-
N. In particular, they must maintain that positive reasons are both necessary and
sufficient for testimonial justification and, therefore, that positive reasons must
carry all of the justificatory burden for testimonial beliefs. This, in turn, lends
itself to the concern expressed in (a): how can we possess enough information to
adequately justify all of our testimonial beliefs? While traveling to London for
the first time, do I, for instance, have enough information about a random local
British newspaper to adequately justify the beliefs that I acquire while reading it?
In contrast to reductionism, dualism has the justificatory work being shared
between the speaker and the hearer, leaving the work for the positive-reasons con-
dition far less burdensome. Specifically, since condition (2) of dualism takes care
of the reliability of the testimony in question, (3) merely has to ensure that the
hearer’s acceptance of the testimony is rationally acceptable. More precisely, on
my view, the positive reasons possessed by a hearer need to be such that they
render it, at the very least, not irrational for her to accept the testimony in ques-
tion. This is a substantially weaker condition than that required by reductionists.
To see this, consider, again, my accepting the reports of a random British news-
paper. Even if I do not have specific beliefs about British newspapers, I have all
sorts of beliefs about England, the people who live there, their government, their
social and political values, and so on. Surely, this information is enough to make
it not irrational to form beliefs on the basis of British newspapers, even if it is not
itself fully sufficient for justifying such beliefs.
Second, and also in response to (a), it is important to notice that there are all
sorts of positive reasons that can have epistemic significance and, therefore, be
relevant to the satisfaction of condition (3) of dualism.³⁶ For instance, suppose
that I know nothing personal about Harold or his testifying habits—I met him
 Jennifer Lackey 173

for the first time on the subway today. I ask him for directions and, while making
direct eye contact with me, he responds in an able and confident manner that
my destination is four blocks to the south. Now, despite the fact that I have no
background with Harold, I may have a substantial amount of inductive evidence
for believing that people are generally both sincere and competent when provid-
ing directions in normal contexts, that reports made with sustained eye contact
are typically sincere ones, or that reports made ably and confidently are typically
competent ones.
More precisely, even if a hearer, B, has not observed the general conformity of
prior reports of a speaker, A, and the corresponding facts, B may have observed
a general conformity of other relevant reports and facts. In particular, there seem to
be at least three classes of inductively based positive reasons that are available to
epistemic agents for distinguishing between reliable and unreliable testimony.
The first class includes criteria for individuating epistemically reliable contexts
and contextual features.³⁷ For instance, one may take a less critical attitude in the
context of an astronomy lecture or a National Geographic report than one does
in the context of an astrology lecture or a National Enquirer report. The explana-
tion for this disparity may appeal not only to the negative evidence that has been
inductively acquired for reports received in the latter contexts, but also to the pos-
itive evidence that has been accumulated for believing that reports received in the
former contexts tend to be reliable. Or consider the different attitudes that may
be taken toward a calm and coherent stranger reporting a robbery a few blocks
away versus an apparently confused person who is smelling of alcohol reporting
the same information. Again, the difference in responses may be explained by
both positive reasons and defeaters: previous inductive evidence indicates that the
contextual features in the first scenario suggest a reliable testifier while the contex-
tual features in the second scenario suggest an unreliable testifier. Similar remarks
can be made about countless other contextual factors, such as facial expressions,
eye contact, mannerisms, narrative voice, and so on.
The second broad class of positive reasons includes criteria for distinguishing
between different kinds of reports.³⁸ So, for example, a hearer may quite reason-
ably take an uncritical stance when a speaker is reporting the time of day, her
name, what she had for dinner, and so on. On the other hand, one may take a
more critical stance when receiving a speaker’s testimony about political matters,
the achievements of her children, alien encounters, UFO sightings, and the like.
Here, prior evidence acquired about subject matters or types of reports provides
recipients of testimony with epistemically relevant positive reasons.
The third class includes criteria for individuating epistemically reliable
speakers.³⁹ For instance, one may have accumulated inductive evidence for
believing that accountants tend to be reliable sources of information about taxes,
while politicians in the middle of their campaigns tend to be unreliable sources
of information about the characters of their political opponents. In such cases, a
pattern of interaction with speakers who fall under various relevant types enables
174 It Takes Two to Tango

hearers to acquire positive reasons for accepting some of the reports that they
are offered.⁴⁰
What these considerations suggest is that ordinary hearers are confronted with
a plethora of epistemically relevant positive reasons that come in a variety of
forms. Such reasons are often not explicitly brought to mind but they nonethe-
less play a crucial role in our epistemic lives, as we tacitly discriminate among
and evaluate pieces of incoming information and compare such input with our
background beliefs.⁴¹ This point is borne out by noticing just how difficult it is
to construct a case in which a speaker truly fails to have any relevant positive reas-
ons for accepting a given report. Indeed, even in alien, the fact that the book in
question is written in what looks like English and appears to be what we would
call a diary may provide Sam with positive reasons for thinking the alien’s society
is similar in some crucial respects to Earth.
The third point I should like to make is in response to (b), the objection that
even when ordinary epistemic agents have enough information to justify par-
ticular reports, the positive reasons themselves are often indebted to testimony.
Here, it is crucial to notice that the positive reasons for accepting a speaker’s testi-
mony can themselves depend on testimony, so long as they are not ultimately and
entirely testimonially grounded. In particular, in order to avoid circular appeals to
testimony, one can reject
(i) For each report, R, the positive reasons justifying R cannot themselves
be acquired from the testimony of others.
and still accept
(ii) For each report, R, the positive reasons justifying R cannot ultimately
be testimonially grounded, where this means that the justificatory or
epistemic chain leading up to R does not ‘bottom out’ in testimony.
To see this, recall the case of nested speaker: through the course of their
friendship, Fred acquired excellent inductively based positive reasons for
believing that Helen is a highly reliable source of information on a wide range
of topics. Because of this, he readily accepts Helen’s testimony that Pauline, a
close friend of hers, is a highly trustworthy person, especially when it comes to
information regarding wild birds. This, in turn, leads him to accept Pauline’s
testimony that albatrosses have the largest wingspan among wild birds. To link
this up with the above distinction, Fred’s positive reasons for Helen’s testimony
satisfy both (i) and (ii), whereas Fred’s positive reasons for Pauline’s testimony
satisfy only (ii). In particular, Fred’s positive reasons for accepting Helen’s reports
have been acquired via sense perception, memory, and inductive inference, and
so they are not based on further testimony in the most direct sort of way.
Regarding Pauline’s testimony, however, Fred’s positive reasons for accepting her
report are based on further testimony in this direct sense; but the epistemic chain
ultimately ‘bottoms out’ in a non-testimonial source, namely, Fred’s inductive
evidence for Helen’s reliability. It is this weaker sense of being non-testimonially
 Jennifer Lackey 175

grounded—expressed in (ii)—that is crucial for avoiding circularity. According

to dualism, the reasons invoked for accepting a speaker’s testimony can, and
often are, themselves acquired from the testimony of others. Indeed, there can
be epistemic chains involving many more speakers than two. For instance, A’s
reason for accepting B’s testimony may be the testimony of C which, in turn, may
be based on the testimony of D, and so on. What is crucial for satisfaction of (2) is
that the final link in the epistemic chain in question is non-testimonially justified.
Perhaps an analogy will help here: in order to avoid both a vicious regress
and the problem of circularity, traditional foundationalism requires that a jus-
tificatory chain ultimately bottom out in foundational, or basic, beliefs. But it
does not require that each justified belief be, in turn, directly justified by a basic
belief. Similarly, in order to avoid both a vicious regress and circular appeals
to testimony, the positive-reasons condition requires that the justificatory chain
ultimately bottom out in a non-testimonial source. But, it does not require direct
non-testimonial grounding for each testimonial belief.
Moreover, even the final link in the chain can be partially indebted to testi-
mony, so long as there is enough non-testimonial support to render it not irra-
tional to accept the report in question. For instance, suppose I form beliefs on the
basis of reading an internet site about Howard Dean because Jack, an acquaint-
ance of mine from the Political Science Department, told me that it is a reliable
source of information. Why do I trust Jack? In part, because my colleague, Jill,
told me that he is an honest, competent, Democrat. However, I have also had
enough personal interaction with Jack to acquire a partial non-testimonial basis
for relying on his testimony. So long as this partial non-testimonial basis is suffi-
cient to render it not irrational to trust the Howard Dean website, Jack can be the
non-testimonial source in which this chain bottoms out.
Once it is clear, therefore, first, that the positive reasons required in (3) of dual-
ism only need to render it not irrational for the hearer to accept the testimony in
question, second, that there are a multitude of kinds of reasons that have epistemic
significance, and, third, that the positive reasons in question can themselves rely on
testimony, so long as they do not do so entirely and ultimately, (a) and (b) simply
lose their force.
There is, however, a further concern that may be expressed about condition
(3) of dualism. To see this, recall that the argument motivating the inclusion of
this condition was alien. But now, one might ask, couldn’t a parallel argument
be constructed regarding other, purportedly more basic, epistemic sources, such
as sense perception, memory, and inference? For instance, suppose that after her
involvement in a car accident, Olivia has complete amnesia with respect to her
perceptual faculties, that is, she remembers nothing about either the workings
or the deliverances of such faculties. Upon waking from the car accident and
seeing the face of her sister, Olivia forms the corresponding perceptual belief.
Intuitively, is such a belief justified? Since this case is similar in all of the relevant
respects to alien, shouldn’t we conclude here, as we did there, that it would be
176 It Takes Two to Tango

irrational for Olivia to hold such a belief in the absence of epistemically relevant
positive reasons on behalf of her perceptual faculties? If so, there seems to be a
problem of overgeneralization here. For now it looks as though positive reasons
are needed to justifiedly hold, not just testimonial beliefs, but any beliefs. And
this, in turn, leads us into all of the problems facing traditional internalist the-
ories of epistemic justification, such as infinite regresses, circularity, foundations,
and so on.
Although a complete response to this concern lies outside the scope of this
paper, I shall here highlight three salient ways in which testimony differs epistem-
ically from other sources of belief. These differences allow us to conclude that,
although alien shows that positive reasons are needed for testimony, no simil-
ar case can be constructed for our other cognitive faculties; the need for positive
reasons thus does not generalize.
First, testimonial beliefs are acquired from persons.⁴² Persons, unlike other
sources of belief, have all sorts of different intentions, desires, goals, motives,
and so on. Some of these desires and goals make it very advantageous to lie,
to exaggerate, to mislead, and to otherwise deceive. Indeed, recall that in the
discussion following alien, some of the considerations motivating the need for
a condition like (3) of dualism were that ‘it may very well be accepted practice
in alien society to be insincere and deceptive when testifying to others. Or,
normal alien psychology may be what we Earthlings would consider psychosis’.
Since both of these possibilities appeal to features distinctive of persons, they are
relevant only in the case of testimony.
Of course, it may be argued that other sources of belief can lead us just as
far astray epistemically as testimony. For instance, aren’t the paradoxes just as
misleading as incompetent testifiers, and aren’t perceptual hallucinations and
illusions just as deceptive as compulsive liars? Given these parallels, it looks like
the mere fact that testimonial beliefs are acquired from persons fails to distinguish
it epistemically from other sources of belief.
By way of response to this point, notice that there are two aspects that are often
involved in rendering a speaker a reliable source of belief: her competence as a
believer and her sincerity as a testifier.⁴³ Accordingly, when a hearer acquires a
false belief from a speaker, one (or both) of these aspects is typically responsible:
either A reports that p when p is false because A herself erroneously believes that
p (i.e., A is an incompetent believer), or A reports that p when p is false because
A intends to deceive her hearer (i.e., A is an insincere testifier). But now notice:
the paradoxes, perceptual illusions, hallucinations, and so on all parallel only the
testimonial case of incompetent believing—there simply is no analogue of insin-
cere testifying with non-testimonial sources of belief.⁴⁴ For insincerity involves
the intention to deceive or mislead, and intentions of this sort are distinctive of
persons. When my rational and perceptual faculties lead me astray epistemically,
they do not intend to do so.⁴⁵ Because of this, failures in the case of testimony are
much more unpredictable than failures in non-testimonial cases.
 Jennifer Lackey 177

A second and somewhat related difference between testimony and other epi-
stemic sources concerns the varying degrees of likelihood that such sources are
unreliable. For instance, the possible worlds in which most of my perceptual
beliefs are indistinguishably false—for instance, worlds in which I am unknow-
ingly a brain-in-a-vat or the victim of an evil demon—are quite distant from
the actual world. Indeed, even possible worlds in which many of my perceptual
beliefs are indistinguishably false are rather far away—worlds, for instance, where
my perceptual faculties frequently malfunction and yet I do not suspect that they
do. In contrast, the possible worlds in which most of my testimonial beliefs are
indistinguishably false—for instance, worlds in which I was raised by parents
who belong to a cult, or worlds in which my government is highly corrupt, or
worlds in which my society is highly superstitious—are much closer. Indeed, for
many people, this is true in the actual world. Given this much greater chance
for error in the case of testimony, the rational acceptance of the reports of others
requires positive reasons in a way that is not paralleled with other cognitive fac-
ulties.
This brings us to the third and, to my mind, most important epistemic differ-
ence between testimony and other sources of belief. To fully appreciate this point,
let us return to our perceptual amnesiac, Olivia, and imagine an average day for
her after leaving the hospital: she stops at the store to buy some groceries, bumps
into some acquaintances on her way home, watches an episode of Seinfeld on TV
while eating dinner, and spends some time on the internet before going to bed.
Along the way, Olivia forms perceptual beliefs about all sorts of things, includ-
ing beliefs about the vegetarian items that Trader Joe’s carries, the kinds of trees
losing their leaves, the number of children her acquaintance now has, which Sein-
feld episode is on, and the color of the background of the MSN website. Now,
because of her perceptual amnesia, Olivia’s acquisition of these perceptual beliefs
is not governed by any acquired principles of perceptual belief formation. But
even in the absence of such principles, it seems reasonable to conclude that the
overall status of Olivia’s daily perceptual beliefs would be very high epistemically.
For Olivia’s beliefs are most likely quite similar to those that would have been
acquired by a subject in the same circumstances who does have acquired prin-
ciples of perceptual belief formation governing her acceptance. Why? Because
sense perception, like other non-testimonial sources, is fairly homogeneous —there
is, for instance, simply not much of a difference epistemically between Olivia see-
ing groceries at Trader Joe’s and Olivia seeing trees without their leaves. Accord-
ingly, when forming non-testimonial beliefs, subjects do not need to be very
discriminating in order to be reliably in touch with the truth.⁴⁶
Now compare Olivia’s day with Edna’s. Edna, Olivia’s best friend, was in the
same car accident that caused Olivia’s perceptual amnesia. In Edna, however,
the accident caused testimonial amnesia: she remembers nothing about either the
workings or the deliverances of testimony. After leaving the hospital, Edna’s day
was nearly identical to Olivia’s. For instance, she stopped at the same grocery
178 It Takes Two to Tango

store, bumped into the same acquaintances on the way home, watched the
same episode of Seinfeld, and visited the same internet sites before going to
bed. Now, because of her testimonial amnesia, Edna’s acquisition of testimonial
beliefs along the way was not governed by any principles of testimonial belief
formation. As a result, Edna trusted to the same extent all of the testimonial
sources she encountered throughout the day—which included a copy of the
National Enquirer that she read at the grocery store, her acquaintance’s 3-
year-old daughter, the characters of Jerry and George on Seinfeld, and an
extremist, evangelical Christian internet site she stumbled upon while surfing
the web—and she believed everything that she was either told or read along the
way—which included testimony that a woman from Georgia was abducted by
aliens, that there are real princes and princesses at Disneyland, that licking the
envelopes of cheap wedding invitations can lead to one’s death, and that those
who are gay will be sent to eternal damnation.
Did Edna fare as well epistemically as Olivia? Not at all. In the absence of
acquired principles governing the acceptance of testimony, Edna was led very far
astray epistemically. She trusted the National Enquirer as much as she would have
trusted the New York Times, she trusted a 3-year-old’s depiction of Disneyland as
much as she would have trusted an adult’s, she trusted the characters on a sitcom
as much as she would have trusted those interviewed in a National Geographic
documentary, and she trusted the rantings of an extremist, evangelical Christian
internet site as much as she would have trusted news found on the MSN website.
Because of this, Edna’s beliefs are very different from those that would have been
acquired by a subject in the same circumstances who does have her testimonial
practices governed by such epistemic principles.
Thus, testimony is quite unlike other sources of belief precisely because it is so
wildly heterogeneous epistemically—there is, for instance, all the difference in the
world between reading the National Enquirer and reading the New York Times.
Moreover, this heterogeneity requires subjects to be much more discriminating
when accepting testimony than when trusting, say, sense perception. Non-
testimonial analogues of alien, therefore, simply fail to motivate a positive-
reasons condition similar to (3) of dualism.

4 . B EYO N D R E D U C T I O N I S M A N D N O N - R E D U C T I O N IS M

In closing, I shall discuss two of the central and most important consequences
that dualism has for the epistemology of testimony.
First, dualism provides easy resolutions to many of the central and
most divisive disagreements between reductionists and non-reductionists. For
instance, because reductionism holds that testimony is reducible to other
epistemic sources, such a view is often attacked for underestimating or devaluing
the importance of testimony. On the other hand, because non-reductionism
 Jennifer Lackey 179

endorses the complete independence of testimony from those sources that

could non-circularly provide positive reasons, such a view is often criticized
for sanctioning gullibility and intellectual irresponsibility. Dualism, however,
avoids both of these objections: testimony is an irreducible epistemic source,
thereby avoiding the charge of underestimating or devaluing the importance
of testimony, and hearers need positive reasons in order to acquire testimonial
justification, thereby avoiding the charge of sanctioning gullibility and
intellectual irresponsibility.
Second, dualism correctly moves the focus of the current debate beyond reduc-
tionism and non-reductionism. As we have seen, a central component of reduc-
tionism—indeed, the one for which the view derived its name—is that the
justification of testimony reduces to the justification of sense perception, memory,
and inductive inference. Testimony does not, therefore, make any justificatory
contributions of its own—whatever appearances there are to the contrary can
ultimately be reduced to the epistemic contributions of these other sources. Oth-
erwise put, if reductionism is true, there is no specifically testimonial justification.
There is only justification for beliefs acquired through testimony. But this view
gives testimony far too little epistemic credit. For, as we saw in nested speaker,
the reliability of the report in question is not something that can be reduced to
perceptual, memorial, and inferential justification. Since the reliability of the
report in a testimonial exchange is a necessary condition of the justification of
beliefs acquired via testimony, testimonial justification itself is not reducible.
Thus, contrary to reductionism, there is justification that is distinctly testimonial
in nature.
Non-reductionists, on the other hand, claim that since testimony is just
as basic a source of justification as sense perception, memory, and inductive
inference, hearers may be justified in accepting the reports of speakers, albeit
defeasibly, merely on the basis of a speaker’s testimony. In this way, testimony is
construed as a completely independent and autonomous source of justification,
needing neither the direct input nor the assistance of any other sources.⁴⁷
But this view gives testimony far too much credit. As we saw in alien, to
accept a speaker’s testimony in the complete absence of positive reasons on
behalf of the testimony in question is to exhibit an epistemically unacceptable
kind of irrationality, one that is incompatible with testimonial justification.
Moreover, since there would be vicious circularity if the positive reasons were
themselves entirely and ultimately acquired via testimony, direct input and
assistance are needed from sense perception, memory, and inductive inference.
Thus, contrary to non-reductionism, testimonial justification depends on the
justificatory resources of other epistemic sources.
In showing the need for positive epistemic work from both the speaker and
hearer, we have seen that testimonial justification is neither reducible to nor com-
pletely independent from sense perception, memory, and inductive inference.
180 It Takes Two to Tango

Thus, insofar as we wish to make genuine progress in the epistemology of testi-

mony, we need to move beyond the debate between reductionism and non-
reductionism.

REFERENCES
Adler, Jonathan E. (1994), ‘Testimony, Trust, Knowing’, Journal of Philosophy, 91:
264–75.
(2002), Belief ’s Own Ethics (Cambridge, Mass.: MIT Press).
Alston, William P. (1989), Epistemic Justification: Essays in the Theory of Knowledge (Ithaca,
NY: Cornell University Press).
Audi, Robert (1997), ‘The Place of Testimony in the Fabric of Knowledge and Justifica-
tion’, American Philosophical Quarterly, 34: 405–22.
(1998), Epistemology: A Contemporary Introduction to the Theory of Knowledge (Lon-
don and New York: Routledge).
Bergmann, Michael (1997), ‘Internalism, Externalism and the No-Defeater Condition’,
Synthese, 110: 399–417.
(2004), ‘Epistemic Circularity: Malignant and Benign’, Philosophy and Phenomeno-
logical Research, 69: 709–27.
Austin, J. L. (1979), ‘Other Minds’, in his Philosophical Papers, 3rd edn. (Oxford: Oxford
University Press).
BonJour, Laurence (1980), ‘Externalist Theories of Epistemic Justification’, Midwest
Studies in Philosophy, 5: 53–73.
(1985), The Structure of Empirical Knowledge (Cambridge, Mass.: Harvard Univer-
sity Press).
and Sosa, Ernest (2003), Epistemic Justification: Internalism vs. Externalism, Founda-
tions vs. Virtues (Oxford: Blackwell Publishing).
Burge, Tyler (1993), ‘Content Preservation’, Philosophical Review, 102: 457–88.
(1997), ‘Interlocution, Perception, and Memory’, Philosophical Studies, 86: 21–47.
Chisholm, Roderick M. (1989), Theory of Knowledge, 3rd edn. (Englewood Cliffs, NJ:
Prentice-Hall).
Coady, C. A. J. (1992), Testimony: A Philosophical Study (Oxford: Clarendon Press).
(1994), ‘Testimony, Observation and ‘‘Autonomous Knowledge’’ ’, in Matilal and
Chakrabarti (1994: 225–50).
Dummett, Michael (1994), ‘Testimony and Memory’, in Matilal and Chakrabarti (1994:
251–72).
Evans, Gareth (1982), The Varieties of Reference (Oxford: Clarendon Press).
Faulkner, Paul (2000), ‘The Social Character of Testimonial Knowledge’, Journal of
Philosophy, 97: 581–601.
(2002), ‘On the Rationality of our Response to Testimony’, Synthese, 131: 353–70.
Foley, Richard (1994), ‘Egoism in Epistemology’, in Schmitt (1994: 53–73).
Fricker, Elizabeth (1987), ‘The Epistemology of Testimony’, Proceedings of the Aristoteli-
an Society, suppl. vol. 61: 57–83.
(1994), ‘Against Gullibility’, in Matilal and Chakrabarti (1994: 125–61).
(1995), ‘Telling and Trusting: Reductionism and Anti-Reductionism in the Epi-
stemology of Testimony’, Mind, 104: 393–411.
 Jennifer Lackey 181

(2002), ‘Trusting Others in the Sciences: A Priori or Empirical Warrant?’, Studies

in History and Philosophy of Science, 33: 373–83.
(forthcoming), ‘Knowledge from Trust in Testimony is Second-Hand Knowledge’,
Philosophy and Phenomenological Research.
Goldman, Alvin I. (1986), Epistemology and Cognition, (Cambridge, Mass.: Harvard
University Press).
(1992), Liaisons: Philosophy Meets the Cognitive and Social Sciences (Cambridge,
Mass.: MIT Press).
(1999), Knowledge in a Social World (Oxford: Clarendon Press).
Greco, John and Sosa, Ernest (eds.) (1999), The Blackwell Guide to Epistemology (Oxford:
Blackwell Publishers).
Hardwig, John (1985), ‘Epistemic Dependence’, Journal of Philosophy, 82: 335–49.
(1991), ‘The Role of Trust in Knowledge’, Journal of Philosophy, 88: 693–708.
Hawthorne, John (2004), Knowledge and Lotteries (Oxford: Oxford University Press).
Hume, David (1967), An Enquiry Concerning Human Understanding, in L. A. Selby-
Bigge (ed.), Hume’s Enquiries (Oxford: Oxford University Press).
Insole, Christopher J. (2000), ‘Seeing Off the Local Threat to Irreducible Knowledge by
Testimony’, Philosophical Quarterly, 50: 44–56.
Lackey, Jennifer (1999), ‘Testimonial Knowledge and Transmission’, Philosophical
Quarterly, 49: 471–90.
(2003), ‘A Minimal Expression of Non-Reductionism in the Epistemology of Testi-
mony’, Noûs 37: 706–23.
(2005), ‘Testimony and the Infant/Child Objection’, Philosophical Studies, 126:
163–90.
(forthcoming a), ‘Learning from Words’, Philosophy and Phenomenological Research.
Lipton, Peter (1998), ‘The Epistemology of Testimony’, Studies in History and Philosophy
of Science, 29: 1–31.
Lyons, Jack (1997), ‘Testimony, Induction and Folk Psychology’, Australasian Journal of
Philosophy, 75: 163–78.
McDowell, John (1994), ‘Knowledge By Hearsay’, in Matilal and Chakrabarti (1994:
195–224).
Matilal, Bimal Krishna, and Chakrabarti, Arindam (eds.), (1994), Knowing From Words
(Dordrecht: Kluwer Academic Publishers).
Millgram, Elijah (1997), Practical Induction (Cambridge, Mass.: Harvard University
Press).
Nozick, Robert (1981), Philosophical Explanations (Cambridge, Mass.: Belknap Press).
Owens, David (2000), Reason without Freedom: The Problem of Epistemic Normativity
(London and New York: Routledge).
Plantinga, Alvin (1993), Warrant and Proper Function (Oxford: Oxford University Press).
Pollock, John (1986), Contemporary Theories of Knowledge (Totowa, NJ: Rowman and
Littlefield).
Reed, Baron (forthcoming), ‘Epistemic Circularity Squared? Skepticism about Common
Sense’, Philosophy and Phenomenological Research.
Reid, Thomas (1993), The Works of Thomas Reid, ed. Sir William Hamilton (Charlottes-
ville, Va.: Lincoln-Rembrandt Publishing).
182 It Takes Two to Tango

Reynolds, Steven L. (2002), ‘Testimony, Knowledge, and Epistemic Goals’, Philosophical

Studies, 110: 139–61.
Root, Michael (2001), ‘Hume on the Virtues of Testimony’, American Philosophical
Quarterly, 38: 19–35.
Ross, Angus (1986), ‘Why Do We Believe What We Are Told?’, Ratio, 28: 69–88.
Rysiew, Patrick (2002), ‘Testimony, Simulation, and the Limits of Inductivism’, Aus-
tralasian Journal of Philosophy, 78: 269–74.
Schmitt, Frederick F. (ed.) (1994), Socializing Epistemology: The Social Dimensions of
Knowledge (Lanham, Md.: Rowman and Littlefield).
(1999), ‘Social Epistemology’, in Greco and Sosa (1999: 354–82).
Sosa, Ernest (1996), ‘Postscript to ‘‘Proper Functionalism and Virtue Epistemology’’ ’, in
J. L. Kvanvig (ed.), Warrant in Contemporary Epistemology (Lanham, Md.: Rowman &
Littlefield: 271–81.
(1999), ‘How Must Knowledge Be Modally Related to What Is Known?’, Philo-
sophical Topics, 26: 373–84.
(2000), ‘Contextualism and Skepticism’, in J. Tomberlin (ed.), Philosophical Issues,
suppl. to Noûs 34: 94–107.
(2002), ‘Tracking, Competence, and Knowledge’, in P. Moser (ed.), The Oxford
Handbook of Epistemology (Oxford: Oxford University Press).
Stevenson, Leslie (1993), ‘Why Believe What People Say?’, Synthese, 94: 429–51.
Strawson, P. F. (1994), ‘Knowing From Words’, in Matilal and Chakrabarti (1994:
23–7).
Webb, Mark Owen (1993), ‘Why I Know About As Much As You: A Reply to Hardwig’,
Journal of Philosophy, 90: 260–70.
Weiner, Matthew (2003), ‘Accepting Testimony’, Philosophical Quarterly, 53: 256–64.
Welbourne, Michael (1979), ‘The Transmission of Knowledge’, Philosophical Quarterly,
29: 1–9.
(1981), ‘The Community of Knowledge’, Philosophical Quarterly, 31: 302–14.
(1986), The Community of Knowledge (Aberdeen: Aberdeen University Press).
(1994), ‘Testimony, Knowledge and Belief ’, in Matilal and Chakrabarti (1994:
297–313).
Williams, Michael (1999), Groundless Belief: An Essay on the Possibility of Epistemology,
2nd edn. (Princeton: Princeton University Press).
Williamson, Timothy (1996), ‘Knowing and Asserting’, Philosophical Review, 105:
489–523.
(2000), Knowledge and Its Limits (Oxford: Oxford University Press).

N OT E S
1. I am here assuming that justification is necessary, and, when added to true belief,
close to being sufficient for knowledge. There are some other kinds of justification
(e.g., justification grounded entirely in one’s subjective perspective) that may escape
some of the arguments I give in this paper. In these cases, my arguments can be read
as targeting reductionism about testimonial warrant or knowledge.
2. This condition is included to avoid circularity, i.e., testimonial beliefs ultimately
justifying other testimonial beliefs.
 Jennifer Lackey 183

3. Proponents of different versions of reductionism include Hume (1967), Fricker

(1987, 1994, 1995, and Chapter 10 in this volume), Adler (1994 and 2002), Lyons
(1997), Lipton (1998), and Van Cleve (Chapter 2 in this volume). For a nice dis-
cussion of Hume’s version of reductionism, see Root (2001). Lehrer (Chapter 7 in
this volume) develops a qualified reductionist/non-reductionist view of testimonial
justification and knowledge.
4. More precisely, this is one version of global reductionism. The other version requires
that ‘a hearer have evidence that most of what she has ever learned through testimony
is true, where this evidence does not in any way rest on knowledge acquired by
her through testimony’ (Fricker 1994: 134). Though this weaker version of global
reductionism avoids some of the objections raised to the stronger one, it faces prob-
lems of its own. Most notably, in the absence of a good reason to believe that most
of what one has ever learned through testimony adequately and non-randomly repres-
ents either testimony in general or testimony that one will encounter in the future, it is
unclear why this reason would justify future acceptances of testimony.
5. For a detailed discussion of some of the specific justificatory issues that arise regard-
ing young children’s acceptance of testimony, see Lackey (2005).
6. A notable exception is Fricker (1994). As she says, ‘looking for generalisations about
the reliability or otherwise of testimony . . . as a homogenous whole, will not be
an enlightening project. Illuminating generalisations, if there are any, will be about
particular types of testimony, differentiated according to subject matter, or type of
speaker, or both. . . . [W]hen it comes to the probability of accuracy of speakers’
assertions, and what sorts of factors warrant a hearer in trusting a speaker, testimony is
not a unitary category’ (Fricker 1994: 139; emphasis added).
7. ‘My reliance on a particular piece of testimony reduces locally just if I have adequate
grounds to take my informant to be trustworthy on this occasion independently of
accepting as true her very utterance’ (Fricker 1995: 404).
8. Strictly speaking, I shall argue that even the weakest version of only the Reduction
Component of reductionism is false, leaving the Positive-Reasons Component of
reductionism untouched. (Indeed, I shall argue in the next section that the Positive-
Reasons Component is correct.) However, since I am taking the Reduction Com-
ponent to be an essential part of reductionism (what, after all, is reductionism with
no reduction?), a view that merely includes the Positive-Reasons Component does
not qualify as a reductionist view.
9. A point of clarification: I refer to this case as nested speaker because the justification
for accepting one speaker’s report is nested within the positive reasons for accepting
another speaker’s report. More precisely, the justification for accepting Pauline’s
testimony is nested within the positive reasons for accepting Helen’s testimony.
10. This is not to say that a belief that fails to be reliably produced or appropriately
connected with the truth will necessarily render a report based on such a belief unre-
liable. Indeed, I have elsewhere argued that the epistemic status of beliefs and reports
can come apart so that an unreliable believer can nonetheless be a reliable testifier.
(See Lackey 1999, 2003, and forthcoming a.) Rather, the problem with Pauline’s
testimony in nested speaker is that her unreliability as a believer leads her to also be
an unreliable testifier.
184 It Takes Two to Tango

11. For discussions specifically about the role of competence and sincerity in testimony,
see, for instance, Welbourne (1979, 1981, 1986, and 1994), Hardwig (1985 and
1991), Ross (1986), Fricker (1987, 1994, 1995, forthcoming, and Chapter 10 in
this volume), Plantinga (1993), McDowell (1994), Audi (1997, 1998, and Chapter
1 in this volume), Root (2001), Owens (2000), and Adler (2002). For indirect
endorsements of competence and sincerity as necessary conditions for testimonial
knowledge (justification), e.g., via the stronger requirement that the speaker have the
knowledge (justified belief ) to which she is testifying, see Burge (1993 and 1997),
Williamson (1996), Dummett (1994), Reynolds (2002), and Schmitt (Chapter 9 in
this volume).
12. See Lackey (1999, 2003, and forthcoming a).
13. A proponent of reductionism may object to the conclusion of nested speaker by
arguing that justification—unlike, for instance, warrant and knowledge—is primar-
ily an internalist notion. For instance, if I were a brain-in-a-vat and had no idea that I
was, one might argue that I would still be justified in believing that I am here typing
at this computer, even if I do not know that I am. Similarly, one might claim that in
nested speaker, the belief that Fred forms on the basis of Pauline’s radically unre-
liable testimony is justified, even if it does not qualify as knowledge. (I am grateful to
Peter Graham for pressing this point.)
By way of response to this objection, I shall make four points. First, it is not at
all uncommon in the literature to find justification being discussed, either entirely
or partially, in externalist terms. To name just a few, see Alston (1989), Goldman
(1992), and BonJour and Sosa (2003).
Second, many reductionists are equally reductionistic about warrant and know-
ledge (though, of course, a truth condition is added when testimonial knowledge is
at issue). Hence, for those who hold that justification is a purely internalist notion,
my arguments in this section can simply be recast as arguments against reductionism
about testimonial warrant or knowledge.
Third, in Section 3, I shall argue that justification has two central components:
(i) a reliability component and (ii) a rationality component. Thus, if I were a brain-
in-a-vat and had no idea that I was, my belief that I am here typing at this computer
would still satisfy the rationality constraint of justification. More precisely, even
though my beliefs in a skeptical scenario would fail the reliability constraint and,
hence, would not be justified, they would nevertheless be rational (indeed, such
beliefs would possess many other positive epistemic properties, such as being held
in an epistemically responsible way, being epistemically virtuous, and so on). So, giv-
en the distinction between (i) and (ii), my view of justification is able to explain the
intuition that skeptical scenario victims both possess and lack something epistemically
important: they possess rationality, but lack reliability and hence justification.
Fourth, if justification is understood as a purely internalist notion, then it is not
entirely clear what connection, if any, it has to knowledge. Traditionally, justific-
ation has been understood as necessary and, when added to true belief, close to
sufficient for knowledge. On this reading of justification—the one that I am here
assuming—internalists and externalists are engaged in a genuine debate about the
same condition for knowledge. Moreover, on this traditional reading, it is clear why
justification has epistemic value: it converts, with some help from a
 Jennifer Lackey 185

Gettier-condition, true belief into knowledge. In contrast, on the purely internalist

conception of justification assumed by this objection, not only does the internalism/
externalism debate turn out to be quibbling over entirely different concepts, it also
becomes unclear what the precise epistemic value of justification is.
14. Let me emphasize that this is how one might plausibly construe the case; it is not,
however, how I would characterize it. Despite the fact that Holly is a generally reli-
able testifier, I would hold that Max’s belief on the basis of her testimony fails to be
justified because the report itself is not reliably produced. (For more on this condi-
tion of testimonial justification, see Lackey (1999, 2003, and forthcoming a).) On
my view, the virtue of nested speaker, in comparison with unnested speaker, is
that it entirely precludes this sort of plausible reading.
15. It is also worth noting that analogues of nested speaker cannot be constructed for
other epistemic sources since it exploits features unique to the case of testimony, i.e.,
the ability to have one’s justification for accepting a speaker’s report nested within
the positive reasons for accepting another speaker’s report.
16. I am grateful for an exchange with Sandy Goldberg that prompted the addition of
both unnested speaker and the discussion contrasting it with nested speaker.
17. A further reductionist strategy for resisting the conclusion of nested speaker is to
make it part of the definition of ‘appropriate’ that the reasons in question lead hear-
ers to reports that bear a reliable connection with the truth. This would make it
impossible for a subject to have appropriate positive reasons for believing a report
that wasn’t reliably connected with the truth. This strategy, however, has at least one
of two unattractive consequences. It either (i) makes the requirement for positive
reasons so externalistic that it undermines the central motivation for endorsing the
Positive-Reasons Component in the first place; namely, to preserve a link between
epistemic justification and subjective rationality, or (ii) it makes the requirements for
testimonial justification so stringent that there would be far less testimonial justifica-
tion than is intuitively acceptable.
18. More precisely, there are two different kinds of defeaters that are standardly taken
to be incompatible with justification (and knowledge). First, there are what we
might call psychological defeaters. A psychological defeater is an experience, doubt,
or belief that is had by S, yet indicates that S’s belief that p is either false or unre-
liably formed or sustained. Defeaters in this sense function by virtue of being had
by S, regardless of their truth value or justificatory status. Second, there are what
we might call normative defeaters. A normative defeater is a doubt or belief that
S ought to have, yet indicates that S’s belief that p is either false or unreliably
formed or sustained. Defeaters in this sense function by virtue of being doubts or
beliefs that S should have (whether or not S does have them) given the presence
of available evidence. For various discussions of what I call psychological defeaters
see, for example, BonJour (1980 and 1985), Nozick (1981), Pollock (1986), Gold-
man (1986), Plantinga (1993), Lackey (1999, 2003, and 2005), Bergmann (1997
and 2004), and Reed (forthcoming). For discussions involving what I call normative
defeaters, approached in a number of different ways, see BonJour (1980 and 1985),
Goldman (1986), Fricker (1987 and 1994), Chisholm (1989), Burge (1993 and
1997), McDowell (1994), Audi (1997 and 1998), Williams (1999), Lackey (1999,
186 It Takes Two to Tango

2003, and 2005), BonJour and Sosa (2003), Hawthorne (2004), and Reed (forth-
coming). What all of these discussions have in common is simply the idea that
evidence can defeat justification (and knowledge) even when the subject does not
form any corresponding beliefs from the evidence in question.
19. For various versions of non-reductionism, see Austin (1979), Welbourne (1979,
1981, 1986, and 1994), Evans (1982), Ross (1986), Hardwig (1985 and 1991),
Coady (1992 and 1994), Reid (1993), Burge (1993 and 1997), Plantinga
(1993), Webb (1993), Dummett (1994), Foley (1994), McDowell (1994), Strawson
(1994), Williamson (1996 and 2000), Goldman (1999), Schmitt (1999), Insole
(2000), Owens (2000), Rysiew (2002), Weiner (2003), and Goldberg (Chapter 6
in this volume). Some phrase their view in terms of knowledge, others in terms of
justification or entitlement, still others in terms of warrant. Audi (1997, 1998, and
Chapter 1 in this volume) embraces a non-reductionist view of testimonial know-
ledge, but not of testimonial justification. Stevenson (1993), Millgram (1997), and
Graham (Chapter 4 in this volume) defend restricted versions of non-reductionism.
Faulkner (2000) develops a ‘hybrid’ reductionist/non-reductionist view of testimo-
nial justification and knowledge, though it differs in some crucial respects from the
dualist theory I suggest later in this paper.
20. It was suggested to me by Lizzie Fricker that another type of example that may make
the same general point would be a person receiving testimony over the internet, with
absolutely no epistemically relevant information about the source of the testimony.
(Fricker also mentions this sort of case in her (2002).)
21. Even outside the epistemology of testimony, this is a very widely accepted view. For
instance, adding a no-defeater condition to theories of epistemic justification that are
otherwise externalist has become the standard response given to the counterexamples
found in BonJour’s (1980 and 1985). See, for instance, Nozick (1981), Goldman
(1986), and Plantinga (1993).
22. I am grateful to comments from Joe Shieber that prompted the addition of this point.
23. For additional arguments showing the inadequacy of standard versions of non-
reductionism, see Lackey (2003).
24. A non-reductionist who appreciates the need for justificatory contributions from
both the hearer and the speaker is Goldberg (Chapter 6 in this volume), who charac-
terizes the hearer’s epistemic contribution in terms of the right she has to rely on the
testimony that is offered. It is precisely the possession of positive reasons, however,
that I argue constitutes this right.
25. I say ‘crudely’ because I do not want to suggest either that the speaker condition is
entirely divorced from questions of rationality or that the hearer condition is entirely
divorced from questions of reliability.
26. Of course, my use of this term has no direct bearing on its use in other contexts, such
as in the philosophy of mind.
27. In Lackey (2003), I argue that there are at least two additional conditions needed
for testimonial justification, one requiring that the hearer is a properly functioning
recipient of testimony and another requiring that the environment is suitable for the
reception of reliable testimony.
 Jennifer Lackey 187

28. A rough taxonomy of the three competing positions, then, is as follows:

Reductionism: Reduction Component and Positive-Reasons Component (PR-N&S)
Non-Reductionism: No-Reduction Component and No-Positive-Reasons Compon-
ent. Dualism: Positive-Reasons Component (PR-N) and No-Reduction Component
29. This example is a slight variation of one found in Audi (1997). For further discus-
sion of the distinction inspired by these sorts of cases, see Lackey (1999).
30. It should be noted that there is a weaker version of (1) (henceforth, (1*)) that
requires only that B believe that p on the basis of the content of A’s testimony, not
that such a belief be based on the content of A’s testimony that p. There are at least
two different kinds of cases that might support including (1*) rather than (1). For
instance, suppose that Ben asks me whether it is snowing outside and I respond by
saying, ‘There is a shovel in the closet’. Here, one might say that Ben is justified in
believing that it is snowing outside on the basis of my testimony that there is a shovel
in the closet. Thus, one might claim that (1) of dualism is, in fact, too strong since
one can be testimonially justified in believing that p on the basis of the content of
a speaker’s testimony that q. The second type of case that may pose a problem for
(1) can be seen by considering the following: suppose that Ruth asks me whether I
bought flowers at the farmer’s market and I respond with a nod. Here, it might be
argued that Ruth is justified in believing that I bought flowers at the farmer’s mar-
ket on the basis of the non-verbal testimony provided by my nod. In this way, one
might claim that while testimonial justification for believing that p is acquired on the
basis of the content of my testimony, it is not based on the content of my testimony
that p.
Although I think both types of examples may provide convincing evidence for
including the weaker (1*), there are also compelling reasons to prefer the stronger
(1). With respect to the first type of case, for instance, one might argue that Ben’s
justification for believing that it is snowing outside relies too heavily on memory
and inference to qualify as testimonial. In particular, one might claim that Ben must
rely on background information stored in memory about the relationship between
shovels and snow and explicitly infer that it is snowing outside from my testimony
that there is a shovel in the closet. And even though memory and inference arguably
play a role in the acquisition of all instances of testimonial justification, one might
claim that here the role of memory and inference is simply too significant for the
resulting justification to properly be regarded as testimonial. Regarding the second
type of case, it may be argued that though I do not explicitly state that I bought
flowers at the farmer’s market, there is a clear sense in which the non-verbal testi-
mony provided by my nod has this as its content. For just as my responding ‘Yes’
in this context amounts to ‘I bought flowers at the farmer’s market’, so too does my
nodding in this context amount to testifying to this proposition.
I do not wish to here choose between (1) and (1*). Though I am leaving the
stronger (1) in my characterization of dualism, this condition can easily be substi-
tuted with the weaker (1*) for those convinced by the cases discussed above.
31. For arguments showing that neither competent believing nor sincere testifying is
necessary for being a reliable testifier, see Lackey (forthcoming a).
188 It Takes Two to Tango

32. See Nozick (1981).

33. See Sosa (1996, 1999, 2000, and 2002). See also Williamson (2000).
34. In a similar spirit, Frederick Schmitt says that ‘[t]he objection [to reductionism] . . . I
regard as most persuasive [is] that we have too little first-hand experience to provide
a nontestimonial basis for induction to the reliability of testimony in all instances in
which testimonial belief is intuitively justified’ (Schmitt 1999: 359).
35. Similarly, Frederick Schmitt claims that ‘[m]y belief in the reliability of . . . testimony
sometimes has what passes for a first-hand inductive base, but this passing first-hand
basis is itself indebted to testimony for its justification’ (Schmitt 1999: 360).
36. This point has been suggested by other proponents of a positive-reasons require-
ment—most notably, by Fricker (1987, 1994, 1995, and 2002) and Faulkner
(2002).
37. Specifically, even if B has not observed a general conformity between A’s reports
and the truth, B may have observed the general conformity of reports delivered in
contexts of kind C and the truth. So, if B believes that A’s report is delivered in a C-
context, then this, combined with B’s inductive evidence regarding contexts of kind
C, may give B an epistemically relevant positive reason for A’s testimony.
38. In particular, even if B has not observed a general conformity between A’s reports
and the truth, B may have observed the general conformity of reports of kind R and
the truth. Thus, if B believes that A’s report is an instance of kind R, this, com-
bined with B’s inductive evidence regarding R-reports, may give B an epistemically
relevant positive reason for A’s testimony.
39. Specifically, even if B has not observed the general conformity between A’s reports
and the truth, B may have observed the general conformity of speakers of kind S
and the truth. Thus, if B believes that A is an S-speaker, then this, combined with
B’s inductive evidence regarding S-speakers, may give B an epistemically relevant
positive reason for A’s testimony.
40. The reasons from each of the three classes discussed may be either at the level of
tokens or types.
41. This point is frequently noticed in the case of negative evidence, that is, it is readily
acknowledged that average epistemic agents have acquired ample inductive evidence
for believing that certain conversational features, such as nervousness, lack of eye
contact, and confused behavior, suggest incompetence, insincerity, and unreliability.
But positive reasons are just the flip side of defeaters. For just as we have accumu-
lated this type of negative inductive evidence, we have also accumulated positive
inductive evidence for believing that certain contextual features are indicators of
competence, sincerity, and reliability.
42. I am using the term ‘person’ broadly so that non-human animals from whom we are
able to acquire testimonial beliefs qualify as persons in this sense.
43. As mentioned in n. 31, I argue in Lackey (forthcoming a) that neither of these
aspects is necessary for a speaker to be a reliable testifier, though they are often con-
stitutive of being a reliable testifier.
44. For a very nice discussion of this point, see Faulkner (2000).
45. A possible non-testimonial exception to this claim is the relationship between self-
deception and certain memories. (I am grateful to Baron Reed for pointing this
out.) A subject may, for instance, deceive herself by calling into question memories
 Jennifer Lackey 189

that are too painful to remember. While this is true, it is quite doubtful whether
self-deception can be overtly conscious and therefore intentional in the way that
insincere testimony is. What is more plausible is that self-deception takes place at
the unconscious level, and indirectly induces false belief. Thus, failures in memory
deriving from self-deception are more analogous to failures in perception regarding
some sub-personal glitch in how the faculty works.
46. My point here is not that all non-testimonial sources are homogeneous to the same
extent —there may, for instance, be more variation between memories from different
stages in one’s life than there is among different kinds of perceptual experiences. My
point is, rather, that there is a striking difference between the amount of variation
found within non-testimonial sources as compared with testimony as an epistemic
source. Moreover, if there are some non-testimonial sources—perhaps, for instance,
inductive inference—that are as epistemically heterogeneous as testimony, then a
positive-reasons condition may be needed to justifiedly accept the deliverances of
these sources as well. What I wish to establish here, however, is that the conclusion
of alien does not generalize to all epistemic sources.
47. There are two qualifications to this thesis: first, though other epistemic sources are
not necessary for justification and knowledge, according to non-reductionism, they
are of course needed for accessing the testimony of others, e.g., sense perception is
required to hear the spoken testimony of others. Second, even if these other epistem-
ic sources do not provide justification for accepting the testimony of speakers, they
can inhibit justification and knowledge by generating defeaters.
This page intentionally left blank
 PART I V
TESTIMONY AND THE EXTENT
O F O U R D E PE N D E N C E O N
OTH E R S
This page intentionally left blank
 9
Testimonial Justification and Transindividual
Reasons
Frederick F. Schmitt

This paper explores one aspect of the general question in what sense knowledge
on the basis of testimony, and more exactly belief justified on the basis of testi-
mony, is social. I take it to be uncontroversial that testimonial knowledge is social
in a fairly strong sense. One might initially express the social claim this way: testi-
monial knowledge of a proposition depends on the knowledge of the testifier.
That is,
I know that p on T ’s testimony that p only if T knows that p.¹
According to this condition, testimonial knowledge is social in the sense that
my having testimonial knowledge entails that there is knowledge belonging (in
the normal case) to an individual other than myself.² This precise condition
of testimonial knowledge has in fact been forcefully challenged (Lackey 1999;
Graham 2000), but I take it to be uncontroversial that something in the vicinity
of this condition is correct, and a strong social claim for testimonial knowledge is
beyond question.³
By contrast, testimonially justified belief is clearly not social in a way closely
analogous to this social condition on testimonial knowledge.⁴ It is not true that
My belief p is justified on the basis of T ’s testimony only if T has a good
reason to believe p.
To be sure, testimonially justified belief is uncontroversially social in a weaker
sense. If my belief p is justified on testimony, it does follow by the meaning of ‘‘on
testimony’’ that there is testimony, hence a testifier, hence (in the normal case)
an individual distinct from me. My belief cannot be justified on testimony that
does not exist. So there being testimonially justified belief entails that there is a
testifier. Nevertheless, my having a testimonially justified belief that p does not
entail that my testifier herself is in any positive epistemic position. I can have a
justified belief that there are strawberries in the refrigerator on Tina’s testimony
even if Tina has no good reason to believe this, if I justifiedly though mistakenly

I would like to thank Jennifer Lackey for help in revising the paper.
194 Transindividual Reasons

believe that Tina has a good reason to believe this. So my belief p’s being justified
on the basis of T ’s testimony does not entail that T has a good reason to believe p.
Testimonially justified belief is not social in this strong sense.⁵
Despite this, I think that a qualified social claim about testimonially justified
belief is defensible. In particular, there are instances of testimonially justified
belief p for which I lack a justified belief that the testifier has a good reason to
believe p, and I lack any other good reason to believe p. In such cases, if there
is to be a relevant good reason for me to believe p, it must be possessed by my
testifier rather than by me. What I wish to defend, then, is the following initially
implausible two-part thesis:
If my belief p is justified on testimony, then my testifier T possesses a good
reason to believe p, unless I possess such a reason.
I can have a belief p justified on testimony even though I possess no good
reason to believe p.⁶
Indeed, I wish to defend a stronger and even less intuitive social thesis, which
I will call the Transindividual Basing Thesis, or the Transindividual Thesis for
short:
(T1) If my belief p is justified on testimony, then it is justified on the basis
of the testifier’s good reason to believe p, unless on the basis of a good
reason to believe p I myself possess.
(T2) My belief p can be justified on testimony even though my belief is not
justified on the basis of a good reason to believe p I possess.⁷
If I fail to base my belief that there are strawberries in the refrigerator on a good
reason I possess, my belief can still be justified on Tina’s testimony, but not if it
fails to be justified on the basis of Tina’s good reason—e.g., because she lacks a
good reason: she did not look carefully enough in the refrigerator this morning.
My testimonial justification for believing p is in this sense hostage to whether my
testifier possesses a good reason to believe p. One might be inclined to dismiss
the Transindividual Thesis on the conceptual ground that, by the very notions
of ‘‘basing,’’ ‘‘reason,’’ and ‘‘reasoning,’’ basing a belief on a reason entails reas-
oning to the belief from the reason; yet, by the very notion of ‘‘testimony,’’ the
chain from the testifier’s reason, through testimony and testimonial uptake, to
my belief does not count as reasoning. But I find ‘‘basing’’ too vague a term to
carry the weight of this short objection.
The case for the Transindividual Thesis, in a nutshell, is this. The first com-
ponent, (T1), follows from the plausible assumption
(G) Any justified belief p of mine is justified on the basis of a good reason to
believe p suitably related to me.
For suppose (G) and suppose my belief p is justified on testimony. If I do not
possess a good reason to believe p, then the only remaining candidate for a reason
on the basis of which my belief is justified is a reason possessed by the testifier.
 Frederick F. Schmitt 195

Hence, my belief p must be justified either on the basis of a reason I possess or

one possessed by the testifier. This conclusion is just (T1). Once (G) is granted,
the case for (T1) is convincing. I will simply assume (G) without argument for
the time being. In Section 2, however, I will argue for (G) restricted to testimoni-
ally justified beliefs. I will do so on the ground that in the case of a testimonially
justified belief, we have no plausible alternative to basing on a good reason as that
in virtue of which the belief is justified.
The second component of the Transindividual Thesis, (T2), is supported in
this way. My testimonially justified belief p cannot always be justified on the
basis of a justified belief that the testifier has a good reason to believe p (alternat-
ively, that the testifier is reliable on the topic of p). For my belief that the testifier
has a good reason to believe p is itself ultimately justified on the basis of testi-
mony—on the basis, say, of my testimonially justified belief q. As it turns out,
this testimonially justified belief q is not in turn nontestimonially justified on
the basis of any good reason I possess—obviously not on the basis of the reason
that the testifier has a good reason to believe p, on pain of circularity, but also
on the basis of no other good reason I possess. Hence, I can be justified in my
testimonial belief even though I do not possess a good reason to believe p. This
is (T2).
While I regard this case for (T2) and thus the overall case for the Transin-
dividual Thesis as promising, I do not claim to be able to develop the case for
(T2) in full. Instead, I will offer a conditional case for the Transindividual Thes-
is, and even that only tempered and in outline. The Thesis is analogous to an
intra-individual basing thesis for memorially justified belief, which I will call the
Transtemporal Thesis:
(M1) If my belief p is justified on (the basis of) memory, then it is justified
on the basis of my original good reasons to believe p, unless on the
basis of a good reason I currently possess to believe p.⁸
(M2) I can be justified in believing p on memory, even though I am not jus-
tified in believing it on the basis of a good reason I currently possess
to believe p.
For example, if my original reason to believe that there are strawberries in the
refrigerator was no good—the product of a hasty look in the refrigerator this
morning—then, assuming I base my belief on no current good reason, my belief
is not now justified. The Transtemporal Thesis is, I take it, intuitively very plaus-
ible and widely accepted, even if not wholly uncontroversial.⁹ My aim in this
paper is to show, albeit in outline, that there is a case for the Trans-
individual Thesis for testimonial justification (in particular (T2) ) analogous to
the obvious and, I assume, widely accepted case for the Transtemporal Thesis for
memorial justification (M2). I will not attempt to argue here that the case for
the Transtemporal Thesis really supports that thesis. My claim is merely condi-
tional and tempered: the case for the Transindividual Thesis about testimonially
196 Transindividual Reasons

justified belief is nearly as strong as that for the Transtemporal Thesis about
memorially justified belief. In other words, if you accept the intuitively plausible
Transtemporal Thesis on the basis of the obvious case for it, then you ought to
find probative the analogous case for the intuitively implausible Transindividu-
al Thesis. (I note at the outset that what I say defends a fortiori the intuitively
more plausible weaker social claim with which we started, that my belief p is
justified on testimony only if either I or the testifier possesses a good reason to
believe p, and I can have a belief p justified on testimony even though I pos-
sess no good reason to believe p. So if you, quite forgiveably, have a special
grudge against the stronger and more counterintuitive Transindividual Thesis
about basing, remember that the discussion can easily be translated into a defense
of the weaker, somewhat more palatable claim about possessing reasons.)

1 THE CONDITIONAL CASE FOR THE

T R A N S I N D I V I D UA L T H E S I S

The case for the Transtemporal Thesis is analogous to the nutshell case I men-
tioned for the Transindividual Thesis. We assume (G), that my justified belief
p must be based on a good reason. The case for component (M1) of the Trans-
temporal Thesis exactly parallels the case already made for (T1) of the Transindi-
vidual Thesis. I take the case for component (M1) to be convincing, once (G) is
granted.
The case for component (M2), that my belief can be memorially justified even
though not justified on the basis of a current reason, is this. It is claimed that
there are beliefs intuitively justified on memory for which there is no available
candidate for a good current reason on which the beliefs are based. There are two
sorts of candidates for a good current reason that need to be ruled out here: a cur-
rent reductive reason and a current nonreductive reason.¹⁰ The ground for (M2),
then, is that there are instances of memorially justified belief for which there are
no good current reductive and no good current nonreductive reasons. Although
many of the points to be made in favor of (M2) will be familiar from the literat-
ure on memorial justification, the reader must indulge my going through them to
bring out a favorable comparison with the case for the testimonial analogue (T2).
Reductive Reasons. Any memorially justified belief p is justified on the basis of
a good current reductive reason only in virtue of being justified on the basis of a
deductive, inductive, or other justifying inference from that reason. Such a reas-
on is a belief and must itself be justified in order to provide justification for my
memorially justified belief p. The relevant feature of any such current reduct-
ive reason, the feature that prevents my memorially justified belief from being
based solely on a current reductive reason, is that such a current reductive reas-
on depends on memory. If my belief is justified on memory in virtue of being
based on a current reductive reason r, it is so justified by inference from r, and r
 Frederick F. Schmitt 197

must therefore itself be a justified belief. But as it happens, any current reductive
reason that could support my memorially justified belief is itself justified on
memory.¹¹ So (assuming that memorial justification depends only on current
reasons) memorial justification on the basis of a current reductive reason depends
ultimately on a current reductive reason memorially justified on the basis of a
current nonreductive reason, on pain of a regress of memorial justification on the
basis of reductive reasons. My current reductive reason r must ultimately be jus-
tified on memory on the basis of a current nonreductive reason (if on any current
reason at all) in order to stem the regress. In this sense, reductive reasons are para-
sitic on nonreductive reasons. This means that a current nonreductive reason is
the ultimate basis of memorial justification on the basis of a current reductive
reason, and hence the ultimate basis of any memorial justification on the basis of
current reasons. Thus, assuming that my memorially justified belief can be based
on a current reductive reason only if that reason is itself justified on memory,
the case for the Transtemporal Thesis need only show that there are instances of
memorially justified beliefs for which there is no basis in good current nonreduct-
ive reasons. The question arises whether being ultimately justified on the basis
of a current nonreductive reason does not touch off a further regress of reasons
(whether current reductive or current nonreductive reasons); but current nonre-
ductive reasons, unlike current reductive reasons, need not themselves be justified
on memory. The answer is that no such regress need arise. For current nonre-
ductive reasons need not be beliefs that must themselves be justified in order to
provide justification for my memorially justified belief p. Indeed, it is natural to
think of current nonreductive reasons as states other than beliefs, hence not even
susceptible of justification at all.
But why say that my belief p is memorially justified on the basis of a cur-
rent reductive reason only if the reason is itself justified on memory? In lieu of
answering the question for each of the various candidates for current reductive
reasons, I will answer it for one paradigmatic candidate, which I believe to be
the most promising candidate for a current reductive reason—a statistical (or
track-record) reason:
My belief p is justified on memory on the basis of a statistical syllogism
from an inductively justified generalization that memorial beliefs of a cer-
tain type X (where, e.g., type X is recollection of a vivid experience) tend
to be true, together with an introspectively justified belief that my belief p
is a memorial belief of type X .
No doubt in instances in which I have a memorially justified belief, my belief
often can come to be justified on the basis of a current reductive reason of the
specified sort. But the question is whether, if my belief came to be justified on this
basis, the reason would escape being ultimately justified on memory. The answer
to this question is No. For the premise that memorial beliefs of type X tend to be
true is inductively justified. But my belief in this premise is inductively justified
198 Transindividual Reasons

on the basis of justified beliefs that given memory beliefs of type X are true, for
sufficiently many and varied memorial beliefs. Yet in general, my belief that a
particular memorial belief of type X is true is justified on the basis of that and
other memorial beliefs, hence on the basis of memory. So any memorial belief
justified on the basis of the statistical reason must ultimately be justified on the
basis of memorial beliefs. Hence, my belief p is memorially justified on the basis
of a current reductive reason of the statistical sort only if the reason is itself jus-
tified on memory. This means that current reductive statistical reasons do not
terminate a regress of memorial justification. Thus, to avoid a regress of statist-
ical reasons, current reductive statistical reasons must be justified on the basis
ultimately of current nonstatistical reasons. Indeed, these current nonstatistical
reasons must be current nonreductive reasons if other current reductive reasons
suffer the same problem as current statistical reductive reasons. And though I
cannot make the case here, I believe that similar arguments apply to other can-
didates for current reductive reasons. If so, a memorial belief justified on the basis
of a current reductive reason is ultimately justified on memory. Hence, a current
reductive reason does not terminate a regress of memorial justification. A belief
memorially justified on the basis of a current reductive reason must therefore be
justified ultimately on the basis of a current nonreductive reason, if it is to be jus-
tified solely on the basis of current reasons. Hence, the question whether (M2)
holds comes to whether some memorially justified beliefs lack an ultimate basis in
good current nonreductive reasons. We will address that question momentarily.
In the meantime, let us turn to testimonially justified belief and ask whether a
reductive reason I possess for such a belief must also ultimately be justified on the
basis of testimony, hence ultimately justified on nonreductive reasons I possess
if justified solely on the basis of reasons I possess. Let us consider a statistical
basis for testimonially justified belief parallel to the statistical basis for memorially
justified belief we just considered. In the case of a statistical basis,
my belief p is justified on testimony on the basis of a statistical syllo-
gism from an inductively justified generalization that testimonial beliefs
of a certain type X (where, e.g., type X is belief on the testimony of a
socially recognized expert in the field to which p belongs) tend to be true,
and an introspectively or perceptually justified belief that my belief p is of
type X .¹²
I am inductively justified in believing this generalization on the basis of justified
beliefs that given testimonial beliefs of type X are true, for sufficiently many and
varied testimonial beliefs.
Here a problem analogous to the problem for memorially justified belief arises.
In general, I am justified in believing that a particular testimonial belief of type
X is true only on the basis (in part) of that or other testimonial beliefs. There are
two grounds for saying that in general I am justified in believing that a particu-
lar testimonial belief is true only on the basis of that testimonial belief or some
 Frederick F. Schmitt 199

others. One ground is that although I may be able to verify firsthand some of
the relevant testimonial beliefs, I am not able to do so for sufficiently many and
varied testimonial beliefs to afford a basis for induction.¹³ A different ground for
the claim is that in general perceptual beliefs on the basis of which I verify my
testimonial beliefs are themselves justified in part on the basis of prior testimoni-
ally justified beliefs. Although no one has decisively established these points (nor
is it even clear how the issue could be decisively settled), most of those writing
recently on the viability of a statistical reduction for testimonial justification have
endorsed one or both of these points.
If either point is accepted, then a statistical reason for my testimonially justified
belief is itself ultimately justified on testimony. Hence, my belief p is testimoni-
ally justified on the basis of a reductive reason of the statistical sort that I possess
only if the reason is itself justified on testimony. Thus, to avoid a regress of stat-
istical reasons I possess, reductive statistical reasons I possess must be justified
on the basis of nonstatistical reasons I possess. As in the case of memorial justi-
fication, these nonstatistical reasons must be nonreductive reasons I possess. For
there is a parallel regress for all the candidates for reductive reasons.¹⁴ We thus
reach the conclusion that a testimonial belief justified on the basis of a reduct-
ive reason I possess is ultimately justified on the basis of a nonreductive reason I
possess. Admittedly, the case here that a testimonially justified belief is ultimately
justified on the basis of a nonreductive reason I possess (if solely on the basis of
reasons I possess) is not as strong as the parallel case that a memorially justified
belief is ultimately justified on the basis of a current nonreductive reason (if solely
on the basis of current reasons), but it does seem presumptive at the present time.
Nonreductive Reasons. We have seen that testimony and memory are in a sim-
ilar boat regarding reductive reasons: for each of them there is a case that justific-
ation on the basis of reductive reasons is parasitic on justification on the basis of
nonreductive reasons. There is a case that current reductive reasons are justified
on the basis of memory and hence on current nonreductive reasons, and there
is a parallel case for testimony. We may now proceed to the question whether
there are memorially justified beliefs not justified ultimately on the basis of good
current nonreductive reasons. There is a case in favor of an affirmative answer. If
the answer is affirmative, then there are memorially justified beliefs not justified
ultimately on the basis of any current reasons at all—thus, (M2).
There are two sorts of candidates for current nonreductive reasons (or justifica-
tion) that might be taken to justify a belief memorially in a manner that does not
reduce to perceptual, inductive, or other inferential justification—coherentist
justifications and foundationalist reasons.¹⁵ The case for (M2) must show that
there are memorially justified beliefs for which both coherentist and foundation-
alist ultimate justifications are missing.
Regarding a coherentist ultimate justification, plausibly there are many
instances of memorially justified beliefs that are not ultimately justified by
coherence with other current beliefs. I will assume here that justification by
200 Transindividual Reasons

coherence requires that the belief p bears a relation of coherence to some set of
relevant beliefs or to each member of some set of relevant beliefs. Similar remarks
apply, I think, on the view that justification by coherence requires membership in
a set of beliefs each of which coheres with other members of the set.
I remember isolated incidents—e.g., that I sampled octopus at age five. I may
be justified in this belief even though it coheres only very weakly with any set
of current beliefs of mine that are relevant to it. I may be justified in this belief
even though I believe only such things as that I heard the word ‘‘octopus’’ around
this time, that I ate seafood in my youth, that my memory is generally reliable,
and that my memory tells me that I sampled octopus at age five. The relation of
coherence my belief that I sampled octopus at age five bears to any set of such rel-
evant beliefs is far too weak to make my belief justified. This is an intuitive point,
but it may be reinforced with this argument. The following principle would seem
to govern the relation between coherence and justification if coherence is ever a
sufficient condition of justification:

If my belief p coheres sufficiently with a set of relevant beliefs to be justi-

fied in virtue of this coherence, then in a case in which I believe not-p but
maintain the beliefs in this set altered only to accommodate differences in
beliefs as to what my memory tells me and beliefs that follow from such
beliefs, and I hold no beliefs outside the set that would undermine justific-
ation for believing p, my belief not-p would not sufficiently cohere with the
altered set of beliefs to be justified in virtue of its coherence with the set.

This principle of coherence bears on examples of the present sort. Instead of a

case in which I remember that I sampled octopus at age five, imagine a case as
similar to this with respect to my beliefs as possible, but in which what I remem-
ber is the negation of the proposition I remember in the original case. In this
altered case, I remember that I did not sample octopus at age five. I have the
beliefs in the specified set for the belief p, corrected for what memory tells me.
I also have no beliefs outside this set that would undermine my justification for
believing p. It seems that my belief not-p coheres about as well with the altered set
of relevant beliefs as my belief p coheres with the original set of relevant beliefs.
So the coherence of my belief not-p should be sufficient for the justification of
that belief, on a coherence theory of justification, if the coherence of my belief
p is sufficient for the justification of that belief. Thus, the principle of coherence
entails that the coherence of my belief p is not sufficient for the justification of
that belief. A coherence theory is inconsistent with the justification of my belief p,
contrary to intuition. This conclusion reinforces the point that there are memori-
ally justified beliefs that, if based on current reasons at all, are not ultimately
justified on the basis of current reasons in virtue of coherence with those reasons.
There is a parallel case, about as strong as the memorial case, that some testi-
monially justified beliefs, if justified on the basis of reasons I possess, are not
 Frederick F. Schmitt 201

ultimately justified in virtue of coherence with those reasons. For I may be jus-
tified in believing that there are mussels in New Zealand on someone’s say-so,
even though my belief coheres only very weakly with any set of relevant beliefs
of mine. I may be justified in my testimonial belief even though I know nothing
that is relevant to this belief, other than such uninformative generalities as that
the testifier is an adult human being, that adult human beings generally speak
the truth, that New Zealand has a seashore, that people who may have com-
municated with my testifier have visited New Zealand, and that mussels live in
the sea. The relation of coherence that my belief that there are mussels in New
Zealand bears to these relevant other beliefs is far too weak to make my testimo-
nial belief justified. The plausible coherence principle of justification mentioned
above entails this same conclusion. These points are enough to cast doubt on a
coherentist ultimate justification of testimonial beliefs.
Turning now to a foundationalist ultimate justification of memory beliefs, the
question is whether all my memory beliefs not ultimately justified by coherence
are ultimately justified on the basis of these current states, or features of states:
my seeming to remember that p
my feeling that I remember that p
my believing that I originally had a good reason to believe p
the strength of my memory belief p.
(My believing that I originally had a good reason to believe p is understood here
as supplying a noninferential basis for my memorial belief in the way that the
states of seeming and feeling that I remember are supposed to provide such a
basis.) There is a case for a negative answer to our question: a memorially justified
belief need not be ultimately justified by coherence or on the basis of any of these
reasons.¹⁶ This is enough to support (M2).
To see why a negative answer is in order, observe, with regard to the first three
candidates on this list, that these states are implausible candidates for a basis on
which all my memorially justified beliefs are justified. Suppose someone asks
me, ‘‘How many bedrooms does your house have?’’ I reply without hesitation,
‘‘Five.’’ I remember, and am justified on the basis of memory in believing, that
my house has five bedrooms. In the typical case, I do not experience a state of
seeming to remember or feeling that I remember, nor do I form a belief, memori-
al or otherwise, that I had good reason to believe p. Certainly, in some cases, I am
aware that I remember p. But there is no mental state of awareness on which I
could base my belief p. To say that I am aware that I remember p is just to say that
I know that I remember p, and I can readily access this knowledge if prompted
to do so.
Now, in some such cases, I would develop a feeling that I remember p if asked
the question, ‘‘Are you sure you have remembered p?’’ But such a feeling is sub-
sequent to my memorially justified belief, and thus my belief cannot be initially
justified on the basis of the feeling. Moreover, the states of seeming to remember
202 Transindividual Reasons

and of feeling that I remember and the second-order belief typically do not arise
unless I consider whether I remember. If I do not consider whether I remember,
then typically I will have no such states. These states seem to serve a function
when a doubt arises as to whether p, and they help to settle that doubt in favor of
p by confirming p in virtue of justifying a belief that I remember p. Note too that
there seem to be cases in which I do remember p, but when I consider whether I
remember p, I do not seem to remember p, and so I doubt that I do remember.
In these cases, initially at least, I am justified in believing p in virtue of remember-
ing, even though I never seem to remember p. I never develop a state of seeming
to remember p, or a feeling that I remember p, or a second-order belief that I
originally had a good reason to believe p. Colin Radford’s example of my cor-
rectly answering a quiz show question about the year of Queen Elizabeth I’s death
is plausibly a case of this sort. Of course, the points so far show only that there
are memorially justified beliefs that lack a foundationalist ultimate justification
of one of the first three kinds. What we seek is a memorially justified belief that
lacks both a coherentist and a foundationalist ultimate justification. Plausibly,
my memorially justified belief that I sampled octopus at age five lacks a coherent-
ist ultimate justification, and it may lack a foundationalist ultimate justification
of the first three kinds.
The fourth candidate for a current nonreductive reason, the strength of my
memorial belief, avoids the charge that it does not usually accompany the
memorial belief. Plausibly, all memorial beliefs have at least a vague degree of
strength. Moreover, the strength of a memorial belief is simultaneous with the
belief, so that, as far as timing goes, a memorial belief could be based on its
strength after its inception, though not at the moment of its inception, since
basing is plausibly a causal relation that takes time and thus requires that the
state on which the belief is based obtains before the first moment of basing. Even
so, there is something peculiar about saying that a memorial belief is justified only
after the moment of its inception, not at that moment. More importantly, it is
implausible, if it is even intelligible, that my belief is justified on the basis of its
own strength, even if we refer here to its strength immediately prior to the time
of justification. Presumably, in the simplest case, barring overdetermination, my
belief is justified on the basis of something, x, only if, were x not to obtain, my
belief would not obtain. This means that if my belief is justified on the basis of its
own strength, it follows that were my belief not to have its actual strength prior to
a time t, it would no longer obtain at t. But there is no plausibility in the idea that
my belief would cease after inception if its strength at inception were different
from what it actually is (were less than required for justification). So there is no
plausibility in the proposal that my memorially justified belief is justified on the
basis of its own strength. Thus, there do seem to be cases of memorially justified
beliefs that lack both a coherence ultimate justification and any foundationalist
ultimate justification. These memorially justified beliefs are not justified on the
basis of any current reasons. This makes a case for (M2).
 Frederick F. Schmitt 203

Turning now to a foundationalist ultimate justification of testimonial belief,

there are several candidates for nonreductive reasons I possess on which testimo-
nially justified beliefs might be based:

my seeming to believe p on testimony

my believing that the testifier has good reason to believe p
the strength of my testimonial belief p
my seeming to understand the testimony that p.

The question of course is whether all testimonial beliefs not ultimately justified
by coherence are justified ultimately on the basis of one or another of these states.
The case for a negative answer seems nearly as good as the case for a negative
answer to the analogous question for memory. I will treat the first three candid-
ates in this section and the fourth candidate in the next section.
Regarding the first two candidates, they do not always or even usually accom-
pany testimonially justified beliefs. We may restrict our attention to the presence
of these states at the moment I form my belief on testimony, since, on the view
we are challenging here, testimonial justification for my belief later on may be
assumed to depend only on my having had one of these states at some time (at the
moment of forming my belief on testimony) and not on my having one of them
at each moment my belief is testimonially justified. In many cases of testimoni-
ally justified belief, I do know that I believe on testimony at the moment I form
my belief. But my testimonially justified belief p is not always accompanied by
my seeming to believe on testimony or by my believing that the testifier has good
reason to believe p. I pick up things that people say without attending to who
is saying them and even without realizing that anyone has said anything or that
I have picked something up by testimony. If, as a result of being asked whether
I believe p on testimony, I consider the matter shortly after I have formed the
belief, in many cases I will seem to believe p on testimony, and I will believe that
the testifier has good reason to believe p. But this is not necessary for my belief to
be testimonially justified. It is enough if I have picked up the belief in the sorts of
circumstances I often do, however unreflective I may be about the uptake. A view
worthy of serious consideration is that my uptake must be sensitive to certain
considerations: my belief p is not testimonially justified unless I would not have
formed the belief p were I to have reason to believe the testifier is untrustworthy
on the topic of p. But even if this is so, I need not actually have the thought that
the testifier is trustworthy on p.
In addition to this point, there are other points against the first two candid-
ates. Regarding the first candidate, it is not entirely clear that there is any state of
my seeming to believe on testimony, and so not clear that there is any candidate
state of this sort on the basis of which I might believe p. There is no doubt that
it can seem to me that I believe on testimony, but it does not follow from this
that there is any phenomenal or other mental state of my seeming to believe on
204 Transindividual Reasons

testimony. Again, there are phenomenal states that indicate that I believe on testi-
mony, such as a sense that I would answer ‘‘Yes’’ if asked whether I believe that p
on testimony. But these phenomenal states need not be present when my belief p
is justified on testimony. Regarding the second candidate, it seems doubtful that
my mere belief that my testifier has a good reason to believe p would be enough
to make my belief p justified (in the absence of a coherence justification for my
belief p). To be sure, my belief p might be justified on the basis of a justified belief
that my testifier has a good reason to believe p. But that is not the proposal in
question here. The proposal is that my mere belief that my testifier has a good
reason to believe p is enough to make my belief p justified, whether or not I am
justified in believing that my testifier has a good reason to believe p. This seems
doubtful.
The third candidate, the strength of my testimonial belief, falls to a criticism
exactly like the one we have already brought against the analogous candidate for
memorially justified belief.
These reflections lead to the conclusion that testimonially justified belief need
not be based on any of these candidate states. The same points apply to a case
like my testimonial belief that there are mussels in New Zealand. This, then, is
a testimonially justified belief lacking both a coherentist and a foundationalist
ultimate justification. Such an example is enough to support (T2). The case for
(T2) seems nearly as strong as that for (M2).

2 BU RG E O N T E S T I M O N I A L E N T I T L E M E N T

In making the case for (T1), I assumed what I take to be a plausible condition on
justified belief, (G):
Any justified belief p of mine is justified on the basis of a good reason to
believe p suitably related to me.
To be more exact, the case for (T1) depends on (G) restricted to testimonially
justified belief. I will not offer a defense of an unrestricted (G) in this paper.
But I do need to defend (G) restricted to testimonially justified belief, since the
case for the Transindividual Thesis depends on it. A conditional defense would
be enough for my purposes: (G) restricted to testimonially justified belief is as
well supported as (G) restricted to memorially justified belief. There is work on
testimonial justification that might be taken to call this into question. In partic-
ular, Burge (1993, 1997) has proposed a view incompatible with (G) restricted
to testimonial justification and with (T1) as well. (Burge at one point endorses
something like (G) restricted to memorial justification, but he also endorses the
contradictory view, as Edwards (2000) notes. I will say only a word about Burge’s
view of memorial justification below and focus rather on his view of testimonial
justification.¹⁷)
 Frederick F. Schmitt 205

(T1), recall, says:

If my belief p is justified on testimony, then it is justified on the basis of the
testifier’s good reason to believe p, unless on the basis of a good reason to
believe p I myself possess.
By contrast, Burge holds the view that my belief can be testimonially justified
though not based on any good reason. This view is incompatible with (T1).¹⁸
The defense of the restricted (G) and of (T1) thus requires undermining Burge’s
view. This defense is relevant not only to the restricted (G) and (T1) but also to
(T2). For undermining Burge’s view also undermines the fourth candidate above
for an ultimate basis of testimonial justification, my seeming to understand the
testimony that p.
Burge expresses his view in this sentence:
in interlocution . . . we have a general a priori prima facie (pro tanto) entitlement to
believe putative assertions that we seem to understand. (1997: 21)
I note before proceeding that Burge distinguishes entitlement from justification:
justification for a belief must be grasped by the subject, while an entitlement need
not be. For present purposes, I assimilate entitlement to justification. I will not
directly address Burge’s claim that my entitlement to believe on testimony is a
priori. (On the Transindividual Thesis this claim is in one straightforward sense
mistaken: our testimonial justification has whatever status—a priori or a posteri-
ori—is possessed by the testifier’s justification for believing p.)
I will focus on this principle, to which I take Burge to be committed:
(U) If p is putatively asserted and I seem to understand the putative asser-
tion that p, then I am prima facie entitled to believe p.
Principle (U) ascribes my testimonial entitlement to believe p to p’s being
putatively asserted and my seeming to understand this putative assertion. Prin-
ciple (U) does not assume that p’s being putatively asserted or my seeming to
understand the assertion is a state on the basis of which I might be entitled to my
testimonial belief.
I observe that we cannot clearly object to principle (U) in a way parallel to the
way we objected earlier to the view that testimonially justified beliefs are always
justified on the basis of my seeming to believe on testimony, by pointing out
that my belief p can be justified on testimony even though I do not realize that I
believe p on testimony. If what entitles me to believe p is my seeming to under-
stand a putative assertion that p, this apparently does not require that I realize
that I understand p as a result of uptake from testimony or from putative asser-
tion. So the fact that my belief can be entitled on testimony without my realizing
that I understand p as a result of uptake from testimony does not preclude satisfy-
ing the antecedent of principle (U).
The objection to principle (U) must rather be that p’s being putatively asserted
and my seeming to understand the assertion is too slim a basis for testimonial
206 Transindividual Reasons

entitlement. This objection seems to me intuitive and compelling. It is very im-

plausible to suppose that the mere fact that p is putatively asserted and I seem to
understand the assertion, is enough to make me entitled to believe p (on the basis
of testimony). The implausibility is great enough that I feel no need to press a
case against principle (U). We may instead ask why Burge would think that the
mere fact that I understand what someone has said entitles me to believe it. It will
suffice to criticize Burge’s argument for principle (U).¹⁹
I take memorial and testimonial entitlement to be in the same boat here. The
memorial analogue of (U) has no intuitive plausibility, and so the question is
whether Burge can make a case for this analogue. As it happens, Burge’s argu-
ment for the memorial analogue of (U) fares better on principle (b1) discussed
below than his argument for (U) does. This means that my defense of (T1)
against Burge’s account of testimonial entitlement is stronger than the analogous
defense of (M1) from Burge’s account of memorial entitlement. This difference
favors a conditional defense of the Transindividual Thesis. So I need only treat
the testimonial case here.
Burge’s exposition of his argument for principle (U) (1993: 469) requires some
interpretation. Lacking room for exegesis, I offer the following reading without
further apology:
(a) If a presentation that p is a seemingly intelligible presentation as true,
then it is prima facie a presentation of a rational source;
(b) If a presentation that p is prima facie a presentation of a rational source,
then it is prima facie a presentation of a source of truth (i.e., a source
that tends to present truths);
(c) If a presentation that p is prima facie a presentation of a source of truth,
then I am prima facie entitled to believe p.
Conclusion: If a presentation that p is a seemingly intelligible presentation
as true, then I am prima facie entitled to believe p.
The Conclusion of this argument is equivalent to principle (U) on the assump-
tion that a presentation that p is a seemingly intelligible presentation as true if and
only if p is putatively asserted and I seem to understand the assertion.
The weakest premise in the argument is premise (b).²⁰ I propose to evalu-
ate Burge’s support for (b) and consider the plausibility of (b) in the course of
doing so.
Burge supports premise (b) on the ground that a rational source has a veritistic
function—it functions to produce true presentations.²¹ Evidently Burge infers
from this ground that
(b1) If a presentation that p is prima facie a presentation of a rational
source, it is prima facie a presentation of a source having a veritistic
function (i.e., the function of producing true presentations).
 Frederick F. Schmitt 207

He then assumes:
(b2) If a presentation that p is prima facie a presentation of a source having
a veritistic function, it is prima facie a presentation of a source of truth.
These two assumptions, (b1) and (b2), together yield (b). This, on my reading,
is Burge’s support for (b). There are, however, two difficulties with this appeal to
the function of producing true presentations.
First, (b1) is implausible. We can grant that rational sources can and some-
times do have the function of producing true presentations. But rational sources
do not necessarily have the function of producing true presentations. It is plaus-
ible that rational belief-formation (or other cognition) has the function of pro-
ducing true beliefs. Thus, (b1) is plausible for memory conceived as a rational
source of memorial beliefs. But what Burge needs to support (b) for testimonial
entitlement in such a way that it can be combined with a plausible (c) to yield
the desired conclusion, is not this innocuous claim, but rather the tendentious
claim that rational interlocution has the function of producing true presentations
of the source. It is far from obvious that rational interlocution has the function
of producing true presentations. What is plausible is that rational interlocution
has the function of furthering the welfare and the goals of the interlocutors (as
Faulkner (2000) has emphasized against Burge). Sometimes truth in presentation
is a means to this end. But it does not follow that rational interlocution always or
ever has the function of producing true presentations. It merely follows that truth
in presentation is sometimes a means to fulfilling the function of rational inter-
locution. This undermines Burge’s support for (b). It also reveals that (b) itself is
questionable.
A second difficulty with Burge’s support for (b) is that (b2) is dubious. Even
if a rational source has a veritistic function (has the function of producing true
presentations), it clearly does not follow that it is a source of truth—that it really
does tend to present truths. The source might have this function but perform it
poorly. So why should it follow from the fact that a presentation is prima facie a
presentation of a rational source having a veritistic function, that it is prima facie
a presentation of a source of truth? Apparently Burge assumes that if one has a
prima facie reason to think that a source has a veritistic function, this gives one
prima facie reason to think that the source fulfills its function.
I concede that if one had a prima facie reason to think that generally things
that have a function fulfill the function, and one had no reason to think that
a prima facie presentation of a rational source having a veritistic function fails
to fulfill its function, then Burge’s assumption would hold. But first, we have
no a priori prima facie reason to think that generally things having a function
fulfill them. It seems conceivable that things do not generally fulfill the functions
they have. This seems conceivable for artifacts that have assigned functions. We
could assign functions to clocks and other devices we design, but many or most
of these devices could fail to fulfill the functions we assign them either through
208 Transindividual Reasons

malfunction that results from unexpected interference from the environment or

through inept design. (Admittedly our tools for building the devices would have
to function well enough and often enough to enable us to build the devices, but
this does not entail that generally artifacts fulfill the functions they have.) But it
also seems conceivable that biological organisms, organelles, and organs could fail
generally to fulfill the biological functions they have. This is clearly conceivable
on an etiological account of biological function, on which something x has the
function of doing A when x exists only because ancestors of x did A. Such an
etiological account requires that at one time ancestors of x did A, but it does not
entail that x does A or even that in doing A ancestors of x were fulfilling the
function of doing A. So it does not entail that generally biological entities fulfill
the functions they have. It is also clearly conceivable that natural things could fail
generally to fulfill the functions they have on the analogical account according to
which something x has the function of doing A when it is as if x were designed
to do A. For it may be as if x is designed to do A even though it is also as if x is
ineptly designed to do A, or even though it is also as if x fails to do A as a result
of environmental interference. In short, it seems conceivable that things do not
generally fulfill functions they have. If so, I see no a priori prima facie reason to
think that generally things that have a function fulfill the function. So a prima
facie reason to think that generally things that have a function fulfill the function
cannot give Burge the a priori entitlement for testimonial belief he seeks.
A second point is that it is doubtful that there is even an a posteriori prima
facie reason to think that generally things that have a function fulfill them. The
biological domain is full of examples of organisms and organs that have functions
they generally fail to perform. Antibodies have the function of killing viruses,
but most never get the chance. Sperm have the function of fertilizing eggs, but
most don’t. It is difficult or impossible to figure whether in the majority of cases,
organisms, organs, and organelles fulfill the functions they have. This is difficult
if only because we have no ready way to individuate organisms and functions.
Without such an individuation, or perhaps even with one, there is no good reas-
on to expect that generally biological entities fulfill the functions they have. Burge
would have to claim that we have a prima facie reason to believe that cognitive
functions are superior to biological ones in how generally they fulfill the func-
tions they have, but I can see no warrant for such a claim. We are left with no
secure a priori or a posteriori prima facie reason to think that a source having a
veritistic function fulfills its function. So we have no reason to think (b2) holds.
And without support for either (b1) or (b2), Burge has no support for premise
(b), hence none for principle (U).
The upshot is that we have no support for Burge’s view that I am prima facie
entitled to believe p if p is putatively asserted and I seem to understand the asser-
tion. So we have no support for the view that I can be justified in a belief on
testimony without being justified on the basis of a good reason to believe p. Nor
do we have support for the view that my seeming to understand the testimony
 Frederick F. Schmitt 209

that p can supply a good reason in virtue of which I am justified in believing p on

testimony.
As we saw in Section 1, we also have no support for an account of testimo-
nial justification that appeals only to reasons possessed by me on the basis of
which my testimonial belief is justified—indeed, we have reason to reject such
an account. The case against reasons of mine, both reductive and nonreductive,
as a basis for testimonially justified belief seems nearly as strong as the case against
current reasons, both reductive and nonreductive, as a basis for memorially justi-
fied belief. And there is no case for the intuitively unappealing claim that seeming
understanding makes me prima facie justified in my testimonial beliefs. In the
case of memory, we are led back to the thought that in some instances of memori-
ally justified belief, I am justified in believing p on the basis of my original good
reason to believe p. I am so justified whenever I am not justified on the basis of
a current good reason to believe p, as often happens. The analogous thought for
testimony is nearly as well supported: in some instances of testimonially justified
belief, I am justified on the basis of the testifier’s good reason to believe p. I am
so justified whenever I am not justified on the basis of my own good reason to
believe p. This is the Transindividual Thesis.

3 ARE THERE ASYMMETRIES BETWEEN TESTIMONIAL

A N D M E M O R I A L J U S T I F I C AT I O N T H AT D I S FAVO R
T R A N S I N D I V I D UA L B A S I N G ?

Are there any differences between testimonial justification and memorial justific-
ation that undermine the conditional case for the Transindividual Thesis given
the Transtemporal Thesis, or that make the Transindividual Thesis less plausible
than the Transtemporal Thesis? In this last section of the paper, I examine three
suggestions of such differences and find them wanting.
(1) Sensitivity to Reasons in the Basing Relation. There is some plausibility to
the suggestion that basing requires a sensitivity to reasons for belief. One such
sensitivity requirement defines sensitivity as counterfactual dependency. In the case
of intra-individual basing, the counterfactual dependency sensitivity requirement
is this:
My belief p is based on reason r only if it is sensitive in the sense that I
would not believe p if I did not possess reason r.
One might generalize this sensitivity requirement to cover not only intra-
individual basing but also transindividual basing:
My belief p is based on reason r only if I would not believe p if reason r did
not obtain (or exist).
One might worry that this generalized counterfactual dependency sensitivity
requirement poses a problem for the Transindividual Thesis. Perhaps it does, but
210 Transindividual Reasons

it poses no problem for the Transindividual Thesis that it does not also pose for
the Transtemporal Thesis.
To begin, note that, for purposes of defending the Transindividual Thesis,
I need claim only that testimonially justified beliefs are based on the testifier’s
reason for believing p, and only that they are so based when the subject possesses
no good reason to believe p. I need not claim that unjustified testimonial beliefs
can be based on the testifier’s reason to believe p, or that testimonially justified
beliefs for which the subject possesses good reason to believe p can be based on
the testifier’s reason. Note next that in the typical case, testimonially justified
beliefs are counterfactually sensitive to reasons possessed by the testifier. I am
testimonially justified in believing that Napoleon was exiled to St Helena. And I
would not believe that Napoleon was exiled to St Helena if my sixth grade teacher
had not had her reason to believe the proposition. In the nearest world in which
my teacher lacks her reason, she also fails to testify that Napoleon was exiled to
St Helena—in which case I fail to believe the proposition. The sensitivity of the
testimony to the testifier’s reason ensures that I am sensitive to the testifier’s reason.
Admittedly, there are also many cases of testimonially justified belief in which
the counterfactual sensitivity fails—I would believe p on a certain testimony even
if the testifier lacked her reason to believe p, because the testifier would testify that
p even if she lacked this reason. She might do so because she has practical motives
for testimony that would cause her to testify that p even if she lacked reason r,
or she might do so because she would believe p for reasons other than reason r if
she lacked that reason. Now, not all of these cases of testimonially justified beliefs
are inconsistent with the counterfactual dependency sensitivity requirement. For
some of them are cases in which I possess a good reason to believe p, and the
Transindividual Thesis is not committed to my testimonial beliefs being based
on the testifier’s reason in these cases. On the Transindividual Thesis, all cases
in which my testimonial belief is justified but the testifier lacks a good reason to
believe p would fall into this irrelevant category. Nevertheless, there are bound
to be some cases that force the proponent of the Transindividual Thesis to reject
a strict version of the generalized counterfactual dependency sensitivity require-
ment, in favor of some subtly qualified version that will no doubt be difficult to
formulate.²²
But memory is in nearly the same boat. In some instances in which I am jus-
tified in believing p on memory, I would still believe p even if I had lacked my
original reason. There are many ways this can happen. Imagine that I believe
that I sampled octopus at age five on the basis of an original sensory experience.
But even if I had lacked my original reason, I would still have believed the pro-
position because I would have done so on the basis of a different experience of
sampling octopus at age five. Or imagine that I would still have believed the pro-
position if I had lacked my experience because I would have seemed to remember
the experience when in fact my belief was really conjecture. These are common
enough cases. Some of these cases are bound to run afoul of the counterfactual
 Frederick F. Schmitt 211

dependency sensitivity requirement. These instances show that the counterfactu-

al dependency sensitivity requirement will need qualification for memorial beliefs
if it is be consistent with the Transtemporal Thesis. It remains to be seen whether
the qualification needed to reconcile the sensitivity requirement with the Trans-
temporal Thesis can motivate whatever qualification is needed to reconcile the
requirement with the Transindividual Thesis. But for now the need for a qual-
ification prevents the counterfactual dependency sensitivity requirement from
posing a clear threat to the Transindividual Thesis.
Despite this, one might hold that a different, probabilistic sensitivity require-
ment poses a greater threat to the Transindividual Thesis:
My belief p is based on reason r only if it is probabilistically sensitive to r,
in the sense that the conditional probability that I do not believe p, given
that r does not obtain, is sufficiently high.
For one might propose that even in cases of testimonially justified belief where
the testifier has a reason r to believe p, and the counterfactual dependency sensit-
ivity requirement is satisfied, the conditional probability that I do not believe p,
given that r does not obtain, is fairly low. One might propose that for any testi-
monial belief for which the testifier has a reason r, the conditional probability
that I do not believe p, given that r does not obtain, is the frequency with which
I do not believe q given that the testimonial reason s fails to obtain, over all cases
of testimonial belief q for which the testifier has reason s. But it happens often
enough in cases of testimonial belief that the testifier has motives against speaking
the truth and testifies that p even when he lacks a reason to believe p. Testifiers
often enough have motives for lying and act on them, testifying to propositions
for which they have no reason. In most of these cases, I believe p even though
the testifier lacks a reason to believe p. If these cases figure in the frequency with
which the conditional probability for any given belief is identified, then this fre-
quency is fairly low. So for any given case of testimonially justified belief p, the
conditional probability that I do not believe p given that the testifier’s reason
fails to obtain is fairly low. Thus, on the probabilistic sensitivity requirement for
basing, my testimonially justified belief cannot be based on the testifier’s reas-
on. All testimonially justified beliefs fail the probabilistic sensitivity requirement.
Thus, testimonially justified beliefs for which I lack a reason fail to be based on
the testifier’s reason, contrary to the Transindividual Thesis.
There is nothing in the case of memorial belief comparable to nonveritist-
ically motivated testimony. So there is nothing about memorial belief that would
comparably reduce the probabilistic sensitivity of memorially justified beliefs to
reasons. This suggests that, if conditional probability is defined as the frequency
specified above, the degree of sensitivity of my testimonially justified belief is sig-
nificantly lower than the degree of sensitivity of my memorially justified belief.²³
If the conditional probability for my testimonial belief is enough lower than the
conditional probability for my memorial belief, then the Transindividual Thesis
212 Transindividual Reasons

will find no protection against the problem posed by the probabilistic sensitivity
requirement by comparing the case of testimony with that of memory.
In defense of the Transindividual Thesis, I grant that there may be a differ-
ence in the specified frequencies for testimonial beliefs and for memorial beliefs.
But I would respond that this fact tells against the Transindividual Thesis only if
the conditional probability for a given testimonial belief required for probabilistic
sensitivity to reasons is identified with the specified frequency. On this identity,
the beliefs on testimony that are nonveritistically motivated reduce the degree
of sensitivity for every testimonial belief, including those that are testimonially
justified on veritistically motivated testimony. But why should the degree of sens-
itivity for each testimonial belief be defined as the frequency over all my testi-
monial beliefs? Why should instances of nonveritistically motivated testimony be
relevant to instances in which testimony is veritistically motivated? On a more
discriminating account of the degree of sensitivity, my testimonial beliefs in cases
of veritistically motivated testimony would exhibit a high degree of sensitivity. I
see no reason to prefer the less discriminating account to the more discriminat-
ing one, and so I see no ground for thinking that testimonially justified beliefs
must fail the probabilistic sensitivity requirement for basing on the testifier’s reas-
ons. There is no clear case against the Transindividual Thesis by appeal to the
probabilitistic sensitivity requirement.
To make an additional point in defense of the Transindividual Thesis, there is
a factor favoring the probabilistic sensitivity of testimonial beliefs that may coun-
tervail any reduction in probabilistic sensitivity from instances of nonveritistically
motivated testimony, even if conditional probability is defined as the specified
frequency. Testifiers generally have strong social incentives to present only truths,
and they consequently have a motive for testifying to p only when they possess a
reason to believe p. There is no comparable social incentive for remembering only
when one has a reason. The social incentives to present truths may compensate
for nonveritistic motivations. So even if we identify the conditional probability
with the frequency over all testimonial beliefs, the conditional probability for a
testimonially justified belief may not be substantially lower than the conditional
probability for a memorially justified belief. The upshot is that there is no clear
warrant for saying that testimonially justified beliefs fail the probabilistic sensit-
ivity requirement while memorially justified beliefs satisfy the requirement. This
point is enough for a conditional defense of the Transindividual Thesis.
(2) The Role of Reasons in Reasoning. One might object to the Transindividual
Thesis on a second ground. This ground appeals to a principle of basing: when I
base my belief p on a reason, my doing so entails a normative constraint on my
reasoning. The constraint is that I ought to reason in light of this reason when I
engage in further reasoning. According to this constraint, I ought to take the reas-
on into account in an appropriate way where it is relevant. In the case of a reason
to believe having a propositional content, reasoning in light of my reason means
reasoning from the assumption that its content is true. Since I am not typically
 Frederick F. Schmitt 213

aware of the content of reasons on which, according to the Transindividual Thes-

is, my belief p is based, I cannot conform to this normative constraint. But if
I cannot conform to the constraint, for a given reason r, then, since ‘‘ought’’
implies ‘‘can,’’ it is not the case that I ought to reason in light of r. But then,
by the principle that basing on r imposes the normative constraint, my belief p
is not based on r. The Transindividual Thesis thus runs afoul of this constraint
on reasons.
Testimony and memory differ here. In many instances of memorial belief, I
can and do recover the original reason on which I based my belief (or a reas-
on with the same content), and I can and do reason in light of the reason. But
in testimonial belief, it would defeat the economy that motivates reliance on
testimony if I were often to acquire the reason on which my testimonial belief
is based. Moreover, I am often unable to acquire the testifier’s reason, either
because I am unable to interrogate the testifier, or because I am unable to under-
stand the testifier’s reason or its bearing on p.²⁴ Thus, in many instances of
memorial belief, I can and do conform to the normative constraint that I ought to
reason in light of my original reason, whereas in most cases of testimonial belief, I
do not do so, and should not do so (on pain of defeating the motivation for testi-
monial belief ), and in many cases I cannot do so. The principle of basing does not
hold for most cases of testimonially justified belief.
The proponent of the Transindividual Thesis may reply to this objection
simply by observing that, while in many instances of memorial belief, I can and
do recover my original reason, I cannot do so in many other instances. These
instances plausibly include instances of memorially justified belief in which I
lack current reasons for belief. I cannot reason in light of my original reason to
believe that I sampled octopus at age five. So the principle of basing does not
hold in many instances of memorial beliefs, including many intuitively justified
instances. Thus, the principle is at odds with the Transtemporal Thesis as well
as the Transindividual Thesis. Moreover, the economy of reliance on memory is
also compromised if we require reasoning in light of original reasons for memory
beliefs. The principle of basing must therefore be rejected if we accept the Trans-
temporal Thesis. Rejecting the principle on this ground is, however, compat-
ible with accepting certain weaker principles of basing. These weaker principles
include the following:
When I base my belief on a reason, then I ought to reason in light of that
reason if:
I am currently in possession of the reason; or
I ought to be aware of having the reason.
A weaker principle specific to testimonial belief is this:
When I base my belief on a testifier T ’s reason, then I ought to reason in
light of that reason if I am aware that my belief is based on T ’s reason, and
I am aware of what T ’s reason is.
214 Transindividual Reasons

If I base my belief that Napoleon was exiled to St Helena on T ’s testimony, and I

am aware that I base my belief on T ’s reason, as well as aware what T ’s reason is,
then I ought to reason in light of T ’s reason. The Transindividual and Transtem-
poral Theses are clearly compatible with these weaker constraints, even though
not with the stronger constraint. This is enough for a conditional defense of the
Transindividual Thesis.
(3) Responsible Belief. The Transindividual Thesis is in fact inconsistent with a
responsibilist conception of justified belief. According to individualistic responsib-
ilism,
I am justified in believing p just in case I act in an epistemically responsible
manner in believing p.
For present purposes, we may say that my acting in an epistemically responsible
manner in believing p is a matter of making my best effort to believe what is
true in believing p, in the specific sense that in believing p I act in a manner that
would reflect my best effort were I concerned solely to believe what is true. (I
leave ‘‘epistemically’’ tacit from here on in.)
Individualistic responsibilism is inconsistent with the Transindividual Thesis
for the following reason. In some instances of belief on testimony, making my
best effort to believe what is true leads me to believe p, even though both the
testifier and I lack a good reason to believe p. I rely on T ’s testimony but despite
appearances T has no good reason to believe p. On individualistic responsibil-
ism, I am justified in believing p. But on the Transindividual Thesis, I am not
so justified, since both the testifier and I lack a good reason to believe p. Thus,
the Transindividual Thesis is inconsistent with individualistic responsibilism.
Individualistic responsibilists will regard this as an adequate reason to reject the
Transindividual Thesis.
One might think that testimony and memory differ here. For in the case of
memorial belief, unlike testimonial belief, I am the subject of the original belief
p. This, one might think, means that the Transtemporal Thesis can be recon-
ciled with a longitudinal version of individualistic responsibilism. It is true that I
can make my best effort to believe p on memory—my best current effort—even
though I had no good original reason to believe p. And this does show that mak-
ing my best current effort does not entail that I have a good original reason. So
the Transtemporal Thesis is inconsistent with current individualistic responsibil-
ism:
I am justified in believing p on memory just in case I currently make my
best effort to believe p on memory.
But, one might think, there is no inconsistency between the Transtemporal Thes-
is and longitudinal individualist responsibilism:
I am justified in believing p on memory just in case I currently make my
best effort to believe p on memory, and I also made my best effort to
believe p in the original episode of belief-formation.
 Frederick F. Schmitt 215

For, one might think, if I made such an effort in the original episode, and not
just currently, I must have had a good reason to believe p. My being justified in
believing p on memory entails that I had a good reason to believe p—enough
for justified belief on memory according to the Transtemporal Thesis. So the
Transtemporal Thesis is, and the Transindividual Thesis is not, consistent with a
suitably formulated individualistic responsibilism.
I think the proponent of the Transindividual Thesis should reply to the objec-
tion by maintaining that even longitudinal individualistic responsibilism is really
inconsistent with the Transtemporal Thesis. So individualistic responsibilism has
no more force against the Transindividual Thesis than against the Transtemporal
Thesis. The point against the consistency of longitudinal individualistic respons-
ibilism and the Transtemporal Thesis is the familiar one that making my best
effort in originally believing p does not entail that I had a good reason to believe p.
I will not make the case here, since it is well enough established elsewhere (Alston
1989; Schmitt 1993).
To see the parallel point regarding testimony, let us consider an attempt
by Owens (2000: chs. 9, 11) to reconcile the Transindividual Thesis with a social
responsibilism for testimonial belief that parallels longitudinal individualistic
responsibilism for memory. On Owens’s proposal, as I read him, a subject’s
acting responsibly in believing p on testimony is not, contrary to individualistic
responsibilism, sufficient for being justified in believing p. But responsibility still
accounts for the epistemic value involved in justified belief. More exactly, Owens
proposes that, although I fail to be justified in a testimonial belief p when the
testifier lacks a good reason to believe p, and although this may be so even when
I act responsibly in believing p, my lack of justification for my belief in such cases
is attributable to a failure of responsibility—not on my part, but on the part of
the testifier. A testimonially justified belief p does not require merely that I act
responsibly in believing p. It requires in addition that the testifier acts responsibly
in testifying to p. Not only must I act responsibly in believing p, but also a
rational source to whom I can transfer responsibility for the truth of my belief
p must act responsibly in testifying to p.²⁵ It can happen that I act responsibly in
believing p, but my belief is unjustified because the testifier, to whom I transfer
responsibility in believing p, does not act responsibly in testifying to p.²⁶
Owens, then, proposes a socialized version of responsibilism for testimony
parallel to longitudinal responsibilism for memory. We may call it social respons-
ibilism:
A belief is justified if every rational agent to whom responsibility for the
belief applies or can pass acts responsibly with regard to the belief.
Acting responsibly with regard to the belief in the case of the subject means act-
ing responsibly in believing p, while acting responsibly with regard to the belief
in the case of the testifier means acting responsibly in testifying. Owens takes
social responsibilism to be compatible with the Transindividual Thesis, just as
216 Transindividual Reasons

longitudinal individualistic responsibilism is said to be compatible with the Tran-

stemporal Thesis.
The difficulty for Owens’s position is that social responsibilism is not in fact
consistent with the Transindividual Thesis.²⁷ Here too we may note that my test-
ifier’s belief may be responsible even though she lacks a good reason to believe p.
It is possible to make a forceful case for the inconsistency between the Transin-
dividual Thesis and social responsibilism by considering features of testimony.
Begin by noting that on the Transindividual Thesis, my testimonially justified
belief is ultimately justified on the basis of some ultimate testifier’s nontestimo-
nially justified beliefs. For example, the justification of my testimonial belief that
Napoleon was exiled to St Helena ultimately traces to the perceptually justified
beliefs of eyewitnesses. This seems inherent in the idea of the Transindividual
Thesis. The Thesis, worked out in full, would allow a testimonially justified belief
to be justified on the basis of a chain of testimonially justified beliefs of testifiers, a
chain initiated by the testimony of an ultimate member possessing a good reason
to believe p. The initial good reason must itself be nontestimonial, on pain of a
regress of testimonial justification.
Now, even the perceptually justified beliefs of eyewitnesses on which my belief
is based are themselves based partly on testimony. An eyewitness knows that these
are British troops in part on their say-so. And a witness’s belief is based on pri-
or testimony about such matters as what a uniform is, what the color red is, etc.
This shows that, on the Transindividual Thesis, each testimonial belief justified
on the basis of the testifier’s reasons is ultimately based on nontestimonial reas-
ons distributed across an entire earlier population of testifiers. In many cases, a
subject’s believing only a proposition for which there are good distributed reas-
ons is certainly not under the subject’s control (or even sensitive to the testifier’s
reasons for belief) in the manner that would seem to be required for blaming
a subject for believing a proposition for which distributed reasons are lacking.
Often enough, subjects cannot avoid picking up testimonial beliefs when good
distributed reasons are lacking. In these cases, they cannot be blamed for believ-
ing as they do despite there not being distributed good reasons for belief. Let us
also note that often enough, testifiers cannot refrain from contributing their share
of the distributed basis for testimony, short of standing permanently mute. They
have no choice but to add their small contribution to the distributed reasons; they
must do so whether or not their contribution contributes to the goodness of the
distributed reasons. Thus, they cannot be blamed for testifying to propositions
as they do, any more than they can be blamed for the reasons they have. So it
often enough happens that I inherit a belief ultimately based on bad distributed
nontestimonial reasons, but neither I nor my sources engage in any blameworthy
acts. People inherit bad systems of classification and explanation for which no
individual, or for that matter community, is blameworthy. Nothing guarantees
that people have the resources to produce well-thought-out systems of these sorts.
No individual or community can be blamed for belief in the demon theory of
 Frederick F. Schmitt 217

disease, though we do not think that there were ever good distributed reasons to
believe the theory. So if a subject is a member of a community encumbered with
belief in that theory and believes on testimony that her niece caught a fever from
possession by demons, there is no one to whom she can transfer blame for her
belief ’s being based on a bad reason. A social responsibilist will have to allow that
her belief is justified, contradicting the Transindividual Thesis.
The upshot of these reflections is that we cannot reconcile social responsibilism
with the Transindividual Thesis. Owens is mistaken in thinking there is any pro-
spect of reconciliation. But this is not a genuine problem for the Transindividual
Thesis. It introduces no asymmetry between testimony and memory. For consid-
erations similar to those regarding testimony apply to memory. I can make my
best effort in my original belief-formation and fail to come up with a good reas-
on to believe p. So in the case of memorial justification, it does not help to move
to longitudinal responsibilism, extending my responsibility to my original belief-
forming episode. That will not ensure that I have a good reason. Longitudinal
responsibilism is thus inconsistent with the Transtemporal Thesis. Testimony
and memory are in the same boat once again.
The foregoing replies seem enough to fend off the threat of an asymmetry
between testimonial and memorial justification and to leave intact my condi-
tional claim that the case for the Transindividual Thesis is nearly as strong as
that for the Transtemporal Thesis.

My conclusion in this paper is that there is a case for the Transindividual Thesis
for testimonial justification nearly as strong as that for the Transtemporal Thesis
for memorial justification. So if we accept the memorial Thesis on the basis of
the latter case, we should also accept the testimonial Thesis with nearly as much
enthusiasm, counterintuitive though it may be.
Two questions remain. One question is why, if there is a case for the Transin-
dividual Thesis nearly as strong as that for the Transtemporal Thesis, we find it
intuitive to say that my memorially justified belief is based on my original reas-
on, but we find it counterintuitive to say that my testimonially justified belief is
based on my testifier’s reason. Perhaps the answer begins by saying that we tend
to think of the basing relation as a psychological relation, and thus we find it
more natural to ascribe the basing relation within a single subject than to ascribe
it across interlocution, which involves a nonpsychological component. But of
course this only trades one question for another: why do we think of the basing
relation as psychological? Since the current state of our thinking about the basing
relation leaves us with only a dim idea of what the basing relation is, we currently
have little satisfactory theory to guide us in trying to answer this question.
The more important question is what implications my conclusion has for the
basing relation. I must leave it at this: the inclining case for the Transindividual
Thesis makes an equally inclining case to characterize the basing relation in such a
way that it can hold between my belief and a reason possessed by my testifier. On
218 Transindividual Reasons

such a characterization, the basing relation is not a pure psychological relation

and in some cases involves transmission of information over a nonmental com-
municative channel, in particular an interlocutionary channel. This point fits
with a characterization of the basing relation that tries to capture its value accord-
ing to the view that the basing relation has value because it affords an alignment
of one’s belief p with the preponderance of existing reasons for and against p.
To the extent that testifiers’ reasons align with the preponderance of existing
reasons, and testifiers testify to propositions that align with the reasons they pos-
sess, and recipients believe the propositions to which testifiers testify, testimonial
beliefs will align with the preponderance of existing reasons. To this extent, then,
the relation between testimonial beliefs and testifiers’ reasons will function to do
what the basing relation does in virtue of which it has value: afford the alignment
of one’s belief p with the preponderance of existing reasons. These remarks may
be enough to vindicate applying the term ‘‘basing’’ to the relation between testi-
monial beliefs and testifier’s reasons, allowing a nonpsychological component in
some basing relations. I note, however, that these remarks suggest that the rela-
tion between testimonial beliefs and testifier’s reasons functions to do what the
basing relation does even when the subject possesses reason to believe p. So they
vindicate applying the term ‘‘basing’’ to this relation even in cases where, on the
Transindividual Thesis, testimonial justification does not require basing on the
testifier’s reason.

REFERENCES
Alston, William (1989), ‘The Deontological Conception of Epistemic Justification’, in
William Alston, Epistemic Justification (Ithaca, NY: Cornell University Press), 115–52.
Audi, Robert (1996), ‘Memorial Justification’, Philosophical Topics, 23: 31–45.
(1997), ‘The Place of Testimony in the Fabric of Knowledge and Justification’,
American Philosophical Quarterly, 34: 405–22.
Burge, Tyler (1993), ‘Content Preservation’, Philosophical Review, 102: 457–88.
(1997), ‘Interlocution, Perception, and Memory’, Philosophical Studies, 86: 21–47.
Coady, C. A. J. (1992), Testimony: A Philosophical Study (Oxford: Oxford University
Press).
Conee, Earl, and Feldman, Richard (2001), ‘Internalism Defended’, in Hilary Kornblith
(ed.), Epistemology: Internalism and Externalism (Oxford: Blackwell), 231–60.
Edwards, Jim (2000), ‘Burge on Testimony and Memory’, Analysis, 60: 124–31.
Faulkner, Paul (2000), ‘The Social Character of Testimonial Knowledge’, Journal of
Philosophy, 97: 581–601.
Foley, Richard (1994), ‘Egoism in Epistemology’, in Frederick F. Schmitt (ed.), Social-
izing Epistemology: The Social Dimensions of Knowledge (Lanham, Md.: Rowman and
Littlefield), 53–73.
Gibbard, Allan (1990), Wise Choices, Apt Feelings: A Theory of Normative Judgment (Cam-
bridge, Mass.: Harvard University Press).
 Frederick F. Schmitt 219

Goldberg, Sanford (2001), ‘Testimonially Based Knowledge from False Testimony’,

Philosophical Quarterly, 51: 512–26.
Goldman, Alvin (1999), ‘Internalism Exposed’, Journal of Philosophy, 96: 271–93.
Graham, Peter (1997), ‘What is Testimony?’, Philosophical Quarterly, 47: 227–32.
(2000), ‘Transferring Knowledge’, Noûs, 34: 131–52.
Korcz, Keith Allen (1997), ‘Recent Work on the Basing Relation’, American Philosophical
Quarterly, 34: 171–91.
Lackey, Jennifer (1999), ‘Testimonial Knowledge and Transmission’, Philosophical
Quarterly, 49: 471–90.
(forthcoming), ‘Memory as a Generative Epistemic Source’, Philosophy and Phe-
nomenological Research.
Lehrer, Keith (1997), Self-Trust: A Study of Reason, Knowledge, and Autonomy (Oxford:
Oxford University Press).
Lyons, Jack (1997), ‘Testimony, Induction and Folk Psychology’, Australasian Journal of
Philosophy, 75: 163–77.
Owens, David (2000), Reason Without Freedom: The Problem of Epistemic Normativity
(London: Routledge).
Roth, Abraham (2003), ‘Practical Intersubjectivity’, in Frederick F. Schmitt (ed.) Social-
izing Metaphysics: The Nature of Social Reality (Lanham, Md.: Rowman and Little-
field), 65–91.
Schmitt, Frederick F. (1987), ‘Justification, Sociality, and Autonomy’, Synthese, 73:
43–85.
(1992), Knowledge and Belief (London: Routledge).
(1993), ‘Epistemic Perspectivism’, in John Heil (ed.), Rationality, Morality, and
Self-Interest: Essays Honoring Mark Carl Overvold (Lanham, Md.: Rowman and Little-
field), 3–27; repr. in Hilary Kornblith (ed.), Epistemology: Internalism and Externalism
(Oxford: Blackwell, 2001), 180–206.
(1999), ‘Social Epistemology’, in John Greco and Ernest Sosa (eds), The Blackwell
Guide to Epistemology (Oxford: Blackwell), 354–82.
(2002), ‘Testimonial Justification: The Parity Argument’, Studies in History and
Philosophy of Science, 33: 385–406.
Senor, Thomas (1993), ‘Internalistic Foundationalism and the Justification of Memory
Belief ’, Synthese, 94: 453–76.

N OT E S
1. It follows from this condition that justified true belief is not sufficient for knowledge
on testimony, since it does not entail that the testifier knows. A weaker condition
requiring that the testifier be in some positive epistemic position with respect to p
will have the same consequence.
2. I say belonging to another individual in the normal case because it would seem to be
possible for my past self to testify to my current self—through a diary, for example.
Indeed, I might not even realize that I am testifying to myself.
3. Neither Lackey nor Graham denies that testimonial knowledge is social in requiring
that the testifier be in some positive epistemic position. Goldberg (2001) has giv-
en a Gettier-inspired counterexample purporting to show that I can know on the
220 Transindividual Reasons

basis of false testimony. Goldberg’s counterexample does not in fact strike down the
condition stated in the text, but only the condition: I know that p on the basis of
T ’s testimony that q only if T knows that q. Indeed, the counterexample has force
against the latter condition only on the assumption that the testifier knows that p; so
it actually assumes the condition stated in the text.
4. I use ‘‘testimonial belief,’’ ‘‘testimonially justified belief,’’ and ‘‘testimonial know-
ledge’’ to mean, respectively, belief entirely based on testimony, belief justified
entirely on the basis of testimony, and knowledge on the basis entirely of testimony.
(Similarly for ‘‘memorial belief ’’ and ‘‘memorially justified belief.’’) I use ‘‘justified
in believing p’’ in whatever sense being justified in believing p is required for know-
ing p. Knowing p requires being doxastically justified in believing p, which entails a
belief p and the belief p’s being based on good reasons. Being doxastically justified in
believing p in turn requires being propositionally justified in believing p, which entails
neither a belief p nor a basing relation, but only possessing good reasons for believ-
ing p. I focus here on doxastic justification, rather than propositional justification,
for two reasons. First, if doxastic justification is social in the sense we are consid-
ering, propositional justification will be as well. For essentially the same case can
be made for the sociality of testimonial propositional justification as for testimonial
doxastic justification. Second, doxastic justification is required for knowledge, and
much of my knowledge depends on my testimonial knowledge. Similarly, any of my
beliefs that are justified on the basis of testimonially justified beliefs must be justified
on the basis of my doxastically and not merely propositionally justified testimonial
beliefs. In these respects, doxastic justification is more important than propositional
justification.
5. Two remarks on testimony and testimonial justification: (a) I assume that T testifies
that p when T presents p as true. I do not think that the presentation required for
testimony must be linguistic or even symbolic (it could be ostensive). Testifying
does not entail that T presents p to an audience intentionally. Simon could speak to
what he thinks is an empty desert, but if a hearer overheard his speech, the hearer
could know things on testimony. I could write a diary without intending anyone to
read it; but if someone did read it, they might believe propositions expressed in the
diary on testimony. Moreover, I can pick up testimony without realizing that I am
doing so. I suspect we pick up a fair bit of what we know in just this way. (b) I am
focusing on this condition: My belief p is justified on the basis of T ’s testimony.
But the epistemology may be the same for my uptake from T ’s mere belief, without
testimony or a presentation as true. For further discussion of what is involved in
testimony, see Coady (1992: ch. 2) and Graham (1997).
6. This condition involves a simplification. It seems possible for me to believe p on
T ’s testimony even though T testifies to some q from which I infer p, rather than
to p itself. A fully general principle would have to allow for such a case, but I will
not attempt to construct such a principle here. I will make the simplifying assump-
tion that in all cases in which I believe p on testimony, I do so on the basis of T ’s
testifying that p.
7. I would expect the Transindividual Thesis to be denied by almost everyone. See, for
example, Audi (1997). The Thesis is endorsed by Owens (2000: chs. 9, 11). And
as Edwards (2000) argues, Burge (1993) ought for consistency to accept the Thesis.
 Frederick F. Schmitt 221

For defense of a parallel thesis about reasons for action, that I can have practical
reasons for action that belong to others, see Roth (2003).
8. Lackey (forthcoming) has argued that (M1) needs qualification. In particular, there
are cases in which a subject comes to be justified in a belief on memory even though
she lacks current and original justification. For, although her belief would otherwise
originally be justified, its justification is defeated by a misleading belief she possesses;
yet she has since lost this defeating belief. I am not entirely certain that there are such
cases because I am not entirely certain that a belief that would otherwise be justified
can be defeated merely by the possession of a misleading background belief. How-
ever, anyone persuaded by the example will need to qualify (M1) so that it requires
only that, in instances in which current justification is lacking, the subject would
otherwise have had original justification if she did not possess a misleading defeater.
Lackey also presents an example that shows that original belief can be lacking in cases
of memorial justification. This shows that the original justification, if it is required,
may be merely propositional rather than doxastic justification.
9. The Transtemporal Thesis is not entirely uncontroversial because it is inconsist-
ent with typical versions of accessibility internalism (see Schmitt 1992: ch. 4 for
discussion of accessibility internalism). For a case against accessibility internalism,
see Goldman (1999). For defense of internalism from Goldman’s criticisms, see
Conee and Feldman (2001).
10. One might claim that we can rule out all current reasons as the basis for a memori-
ally justified belief quite easily, on the conceptual ground that, by the very notion
of memory, no belief justified on the basis of current reasons can be a memorially
justified belief. One might then appeal to this argument in support of (M2): we have
memorially justified beliefs, but none can be based on current reasons. I find this
conceptual case for (M2) too facile to carry conviction. I note, however, that if it is
a good argument for (M2), there is equal plausibility in the analogous ground for
(T2). Since this would make the conditional case for (T2), we may therefore ignore
the argument here.
11. There are other arguments for the same conclusion. First, any current reductive
reason that could support my memorially justified belief must itself be justified on
memory, simply because, by the very notion of memory, my belief p would not
count as justified on memory in virtue of being justified on the basis of a justifying
inference from a current reason unless that reason is itself justified on the basis of
memory. Second, one might ask just what is required for a reason r to be current
when the belief p is justified on the basis of a justifying inference from r. Any infer-
ence takes time. Presumably the reason r must itself be preserved in memory if it is
to count as current. So for purposes of the inferential justification, r must be justified
partly on the basis of memory. I find these arguments too swift to carry weight, and I
will not attempt to assess them but instead rely on the argument I give in the text.
12. It is worth noting that there is a special problem for the statistical reduction of
memorial justification that does not apply to the statistical reduction of testimonial
justification. For the degree of justification supplied by a statistical syllogism is indif-
ferent to how well memory tracks what the subject has past reasons to believe, once
the inductive premise that memorial beliefs of type X tend to be true is justified.
But the degree of memorial justification does intuitively diminish with a reduction
222 Transindividual Reasons

in such tracking. There is no analogous problem for testimonial justification. In this

regard, memorial justification on the basis of current reductive reasons is in worse
shape than testimonial justification on the basis of possessed reductive reasons.
13. For criticism of the statistical basis along these lines, see Schmitt (1987, 1999)
and Coady (1992). For a defense of the statistical basis, see Lyons (1997).
14. For reduction by appeal to charity, see Coady (1992: ch. 4). For criticism of reduc-
tion by appeal to charity, see Lyons (1997). For reduction by the argument from
parity, see Gibbard (1990), Foley (1994), and Lehrer (1997). For criticism of that
argument, see Schmitt (2002).
15. One might allow that a coherentist justification is a justification in virtue of which I
have a reason for my belief, but deny that it is a justification in virtue of which my
belief is based on any reasons. My attitude toward this, for present purposes, is that
a coherentist account of the justification for my memorially or testimonially justified
belief will not be very plausible unless we can treat coherence relations as entailing
or otherwise conforming to a basing relation. If this is not so, then a coherentist
justification is ruled out of court by (G) on the assumption that the suitable relation
must be a basing relation.
16. For a defense of a nonreductive foundationalist account of memorial justification,
see Audi (1996). For criticism of foundationalist accounts along lines similar to
those scouted here, see Senor (1993).
17. Edwards points out that Burge contradicts himself. Burge’s Acceptance Principle
clearly entails that I am a priori entitled to my memory belief even if my belief
originally lacked entitlement, but Burge explicitly says that my entitlement to my
memory belief presumes an original entitlement (1997: 44 n. 2). The latter claim
is most naturally interpreted as an endorsement of our Transtemporal Thesis for
memorial justification. Burge also seems to contradict himself in the parallel way
in the case of testimony (1997: 39–40). Edwards suggests, correctly in my opin-
ion, that Burge should endorse what I am calling the Transindividual Thesis for
testimonial justification.
18. I note, however, that Burge’s view does entail (T2), that my belief can be justified on
the basis of testimony even though not justified on the basis of any reason I possess.
19. More exactly, I will criticize one of Burge’s two arguments. The argument I leave
out is that ‘‘if p is the content of an intelligible message, it must bear a relation to
a subject matter, and thus it is prima facie true’’ (1993: 476). The argument here, I
surmise, assumes a principle of charity about content: an intelligible message gets its
content from a relation to a subject matter and so belongs to a set of propositions
about that subject matter most of which are true. I can see that one who is entitled to
believe this principle of charity can be entitled to believe p by arguing that p from the
principle as a premise. But most people do not believe the principle of charity, are
not entitled to believe it, and do not argue from it. So appeal to the principle cannot
explain their entitlement to testimonial beliefs. The mere availability of an argument
from the principle is not enough to entitle me to believe p. Moreover, the degree
of frequency of true testimonial belief entailed by charity (namely, a frequency of
a mere fifty percent or better) is not sufficient to make the process of testimoni-
al uptake reliable enough for entitled belief, on any plausible reliabilist account of
entitled belief.
 Frederick F. Schmitt 223

20. There is another difficulty with Burge’s argument for the Conclusion, a serious one.
‘‘Prima facie’’ is always prima facie to a subject. Trouble reveals itself when we ask:
to which subject is p prima facie a presentation of a rational source or source of truth
in (a), (b), and (c)? Is the subject me, or is it a third party theorist? Arguably, p is not
prima facie to me a presentation of a rational source or source of truth, at least not
in a sense that entails that I am prima facie entitled to believe that it is a presenta-
tion of a rational source or source of truth. The mere fact that a source is seemingly
intelligible to me can hardly make me entitled to believe that the source is ration-
al. Surely it doesn’t do so if I lack good reason to believe that there is a connection
between intelligibility and rationality. At the very least, we would have to add that I
have good reason to believe that there is such a connection. But then Burge’s account
of entitlement would apply only to subjects with enough sophistication to believe
that there is a connection between intelligibility and rationality. His account would
lose generality. If, by contrast, p is taken to be prima facie, to a third party theor-
ist, a presentation of a rational source or source of truth, then (a), (b), and (c) have
no plausibility. For although a third party theorist may have good reason to believe
intelligibility and rationality are connected, and can therefore infer from p’s being a
seemingly intelligible presentation to its being a presentation of a rational source, the
availability of this inference to a third party theorist would hardly make me entitled
to believe p. This is a difficulty even for a view that, like Burge’s, allows for entitle-
ment when the subject does not grasp the reason. There just isn’t any plausibility to
the idea that an argument from this kind of principle of charity could entitle me to a
belief when I have no notion of charity. It isn’t like the case in which I am entitled by
the reliability of my perceptual process even though I am unaware of the process or
its reliability.
21. Burge suggests supporting the claim that if a presentation p is prima facie a present-
ation of a rational source, then I am prima facie entitled to believe p (a consequence
of (b) and (c) ) by appeal to the function of reason: presentation p ‘‘is prima facie
preserved (or received) from a rational source, or resource for reason; reliance on
rational sources—or resources for reason—is, other things equal, necessary to the
function of reason’’ (1993: 469). But then Burge appears to drop the idea of appeal-
ing to what is necessary to the function of reason. So I will not discuss it here.
22. Note that there is a similar difficulty for the Transindividual Thesis posed by the
plausible weaker requirement of sensitivity to good reasons for testimonially justified
belief: I am testimonially justified in believing p on the basis of reason r only if I would
not believe p if the testifier lacked a good reason to believe p. The temporal analogue
of this requirement poses an analogous problem for the Transtemporal Thesis.
23. I note that the fact that testimonial uptake occasionally involves misunderstanding
does not further detract from the conditional probability for testimonial belief. True,
there is no phenomenon of memory comparable to misunderstanding. But misun-
derstanding is not a case in which the testifier lacks a reason to believe p, testifies that
p, and I believe p. So it does not lower the specified frequency. The same goes for
misspeaking and mishearing in testimony.
24. This does show that on the Transindividual Thesis, I can base my belief on a reason
I don’t understand. A proponent of the Transindividual Thesis has to live with this
consequence.
224 Transindividual Reasons

25. Acting responsibly in testifying to p is presumably the testifier’s making her best
effort in testifying in the sense of acting in a manner that reflects her best effort were
she concerned solely to testify to the truth.
26. Owens speaks of transferring responsibility for the truth of p, but perhaps it is better
to speak of transferring responsibility for there being good reasons for p available to
appropriate subjects.
27. Owens in fact observes that social responsibilism as characterized here does not
really supply a sufficient condition of testimonial justification compatible with the
Transindividual Thesis. I cannot, merely by acting responsibly in believing p, and
relying on a testifier who acts responsibly in testifying that p, ensure that the testi-
mony and its uptake are free of all the defects that could prevent my belief from
being justified (on the Transindividual Thesis). For neither the testifier nor I need
be blameworthy for her misspeaking, my mishearing, or my misunderstanding the
testimony—perhaps a loud noise startled one or the other of us, but having had
no reason to expect it to cause any of these defects, we proceed with testimony and
uptake. But these mishaps prevent my belief from being justified nonetheless. It is
unjustified simply because I believe p, but the testimony is that q, and the testifier
has no reason for p, so that my belief is baseless. Owens thus recognizes that acting
responsibly is not the whole story of justification, once the Transindividual Thesis is
accepted. However, Owens seems to regard this as a relatively minor qualification of
social responsibilism. My point in the text is that social responsibilism is more deeply
incompatible with the Transindividual Thesis than Owens recognizes.
 10
Testimony and Epistemic Autonomy
Elizabeth Fricker

1 . D I V I S I O N O F E PI S T E M I C L A B O U R V E R S U S T H E
I D E A L O F I N D I V I D UA L E PI S T E M I C AU TO N O M Y

A reference point in philosophical investigation of knowledge from testimony

is the ideal of the ‘autonomous knower’. This ideal type relies on no one else
for any of her knowledge. Thus she takes no one else’s word for anything,
but accepts only what she has found out for herself, relying only on her own
cognitive faculties and investigative and inferential powers. Descartes explicitly
espoused this ideal, and method, in his Meditations (Descartes 1641). Locke
equally rejected ‘other men’s opinions floating in one’s brain’ as not constituting
knowledge (Locke 1690). The wholly autonomous knower will not accept any
proposition, unless she herself possesses the evidence establishing it. Thus she
will not accept anything on the basis of another’s word for it, even when she has
evidence of their trustworthiness on the topic in question.
Such extreme purism restricts how much one can come to know very severely.
We humans are essentially social creatures, and it is not clear that we do or
could possess any knowledge at all which is not in some way, perhaps obliquely,
dependent on testimony. How exactly does the system of empirical belief—
hopefully knowledge—of each of us depend on others’ testimony? There is cer-
tainly massive causal reliance on testimony in the process by which each of us
develops into a language-user and thinker, ‘grows into possession of a world’.¹
The initial stages of language acquisition by a child inevitably occur through a

Earlier versions of this paper were given at a workshop on ‘Testimony, Trust and Action’
in King’s College Cambridge in September 2003, at a conference on ‘Moral Testimony’ in the
Philosophy Department at Birmingham University in March 2004, and at a conference at the
Inter-University Centre in Dbrovnik, Croatia, in May 2005. I received very useful comments from
audiences at these events, in light of which I corrected various errors. I am also very grateful to both
John Hawthorne and Stephen Schiffer for valuable comments and discussion on an earlier draft.
The research for this paper was done between January and June 2002, during a period of leave
funded by my employers, Magdalen College and Oxford University, and by a Fellowship from the
Mind Association. My thanks for their support.
226 Testimony and Epistemic Autonomy

process of simple trust² in its teachers—parents and other carers. In this cog-
nitive developmental process learning meanings is not separable from coming to
grasp and accept our shared basic world picture, the common-sense theory which
structures and frames our empirical thought. There is, for instance, no distinction
to be drawn between learning the meanings of ‘chair’, and ‘horse’, and ‘jump’,
and ‘cook’, and learning about chairs, and horses, and jumping, and cooking.³
The fact that each of us is causally reliant on others’ testimony in the historical
process by which she acquires her system of concepts and beliefs does not entail
that, once adult, each of us remains epistemically dependent on testimony for
her empirical knowledge. Perhaps each of us can afterwards push away the ladder
of trust in others, up which she has climbed into possession of a world. Beliefs
which were first acquired through a process involving simple trust in testimony,
and were initially epistemically based on testimony (as we may say once core
normative epistemic concepts become applicable to the developing child, viz.
when she becomes a thinker capable of epistemic self-criticism), may later acquire
an alternative basis. It may be that beliefs from the epistemic source of percep-
tion, linked by memory and extended by inference, can take over, together with
support from inference to the best explanation and broader coherence. Suppose
one could, once epistemically matured, thus push away the ladder of testimony,
retaining only the portion of one’s beliefs which remain epistemically supported
without reliance on it. In maintaining the ideal, one would then be restricted to
what one learns from one’s own senses and preserves in memory, plus whatever
one can get to by use of one’s own inferential powers from that base—with a ban
on even reasoned, empirically backed trust in the word of others!
There is reason to doubt that one can in that way eliminate all epistemic
dependence on testimony in one’s mature system of empirical belief, even if pre-
pared drastically to prune it. For one to do so, her original epistemic dependence
on testimony would have, everywhere, to be replaced by adequate support from
other epistemic sources, or the belief in question dropped. Now of course it often
happens in particular cases that one first learns of something through another’s
testimony, and then is later able to confirm it for oneself through perception,
perhaps combined with memory and inference. My daughter tells me her new
teacher wears glasses; later I see the teacher for myself. The weather forecaster on
Tuesday predicts that it will rain on Wednesday; Wednesday proves wet. Facts
about a foreign country known to one at first only through travel literature and
friends’ reports are confirmed by perception, when one travels there oneself. In
these and countless similar cases one later gets first-hand perceptual evidence of
what one first believed on testimony. In such cases contrary perceptual evidence
would decisively falsify the testimony.
There are other ways, less direct but no less powerful, in which alternative
grounds for belief can grow strong enough to take over the support of a
belief originally acquired from and based on testimony. Inference to the best
explanation and explanatory coherence more broadly can take over the support
 Elizabeth Fricker 227

of many beliefs originally based solely on others’ testimony. It is plausible, for

instance, that one’s implicit beliefs about what the words of one’s language mean
no longer rest on the past trusted testimony through which they were learned.
One’s linguistic interactions with others would not run as smoothly as they do, if
one’s first teachers had deceived one! (See Adler 1994; Lyons 1997.)
Nonetheless the role of past trusted testimony in the system of empirical belief
of each of us is fundamental, because it has shaped the conceptual frame with-
in which current individual perceptions are made—how the sensory given is
conceptualized to yield perceptual experience and belief.⁴ Thus, in our three
examples above, while it is true that subsequent contrary perception would refute
the earlier testimony, these perceptions are themselves subtly dependent on a
framework of concepts shaped in part by earlier trust in testimony. I see the
teacher’s reading glasses, but that she is my daughter’s teacher I know only
through a set of background beliefs in which testimony is inextricably involved.
When it rains on Wednesday I see and hear the rain; but my knowledge that it
is Wednesday is testimony-infected, and that very concept and its application is
one constituted by human consensus involving testimony. Similarly, when I vis-
it Australia for the first time, in one way I gain personal confirmation of what I
had previously known of only through testimony; but my knowledge that I am
in Australia at all depends on testimony in multifarious and hard to pin down
ways: initially I knew where my flight landed only through trusting the testimony
of travel agent and airline personnel, and though the evidence of roadsigns and
so forth may take over, these are all put there by human agency, and constitute
a kind of testimony. Moreover, the controlling idea in terms of which I con-
ceptualize and slot in all my own personal experiences—of the spherical planet
earth with its land masses and seas, its countries, nations, and other geopolitical
institutions, its history and prehistory—was acquired from testimony.
This brief sketch has shown how the epistemic dependence on testimony in the
beliefs of each of us socially embedded twenty-first-century individuals is subtle
and widespread, if not all-pervading. A more sustained enquiry is needed to see
whether it could in principle be eliminated by such an individual while leaving
her any beliefs at all. Worse yet for the would-be epistemic autonome: it may
not be sharply determinable at all, whether and when freedom from oblique epi-
stemic dependence on testimony is achieved, since isolating the contribution of
testimony from that of other sources of support, in a system of belief with rich
explanatory coherence, is not a clear-cut matter. It is at any rate certain that, in
order to live up to the ideal of individual epistemic autonomy, a very great deal
of what is believed by a normal member of a modern society, with its exten-
ded division of epistemic labour, would have to be bracketed, given up—most
of geography, history, the natural and social sciences including medicine, and
so forth.
Giving all that up is no more a serious practical possibility than living out the
life of a more thoroughgoing sceptic—one who doubts even the evidence of the
228 Testimony and Epistemic Autonomy

senses as indicators of a perceptible external world. Who would really give up the
fruits of the sciences including all technology, medicine, dentistry, foreign travel,
as well as historical understanding and knowledge—and so on? The epistemic-
ally autonomous individual could not trust an electrician to wire her (self-built!)
house for her, since she would not accept his testimony about what he was going
to do, and that it would work safely; nor her doctor to prescribe medicines; nor
would she try ski-ing because her friends (she could not have many!) told her it
was fun.
We have found that testimony, for each of us in our modern social and epi-
stemic predicament in which division of epistemic labour along with other sorts
is the rule, is an essential source of empirical grounding for her beliefs about the
world she finds herself in, and her own place in it. This system of empirically
based belief is richly coherent, including its ability to explain its own sources.
Notice that the trust in testimony of which I am stressing the ubiquity need
not however be given uncritically, without empirical grounds. I have argued else-
where that a mature recipient of testimony need and should not trust another’s
word without adequate empirically based warrant to do so. We have seen that it
is impractical to live up to the supposed ideal of individual epistemic autonomy.
One cannot live in a modern scientifically and technologically sophisticated soci-
ety, nor have any social life at all, without trusting others in almost one’s every
action. But this is not to say that one’s trust in the vast heritage of knowledge and
know-how built up from others’ investigations, expertise, and experience must
be blind—uncritical and undiscriminating. Good empirical grounds for taking a
fresh instance of testimony to be sincere and reliable—or for being distrustful of
it—are often to be had; and inference to the best explanation and rich coherence
within one’s accumulated system of belief can support, ex post, one’s reliance on
some earlier pieces of testimony, while equally discrediting others. (I here barely
touch on issues which need much fuller discussion. See Adler 1994; Coady 1992;
Fricker 1994, 2002, 2005.)
Still, one who trusts testimony discriminatingly, only when she has an
adequate empirical basis to do so, and whose past trust is now vindicated through
support from explanatory coherence, is yet dependent on testimony in her beliefs,
and actions based on them. If I take others’ word for things, I extend my
knowledge far beyond the range I could achieve on my own, but by this very fact
I am not epistemically autonomous. I believe many things for which I personally
do not possess the evidence, and my believing is premised on the supposition that
some other person or set of persons jointly has, or had, access to that evidence,
and evaluated it correctly. (These points are expanded in Section 4, below.)
We have seen that even if one could, by a heroic effort of epistemic recon-
struction, push the ladder of past trusted testimony away, the project of attaining
and maintaining the ideal of complete individual epistemic autonomy is not
an attractive or feasible one—one would forgo too much! Is there reason to
regret this? In this paper I shall respond to this question by addressing a closely
 Elizabeth Fricker 229

related one: In what way exactly is one’s epistemic self-governance necessarily

compromised, by one’s practically inevitable dependence on others’ testimony?
Putting the question the other way about: In what way, and to what extent, can
one maintain one’s epistemic self-governance despite one’s inevitable reliance on
others’ testimony, and the technological fruits of others’ knowledge and expert-
ise, in almost every area of one’s life?⁵ I will first prepare the ground to address
this question by considering another: In what circumstances, and on what topics,
may one properly accept and learn from another’s testimony? In developing an
answer to this question we will find material relevant to answering our first one.

2 . T H E C I RC U M S TA N C E S A N D TO PI C S O F P RO PE R
AC C E P TA N C E O F T E S T I M O N Y

In what circumstances and on what topics may one person with epistemic pro-
priety accept the testimony of another and by so doing learn, acquire know-
ledge, from her? Conversely: What are the circumstances in which, and topics
on which, one person may tell something to an audience, thereby expressing her
knowledge, and reasonably intend and expect to be believed, trusted—to have
her word on the matter accepted? These are distinct questions, but the mutuality
of the illocutionary act of telling means that their answers will coincide, where the
expectation of being trusted is well founded.⁶

2.i A Precondition for Testimonial Spreading of Knowledge

As a preliminary I note a precondition for testimony to be given and received at
all. For an act of telling to succeed there must be mutual understanding. A mes-
sage must be got across and accepted. So there must be a proposition which the
teller intends by her action to present as true, and this must be identical with the
one grasped by her audience as so presented, and accepted by her.⁷ This does not,
in itself, entail a shared language. Nor does it entail that, when a shared language
is employed, the message conveyed is what the speech act literally means—a pro-
position which the sentence used is conventionally apt to convey, and is plausibly
interpreted as specifying on that occasion.⁸ But communication of a message is
most commonly effected by use of sentences of a shared language in accord-
ance with the constraints of their literal meaning, to make explicit assertions.
The occasions on which other media are feasible vehicles are relatively few and
far between. This being so, the spreading of knowledge by means of testimony,
or something like it, is possible to any significant extent only when there is a
shared language, and mutual understanding of speech acts between speaker and
audience.
Our common-sense view—which I shall not question here—is that we do
indeed share a language, including its semantics, with our co-speakers to an extent
sufficient for mutual understanding and successful communication. But one can
230 Testimony and Epistemic Autonomy

learn from another’s testimony only when one does not already know what she
tells one. Hence, if difference of opinion regarding the truth value of some sen-
tence S entailed difference of meaning attached to S, there would be no learning
from others regarding the proposition expressed by S.⁹ This observation is, I
think, enough to discredit extreme ‘holistic’ theories about the fixation of mean-
ing (already implausible). But there are certain areas of discourse where disagree-
ment might be thought to undermine the supposition that meaning is shared. If
this suspicion were confirmed, then learning from testimony in the strict sense
(as opposed to changing one’s language to conform more with others) would be
shown to be impossible in these areas.
Difference of opinion due to ignorance, where one party simply lacks firm
belief either way on the topic, is unproblematic. Equally, disagreement in the
strong sense of conflict of opinion is unproblematic, when its origin is traceable
to different access to evidence. (In such a case pooling of evidence will produce
convergence of opinions.) It is when disagreement in judgement persists despite
similar access to evidence that, in certain areas, the supposition of shared mean-
ing may be threatened. If there are certain subject matters where disagreement of
judgement in response to the same evidence entails difference of meaning, then
there can be no learning from testimony in the strong sense of deferring to others’
judgement, letting it override one’s own, on those topics.
It is a point familiar from a certain style of philosophical account of how
meanings and beliefs are simultaneously attributed to someone, that a tentat-
ive interpretation of an utterance which yields a difference of opinion between
interpretee and interpreter not explained by differential access to evidence, is
thereby thrown into doubt. Other aspects of the total interpretation being equal,
it is more ‘charitable’ hence a priori better warranted, to interpret the other as
meaning something else (see Davidson 1984). But other aspects may well not be
equal, and so the defeat of the assumption of shared meaning is generally not
instantaneous. We all have had futile arguments—‘‘It’s green’’; ‘‘No it’s not, it’s
yellow’’—where the suspicion lurks that there is not really a substantial matter at
issue, rather than a non-concordance of linguistic usage at its vague edges, com-
pounded perhaps by a pig-headed refusal of the out-of-line debater to adjust her
usage. Equally we all have had arguments where it seems certain that there is a
substantial, not merely a semantic matter at stake—‘‘It’s unfair that you let Juli-
an go in the front of the car, but you never let me’’, although progress towards
agreement may seem no less hard to achieve.
Colour concepts, and other simple perceptually applied concepts; plus moral
and also aesthetic concepts, are ones where sorting out substantive from merely
linguistic disagreement on particular occasions is difficult; no less difficult than
giving an account of how the precise content of those concepts is fixed. In these
and some other cases, there really may be no way to distinguish between defer-
ring to others’ judgement about the application of an already shared concept, and
adjusting one’s concept. (We remarked earlier that, in one’s initial acquisition of
 Elizabeth Fricker 231

one’s language, there is no sharp line between acquiring new information, beliefs,
about things one already has a concept of; and acquiring those concepts.)
These considerations will be relevant in a full investigation of the possibilit-
ies for learning from testimony about these topics. It will be important to bear
them in mind, when considering to what extent one can defer to others’ judge-
ment on moral and aesthetic matters. There may prove to be limitations on this
grounded in considerations about meaning, for aesthetic judgement in particular,
I suspect.¹⁰ Having noted this, I will not explore it more fully here.

2.ii A Principle Concerning Deferential Acceptance

With these points about the need for shared meaning made, we can proceed with
the main positive idea. We can formulate a general principle:
Testimony Deferential Acceptance Principle (TDAP 1): For one
properly to accept that P on the basis of trust in another’s testimony that
P—her word that P¹¹—requires that she be epistemically well enough
placed with respect to P so that were she to have, or make a judgement
to form a conscious belief regarding whether P, her belief would almost
certainly be knowledge;¹² and that she be better epistemically placed with
respect to P than oneself; and that one recognize these things to be so.
TDAP1 specifies a condition necessary for epistemically proper trusting
acceptance of another’s testimony on some topic. It is not sufficient, because
while the hearer’s cognizance of the testifier’s strong epistemic position vis-à-
vis the topic makes it rational for her, other things being equal, confidently to
expect the testifier’s judgement about the matter in question to be correct—to
deem her competent about the matter in question—TDAP1 does not speak
to the question of the testifier’s sincerity. As I have argued elsewhere, the
overall trustworthiness of a speaker’s testimony breaks down into these two quite
separate components. In this investigation I concentrate on the circumstances in
which deferential acceptance of another’s judgement, as expressed in her sincere
testimony, is epistemically proper. Thus, having noted the need for adequate
warrant to believe the speaker sincere, I put further consideration of sincerity
aside, assuming in what follows that insincerity is not an issue in whether to trust
the other’s testimony.¹³
The matter of sincerity is one reason why TDAP1 specifies only a necessary,
not a sufficient condition, for epistemically proper deference to another’s
testimony. The only other reason I can think of why other things would
not be equal, regarding the hearer’s expectation of correctness of the testifier’s
expressed judgement, is if she were also aware of significant contrary testimony.
Contrary testimony will be epistemically significant if it either comes from
another equally well-qualified expert; or, in some cases, if it is from many
mutually independent sources, albeit not especially expert ones.¹⁴ A more refined
232 Testimony and Epistemic Autonomy

condition incorporating these two factors, which is normatively both necessary

and sufficient for deferential acceptance is:

TDAP 2: One properly accepts that P on the basis of trust in another’s

testimony that P—her word that P—just if she speaks sincerely, and she is
epistemically well enough placed with respect to P so that were she to have,
or make a judgement to form, a conscious belief regarding whether P, her
belief would almost certainly be knowledge; and she is better epistemically
placed with respect to P than oneself; and one recognizes all these things to
be so; and one is not aware of significant contrary testimony regarding P.

TDAP 2 specifies when it is proper to accept another’s testimony that P out-

right. There will also be situations where neither party is in a position to make a
knowledgeable judgement as to whether P, but one is better epistemically placed
than the other. In these cases it will be epistemically rational for the worse-placed
person to defer to the other’s opinion, while falling short of taking her utterance
as an expression of knowledge; hence forming only a tentative belief regarding
P. This is required when, for instance, an informed decision about how to act
is urgently needed. There may be some topics for which this situation is the
rule—that is, where knowledge as opposed to more or less well-grounded spec-
ulation is very hard to come by.¹⁵ It remains true that one should not accept
outright another’s testimony that P, unless one reasonably believes her to be so
placed as to (almost certainly) form knowledgeable belief regarding P. Hence
TDAP2 is the correct general principle governing the outright acceptance of
another’s testimony.
An explanatory comment is needed on the role of the complementary ‘intern-
al’ and ‘external’ components of epistemic propriety in TDAP1 and 2. I have
formulated TDAP1 and 2 incorporating both internal and external components,
because I am concerned to describe what happens when things go right, and thus
how knowledge is spread by means of testimony. Externally, things are going
right when the testifier speaks from her expertise-generated knowledge, and is
sincere. But epistemic rationality has a key internal component: it is not rational
to accept unquestioningly the testimony of an expert who, so far as one knows,
is no such thing. And, though not all-in epistemically proper, it is subjectively
rational and epistemically blame-free to accept another’s testimony, when one
falsely but justifiedly believes her to be an expert about the topic, being deceived
about this through no fault of one’s own. I am against purely exernalist accounts
of when acceptance of testimony is epistemically proper. These fail to incorporate
the requirement that the subject maintain epistemic responsibility for her own
beliefs. In Section 3 I spell out the implications, and means of satisfying, this
requirement.¹⁶
We may distinguish between a weak and a strong form of deference to anoth-
er’s testimony:
 Elizabeth Fricker 233

Weak Deferential Acceptance occurs when I form belief that P on the basis
of trust in another’s testimony that P, when I myself have no firm pre-
existing belief regarding P; nor would I form any firm belief regarding P,
were I to consider the question whether P using only my current epistemic
resources, apart from the current testimony to P.
Strong Deferential Acceptance occurs when I let another’s trusted testimony
regarding P override my own previous firm belief, or disposition to form a
firm belief, regarding P.
The distinction between strong and weak deferential acceptance may or may
not turn out to be important. First off, it seems that there could be subject
matters where strong deferential acceptance is never epistemically appropriate,
although weak deference can be. This fact may illuminate the nature of that sub-
ject matter.
Whether for weak or strong deferential acceptance, it seems that TDAP2 is the
correct normative principle: her sincerity not being in question, and my being
aware of no significant contrary testimony, it is epistemically proper that I defer
to another’s testimony in forming belief regarding P, or in overriding my own
previous belief regarding P, just if I recognize that she is better epistemically
placed than I am to determine whether P; and it is epistemically proper that I
accept her testimony outright just if I recognize this, and also that she is so placed
as to form (almost certainly) knowledgeable belief regarding P. We may intro-
duce a thin and inclusive sense of ‘expert’ capturing this core normative necessary
condition for deferential acceptance expressed in TDAP1 (which is also norm-
atively sufficient, apart from the matters of sincerity and absence of significant
contrary testimony):
S is an expert about P relative to H at t just if at t, S is epistemically well
enough placed with respect to P so that were she to have, or make a judge-
ment to form a conscious belief regarding whether P, her belief would
almost certainly be knowledge; and she is better epistemically placed than
H to determine whether P.

2.iii Bases of Expertise

We can now explore the different possible bases of such relative epistemic expert-
ness of S over H regarding some collection of propositions P comprising a subject
matter W. In so doing we will be developing a description of the various cir-
cumstances in which it is epistemically proper, when she knows them to obtain,
for one person deferentially to accept another’s testimony regarding some sub-
ject matter. The idea of someone’s being epistemically ‘well placed’ regarding P is
used so far in a broad catch-all sense. We will now see how a variety of specific cir-
cumstances may contribute to this. One is literally the spatio-temporal location
of the person; another is particular skills and perceptual and cognitive equipment
she possesses.
234 Testimony and Epistemic Autonomy

There are various kinds of situation in which it is obvious and unproblematic

that S will at that time be epistemically expert relative to H regarding some sub-
ject matter W.
Case One: W is an observable event or state of affairs, and S is or was at the
time of its occurrence so positioned as to be able to observe it, whereas H
was not.
Suppose, for instance, that Natalie went to the RadioHead concert in South
Park, while I did not. Then—assuming she has normal observational and
memory capacities—she knows quite a bit of what happened during it, and I
can learn from her telling me about it. Without access to some such eyewitness
account (written or spoken), I cannot know very much about what happened.
Of course Natalie is not my only possible informant. And there may be a limited
amount I can infer from other sources of evidence—walking there the next day
I observe huge numbers of empty drinks cans scattered around, and see the stage
being dismantled. But I cannot learn any detail except from testimony; and only
hers is conveniently available.¹⁷
Natalie is an expert on what happened at the concert, relative to me.
Hence—if she is truthful—I can learn from her. But her expert status on the
topic relative to me is highly accidental. It is based in a mere happenstance
about our locations on one particular day, not on any more stable and intrinsic
epistemic talent, skill, or base of knowledge that she possesses whereas I lack.
Had I possessed a ticket and gone to the concert, while Natalie stayed at home,
I would instead have been the expert vis-à-vis her. Such merely accidental and
extrinsic expert status is often brief and transient. Emily is momentarily an expert
relative to me about what is in the fridge; but only until I take a look for myself.
Case Two: Superior perceptual skill of S over H.
Now suppose Natalie and I are both at the concert, and are trying to make out
what is happening on the stage from some distance. I am shortsighted, whereas
Natalie has excellent distance vision. She reports ‘the supporting band is coming
on’ and I accept her report, not being able to see anything specific for myself. Or I
may think I can see something different, but I allow my visually based judgement
to be overruled by hers, in the knowledge that she has better distance vision than
me.¹⁸ This case is more interesting. Natalie is an expert relative to me about what
is happening in the distance not because of an accidental difference in our loca-
tions, but due to a superior epistemic skill she has relative to me, which is (in a
relaxed sense) intrinsic, and fairly stable. She is not just accidentally better placed,
that is spatio-temporally located, than me, regarding the topic; she is better epi-
stemically equipped than me to make judgements of a certain kind—namely,
judgements about events occurring in the visual perceptual distance. Let us say
that she is not only currently an expert relative to me about the happenings on the
stage, in the thin sense defined above; but that this is due to an epistemic expertise
 Elizabeth Fricker 235

she possesses relative to me, regarding a range of matters-in-circumstances. Spe-

cifically:
S has an expertise relative to H on some subject matter W at a time t just if
S has a superior ability at t to determine the truth of propositions in W
which is based in superior perceptual and/or cognitive skills and know-
ledge, and is hence (in a relaxed sense) intrinsic, or has a crucial intrinsic
component.
Exercise of an expertise will almost certainly require that the environment be
normal in various respects—as with perceptual skills—and so is intrinsic only in
a sense which is relaxed, though surely intuitive. Exercise of specialized cognitive
skills may require access to equipment, even laboratories; but has a crucial intrins-
ic component. An expertise is, in this lenient sense, a superior epistemic power
possessed by a person due to her specific differentiating characteristics, such as
superior perceptual skills, or specialized field of training and knowledge. Her
expertises are relatively stable properties of a person, since they are not owed to
mere accidents of spatio-temporal location, but are more deep-seated properties
of that person; some owed to genetic endowment, but many acquired through
special training or education.¹⁹
S’s possessing a superior perceptual ability to H is one kind of expertise which
S may have relative to H. This may be due to native differences in perceptual
equipment, as with acute versus poor distance vision. But differential percep-
tual ability may also be due to training and background knowledge. An expert
at cricket can see and describe what is taking place during the game—‘‘It was
a fast ball that moved in from outside the off stump, and the batsman caught
an edge on it, and was caught behind by the wicket-keeper’’—when a novice
will have discerned almost nothing specific at all. The same goes for aural per-
ceptual abilities, for instance to discern and describe the harmonic progressions
in a complex piece of music; or to catch and understand the words of speech
in a particular language. Because background knowledge and skills inform and
shape perception in this way, there is no sharp distinction between perceptu-
al versus knowledge-based expertise. Many bases of expertise involve both in
inextricable combination. The complex perceptual-cum-knowledge-based skill
provides a superior ability to determine the truth of a range of propositions in
certain circumstances. Other bases of expertise are more heavily grounded in spe-
cialized knowledge and training, with a lesser role for associated perceptual skills.
This includes scientific knowledge and skills in experimental procedures and the
evaluation of data; and technological knowledge and know-how, such as that of
the garage mechanic, builder, or computer technician. Most purely cognitive is
superior ability at reasoning in a particular abstract domain such as mathematics.
When another has expertise relative to me in a certain, perhaps esoteric, field
of knowledge, it is clear that I can and should—assuming I trust in her sincer-
ity—defer to her in forming beliefs about the domain in question. Where the
236 Testimony and Epistemic Autonomy

field is sufficiently advanced and complex, I may not even be able to evaluate the
arguments, nor the significance of the evidence, myself; and I may lack the nat-
ive talents to acquire the skills to do so, even if I had both time and inclination
(see Hardwig 1991). We will draw out the significance of these facts in the next
section.
We can make some observations about the relations between weak and strong
deference, and expertise. Deference to another is appropriate (assuming that she
has, like me, a normal endowment of perceptual and cognitive skills) when she
but not I has had access to the relevant evidence—for instance, when she but
not I has had opportunity to exercise normal perceptual judgemental abilities,
as in Case One. Since I have no basis for firm belief in such cases, this will be
weak deference. I learn from the other about something of which I would oth-
erwise be ignorant. Her report informs me, rather than overriding my own prior
firm belief. Thus we can conclude that: Weak deference is often appropriate,
even when the other has no superior expertise to me regarding the topic, she is
merely contingently more expert than me, at this moment.
In contrast, when I and the other both have access to relevant evidence,²⁰
deference to her will be appropriate only if I accept that she has a relevant epi-
stemic power, an expertise, which is superior to mine. Since ex hypothesi we each
have access to relevant evidence, I also have a basis for firm belief myself; so
this will generally be strong deference. In this type of situation a stronger kind
of deferring to another’s epistemic power, her superior authority, is involved. I
accept the other’s judgement as overruling my own, in light of my acknowledge-
ment of her superior epistemic power regarding the matter in question. Our Case
Two above instances such strong deference. Natalie has an expertise relative to
me on events going on in the perceptible distance. Thus she is better epistem-
ically equipped than me to make judgements about what is taking place on the
stage in the park, even though we are standing next to each other, and each able
to look towards the stage. If my vision is as good as Natalie’s, then I will ration-
ally defer to her testimony regarding what went on at the concert only if I was
not there myself. But if I know her sight is better than mine, I may and should
rationally allow her reports of her perceptual judgements to overrule my own
perceptually based judgements, when we are similarly spatially located.
We may thus conjecture that: Strong Deference is appropriate only when
the other has a superior expertise—an intrinsic epistemic power—to me.²¹
This is largely true, although there are two counter-cases. First, it can be that
I have a basis adequate for firm belief, but S, while having no relevant greater
epistemic powers than me, has a stronger one which trumps it. For instance, I
believe that Tom is away on holiday on the basis of my memory of his testi-
mony of three weeks ago; but Chloe testifies to having seen Tom in town today.
It is not determined whether I should accept Chloe’s testimony in these circum-
stance, without further details. But there surely will be some cases of this kind,
where it is right to accept another’s testimony overriding one’s own previous firm
 Elizabeth Fricker 237

belief, because she has had access to fresh evidence, though her relevant epistem-
ic powers are no greater than mine. One factor which would do the trick is if
Chloe’s testimony was independently corroborated by many others. This is our
second type of counter-case: I should bow to others’ testimony about some mat-
ter, even if their skills and evidential position severally are not superior to mine,
if weight of numbers is massively on their side. Notwithstanding these counter-
cases, our conjecture captures a general tendency.

3 . D E F E R E N C E O N M O R A L A N D A E S T H E T I C M AT T E R S ?

We started with the idea that it is rational to defer to another’s apparently sincere
testimony on some topic P just if I recognise that she is better placed than me to
judge whether P (and I am aware of no significant contrary testimony regarding
P). This being so, learning from testimony is possible only in domains where it
makes sense to think that one person can be better placed than another to make
judgements. This in turn requires some notion of objective standards of evid-
ence and correct judgement for the domain in question. If any basis whatever
on which a judgement is made is as good as any other, then the idea of anoth-
er’s being better placed than me does not apply. In fact this restriction imposes
little more than the very idea of judgement imposes in the first place. There is a
determinate content to judge only if there are standards for correct judgement,
independent of what seems to any particular individual to be correct. However, it
could perhaps be that this minimal notion of objectivity applies in some domain,
but for some reason someone else can never be better placed than me to make
judgements about it, or at least it could never be rational for me to believe this.
Some accounts of self-ascriptions of certain conscious mental states would place
them in this category. Exploring the possibility or otherwise of rational deference
to testimony may give fresh insights onto this topic, as well as others, though I
cannot pursue this thought further here.²²
We are investigating the circumstances in which, and topics on which, it can
be rationally permissible, indeed mandatory, deferentially to accept another’s
testimony. Whether and if so in what circumstances deference to others’
testimony on moral and aesthetic matters is ever rational—epistemically and
morally proper—is a large topic, an adequate discussion of which would require
a separate paper. But I shall make a key preliminary point. The kind of objectivity
in standards of judgement which we have just seen to be required—the idea that
there are better and worse ways of arriving at judgements in the domain—is
relatively unproblematic, and no more than common sense, for both moral and
aesthetic judgements. Thus it is only a very moderate thesis to hold that there is
such a thing as superior expertise on moral and aesthetic matters. This view can
be held without commitment to any metaphysically outlandish and epistemically
problematic form of moral or aesthetic realism. The notion of objectivity which
238 Testimony and Epistemic Autonomy

must be invoked need not be understood in terms of correspondence with

wholly mind-independent facts, and more or less accurate means of homing
in on those. (There may or may not be any domains in which we wish so
to understand it!) The required notion of objectivity of judgements in the
moral and aesthetic domains can be explained in terms of better and worse
ways of arriving at judgements of the class in question, with better and worse
defined otherwise than in terms of homing in on mind-independent truth.²³
This being so, there is nothing immediately incoherent, or metaphysically neck-
out-sticking, about the idea of moral and aesthetic experts. The idea accords
with common sense, and normal practices of deference. Aesthetic experts are
those who are specially trained, experienced, and knowledgeable in a certain
area—say Baroque music, or Renaissance painting. One would defer to them
about the qualities, including aesthetic ones, of items in their field of expertise.
The idea of moral experts is equally valid. Some people are especially trained,
experienced, and knowledgeable in the kinds of considerations involved in
making moral judgements. Such expertise may be primarily in a specific field,
involving a particular kind of empirical matter.²⁴ A quasi-realist about the moral,
or aesthetic, can make perfect sense of expertise in these domains. I may defer to
an expert in an aesthetic field because I know my own opinion-forming processes
are crude and uninformed, untrained, in the relevant aesthetic—which amongst
Caravaggio’s paintings are the greatest masterpieces; or what is a specially good
example of an early nineteenth-century English transfer-printed cup and saucer.
(Learning through deference, I may come in time to be a bit of an expert myself!)
I might defer for similar reasons in a moral matter. Or I might defer, or seek
advice here, not because of a general lack of expertise, but because my consultant
is better placed than me—a relative expert—regarding the current matter. She
may know more of the relevant background facts about a difficult case regarding
custody of children in divorce proceedings, or she may unlike me be impartial,
not being involved in the situation as I am. Or I may just want a second opinion,
or to talk the matter over with someone else, as part of the process of forming my
own judgement. Would it be just, or cruel, to carry out my threat and deprive
my son of his Beano,²⁵ for getting into trouble at school for fooling about in class
again? This is the kind of situation where one may want to confer, and maybe
defer to another’s judgement.
A full investigation of the possibilities for, and constraints on, rational
deference on moral and aesthetic matters must canvas more considerations
than those raised here. I suggested earlier that there may be constraints deriv-
ing from the meaning of aesthetic predicates—their tie to a specific non-
judgemental cognitive-cum-affective response in the subject—on the extent to
which deference on aesthetic matters is possible. (One possibility is that only
weak, never strong deference, is rationally possible.) For deference on moral
judgements, there are important ties with the idea of individual autonomy and
responsibility which may place limits. In the present paper I merely wish to
 Elizabeth Fricker 239

point out that the idea of expertise on these topics is an everyday and apparen-
tly sensible one, and thus that deferential acceptance of testimony on these
matters is prima facie rationally possible, as well as being a common occurrence
(see Jones 1999).
We have briefly reviewed the various bases on which another person may some-
times be far better placed than oneself to make judgements about a certain subject
matter. We saw that such superior epistemic status is sometimes based in acci-
dents of location, and may be short-lived; but is sometimes based in intrinsic and
relatively stable differences in epistemic powers between two individuals. When
I appreciate that another person is thus expert relative to me, it is not merely
rationally permissible, but rationally mandatory, to defer to her judgement over
my own conclusions, regarding the subject matter in question. This being so, one
may question whether the supposed ideal figure of the autonomous knower, who
refuses ever to trustingly accept another’s testimony, a fortiori will never allow
her own judgement to be corrected by another’s, is really such an ideal after all. I
will return to this question in my final section. First I address my earlier question:
To what extent can one maintain one’s epistemic self-governance despite one’s
inevitable reliance on others’ testimony, and the technological fruits of others’
knowledge and expertise, in almost every area of one’s life?

4 . R E L I A N C E O N OT H E R S ’ WO R D A N D E PI S T E M I C
S E L F - G OV E R N A N C E

We have seen, as encapsulated in TDAP2, how it is rational to accept another’s

word on a topic, and even to allow her expressed judgement to override one’s
own prior opinion, when one knows that she is strongly placed epistemically,
and better placed than oneself, regarding the matter in question. For each of us,
her appreciation of her own circumscribed and feeble epistemic powers and small
position in the larger scheme of things, together with her grasp of folk psycho-
logy, including where applicable appreciation of others’ superior expertise and
epistemically more advantageous position, entails that deference to others’ opin-
ions is rational, in these circumstances. Lack of such appreciation of one’s limited
powers and others’ superior ones, and an accompanying refusal to bow to others’
judgement or advice even when they are clearly relatively expert, is pig-headed
irrationality, not epistemic virtue or strength.
Does this mean, then, that there is after all no loss of epistemic autonomy
incurred by the way in which, in our modern condition, we rely on others’ know-
ledge and its technological fruits for whole swathes of our fabric of knowledge
and in our daily lives (as sketched in my introduction)? It does not. It is cru-
cial for the maintenance of epistemic self-governance that our trust in the word
of others is given not blindly and universally, but discriminatingly. By trust-
ing only cannily, and with good grounds, we can do much to retain epistemic
240 Testimony and Epistemic Autonomy

self-governance. I shall return to this theme shortly. But there is still an important
loss of autonomy, as I will now explain.
I mentioned our awareness of our own cognitive limitations, our feeble powers.
We can only see what is here and now, and that only to a limited extent. Our
memories even of this are less than total and often corrupt, and our inferential
powers are feeble. A superior being, one who lacked our cognitive limitations,
and could do all the work herself, in finding out about the universe, could be epi-
stemically autonomous in a way that no one of us, with our limited research time
and processing capacities, is able to be. She would not need to take anything on
trust from another’s word, because she would have the epistemic power to check
up, to find out for herself about everything she wanted to know, without reliance
on others. We are not such beings, and so we can extend our knowledge beyond
a small base only through rational trust in the spoken or written word of oth-
ers. My trust in another’s word is rational when I have good grounds to believe
her competent about her topic and sincere, and by this means I can know about
all kinds of matters which I lack the time or talents to find out for myself. But
this knowledge from trust in testimony is knowledge at second hand (or third, or
fourth . . .), and as such my epistemic position vis-à-vis what I know is in at least
one respect inferior to when I know at first hand.
When I form belief that P through my trust in a speaker’s word given to me
that P, her testimony that P, I take her to speak from knowledge. That is, this
is a normative commitment of my accepting her utterance at face value, as an
expression of knowledge. If I come to know she does not speak from know-
ledge, this is a normative defeater for my belief. Additionally, in my own view
of knowledge as requiring adequate grounds, I must be disposed upon reflec-
tion to form the belief that she speaks from knowledge. This belief is an essential
justifying ground for my belief in what I am told and trustingly accept, and so
must itself be knowledge. In short: my reason for believing P true is because
I believe, or am disposed to form belief upon reflection, that my informant is
telling me what she knows. This being so, I know only because someone else’s
knowledge has been passed on, spread to me by the mechanism of telling, of
testimony.²⁶
Knowledge can be passed on in this manner through many links in a chain of
trusted testimony. But the regress must stop eventually with someone who knows
that P not from trust in testimony. The following axiom holds:

T: If H knows that P through being told that P and trusting the teller,
there is or was someone who knows that P in some other way—not in
virtue of having been told that P and trusting the teller.

It is a consequence of T that if someone knows that P through trust in testi-

mony, there must be some other way in which P is or once was known. Hence T
has the corollary:
 Elizabeth Fricker 241

T corollary: For any proposition P that can be known, there must be some
way other than trust in testimony through which P can or once could
be known.²⁷

Why cannot a chain of trusted testimony go in a circle, falsifying T? The

regress must end with someone who knows that P in some other way, because
knowledge requires evidence or grounds. When I know that P from someone’s
testimony, my personal ground for my belief that P, the warrant in virtue of
which I am entitled to it, is my knowledge that my informant knows that P. But
in taking P to be known I am rationally committed to an existential supposition:
that there is, that it is to say that some individual or group of persons between
them possesses, evidence or warrant for P, which is not just that someone they
trust has told them that P. As T expresses, knowledgeable belief based on trusted
testimony implicitly refers back to the existence of a non-testimonial ground or
warrant for what is testified to: the ground or warrant in virtue of whose posses-
sion the original teller(s) spoke from knowledge.²⁸ Hence there cannot be a state
of affairs that is known of only through trust in testimony. A chain of testimo-
nially spread belief which went in a circle would lack any empirical grounding,
and what is believed would not be true unless by luck.²⁹ Consonant with this
fact, there is a sense of ‘the evidence for P’, used in scientific-style discourse, when
it is asked: ‘‘What is the evidence for P?’’, in which someone’s testimony that
P is not evidence for P at all. For instance the question: ‘‘What is the evidence
that smoking causes lung cancer?’’ is not answered by responding: a lot of distin-
guished scientists have asserted that it does. The question asks for an account
of the real evidence, the evidence on which the experts’ conclusion is based.
The well-groundedness of belief spread about through testimony depends on the
existence of such non-testimonial evidence for P—that is, on its possession, per-
haps distributedly, upstream in the chain of informants.
Now we see the respect in which knowledge from trust in testimony is in one
way inferior owing to its being at second hand. When I know that P solely from
trust in testimony, I do not possess the evidence for P. Instead, my knowledge
is premissed on the existential supposition that there is non-testimonial evidence
for P, although I myself do not possess it. I am rationally committed to the pro-
position that a person or persons upstream in the chain of informants between
them possess that evidence—the grounds for believing P true. Where the propos-
ition is an empirical one that is part of a theory, I am also rationally committed to
the proposition that these others have evaluated the evidence and drawn conclu-
sions from it correctly (often, this ability is a large part of a special expertise the
others possess vis-à-vis me).³⁰ Is it a weakness in my epistemic position regarding
P that my ultimate ground for believing P is this derivative second-order one, the
proposition—which I must be disposed to form belief in—that there is empir-
ical warrant,³¹ though unknown to me, for believing P? Where my informant’s
expert status vis-à-vis me is accidental, it does not seem a worry.—My son tells
242 Testimony and Epistemic Autonomy

me there is still some milk left in the fridge, and I believe him. But if it mattered
a lot I could easily check up for myself, and if what he told me were false I would
quickly find out. I can get to the first-hand evidence, if need be, and I can eval-
uate it correctly. But where my reliance on others depends on an expertise they
possess relative to me which is more deep-seated, and I lack the ability to check
up for myself if it seems worth it, the existential supposition and dependence on
others’ epistemic skills and truthfulness is more troubling.³²
Epistemic dependence on others is troubling first because it is risky—there are
many motives for deceit, and causes of honest error, on the part of each of us;
and while each can try to trust only where there is ground to expect sincerity and
competence, as elaborated below, each link in a chain of testimonial transmission
incurs its own risk of error. It is troubling second, because along with the epi-
stemic dependence on others comes a no less risky practical dependence on them,
in many areas—for instance, for maintenance of all the technological devices on
which one depends every day, from electric lighting to computer to driving one’s
car, and so forth. Third, epistemic dependence on others, while it extends one’s
knowledge base so enormously, also lessens one’s ability rationally to police one’s
belief system for falsity. There are many things a layperson believes for which she
would not know how to assess the scientific evidence which supports them, even
if presented with it. This being so, these beliefs of one will lack the characteristic
sensitivity to defeating evidence, should it come along, which is usually taken to
be a hallmark of belief which amounts to knowledge.³³
I have spelled out the bad news for epistemic self-governance entailed by our
dependence on the word of others. The good news is that—as I already emphas-
ized—our trust in others need and should not be given blindly, but cannily, only
where it is due. Although cognitively limited beings as we are, we must perforce
rely on others if we want to enjoy the epistemic and technological riches of mod-
ern society, we can take care only to trust those we have good reason to hold
worthy of our trust.³⁴ Fortunately we all have some basic cognitive equipment to
help us assess both the sincerity and competence of others in many, though by
no means all circumstances. This is because we are all experts (though of varying
degrees of skill) in one special topic, namely that of folk psychology. Thus, where
we do not have access to or cannot evaluate the evidence for propositions in some
domain ourselves, we move one level up, and instead evaluate the experts, our
human sources of knowledge about this domain.³⁵ But assessing an informant’s
trustworthiness is not always easy, and sometimes there are not sufficient epi-
stemic resources available to the layperson to enable a firmly based evaluation to
be made at all. The risks involved in trusting others are considerable, especially
where there are motivations for deception at work. As I have been arguing, there
is often good empirical ground for trusting others, and where so it is consistent
with our maintenance of our epistemic self-governance, our responsibility for our
own beliefs, that we believe on trust in the word of others, relying on their report
for the truth of something where we do not possess for ourselves the evidence,
 Elizabeth Fricker 243

and may not even be capable of appreciating its significance. Moreover, as we saw
in Section 2, where I know another to be epistemically expert relative to me on
a topic, it is not just rationally permissible, but rationally mandatory for me to
accept her judgement in preference to my own, just so long as I have good ground
to trust her sincerity. Where there is not good ground to believe an inform-
ant trustworthy, however, epistemic self-governance entails that we should not
accept the reports of others. Caution and canniness should govern our response
to others’ testimony. Unless we exercise it, we fail to maintain responsibility for
our own beliefs.

5 . T H E I D E A L R EV I S I T E D

I return now to the figure with which I began, the autonomous knower, who
trusts no one else’s word on any matter, hence believes only where she herself pos-
sesses sufficient evidence, non-testimonial grounds, for what is believed. In the
light of the material of the last section we can clarify the autonomous knower
in this way: she never believes on the basis of a second-order warrant for belief,
the belief that someone else knows, someone else possesses evidence showing the
truth of the proposition believed. Is this figure really an ideal? We observed that
a superior being, with all the epistemic powers to find out everything she wanted
to know for herself, could live up to this ideal of complete epistemic autonomy
without thereby circumscribing the extent of her knowledge. Given the risks
involved in epistemic dependence on others we saw in the last part of the previous
section, this superior being is, I suppose, epistemically better placed than humans
are. That is, if she knew at first hand just as much as I myself know in large
part through trust in others’ testimony, she would be epistemically more secure,
hence both practically more independent, and—in some abstract sense—more
autonomous than I am. In the same way that I might regret that I cannot fly, or
live to be 300 years old, I might regret that I am not such a being.³⁶
But what of a human, with no more than human perceptual, physical, and
cognitive powers, who attempted to maintain a regime of complete epistemic
autonomy—that is to say, who never took anyone’s word for anything, and nev-
er deferred to another’s judgement on any matter? We have seen that rational
prudence dictates that one should bestow trust only where it is due; where one
has good grounds to believe one’s informant competent and sincere. But equally,
as encapsulated in TDAP2, where there is good ground to believe another expert
relative to oneself, it is not just rationally permissible, but mandatory, deferen-
tially to accept the other’s judgement. So what would this individual’s beliefs
about others have to be like, for her refusal ever to believe on anyone else’s say-so
to accord with maintenance of a rationally coherent system of beliefs? If ration-
al at all, she would be not an ideal, but rather a paranoid sceptic about others’
intentions and capacities. Or perhaps she would be severely cognitively lacking,
244 Testimony and Epistemic Autonomy

simply lacking any adequate grasp of what other people are, their capacities and
positions in the world—not a master of folk psychology, but an individual sol-
ipsist. She cannot ever admit that anyone else knows anything which she does
not independently know herself since—as we saw—to admit this is to provide
oneself with a second-order warrant to believe that thing oneself (‘A knows that
P’ entails that P, and this entailment is a priori and obvious). One might won-
der also whether she trusts the recorded beliefs of her own past self, as written
down in her personal diaries and other records. The human would-be epistemic
autonome on closer investigation is not an ideal, but either paranoid or severely
cognitively lacking, or deeply rationally incoherent. We all can remember occa-
sions on which someone we know has irrationally refused to change her opinion
in response to testimony from someone evidently better placed to judge of the
matter than she is. The individual autonome carries this irrational tendency to its
irrational extreme.
For each one of us the extent and occasions on which she should accept and
rely on others’ testimony is a delicate matter, decisions about which require
careful assessment on particular occasions. But that there are some occasions on
which it is rational deferentially to accept another’s testimony, and irrational to
refuse to do so, is entailed by her background knowledge of her own cognitive
and physical nature and limitations, together with her appreciation of how other
people are both like and in other respects unlike herself, hence on some occasions
better epistemically placed regarding some matter than she is herself. I may
rationally regret that I cannot fly, or go for a week without sleep without any loss
of performance, or find out for myself everything which I would like to know.
But given my cognitive and physical limitations as parametric, there is no room
for rational regret about my extended but canny trust in the word of others, and
enormous epistemic and consequent other riches to be gained from it.

REFERENCES
Adler, Jonathan (1994), ‘Testimony, Trust and Knowing’, Journal of Philosophy, 91:
264–75.
Blackburn, Simon (1984), Spreading the Word (New York: Oxford University Press).
Burge, Tyler (1993), ‘Content Preservation’, Philosophical Review, 102: 457–88
Clement, Fabrice, Koenig, Melissa, and Harris, Paul (2004), ‘The Ontogenesis of Trust’,
Mind and Language, 19/4: 360.
Coady, C. A. J. (1992), Testimony: A Philosophical Study (Oxford: Clarendon Press).
Davidson, Donald (1984), Enquiries into Truth and Interpretation (Oxford: Oxford Uni-
versity Press).
Descartes, René (1641), Meditations on First Philosophy, in The Philosophical Works of
Descartes, ed. Haldane and Ross vol. i. (Cambridge: Cambridge University Press, 1967).
Fricker, Elizabeth (1994), ‘Against Gullibility’, in B. K. Matilal and A. Chakrabarti
(eds.), Knowing from Words, (Amsterdam: Kluwer Academic Publishers), 125–61.
 Elizabeth Fricker 245

(1998) ‘Self Knowledge: Special Access versus Artefact of Grammar—A Dicho-

tomy Rejected’, in C. Wright, B. Smith, and C. MacDonald (eds.), Knowing Our Own
Minds (Oxford: Clarendon Press), 155–206.
(2002), ‘Trusting Others in the Sciences: A Priori or Empirical Warrant?’, Studies
in History and Philosophy of Science, 33: 373–83.
(2005), ‘Testimony: Knowing through being Told’, in I. Niiniluoto, M. Sintonen,
and J. Wolenski (eds.), The Handbook of Epistemology (Netherlands: Kluwer Academic
Publishers).
(forthcoming), ‘Second-Hand Knowledge’, Philosophy and Phenomenological
Research
Goldberg, Sanford (2001), ‘Testimonially Based Knowledge from False Testimony’,
Philosophical Quarterly, 51: 512–26.
Goldman, Alvin (2002), ‘Experts: Which Ones Should You Trust?’, Philosophy and Phe-
nomenological Research, 63/1: 85–110.
Grice, H. (1957), ‘Meaning’, Philosophical Review, 66: 377–88.
Hardwig, John (1985), ‘Epistemic Dependence’, Journal of Philosophy, 82: 335–49.
(1991), ‘The Role of Trust in Knowledge’, Journal of Philosophy, 88: 693–708.
Jones, Karen (1999), ‘Second-Hand Moral Knowledge’, Journal of Philosophy, 96: 55–78.
Locke, John (1690), An Essay Concerning Human Understanding, abridged and ed. John
(London: Everyman, 1993).
Lyons, Jack (1997), ‘Testimony, Induction and Folk Psychology’, Australasian Journal of
Philosophy, 75: 163–78.
McDowell, John (1994), Mind and World (Cambridge, Mass.: Harvard University Press).
Quine, W. V. O. (1953), ‘Two Dogmas of Empiricism’, in his From a Logical Point of
View (Cambridge, Mass.: Harvard University Press).
Schiffer, Stephen (1972), Meaning (Oxford: Oxford University Press).
Welbourne, Michael (1994), ‘Testimony, Knowledge and Belief ’, in B. K. Matilal
and A. Chakrabarti (eds.) Knowing from Words (Amsterdam: Kluwer Academic
Publishers), 297–313.

N OT E S
1. See McDowell (1994). I am talking here about humans, and how they are psycho-
developmentally able to acquire language-and-thought; not about other logically
possible intelligences, nor the philosophical fiction who springs instantaneously into
existence, a functional replica of a human. ‘Testimony’ here is to be taken broadly,
to include verbal teaching and coaching by others. It would be a mistake to obscure
our dependence on trust in others’ sincerity and competence, in this developmental
process, through a definitional stop.
2. By ‘simple trust’ I mean: trusting response to what others tell or teach us, by one
who as yet lacks the conceptual resources to entertain doubts about the reliabil-
ity of others’ teaching. This is the inevitable initial condition of the infant learn-
ing its first words through interaction with its carers. (However many writers on
testimony exaggerate how long this initial condition persists—don’t underestimate
children—they get wise pretty soon! See Clement, Koenig, and Harris (2004).)
246 Testimony and Epistemic Autonomy

3. A distinction cannot be drawn between analytic versus synthetic, amongst the famil-
iar platitudes involving cluster concepts like these, and so many other of our con-
cepts. See Quine (1953).
4. Mature visual experience is basically the result of the visual system’s best guess as to
what is out there, given the proximal stimulus to the retina. Some of this is hard-
wired and hence culturally invariant, notably the perception of 3-D shaped solid
objects. But perception is also soaked by thicker, culture-specific concepts, so that
its perceptual deliverances to consciousness are much richer: it’s a mobile phone; a
tomato; my daughter—these and their like are typical parts of the content of percep-
tual experience, not inferences from it.
5. Complete epistemic autonomy, as described here, by definition requires not relying
on anyone else’s testimony for any of one’s knowledge. I shall explore whether a
weaker, but crucial, notion of epistemic self-governance—epistemic responsibility
for one’s own beliefs—is consistent with accepting things on other people’s word
for them.
6. That is to say, there is—as will be developed below—a set of conditions regarding
speaker’s and hearer’s circumstances such that both the offering, and the accept-
ance, of testimony on a topic is objectively epistemically appropriate just when
they obtain; so that a speaker gives testimony epistemically properly, and a hear-
er epistemically properly accepts it, when each knows these to obtain. (Where the
speaker or hearer believes justifiedly, but falsely, that they do so, her act or response
is subjectively but not all-in epistemically proper.) See Fricker (forthcoming) for a
supporting account of the speech act of telling. I there show how the nature of the
communicative speech act of telling is crucial to the question when, and on what
basis, the teller may properly be believed—to the epistemology of telling, and testi-
mony more broadly. The qualification ‘epistemic’ to the type of propriety here is not
idle—a telling could be epistemically appropriate, but grossly inappropriate in some
other dimension, e.g. irrelevance, or rudeness.
7. When going for detail some qualifications are needed here. First, for statements
made with sentences containing indexicals, understanding may require grasping an
appropriately related content or proposition, rather than the very same one—same
referent but different senses: ‘‘I’m hungry’’; ‘‘It’s hot here’’, uttered in a telephone
conversation. Second, there can be cases where something is correctly conveyed by
testimony, although the utterance is partly misunderstood; it may be that only the
correctly understood part is believed. See Goldberg (2001).
8. There clearly can be Gricean (Grice 1957; Schiffer 1972) acts of communication
which do not employ language as their medium. There are also non-literal message-
conveying linguistic utterances such as ironic or sarcastic ones. And a speaker may
succeed in getting her message across, be correctly interpreted and believed, despite
using words wrongly in some respect—not in accordance with the constraints and
permissions of the literal meaning of the sentence she mistakenly employs. Even
where communication of what is literally asserted in the speech act occurs, presup-
positions and conventional implicatures may be conveyed too. These acts all share
with paradigm tellings the successful getting across of a message. I shall not investig-
ate here the respects in which they differ; except to say that where what is conveyed
 Elizabeth Fricker 247

is not explicitly asserted there is, I believe, a diminution in the responsibility for
the truth of what is got across incurred by the utterer. This is one reason to reserve
the term ‘telling’ for acts of communication via explicit assertion exploiting literal
meaning, as is done in ordinary parlance.
9. I am being careless here about distinguishing sentence types from particular utter-
ances of them effecting speech acts, and the role of context in fixing what precise pro-
position a particular utterance of a sentence expresses. This matter, though crucial
and pervasive in natural language, is tangential to the current point. The fastidious
reader may imagine the necessary complicating adjustments.
10. ‘‘There are very beautiful pictures in the Uffizi in Florence, though I have never seen
them.’’—this sounds deviant to my ear. As opposed to ‘‘There are said to be very
beautiful pictures in the Uffizi, though I have never seen them’’. On the other hand
‘‘There are famous paintings by Botticelli in the Uffizi, though I have never seen
them’’ sounds fine.
11. I form belief that P on the basis of trust in another’s testimony that P, when I do so
because I take her utterance at face value, as an expression of her knowledge that P.
In so doing I take her word for it that P. There is a variety of other cases where a
hearer forms belief that P in response to observing testimony that P, which are not
cases of trust in that testimony. Fricker (forthcoming) contrasts these cases with the
case of trust in the testimony, and argues that the latter relatively narrow category
is the key epistemic kind to discern, in theorizing how knowledge can be spread by
means of testimony. The condition proposed in TDAP for forming belief in what is
stated would not be correct, for a broader category. Rather, it further characterizes
the narrow category.
12. There is scope for further refinement here: it could be that an informant is very
unlikely to form a belief that P which is not knowledge; but is more prone to error,
or careless judgement, than not-P. This kind of one-sided reliability is quite plausible
in some cases—e.g. someone who is slow to make a judgement of guilt of anoth-
er—and a hearer could be aware of this epistemic disposition of an informant. But
more usually, someone will be in this way reliable regarding P only if she is also
similarly reliable regarding not-P. TDAP as formulated specifies this stronger con-
dition. Perhaps someone could be self-deceived, so that she in some sense ‘really
knows’ that P, while kidding herself, and telling others, that not-P. TDAP concerns
knowledge expressed in conscious judgement, and so excludes repressed knowledge,
if such is possible.
13. In contrast with her competence, or expertness as I am here calling it, I think that
one is entitled to presume a speaker sincere, unless there are specific cues or other
evidence calling this into question. This fact is not an epistemic principle special to
testimony, but is fall-out from correct general principles governing the ascription of
mental states to other persons. See Fricker (1994).
14. The issues here are delicate. Mere weight of numbers of concurring testifiers does not
per se increase the probability of correctness; it depends on the details regarding the
likely explanation of how they have come to hold their expressed beliefs. See Gold-
man (2002) for an excellent discussion of what epistemic resources a layperson may
have, to decide which to trust out of two experts giving contrary testimony.
248 Testimony and Epistemic Autonomy

15. For instance regarding various future matters: the weather, currency and interest rate
movements, etc. Here one should defer to and act on the basis of the best advice;
while being aware that it is not knowledge—hence hedging one’s bets accordingly.
16. An account such as Wellbourne (1994), which holds it sufficient for the recipient
of testimony to come to know, that the hearer speaks from knowledge, is purely
external and as such violates my requirement. However accounts like Burge (1993)
or Coady (1992), which maintain an entitlement to trust testimony as such which is
however defeasible, can be seen instead as proposing a specific thesis regarding how
internal rationality is satisfied, in this case—albeit one with which I disagree. (In the
case of perception, it is plausible that epistemic responsibility permits one to take
one’s senses on trust, unless aware of defeaters.)
17. This is why one knows so little about what goes on during one’s children’s days
at school!
18. Can I know this, without begging the question—ungroundedly trusting her testi-
mony over mine? Certainly: I have found on many previous occasions that what she
has judged from a distance proves correct, as we get nearer. The fact that expert-
ise is time and circumstance relative, often transient, means that another’s epistemic
expertise relative to oneself can often be conclusively established by oneself, despite
one’s own inferior epistemic power.
19. Expertise of S has been defined as relative to another person, H. But we can easily
extract a more general concept of expertise, which is a superior epistemic power
regarding some topic relative to all those without the specialist training or skills in
question—the layperson or non-specialist.
20. The notion of ‘access to relevant evidence’, and certainly of two persons having equal
access to relevant evidence, is fraught with difficulty, given the theory-dependence of
one’s observational powers—as my cricketing example above illustrates. It does not
bear much theoretical weight in the present argument, and all I require is that there
be some cases where it clearly applies, and others where it clearly does not. I intend
that it holds of Case Two, and similar situations.
21. Is superior expertise also normatively sufficient for strong deference? No, since two
people both with superior expertise to me may supply contrary testimony. Apart
from this, I cannot think why else it should fail to be.
22. Fricker (1998) argues, on precisely this point, that accepting the possibility of cor-
rection of one’s self-ascriptions of mental states made through avowal, by other
evidence from one’s behaviour, which might be pointed out to one by others, is a
condition for one to be ascribing a genuine concept in these self-ascriptions.
23. I here make a large, but unoriginal claim, which requires at least a fat book for
adequate defence. I have in mind positions like the ‘quasi-realism’ of Blackburn
(1984).
24. The inextricable interweaving of fact and value in the considerations relevant to a
final conclusion on a complex matter reinforce this point. Consider, for instance, the
members of a panel appointed to draw up proposed legislation controlling research
using human embryos. Both scientific and moral expertise are required, and intel-
ligent conclusions rest on inextricable understanding of both. Another example of
a specific partly moral expertise is making decisions about when children should be
taken away from their parents and into care.
 Elizabeth Fricker 249

25. The Beano is a popular comic-strip magazine for children, in the UK.
26. Knowledge requires grounds, and if I trust a speaker who tells me something true but
does not herself know it, my own belief will be based on a false premiss and so not
be knowledge. This is the general conception of knowledge I favour, and my account
of knowledge from testimony is shaped by it. Even if a different view of necessary
conditions for knowledge is taken, that the speaker knows what she tells is clearly a
rational commitment of a belief based on trust in testimony.
27. The ‘other ways’ may however include deduction, induction, or inference to the best
explanation from premisses some of which were supplied by diverse bits of testi-
mony. See Fricker (forthcoming). The tense qualification is important here—the
original informant may have since died, or simply forgotten what she once knew and
told to others.
28. T and its corollary do not imply the stronger claim: For any P which is known,
there is someone who knows it in a way which has no epistemic dependence on
testimony. This stronger claim is false, as is explored in Fricker (forthcoming). The
source of testimonially spread knowledge that P may have learned some of the
facts from which she inferred P from others’ testimony. Thus the ultimate, non-
testimonial evidence for any complex theoretical proposition may be possessed only
distributedly, by the members of a group. See Hardwig (1991).
29. This remark remains true, but needs careful explanation, when we are dealing with
facts constituted by human practices—the boundary between two countries, what
something’s name is, and so forth. The testimony itself would not make the belief
true, but enough people acting on belief in it would do so.
30. My belief is premissed on these suppositions not in the strong sense that I must
occurrently believe them; rather, they are normative commitments of my form-
ing belief on trust in testimony. As such, I must come to acknowledge them if
talked through it—and my trust is normatively defeated if I come to believe any
of them false.
31. I say ‘warrant’ here rather than ‘grounds’, since there are some types of belief—e.g
some beliefs regarding one’s own mental states, and perhaps basic perceptual
beliefs—which are empirically warranted, but not by grounds for belief.
32. Epistemic dependence of this sort is explored in a series of seminal articles by Hard-
wig (1985, 1991). Hardwig suggests the schema: ‘H has reason to believe that S has
reason to believe that P → H has reason to believe that P.’ The schema only holds of
prima facie reason, however—I could know that S has reason to believe that P, while
myself being aware of defeaters for those reasons. Our present point is that the reasons
in question are different. As I have been emphasizing, the ground for belief supplied
by trust in testimony is a second-order one. My reason to believe is that I believe
that my informant knows that P, hence that she or someone upstream of her has a
non-testimonial warrant to believe that P. My original source’s reason to believe is
this non-testimonial warrant, the evidence for P.
33. A further point is that once the original source of a testimonially spread belief is no
longer available, the original warrant for the belief is no longer retrievable. However
this feature characterizes most of our beliefs. Cognitively limited beings that we are,
we generally form a belief from the evidence, then store the fact in memory and
250 Testimony and Epistemic Autonomy

jettison the evidence. The lack of sensitivity to potentially refuting new evidence is,
in contrast, a risk of testimonial belief only.
34. It should be abundantly clear by now that I am against all accounts of how know-
ledge may be gained through testimony which do not require that the recipient
trusts only where she has good grounds to do so. They are inconsistent with the
requirement of individual rationality, that epistemic self-governance in the sense of
responsibility for policing one’s beliefs for truth, is maintained by the individual,
the thinking, believing, and acting subject. A rational individual cannot delegate this
responsibility to others, although as I am elaborating here, the requirement can be
discharged by moving up a level: evaluating the reporters, when we are unable to test
their reports for truth directly.
35. In Fricker (1994, 2002) I discuss how non-question-begging evaluation of the dual
components of a speaker’s trustworthiness, her sincerity and competence, is often
possible. See also the excellent discussion in Goldman (2002).
36. No heavy commitment to the coherence of the conception of this superior being is
intended or incurred. I use her merely as a heuristic device in the development of my
argument.
 PART V
N EW AR E A S A N D N EW
DIRECTIONS IN THE
E PI S T E M O LO G Y O F T E S T I M O N Y
This page intentionally left blank
 11
Pathologies of Testimony
C. A. J. Coady

I am going to deal with some of the ways in which various forms of telling things
to others come under moral or epistemic suspicion. Since I’m considering this
against the background of what is (according to me) our very deep reliance upon
testimony and the associated trust in others that accompanies it, then I am think-
ing of these ways as pathologies, in that they present as distortions of or diseases
of the normal case of telling and relying on what is told. But, as we shall see, they
have very different morphologies and there is a real question about how much of
a distortion they might be. In particular, there is a question whether they deserve
(all of ) the moral and epistemic odium or suspicion that is their usual lot.
The phenomena I shall discuss are gossip, rumour, and urban myth. Another
obvious candidate is lying but I shall not treat of it here. Lying is an interest-
ing phenomenon with many different facets of philosophical interest, but since
it arises from the deliberate intention to deceive an audience by saying what
the speaker believes to be false it is too obvious a pathology of testimony to be
treated in a paper concerned with more ambiguous candidates for that title. A
more interesting candidate in the present context is the phenomenon of ‘‘spin’’
but reasons of space will prevent my treating of that; it must await another occa-
sion. Gossip, rumour, and urban myth are different in many ways, most notably
perhaps in their relation to the truth and the position of the speaker with regard
to truth. As already noted, lies are distinguished at least by the intention to say
what is believed by the speaker to be false. By contrast, it is plausible to think that
gossip is standardly sincere, and may be true and known to be true. Rumour
may be true and believed to be true but the justificatory base for speaking it
is weak, and urban myth is more legend than rumour—it is more frozen and
immune to refutation, but it can function in a similar way to more ephemeral
rumours. Initially, I shall treat these phenomena as though they involved mere
transmission of propositions, but of course this is a considerable simplification
and even abstraction from the reality. Gossip, rumour, and especially urban myth
are highly narrative in form; they are presented in a dramatic mode, sometimes
even in song or poem, and they often contain, explicitly or implicitly, strong
interpretive and evaluative elements. Here, as elsewhere, the picture of transmis-
sion as the passing on of a lump of information in a single line of transference
254 Pathologies of Testimony

from individual to individual is misleading. Nonetheless, the misleading picture

reflects something right, namely, the fact that testimony begins with some form
of witnessing. This need not be a perceptual encounter though that is a primary
case of witnessing. It can also consist in a proof or an expert judgement, as I
have previously argued (Coady 1992: 48, 51–62). Here I mention the point only
to deal with one claim about gossip that seems wrong. It is made by Laurence
Thomas who gives the following example as a clear case of gossip:
Austin: I can’t believe that they are awarding her the Nobel prize in physics. Must be
because she is a woman.
Lee: To tell you the truth I have been thinking the same thing myself.
(Thomas 1994: 47)

Thomas thinks this is ‘‘surely gossip’’ (Thomas 1994: 48), but I think it is better
described as malicious speculation. I suspect Thomas treats this exchange as obvi-
ous gossip because he thinks that gossip, as in the example, must express negative
feelings about the subject of the gossip. As we shall see, this negativity need not be
a feature of gossip, but, in any case, I am interested in gossip as a possibly degen-
erate form of testimony, and the exchange between Austin and Lee makes no
pretence of being a testimony transaction. Austin’s ‘‘message’’ is a speculation; he
has witnessed nothing relevant in even the most extended sense of ‘‘witnessing’’.
These are points to return to.
Thomas might reply that he doesn’t think of gossip as a form of testimony,
even a degenerate form. It is true that there are usages of ‘‘gossip’’ that are wider,
and perhaps looser, than the use I am interested in, uses in which any casual
exchange or conversation about anything whatever may count as gossip. Some-
times people think this way when they talk of a ‘‘gossip session’’ or ‘‘shooting the
breeze’’.¹ Just chatting about one’s own exploits or commenting, speculating, and
guessing about aspects of the known exploits of others could then be gossip, but
in the sense of the term that interests me you cannot gossip about yourself and
conveying guesswork and so on doesn’t count as gossip. There is then a degree of
stipulation about my approach, but any initial characterization of the topic will
involve some stipulation and will express particular theoretical interests. I claim
on behalf of my account that it captures a great deal of what people normally
mean by gossip (and of rumour and urban myth) and illuminates significant mor-
al and epistemological scenery.
Let us begin with gossip. There is quite an extensive literature on gossip, some
of it philosophical, but a lot of it sociological. There are disputes about how to
define the topic, as we have already seen, so I will not offer a tight definition, but
rather give some necessary marks of the concept of gossip as I understand it. I
shall take it that gossip has the following features: it is usually conveyed by those
who believe it to be true, it can be the transmission of perfectly justified beliefs,
it need not be malicious though it sometimes is. Its subject matter is invariably
personal, though of course it can be about persons who represent or are thought
 C. A. J. Coady 255

to typify groups. It is also transgressive, in at least the sense that the transmission
of the information can be presumed to be unwelcome to the subject of it. Some
theorists claim that gossip is distinguished by the triviality of its subject matter.
It is, they say, essentially ‘‘idle talk’’. It is understandable that this should be said,
but it remains contentious. After all, gossip about someone’s job prospects or
manoeuvrings, about their adulterous conduct, about a political leader’s financial
dealings are all matters that may have momentous consequences for the individu-
al concerned and for others affected by his or her actions. Acknowledging this,
Rosnow has amended the claim to make it more plausible by concentrating on
the style of communication rather than its content. It is the setting and context
of the conversation that must have at least the appearance of the casual. As he
puts it, the exchange should be ‘‘characterized by a kind of belle indifference’’. The
talk should be ‘‘packaged’’ to appear as idle (Rosnow 2001: 210). This is indeed
more plausible and seems to capture something typical of gossip. It is reflected in
the casual, bantering tone adopted by many newspaper gossip columnists. One
philosopher agrees that an idle tone is characteristic of gossip but takes this to
be a moral defect of it. She says, ‘‘It is characteristic then of gossip to fail to give
matters their due regard; gossip often involves a mismatch between the tone and
substance of the discussion. Such a mismatch may simply reveal superficiality or
it may constitute a failure of empathy and moral understanding’’ (Holland 1996:
203). I shall deal with this criticism later, but it is worth noting at this point
that much of the literature on gossip is concerned with its moral status. Most
of the criticism of gossip is moral rather than epistemic. This suggests that gos-
sip exhibits no essential epistemic difference from ordinary testimony; there is no
reason why gossip should be less reliable than ordinary testimony. In what fol-
lows, I shall try to assess the moral objections to gossip, but also raise some points
about its possible epistemological deficiencies. There is one more mark of gossip
that needs to be mentioned, namely, that there is something essentially restricted
or intimate about at least the initial range of gossip. This point is hard to make
clearly, but the basic idea is that the natural home in which gossip begins is that of
a small group, though, of course, the information may spread from small group
to small group until the news becomes widespread. Someone is not gossiping
who shouts the information from the rooftops for all and sundry to hear.
The debate about the moral status of gossip (sometimes cast as a debate about
its positive or negative social impacts) is basically between those who think that
gossip violates some significant moral constraint, such as respect for persons, and
those who think that gossip has an important social role. These latter sometimes
ignore the supposed violation, but sometimes acknowledge it and think that it is
outweighed by the good effects.²
First, the moral objections to gossip. Some of these proceed as though gossip
must always be malicious. If this were true, then there would indeed be a black
mark against it from the beginning. But it is clearly not true, at least if the malice
in question is seen as stemming from a motivation to do harm. Many gossips
256 Pathologies of Testimony

have no interest in maligning the object of their talk, they are just interested in
the buzz of the topic and the status their communication gives them. Indeed,
gossip need not retail anything disreputable about its object. I may gossip about
something you are doing that is either neutral or positive with respect to your
reputation, as when I retail some confidential facts about your soon to be success-
ful house sale or the secret that you are about to be elevated to the peerage. What
seems true is that there must be something unwelcome to the gossipee about the
revelation you are conveying. The subject of the gossip must at least be presumed
not to want the information in question conveyed either generally or to that par-
ticular audience.³
I shall proceed as if this unwelcome aspect is essential to gossip, though there
are in fact several counter-examples to an unqualified version of this claim. Some
very trivial newspaper ‘‘gossip’’, about, for example, a society wedding may totally
lack the unwelcome aspect, but for this very reason we may not want to treat it as
gossip, or not as paradigmatic gossip. It is more like advertising or public rela-
tions offered under the pretence of news. More interestingly, there is a range of
cases where the subject of the gossip may positively welcome the spreading of the
information, and may indeed be the source of it, yet the information be both sig-
nificant and somehow transgressive. Consider the case of academic gossip about
who is to be offered a new position. If we think of the successful candidate as
either the source, or as least the cheerful subject, then the unwelcome thesis seems
imperilled.⁴ There are two possible responses here. One is to shift focus on the
subject of the gossip so that the gossip is about the actions of the appointments
committee, and only secondarily about the status of the successful candidate.
Members of the committee will find the gossip unwelcome—either in reality or
by the conventions of confidentiality to which they are obliged to adhere. Anoth-
er alternative is to say that the candidate may actually welcome the spreading of
the news, but must take the stance of finding the transmission unwelcome. To
adopt a useful test for gossip, we should ask: would the conversation stop or suffer
embarrassment if the candidate suddenly entered the room? If the answer in the
present case is, ‘‘Yes’’ then there seems a sense in which the candidate must be
viewed as finding the transmission unwelcome whatever he privately thinks.
It is this unwelcome feature of gossip that provides one significant purchase for
the moral critic. Granted that gossip need not be retailed with the aim of dam-
aging the subject of it, might it nonetheless count against any proposed good in
gossip that it must run counter to the subject’s desire that the information remain
confidential or at least restricted? One philosopher has argued that, because of
this, gossip violates the Kantian injunction to treat others as ends in themselves
and not merely as means. Prima facie, there is some plausibility in this. The gos-
siper need have no regard for the good of the person whose fortunes they report,
the facts about them are recounted (at best) merely for the pleasure it gives the
speaker and the audience, and possibly for the regard the gossiper achieves by
reason of his/her knowledgability. Does this violate the Kantian injunction? This
 C. A. J. Coady 257

is one of the points at which the injunction is hard to interpret. Am I treating

someone only as a means if I drive tourists by their handsome house to show
them a fine example of the sort of Victorian architecture typical of the area? And
suppose I know that they don’t want their house to be the object of tourist gaz-
ing? What this suggests is that the test of ‘‘failing to treat someone as an end’’
cannot rule out all cases of acting against their wishes, even where these wishes are
concerned with their privacy. But, perhaps we can only allow such disregarding
where the disregard is aimed at achieving some good, and the fact is that gossip
has no such purpose.
There are two problems with this. The first is that gossip arguably does have
such a purpose, or purposes, and the second is that, even if it doesn’t, the subject’s
desire for privacy, confidentiality, or secrecy may be unreasonable. The social
psychologists that give a positive account of gossip often stress that it has a role
in solidifying social norms. The idea seems to be that gossip about the behaviour
of others takes place in an explicit or implicit context of norms and the evalu-
ation commonly invoked (Oooh, she didn’t! Gosh, fancy that! . . .) has the effect
of reinforcing those norms. As Rosnow puts it: ‘‘the consequences of rumour and
gossip also reflect the extent to which social norms can be enforced on individuals
who ostensibly threaten or violate them. In small groups, for example, gossip may
be a way of shepherding the herd by saying ‘These are the boundaries, and you’re
crossing them’ ’’ (Rosnow 2001: 224). Rosnow here seems to be running two dif-
ferent things together. On the one hand, the effect of reinforcing the norms for
the group involved in the retailing and hearing of the gossip, and on the other,
the enforcing of the norms on the person gossiped about. The latter will not be
the direct effect of the gossip since that person, in the first instance, at any rate,
doesn’t hear it. The former may well take place and lead to the latter.
The reinforcement thesis raises two further puzzles. The first concerns the gen-
erally favourable attitude of social scientists to this reinforcement (of either kind),
and the second concerns the empirical claim itself that the attitude rests upon.
On the first, it is strange (as Holland has pointed out) that the cementing of
existing norms is simply taken for granted as a positive outcome. Perhaps this
reflects the persistence of the simple-minded cultural relativisms that have been
so influential, and so damaging in the history of anthropology and related stud-
ies. Whatever the explanation, it is obvious that existing social norms may be
either good, bad, or indifferent from a moral perspective. The reinforcement
provided by gossip may be to solidify widespread immoral prejudices against
blacks, women, homosexuals, foreigners, immigrants, or whatever, and where
this is so, it constitutes an objection to the role of gossip not an endorsement
of it. As to the second, the role of gossip may just as easily be the breaking down
of social norms as the protection of them. In societies with rigid public codes,
gossip may serve to show the hypocrisy of those who proclaim and enforce them,
thus leading to the destruction of respect for those elites and even of the codes
themselves. So, gossip about the sexual behaviour of various influential leadership
258 Pathologies of Testimony

groups may lead to the erosion of strong sexual censorship regimes, and gossip
about the perks and lifestyle of Communist leaders may have played a part in
undermining public confidence in the professed values of fraternity and equal-
ity. It is also possible that both reinforcement and subversion can occur together
(though directed towards different prevailing norms). Pre-Revolutionary France
provides a case in point. In a society in which formal news outlets were stereo-
typed and tightly controlled by those in power, the primary source of inform-
ation about the King and the ruling elites came from gossip and rumour. As
Robert Darnton has argued in his fascinating book, George Washington’s False
Teeth, gossip and rumour about the French Court not only provided a rare source
of personal and political information but thereby possibly played a part in prepar-
ing the ground for the Revolution (Darnton 2003: ch. 2). It did this by exposing
the hypocrisy of the Court and the unworthiness of the King and his dependence
on the intrigues of others. Some of the underground transmission was classic-
al gossip retailed at selected sites such as the Tree of Cracow in the grounds
of the Palais-Royal. Other was more rumour, though sometimes beginning as
gossip, and other again was song and coded poetry circulated by paper notes or
memorized word of mouth. Again, as with the reinforcement story, the question
of whether the undermining is good or bad requires independent adjudication.
More broadly, the question whether gossip has good or bad outcomes cannot be
given a global answer. It all depends on the context and the actual outcomes.
We cannot, however, rule out in advance the possibility that the outcomes will
be good.
Another good outcome claimed on behalf of gossip is the way gossip
can improve individuals’ self-regard. By a process of what some have called
‘‘downward comparison’’ we may come to think better of ourselves by comparing
our behaviour or character to that revealed of some other person by gossip (Suls
and Goodkin 1994: 173; Ben-Ze’ev 1994: 19). The revelation that some famous
person, for instance, has feet of clay may reassure us about the normalcy of our
own feet. Again, the two questions posed earlier arise regarding this claim. First, it
is not clear that bad information about others is a particularly sensible or healthy
way of achieving a sound understanding of one’s own worth. Self-congratulation
can result in self-delusion. Second, the deflating information about the worth
of others may equally well give rise to the sense that one should lower one’s
own standards rather than take pride in maintaining them. Either way of course
there may be some solidifying effect of the gossip exchange; the deluded may be
mutually bound by their delusions and a group may find some agreeable cohesion
in mutually endorsed lowering of expectations of self and other. One possible
connection of gossip with norms is that whereby the conveyors of gossip may be
led to reassess their moral views by finding that their audience is not as shocked
by the gossip as the speakers had expected. Gossip may therefore assist in what
could be a useful form of self-criticism.⁵
 C. A. J. Coady 259

A further potentially positive feature of gossip that is perhaps too obvious to

be remarked upon by social psychologists is the fact that many people find the
telling and receiving of gossip enjoyable. Most people are curious to know the
truth not only about the physical world or mathematics but also about the deeds
and misdeeds of other people. Sometimes this curiosity has a functional point in
orienting us towards the people with whom we are going to interact: this is the
idea that social psychologists are getting at with their somewhat simplified talk of
norm reinforcement and the like. But sometimes the satisfaction of the curiosity
is simply fascinating in itself, even if it can also be useful. We are interactive social
beings who spend a great deal of our lives in conversation, much of it about oth-
er people. News about their journey through life with its pitfalls and triumphs
is intrinsically interesting to most people, and it is often even more interesting
when we know or believe that we are not going to hear it from them.
I have claimed that gossip is pleasurable for many of us, but there are those
who find gossip of any sort uninteresting, even distasteful. They inject a chill of
disapproval into the gossip room and often produce a cessation of the activity
until they depart. Thereafter, the gossip resumes and often expands to include
them as subjects! It would be interesting to explore this difference in attitude
and some of its nuances, such as those who are happy to hear gossip but not
to retail it, but I shall not pursue this matter here. Of course, it is necessary to
stress that when I refer to the pleasure of gossip as good, I do not mean that it
is an unqualified good. Pleasure is at best only one good among others; it can
be outweighed by other goods or harms, and some pleasures may be inherently
illicit, as are those (to take an extreme case) of paedophilia.
This leads us to the claim that there is something intrinsically bad about gos-
sip that may preclude consideration of the good outcomes (if any) or at least may
need to be weighed against them. Here, we should consider two of the issues
raised earlier. First, there is Holland’s claim that ‘‘It is characteristic then of gossip
to fail to give matters their due regard; gossip often involves a mismatch between
the tone and substance of the discussion. Such a mismatch may simply reveal
superficiality or it may constitute a failure of empathy and moral understanding.’’
Gabrielle Taylor makes a similar claim, though less forcefully. ‘‘Gossipers’’, she
says, ‘‘trivialize experience by ignoring the impact with which the author of the
experience will in some way have to cope. Thus they distort and belittle that per-
son’s experience’’ (Taylor 1994: 46). On the face of it, both claims seem excess-
ively high-minded. Gossip is not always concerned with matters of moment, and,
even where it is, the requirement that tone should match subject matter smacks
of puritanism. The idea that serious subjects should only be addressed in a sol-
emn tone of voice presents an affront to the complexity of human communicative
intercourse, especially to the dimensions of humour, irony, satire, and playfulness
that are rightly characteristic of much of it. Of course, tone may reprehensively
mismatch content, as with some sneering discourse, and this may well exhibit
260 Pathologies of Testimony

superficiality or moral failure of some sort. My only point is that, contrary to

Holland, the tone characteristic of much gossip need not display moral fault, nor
need it trivialize what should not be so treated.
Second, there is the issue, already briefly discussed, that gossip is standardly
unwelcome to the gossipee. It is natural to think that this shows it to be in viola-
tion of respect for persons, and hence immoral. As mentioned above this depends
to some degree on why the conveying of the information is unwelcome. There
is no absolute right that information about oneself be withheld from others.
Respect for privacy cannot reach that far. Perhaps my not wanting facts about
myself to be made known to others creates some presumption against doing so,
but it may be rebuttable in a variety of circumstances. Nor need these circum-
stances require some very great good to rebut the presumption. To give a personal
example. I was once in a group of philosophers who were criticizing another more
famous philosopher in ways that seemed to me to impugn his character as well
as his philosophical standing. I had discovered from his publisher that the author
under attack had actually assigned all his royalties from a very successful book to
charity and proposed to do so again for the second edition. When I mentioned
to the author my knowledge of this fact and my respect for his decision, he asked
me not to tell anyone about his altruism. But in the company of his somewhat
carping critics, I thought it right to do so and did. I think I was justified in act-
ing in this way, and, if so, the example shows that the gossipee’s desire to keep
the matter secret need not be determinative of the moral status of the gossip. It
also shows that Bertrand Russell was wrong to claim that ‘‘no one gossips about
other people’s secret virtues’’ (Russell 1994: 50). It further suggests that it does
not require the prospect of some overwhelming good to validate the gossip trans-
action, since the author in question is fairly unconcerned with what people think
of him.
Another interesting phenomenon is that of what might be called remedial gos-
sip. By this I mean that phenomenon where B gossips about A to C, D, and E and
then E tells A that B has been telling them about A. We might be reluctant to call
E’s transmission ‘‘gossip’’ because it is aimed at alerting A to the gossip about him
or her. But it may well be offered in the characteristic tone of gossip and E may be
well aware that B, C, and D would be unhappy to have the matter relayed back to
A. Nor need E’s motives be particularly noble.
I said that gossip is usually conveyed by those who believe it to be true and
who may well be justified in their belief. This feature makes it seem that gossip
is not, from the point of view of epistemology, a ‘‘pathology’’ of testimony at
all. This idea is worth closer examination. One problem with it is that some of
the motives typical of gossip may be fragile with respect to truth-transmission.
The desire to titivate, or to show one’s ‘‘insider’’ status as someone with access to
otherwise unavailable information is not the sober background we expect of sol-
id testimony. Bertrand Russell is thinking in this way in the section from which
the earlier quotation was taken. He says, ‘‘The widespread interest in gossip is
 C. A. J. Coady 261

inspired, not by a love of knowledge but by malice: no one gossips about other
people’s secret virtues, but only about their secret vices. Accordingly most gossip
is untrue, but care is taken not to verify it. Our neighbour’s sins, like the consol-
ations of religion, are so agreeable that we do not stop to scrutinise the evidence
closely. Curiosity properly so-called, on the other hand, is inspired by a genuine
love of knowledge.’’
I have argued that gossip is not necessarily malicious. If it were, then the epi-
stemic objection would carry some weight since an informant who is inspired
by malice will be seeking to put another’s behaviour in a bad light and this
may well bias their testimony. But even gossip that conveys negative informa-
tion about a person need not be inspired by emphatic malice; it may be mere
excitement at discovering the faults or follies of another and an interest in being
the one to know and convey such stuff. Curiosity can be as lofty a disposition as
Russell claims, but it need not be. It can be more earthy without becoming posit-
ively malicious and it may still retain an orientation to knowledge. Leaving aside
malice, a disdain for other common motives for gossip faces two problems. One
is that various forms of legitimate testimony, even of formal testimony in courts,
may be subject to similar motivations. Witnesses conveying information that has
no taint of violation may nonetheless be moved by the desire to cut a figure in
great events and may delight in the excitement their news will create. They may
also be moved by a concern to get things right—any account of what inspires
reliable testimony must allow for the existence of mixed motives. Second, even
if the motivations associated with gossip are not of the loftiest, they need not be
destructive of a concern for the truth. The desire to impress may indeed make one
more scrupulous in conveying accurate information since one’s status as a reliable
informant may be crucial to the positive appreciation one seeks.
Another epistemic objection to gossip is that it excludes by its very nature an
important source of epistemic reliability, namely, confirmation, falsification, or
correction by the subject of the gossip.⁶ If we accept that absence of the subject is
a necessary concomitant of gossip then it is indeed true that the gossip situation
deprives the gossipers of an additional direct checking resource for the informa-
tion. But I doubt that this is as drastic an epistemic flaw as it initially seems. In the
first place, where the information originates from the subject, and our informant
is known to be close to the subject and a generally reliable witness who has no
particular stake in lying, it may simply be redundant, verging on the neurotic, to
insist on checking with the subject. (Consider the parallel with those in the grip
of mild obsessive-compulsion who need to return home to check by perception
their firm memory that they have turned off the iron or locked the back door.)
There are always further checks one can do on testimonial information, just as
there are with observational or remembered information, but it is epistemically
redundant to do so in many cases. In the second place, there will usually be little
reason to believe that the subject of the gossip can be relied upon to provide valu-
able confirmation or disconfirmation. After all, since the subject does not want
262 Pathologies of Testimony

the information spread at all, he or she is most likely to deny it or refuse con-
firmation, whether the facts are as reported or not. In the third place, where the
gossip does not originate from the subject, it may well be that the subject is in
no position to confirm or deny the information. There is an interesting range of
cases here. Consider the case where information is that you are to be awarded
some honour and the source is an ‘‘insider’’ in the conferring process. You may
know nothing of the fact at the relevant stage, or you may know only that there is
some prospect of the honour. So your exclusion from the gossip circle creates no
epistemic problem for the reliability of the information.
This example is also interesting as a problem for the unwanted criterion of gos-
sip since it is presumably the people in the honouring process (or most of them)
who don’t want the information spread, even though it is information about you.
Nonetheless, you are likely to want to know about it before the gossiper and audi-
ence get to hear about it. If, however, you are so unusual as not to care about
that, then you would not be discomforted at being present during the gossip
session. This, I take it, counts strongly against the transmission being gossip, at
least about you—it may count as gossip about the honour process and honour
bestowers.
I conclude that this epistemic objection to gossip fails as a general objection,
though there may indeed be occasions on which closing off access to the sub-
ject’s own testimony has epistemological disadvantages. More generally, gossip as
I have defined it is not a pathological form of testimony but a normal form of
it. Whatever its moral standing, and I have argued that this may be less dismal
than usually thought, its intellectual status is reasonably respectable. In epistem-
ic terms it may be likened to whispered information rather than openly spoken
word of mouth.

RU M O U R

Gossip and rumour are often run together in social science treatments, but they
are basically different in kind though there are some areas of overlap. At any rate,
I shall treat them as differing in the way suggested at the outset, namely with
respect to the justificatory base of the information conveyed. Gossip may be true
and known or justifiably believed to be so, rumour has by (my) definition no
strong justificatory base. The typical way to introduce a rumour is to say, ‘‘Have
you heard . . .?’’ whereas the typical introductory mode of delivering gossip is,
‘‘Did you know . . .?’’ The gossiper may, of course, convey false information and
be mistaken about the strength of the justification he has for the information.
But he must present himself as being in an authoritative position with respect
to the information. Must he always believe that he is telling the truth and in an
epistemic position to do so? Can’t there be deliberately deceitful gossip? Sissela
Bok thinks so, indeed she lists ‘‘gossip that is known to be false’’, as one of the
 C. A. J. Coady 263

three main categories of gossip (Bok 1983: 98). But she, like many others, makes
no real effort to distinguish gossip from rumour. To my ear, the phrase ‘‘lying
gossip’’ does not ring true; the more accurate description of the activity is just
‘‘spreading lies’’. In any case, it is untypical of gossip. Rumour-mongers, by con-
trast, may well deliberately create false rumours. In wartime, the spreading of false
rumours can be a crucial political weapon, as can be the combating of damaging
rumours. Moreover, rumours may well begin in sheer speculation, though they
will usually mimic testimonial transmission by conveying the idea that someone
somewhere is a witnessing source.
A further difference is that rumour is not restricted by topic to the personal. It
may be about institutional, political, religious, or physical events. You can spread
a rumour about an earthquake but you cannot gossip about it. A further differ-
ence is that gossip is normally quite restricted in its circulation. A small group
may get together to gossip about a colleague or a boss or whomever, but have
no desire to spread the information further. (Gossip columns in newspapers are
an exception to this, though even here there lingers some element of the closed
group point, inasmuch as the readers of the column are meant to have a sense of
being inducted into a privileged, if rather large, group.) Rumour, however, seems
essentially prone to run abroad, ‘‘to spread like wildfire’’. Of course, one area of
overlap is that something that begins as gossip may well continue as rumour. This
is probably one of the reasons for the widespread failure to distinguish between
the two.
Given rumour’s indifference to a secure basis for its reports, can anything
positive be said on its behalf? Even raising this question may seem perverse, in
the face of the bad reputation so widely enjoyed by rumour. But it should be
remarked that the enemies of rumour sometimes have an agenda that is not
itself entirely respectable. I mentioned earlier the attempt to control and coun-
teract rumours during wartime. The American psychologist Gordon Allport, for
instance, was one of those who set up rumour clinics to help curb the spread of
rumours in Massachusetts during World War II. He was later the co-author of
an influential study of rumour (Allport and Postman 1965). Allport was helped
in the anti-rumour campaign by his doctoral student Robert H. Knapp who
defined rumour as ‘‘a proposition for belief of topical reference disseminated
without official verification’’ (Knapp 1944: 22). This definition would count
every newspaper exposé of government misdeeds or secrets that met with offi-
cial silence or denial as rumour even if they were true and thoroughly justified.
This makes nonsense of our ordinary understanding of rumour and reflects an
ideal of official control of information that is, to put it mildly, undemocratic. It is
not surprising that it might arise in a climate of war, but those of us who believe
that war is frequently unjustified can hardly take consolation from this. Where
governments are bent upon the suppression of information, for whatever reason,
then ‘‘propositions for belief’’ unendorsed by the authorities will assume greater
significance.
264 Pathologies of Testimony

One thing that might be noted as an epistemic merit is the power of rumour
in providing hypotheses for further exploration. By itself a rumour may be poor
epistemic coin, but investigating it may lead to expanding one’s knowledge in
direct and indirect ways.⁷ The direct route is that of confirmation or falsification
of the rumour’s contents. The indirect route may be the discovery of interesting
information that explains the rumour’s genesis, or it may be the discovery of a
genuine truth that the rumour misleadingly presents. So we may discover that
a rumour to the effect that a rogue politician is selling nuclear secrets to North
Korea against his own government’s policies and desires is false, but also discover
the related truth that a respected senior scientist is selling the secrets with his own
government’s connivance. Dismissal of reports because they are ‘‘mere’’ rumours
can sometimes block opportunities for discovering important truths. One of the
most striking instances of this was the rumour emanating from Tokyo in 1941
concerning an impending attack on Pearl Harbor. The US Ambassador Grew
reported in January 1941 from Tokyo a rumour that the Japanese Navy was plan-
ning a surprise attack on Pearl Harbor. It was attributed to the Peruvian Embassy
which was regarded by the ambassador and others as ‘‘a not very reliable’’ source.
The rumour was dismissed by one and all, including the Ambassador, as fantastic.
As it turned out, the date the rumour began coincided with the inception of the
Yamamoto plan for bombing Pearl Harbor (Shibutani 1966: 73–4).
In light of these positive aspects to rumour, it might be claimed that those who
spread a rumour are not so dastardly as they are usually thought to be. Indeed,
with all its faults, the poor coinage of rumour may often be the only inform-
ation currency available. Some social scientists exploring rumour argue that it
commonly arises in response to alarming situations where reliable information
is cut off, either deliberately by censorship or repression, or accidentally by con-
fusion or disaster. In such situations, rumour tends to be a focus of high social
interaction. The ‘‘news’’ conveyed is not simply passed on from one voice to
another, like the passage of a ‘‘brick’’ of information. The item is often specu-
lated upon, criticized, amplified, compared with other related items of rumour,
and with what pieces of hard news or testimony are available. In addition the ori-
ginator of the rumour, if indeed there is only one such source, need not have had
anything like observational access to the supposed fact that is spread. He or she
may have produced the information as a speculation in the absence of any reli-
able testimony or observation. Even so, the rumour mill will usually present the
rumoured facts as something that is somehow sourced in someone’s observations,
though it will usually be vague about who that is. Shibutani’s discussion of the
World War II rumours about the reduction in training and furlough time for US
troops in Georgia in late 1944 exhibits this sort of pattern. The rumours were
fuelled by the hard news of the increased German counter-offensive in Europe
and the absence of any real information about the effect this might have on these
newly inducted troops (Shibutani 1966: 9–14).
 C. A. J. Coady 265

Psychological investigations of rumour tend to divide between those with an

individualistic and rather static picture of transmission and social psychologists,
like Shibutani, who see the picture in more dynamic and social functionalist
terms.⁸ The former tend to be hostile to rumour and the latter more sympath-
etic. The former proceed by setting up highly artificial experimental scenarios
in which a piece of information is to be passed on to specified others (often
the experimenter) and changes in the literal form of the message are noted for
inaccuracy. This way of proceeding is open to a number of objections, some of
which I have raised elsewhere (Coady 1992). The artificiality of the experimental
situations rules out many of the standard ways of confirming, sifting, and inter-
preting the original informational setting; the emphasis on the literal preservation
of a sentence or set of sentences is an unnecessary restriction on the conveying of
information; the message in question often has little significance for the witness,
and so on. By contrast, the social psychologists emphasize the fluidity of real life
situations, the need for information in information-starved settings, the relation
of rumour to practical thinking and decision-making rather than speculative reas-
on and abstract truth. Indeed, they are inclined to minimize the issue of truth.
So Shibutani says: ‘‘rumor is not so much the dissemination of a designated mes-
sage as the process of forming a definition of a situation’’ (Shibutani 1966: 9).
The cultural historian Hans-Joachim Neubauer describes it as ‘‘a mirror in which
society catches sight of its hidden self’’ (Neubauer 1999: 174). These emphases
are no doubt important for the empirical study of rumour, but it should be recog-
nized that the dissemination of a designated message is perfectly compatible with
forming a definition of a situation or mirroring social psychology. The interest
in realistic social settings should not be seen as excluding the concern for the
relation of rumour to truth or to epistemic justification.
So is rumour pathological testimony? It seems that it is a form of testimony
because it involves the transmission of propositions from one or more persons to
others, but it often lacks what I have elsewhere claimed to be definitive of testi-
mony. There are sometimes no original sources in even the attenuated senses that
I noted in my book since rumour can arise from the merest speculation. Further-
more, the speaker of rumour will often have no competence with regard to the
‘‘information’’ conveyed and may be well aware of that. If we think some degree
of authority or competence, no matter how minimal, is a precondition for giv-
ing testimony then quite a lot of rumour will be disqualified as testimony. There
is a very amusing Australian radio show called ‘‘The Coodabeen Champions’’ in
which several witty and very well-informed people comment on sporting events,
especially Australian Rules football. In one segment of the show there is a ‘‘talk-
back’’ session in which members of the public (in fact, the commentary team
with disguised voices) raise topics for discussion. One regular always rang with a
rumour, this being the show’s acknowledgement of the fact that sporting culture
is rife with rumours about secret injuries, the sacking of coaches, the transfer of
266 Pathologies of Testimony

players, and so on. ‘‘Tony, have you heard the rumour’’, he would begin. When
pressed on his credentials, it would emerge that his brother’s wife’s uncle knew a
bloke who had a neighbour who heard someone in his dentist’s waiting room say
he thought he’d heard that if St Kilda suffered one more loss, the coach would be
sacked. With credentials as thin as this, we may well want to treat such rumours
as not testimony at all.
Nonetheless, they have at least the superficial appearance of testimony. They
are not completely unrelated to it, so if we decide that rumours are not really
or fully forms of testimony we will be treating them as what J. L. Austin called
‘‘misfires’’ rather than ‘‘abuses’’ of the speech acts of testifying. Austin’s early dis-
cussions of ‘‘performative utterances’’ identified a range of infelicities to which
they could be prone in contrast to their failing to be true or false. The idea of
infelicities carried over to his more sophisticated discussion of illocutionary and
other speech acts. The concepts of misfires and abuses were forged to show differ-
ent ways in which speech acts could be infelicitous. Misfires make the purported
speech act go wrong in a way that nullifies it, whereas abuses constitute real but
irregular performances of the act. To use one of Austin’s examples, it is a plausible
precondition of successfully marrying that there exists an institutional framework
within which the words ‘‘I do’’ serve to effect a marriage. But if, for example,
you are already married (in a monogamous society) or the officiating officer is not
really a licensed clergyman or appropriate civil official then you have merely gone
through a form of marriage. The act has misfired, the marriage is void (Austin
1961: 223–8). By contrast, if you utter the vows of fidelity in the appropriate
circumstances with no intention of keeping them you are still married, even if
you are a genuine cad and the marriage hardly a paradigm of what it should be.
Similarly an insincere promise may not be the best sort of model for promising,
but it is still a promise and your commitment can be held against you when you
fail to perform. Later, in How to Do Things with Words, Austin elaborated the
distinction and produced sub-groups within misfires and abuses. Similarly, we
might say that the rumour-monger’s lack of credentials makes his testimony void,
as testimony, but he has nonetheless gone through a form of testimony; com-
pare the marriage ceremony performed by a bogus clergyman or official. This case
of misfired testimony would count as a misinvocation in Austin’s terms (Austin
1962: 14–20).
This gives us grounds for treating rumour as a pathology of testimony and
giving a more precise sense to what that means.

URBAN MYTH

Urban myths have much in common with rumours. Indeed, in many respects,
they could be seen as a type of rumour. But they have many distinctive features
that make it plausible to treat them separately. They have, for instance, much
 C. A. J. Coady 267

higher levels of narrative complexity. The ‘‘testimony’’ they offer is presented in

the form of a fully fledged story where a rumour is usually rather less developed or
embellished. A further difference is the abiding nature of the urban myth. Urban
myths tend to survive their implausibility and even refutation in specific cases,
and re-emerge, sometimes in slightly different dress, at another time and often
place. Quite frequently, they have a kind of moral lesson attached to them, like
Aesop’s Fables though not as explicit, and their revival is often triggered by issues
of contemporary interest. In December 2002 there was an interesting episode in
Tasmania, Australia, where it was widely reported that a truck driver had given
a lift to a man of Middle Eastern appearance who left his wallet in the cabin
after alighting in the capital city, Hobart. The honest truck driver called after the
man who said to him in a rush of gratitude, ‘‘Don’t go near Salamanca Place on
Saturday.’’ Salamanca Place is a popular market area near the waterfront and is
always crowded on Saturday mornings. There was an American naval ship in the
harbour. My son, David Coady, a philosopher at the University of Tasmania, was
contacted by the media for comment and pointed out that this was a classic urban
myth that had already surfaced elsewhere in slightly different forms. His view was
publicized widely, panic was averted, and (fortunately for him and the reputa-
tion of philosophers) nothing alarming happened on the morning in question.
The truck driver story combines a moral element of virtue rewarded with a more
general threatening element.
Then there are the many instances of humorous tales of vindication or revenge.
A classic is the one that the avid collector of urban myths, Jan Harold Brunvand,
has christened ‘‘the $50 Porsche’’ though the type of car varies with different
national and cultural settings (Brunvand 1999: 77; 1981). One version of the
story (in bare bones) goes as follows: Someone is reading the newspaper ads for
used cars and sees an ad for a very recent model Mercedes (Porsche, Rolls Royce
. . .) available for sale for $50. He rings the number and a woman answers and
says that it is truly available at the price. He goes to the address given and a beau-
tiful woman comes out of an expensive house and shows him a spotless, perfectly
maintained Mercedes only two years old. As he hands over the cheque, he asks
how it could possibly be sold at such a price. She replies that her husband has
recently left her for a younger woman but contacted her just the other day, in a
spirit of ‘‘surely we can still be friends’’, to ask her to sell his posh car since he
needed an injection of funds, and to send the money on to him. This she was now
doing!
There are other tales that build on primal fears. One is the spider’s eggs myth.
Invariably the victim in the story is a woman. In basic outline the story goes that
she has gone far off for a holiday, but before (or during) the holiday notices a
small sore or lump on her face. At some time on the holiday, the sore swells and
bursts, and a whole lot of baby spiders come out. Various embellishments con-
cern whether she knows she was bitten by a spider or it happened while she was
asleep, etc. Sometimes she dies of a heart attack. This story seems to have begun
268 Pathologies of Testimony

in Northern Europe as late as 1980, but spread to venues in North America and
elsewhere.
The field of stories that are called urban myths is very wide and contains much
variation, but I shall take it that what I call ‘‘urban myth’’ is invariably false
and ill-founded, though commonly enough believed to be true. One qualifica-
tion that is needed to this is that a true but dramatic and surprising story, based
on reliable testimony, may transmute into an urban myth with a life of its own.
A case in point is the mathematician’s story: ‘‘The Unsolvable Math Problem’’.
The original true story concerned an episode in the career of the mathematician
George B. Dantzig who, as a student at Berkeley in 1940, arrived late for a class
given by Jerzy Neyman and noticed that there were two problems on the black-
board. Thinking they had been assigned for homework, he copied them down
and went away and worked on them for several days. He then returned to his pro-
fessor and apologized for taking so long, but they had been rather hard. He asked
whether his teacher still wanted them and was told to put them on the desk. Six
weeks or so later, he and his wife were awakened by banging on his door early
one Sunday morning. It was Neyman saying that he had just written an introduc-
tion to the homework and wanted Dantzig to read it so he could send the lot off
to a journal for publication. The homework had in fact been put on the board
merely to show the class two famous, unsolved mathematical problems in statist-
ics. Various versions of the true story were then in circulation for years afterwards
as, for instance, the subject of sermons showing something or other about God
or human character. In the course of this, they changed shape in all sorts of ways,
even appearing in the film ‘‘Good Will Hunting’’. So this is a case of numer-
ous fictional accounts having sprung from one apparently genuine happening
faithfully recorded by Dantzig and, in part, easily checkable against the public-
ation by Datzig of his solutions.⁹ Another form of truth for urban myth might be
provided by life imitating fiction, if, for example, some deserted wife found her-
self in the advantageous position described in the ‘‘$50 Porsche’’ story and took
instruction from it! But with these exceptions noted, it is part of qualifying for
the title ‘‘urban myth’’ that the tale is false, though for any given story we may
not know on hearing it whether it is urban myth or not. Being familiar with a
number of urban myths, however, is a good recipe for recognizing others. This is
partly because many of them merely ring changes upon the familiar ones, but also
because there is a certain narrative style common to many urban myths. I have
mentioned Aesop’s Fables and another analogy is with biblical parables except
that they are not intended to be believed as historical accounts and are often
prefixed by phrases that indicate as much, such as ‘‘the kingdom of heaven is
like . . .’’ Often, urban myths play upon certain widespread fears or phobias, and
there are some types of myth that invoke the wildly improbable or supernatural.
Like rumours, it is usually quite unclear how one could check on the sources of
the story. Where there are links in the chain of apparent testimony they quickly
come to a halt well short of a reliable witness.¹⁰ Is urban myth a pathology of
 C. A. J. Coady 269

testimony? Well, that depends on the degree to which it is presented as testimony

or as fiction. Usually, the stories will be told as if true and based somehow on
report, but sometimes they flourish merely because they are a good story. So, urb-
an myth is a pathology of testimony in much the same way as rumour, though
perhaps closer to fiction.
What general lessons should be drawn from this discussion? The first is that
whether some communication is a degenerate form of testimony cannot be
simply read off from the form or content of its telling, though either may give
clues to this fact. A communication may have the usual hallmarks of an urban
myth, for instance, and we may be wise to treat it as such, but it may turn out
to be a case where truth is stranger than fiction and there is a reliable testimonial
path to that truth. It seemed like an urban myth but it wasn’t. The second is that
a pathological testimony need not be altogether worthless. The communication
may rightly be judged unreliable but turn out to be true, as with certain rumours
or lies. Liars can accidentally say what is true while meaning to say what is false
and our knowledge of the background to the liar’s performance may enable us to
spot this and profit from it. Somewhat similarly, a rumour-monger can hit upon
the truth without being entitled to any confidence that it is true. This may indeed
have been the case with the rumour emanating from the Peruvian Embassy about
the planned attack on Pearl Harbor; the ambassador could have been right in
his assessment of the general reliability of sources in the Peruvian Embassy. On
the other hand, it might have been his assessment of their credibility that was at
fault and the communication may not have deserved the title ‘‘rumour’’. Once
we have reason to characterize some communication as rumour, then equally we
have reason to treat it as lacking credibility. It may, nonetheless, deserve further
investigation, if the stakes are high enough, just in case it is true, or with the
prospect of discovering some truth that it distorts. Once we establish that some
communication is urban myth then we can be sure that it is false (with the
minor qualifications noted earlier) but it remains possible that it is a revealing
fiction. The matter is different with gossip. Once we have established that some
communication is gossip, the question of its reliability and truth is (pace Russell)
much the same as that of any other piece of testimony.

REFERENCES
Allport, Gordon W., and Postman, Leo (1965), The Psychology of Rumor, 2nd edn. (New
York: Russell & Russell).
Austin, J. L. (1961), ‘Performative Utterances’, in his Philosophical Papers (Oxford:
Oxford University Press).
(1962), How To Do Things with Words, lecture II (Oxford: Oxford University
Press).
Ben-Ze’ev, Aaron (1994), ‘The Vindication of Gossip’, in Robert F. Goodman and
Aaron Ben-Ze’ev (eds.), Good Gossip (Lawrence, Kan: University Press of Kansas).
270 Pathologies of Testimony

Bok, Sissela (1983), Secrets (New York: Vintage Books).

Brunvand, Jan Harold (1981), The Vanishing Hitchhiker: American Urban Legends and
their Meaning (New York: Norton).
(1999), Too Good To Be True: The Colossal Book of Urban Legends (New York:
Norton).
Coady, C. A. J. (1992), Testimony: A Philosophical Study (Oxford: Oxford University
Press).
Darnton, Robert (2003), George Washington’s False Teeth: An Unconventional Guide to
the 18th Century (New York: W. W. Norton and Co.).
Holland, Margaret G. (1996), ‘What’s Wrong with Telling the Truth? An Analysis of
Gossip’, American Philosophical Quarterly, 33/2.
Knapp, Robert H. (1944), ‘A Psychology of Rumour’, Public Opinion Quarterly, 8.
Neubauer, Hans-Joachim (1999), The Rumour: A Cultural History, trans.
Christian Braun (London: Free Association Books).
Rosnow, Ralph L. (2001), ‘Rumour and Gossip in Interpersonal Interaction and Bey-
ond: A Social Exchange Perspective’, in Robin M. Kolwaski, (ed.), Behaving Badly:
Aversive Behaviors in Interpersonal Relationships (Washington, DC: American Psycho-
logical Association).
Russell, Bertrand (1994), On Education: Especially in Early Childhood (London: Rout-
ledge) (first published 1926).
Shibutani, Tamotsu (1966), Improvised News: A Sociological Study of Rumor (New York:
Bobbs-Merrill).
Sousa, Ronald de (1994), ‘In Praise of Gossip: Indiscretion as a Saintly Virtue’, in
Robert F. Goodman and Aaron Ben-Ze’ev (eds.), Good Gossip (Lawrence, Kan Uni-
versity Press of Kansas).
Stern, William (1902), ‘Zur Psychologie der Aussage’, Zeitschrift für die gesamte Stra-
frechtswissenschaft, 22.
Suls, Jerry, and Goodkin, Franklin (1994), ‘Medical Gossip and Rumor: Their Role in
the Lay Referral System’, in Robert F. Goodman and Aaron Ben-Ze’ev (eds.), Good
Gossip (Lawrence, Kan University Press of Kansas).
Taylor, Gabrielle (1994), ‘Gossip as Moral Talk’, in Robert F. Goodman and Aaron Ben-
Ze’ev (eds.), Good Gossip (Lawrence, Kan University Press of Kansas).
Thomas, Laurence (1994), ‘The Logic of Gossip’, in Robert F. Goodman and
Aaron Ben-Ze’ev (eds.), Good Gossip (Lawrence, Kan University Press of Kansas).

N OT E S

1. Tom Campbell tells me that there is a Glaswegian phrase for this: ‘‘having a good
hing’’.
2. A few are positively euphoric about the merits of gossip, most notably Ronald de
Sousa (1994: 25–33). He seems to me far too dismissive of the downside of gossip
and he exhibits a sort of romantic exaggeration of its benefits. He says ‘‘If all truths
became public we would approach utopia’’ (1994: 31). This blatantly defies the fact
that there is sometimes a significant value in privacy and confidentiality.
3. What I am calling the ‘‘unwelcome’’ aspect of gossip might be captured in other
ways. It might be thought that my criterion turns too much on the psychology of
 C. A. J. Coady 271

the gossipee or on what the gossips know of that psychology. David Rodin sugges-
ted instead that the information should be ‘‘unauthorized’’. That might do, though
there are then problems about who or what does the authorizing. I prefer to stick
with ‘‘unwanted’’ and I note in the text and try to deal with some of the difficulties
this choice raises. I suspect that there may be other ways of making the transgressive
point that would demarcate a closely similar conceptual territory for our discussion.
4. This objection was put to me by Margaret Coady.
5. I owe this point to Ruth Zimmerling.
6. I owe this perceptive objection to John O’Neill of Lancaster University.
7. Somewhat similar points could be made about ordinary lies.
8. For the more static picture see William Stern ‘Zur Psychologie der Aussage’, in Zeits-
chrift für die gesamte Strafrechtswissenschaft, 22 (1902) discussed in Neubauer (1999:
157–8) and see also Allport and Postman (1965).
9. According to Brunvand, one of them is in the Annals of Mathematical Statistics,
1951. See Brunvand (1999: 452–6) for the full details of what he has christened,
in a slightly misleading way, ‘‘The Unsolvable Math Problem’’.
10. An exception is the virgin birth tale from the American Civil War (see Brunvand
(1999: 469–72) where the story is called ‘Bullet Baby’) but it originated in an
ingenious hoax by a reputable medical practitioner and is therefore not a typical
urban myth.
 12
Getting Told and Being Believed
Richard Moran

Recent interest in the epistemology of testimony has focused attention on what

justification we may commonly have in the vast areas of life where we are depend-
ent on what other people tell us. This dependence is not restricted to what we
are told in face to face encounters, for we also take ourselves to know all sorts of
things that only reached us through a long chain of utterances and documents,
whose evidential status we have never investigated for ourselves and which we will
never be in a position to investigate. And the content of such knowledge is not
confined to the arcana of specialized studies, but includes such mundane matters
as the facts of one’s own birth and parentage, the geographical and institutional
facts of one’s immediate environment, and the ‘general facts’ that make up one’s
basic sense of what the world is like.
In part it is the enormity of this dependence that makes for the interest in
the subject of testimony, combined with the apparent clash between the kind
of epistemic relations involved here and the classic empiricist picture of genuine
knowledge basing itself either on direct experience of the facts, or on working
out conclusions for oneself.¹ It isn’t just that the bulk of what we take ourselves
to know is so highly mediated, as even knowledge gained through a microscope
or other scientific instrument must be; rather it is that the vehicle of mediation
here—what other people say—seems so flimsy, unregulated, and is known in

This paper has a long history and I have accumulated more than the usual share of debts along the
way. In presenting both this and related material, I have benefited from responses of audiences at:
Wake Forest University, Arizona State University, University of Wisconsin, Milwaukee, New York
University; the Center for Human Values at Princeton University, Vassar College, University of
Connecticut, Storrs, Rice University, Johns Hopkins, Stanford University, Macquarrie University,
Sydney, University of Pennsylvania, Amherst College, University of Virginia, Columbia University,
University of Chicago, and the University of Minnesota. Earlier versions of this material were
presented at the Third Italian American Philosophy Conference, Rome and Frascati, June 2001,
and at the Instituto de Investigaciones Filosóficas, UNAM, Mexico City. I am grateful to my hosts
on those occasions.
I also received very helpful comments from Jonathan Adler, Tyler Burge, Kate Elgin, Paul
Faulkner, Luca Ferrero, Amy Gutmann, Paul Harris, Pamela Hieronymi, Karen Jones, Adam
Leite, Wolfgang Mann, Peter de Marneffe, Ed Minar, David Owens, Connie Rosati, Angus Ross,
Nishi Shah, Martin Stone, David Sussman, Michael Thompson, David Velleman, Jonathan Vogel,
Matthew Weiner, Daniel Weinstock, as well as the comments of two anonymous referees.
I am grateful to the editors of The Philosopher’s Imprint for permission to reprint.
 Richard Moran 273

plenty of cases to be unreliable, even deliberately so. People do lie, get things
wrong, and speak carelessly. And while we may realistically hope for continued
improvement in the various technical means of epistemic mediation (advances
in scientific instrumentation are part of the history of scientific progress, after
all), there is little reason to expect that the fallibility and mendacity associated
with human testimony will one day be overcome. So in this light, reflecting on
just how much we rely on the word of others, we may conclude that either we
are very careless believers indeed, with no right to claim to know more than a
fraction of what we think we know, or some great reductionist program must be
in the offing, tracing this chain back to something resembling the classic picture
of knowledge by acquaintance.²
Hume’s famous discussion of the believability of reports of miracles is the locus
classicus for attempts to understand the epistemic status of testimony as ultimately
the same as any other reliable evidence.³ And part of what is meant by this claim
is that the basis we may have in any given case for believing what we hear can
only be an a posteriori judgement to the effect that in this case there is a reli-
able evidential correlation between the statement we are being offered and the
facts themselves. Several recent writers, most notably C. A. J. Coady in his book
Testimony,⁴ have argued that the Humean picture cannot succeed in reconstruct-
ing our actual basis for believing what people say, and that our entitlement to
believe what we are told must have, in part, an a priori basis. Somewhat lost in
much recent discussion, however, is attention to the basic relationship between
people when one person tells a second person something, and the second per-
son believes him. This is the primary everyday occurrence, and it is the basic
way knowledge gets around. Or at least, so we say. For normally (though not
without exception) we take it to be sufficient for bringing someone to know that
P that they were told by someone who knew, and they believed him. And now,
of course, if this second person is taken to know that P, he may tell another per-
son, and so on. This may seem absurdly simple and unreflective, and to be at
odds with an earlier picture of genuine knowledge as being more of an achieve-
ment lying at the end of an arduous path from belief or opinion. My concern in
this paper, however, is not so much with the conditions for knowledge as with
the nature of the two sides of the relationship described here. One person tells the
other person something, and this other person believes him. I want to understand
what ‘telling’ is, especially as this contrasts with other things done in (assertor-
ic) speech such as persuading, arguing, or demonstrating; activities which may
also lead to belief or knowledge for the interlocutor, but in importantly different
ways. And primarily I want to examine the relation of believing where its direct
object is not a proposition but a person. For in the basic case described above, it
is the speaker who is believed, and belief in the proposition asserted follows from
this. These are different epistemic phenomena. For the hearer might not believe
the speaker at all, taking him for a con man, but yet believe that what he has said
274 Getting Told and Being Believed

is in fact true. Whereas when the hearer believes the speaker, he not only believes
what is said but does so on the basis of taking the speaker’s word for it. I don’t
mean to suggest that this distinction has been wholly ignored in the literature of
testimony, and I will soon come to discuss what I think is the best recent dis-
cussion of it. But both it and the distinction between the speech act of ‘telling’
and other things done with assertion have not been given a central place in the
discussion of what is distinctive about the epistemic dependence on testimony.
Specifically, I wish to argue that any account of testimony that seeks to resist the
(Humean) assimilation of its epistemic status to that of an evidence-like correla-
tion between one set of phenomena and another will have to give a central place
to the distinctive relation of believing another person.⁵ Only in this way can we
account for what is distinctive about acquiring beliefs from what people say, as
opposed to learning from other expressive or revealing behavior of theirs. The
hope is to show that the paradigmatic situations of telling cannot be thought of
as the presentation or acceptance of evidence at all, and that this is connected
with the specifically linguistic nature of the transfer of knowledge through testi-
mony (which will take us through an epistemological reading of Grice’s original
account of non-natural meaning).

EV I D E N T I A L R E L AT I O N S A N D T H E A P R I O R I

It is in part due to the epistemological context of recent discussion of testimony,

that argument has focused on the question of the a priori or a posteriori status
of our justification for beliefs acquired in this way. In recent work, both C. A.
J Coady and Tyler Burge have argued against a broadly Humean picture, by
attacking the idea that we could only have a posteriori justification for believ-
ing what others tell us.⁶ Coady presents more than one argument against the
Humean idea, but several of them begin with the following strategy. If we can
only have a posteriori grounds for taking what people say to be a reliable guide to
the facts, then on such a view it must be conceivable for there to be a community
of speakers whose assertions bore no reliable relation to the facts. If we are to be
in a position to deploy an a posteriori argument for the existence of such a cor-
relation, it must be possible for us to begin confronting the linguistic evidence
without begging that very question. Coady presents a powerful and connected set
of arguments for the conclusion that this is not, in fact, a coherent possibility.
Not only would the practice within the community of making or accepting asser-
tions soon break down on such assumptions, but from a perspective outside the
community there are deeper reasons connected with the interpretation of speech
which prevent the Humean scenario from being realizable. For assigning con-
tent to the utterances of the hypothetical speakers requires, for familiar reasons,
regular correlation between assertive utterances and the conditions under which
they would be true. Massive disparity between the content we assign to utterances
 Richard Moran 275

and their truth or rationality would oblige us as interpreters to revise our original
assignments of content to them. So there is, in fact, no genuine possibility of a
community of speakers whose assertions failed, as a general rule, to correlate with
the facts. And thus, contrary to the Humean picture, our general justification for
believing what people say cannot be a purely empirical, a posteriori one.⁷
Both arguments direct themselves against the idea that we have, at best, empir-
ical, inductive grounds for believing what people say. I don’t dispute this general
point or the particular way it is argued for in these two instances, but do want
to point out that this general form of argument describes no particular role for
the notions either of a speaker telling someone something, or of believing that
speaker. What the generality of such arguments provides is a defeasible a priori
warrant for believing that what other people say will normally be true. But any
argument pitched at that level of generality will leave untouched the question of
whether believing the person (as opposed to believing the truth of what is said)
is a legitimate, and perhaps basic, source of new beliefs. For we might well have
an a priori defeasible warrant for accepting the beliefs we gain through observing
the behavior of others (verbal and otherwise) without this warrant involving the
concepts of ‘saying’ or ‘telling’ at all. By itself, such a justification is no different
from the presumptive right we may have (ceteris paribus) to rely on the deliver-
ances of the senses or of memory. At this level of argument, the speech of other
people could still be something which is treated as evidence for the truth of vari-
ous claims about the world, the difference would only be that here we may have
some non-empirical right to treat this phenomenon as evidence, perhaps even
very good evidence.⁸ This general line of thought begins, then, to look more like
a non-skeptical version of the basic Humean view, and less like a vindication of
testimony as a distinct source of beliefs, one not reducible to a form of evidence.
And yet it is the special relations of telling someone, being told, and accept-
ing or refusing another’s word that are the home of the network of beliefs we
acquire through human testimony. And these relations, I hope to show, provide
a kind of reason for belief that is categorically different from that provided by
evidence.
Another way of putting this criticism would be to say that arguments of the
generality of Coady’s do not address the question of what is distinctive about
acquiring beliefs from what people say, as compared with other things people
do. At bottom, the epistemological role of communicative speech is not seen as
essentially different from that of other behavior. But the observable behavior of
other people may be a source of true beliefs in all sorts of ways, which need have
nothing to do with believing the other person. I may look out my window on a
sunny day and see people bundled up against the cold, and then reliably conclude
from this that it must be colder outside than it otherwise looks. This transition in
thought is not essentially different from the picture according to which I observe
the verbal behavior of some exotic community, and in seeking to understand
what it means, I necessarily rely on various assumptions about their rationality
276 Getting Told and Being Believed

and general awareness. And here one could point out that the same ‘rationalizing’
or charitable constraint on understanding what these people say also provides a
defeasible warrant for taking what we understand them to say to be true. This is
because we take their speech normally to express their beliefs, and we take their
beliefs (as interpreted by us) normally to be true. This familiar, general scheme
applies in the same way to the behavior of the people I see bundled up against the
cold, and to the verbal behavior observed by the Radical Interpreter. Pictured in
this way, one’s relation to the exotic speech community does not involve being
told anything at all, or believing them, any more than it does in the case of the
people observed from the window. In both cases it is just a matter of an inference
from behavior which is seen as rational to some conclusion about the state of the
world. So nothing along these lines, justifying the beliefs we acquire from other
people, can count as a vindication of our reliance on testimony, since it is not a
vindication of what we learn through believing other people.
This is, of course, the familiar role of speech and its relation to belief in con-
temporary philosophy of mind, and it should not be surprising to see it exerting
a degree of control over the recent discussion of testimony. Within this discourse,
speech is seen as a kind of interpretable human behavior like any other. When
we interpret such behavior, we seek to make it understandable within the rational
categories of what is called ‘folk psychology’, and ascribe beliefs and other atti-
tudes which will be reasonable approximations to the True and the Good. And
this picture of our relation to the speech of other people leads almost impercept-
ibly into a view about testimony. For we can argue from here: when we interpret
the speech of another we do not only learn about the speaker, we also learn about
the world. Most obviously, when someone makes an assertion we may not only
learn about what he believes, but if the assertion is true, we may also learn the
truth of what is asserted. And if our interpretation is guided by principles of char-
ity, we will indeed take most of what people say to be true, even in cases where we
have no independent reason for thinking it true. In this way, the fact of the other
person’s belief (as interpreted by us) may function as our reason for believing the
same thing. We thus gain true beliefs about persons as well as about the world
they are talking about.

PE RV E R S I T Y, D E PE N D E N C E , A N D R I S K

What this general scheme provides us with is a presumptive right to share the
beliefs we take the speaker to have. But, other things being equal, we would have
the very same right however we learned of that person’s beliefs.⁹ This epistemic
warrant described in this scheme need not involve a dependence on speech any
more than it did when I learned about the weather by seeing how people outside
were dressed. Speech, of course, can be an especially revealing and fine-grained
basis for belief-ascription, but from this perspective it is but a particular instance
of the more general scheme of interpreting behavior.
 Richard Moran 277

Since it is knowledge of the other person’s beliefs that is doing all the epistemic
work on this picture, we should note that while speech is in some obvious ways
a privileged route to such knowledge, it is also one which subjects the interpret-
er to special risks which are not shared by other possible ways of coming to this
same knowledge. When I learn of someone’s beliefs through what they tell me, I
am dependent on such things as their discretion, sincerity, good intentions—in
short, on how they deliberately present themselves to me—in a way that I am
not dependent when I infer their beliefs in other ways. People are known to lie,
exaggerate, and otherwise speak in ways that do not express their genuine beliefs.
Thus, in relying on what a person says, I am incurring an additional risk that
the behavior he is manifesting may be deliberately calculated to mislead me as to
what he believes. I am here dependent on him, and his intentions with respect to
me, and not just on my own abilities as an interpreter of the evidence. This source
of error is a much more remote possibility in the case of inferences drawn from
the private observation of someone’s behavior. The people bundled up against
the cold could be dressed up like that just so as to fool me, but this is hardly the
everyday occurrence that lying and misrepresentation is. And that risk of error
is not a possibility at all for those ways, real or imaginary, of learning someone’s
beliefs directly and without the mediation of voluntary expression or behavior at
all (i.e., whatever is imagined in imagining the effects of truth-serum, hypnotism,
or brain-scans). If the epistemic import of what people say is at bottom that of
an indication of what they believe, it would seem perverse for us to give any priv-
ileged status to the vehicle of knowledge (speech and assertion) where we are most
vulnerable, because most dependent on the free disposal of the other person. And
if we are considering speech as evidence, we will have eventually to face the ques-
tion of how recognition of its intentional character could ever enhance rather than
detract from its epistemic value for an audience. Ordinarily, if I confront some-
thing as evidence (the telltale footprint, the cigarette butt left in the ashtray), and
then learn that it was left there deliberately, and even with the intention of bring-
ing me to a particular belief, this will only discredit it as evidence in my eyes.
It won’t seem better evidence, or even just as good, but instead like something
fraudulent, or tainted evidence.
Insofar as speech does occupy a privileged place in what we learn from other
people, this sort of view seems to picture us as perversely preferring to increase
our epistemic exposure, by placing ourselves at the mercy of the free disposition
of another, according a privileged place to human speech, which is here construed
as a kind of evidence that has been deliberately tampered with. On the ‘evidential’
reconstruction of testimony, speech functions as no more than a very possibly
misleading way of learning the speaker’s beliefs. Other things being equal, some
more direct way of learning would be better; and in particular we should prefer
any way of learning the speaker’s beliefs that was not wholly dependent on his
overt, deliberate revelation of them. Anything that necessarily involved his free
action in this way, and thus brought with it the possibility of deliberate deceit,
278 Getting Told and Being Believed

could only be a less reliable way of learning his beliefs than some otherwise com-
parable way that involved going behind his back (mind-reading, brain-scans,
private observation of his behavior). If speech is seen as a form of evidence, then
once its intentional character is recognized (that is, not just as intentional beha-
vior, but intentional with respect to inducing a particular belief ) we need an
account of how it could count as anything more than doctored evidence.

A S S E RT I O N A S A S S U R A N C E

Let us contrast this view with another picture of how what another person tells
me may contribute to my belief, a picture that will give central place to the act
of saying something and the response of believing or disbelieving the person. On
a genuinely non-Humean account, when someone tells me it’s cold out I don’t
simply gain an awareness of his beliefs, I am also given his assurance that it’s cold
out. This is something I could not have gained by the private observation of his
behavior. When someone gives me his assurance that it’s cold out he explicitly
assumes a certain responsibility for what I believe. What this provides me with
is different in kind, though not necessarily in degree of certainty, from beliefs I
might have read off from his behavior, just as what I gain from his declaration of
intention differs from the firm expectation I may form from knowing his habits.
On the evidential picture, by contrast, the speaker’s assurance as such just clouds
the issue, since all the verbal expression of assurance can do is interpose an addi-
tional piece of (possibly misleading) evidence between me and what I really want
to know. I now have some more behavior to interpret, verbal this time, which
brings with it special new possibilities for being misled. From my role as inter-
preter of others, my ultimate destination is the truth about the world, but often
I must pass through the beliefs of another person as my only (fallible) access to
this truth. And now relying on what he deliberately says provides me with at best a
distinctively fallible way of learning what his beliefs are.
On both views, when I take someone’s word for something I am peculiarly
dependent on the will or discretion of the speaker, in a way that I would not be
in the situation of interpreting the evidence of his behavior. But they view this
dependence differently. On the Assurance view, going behind his back to learn
his beliefs would not be better, or even just as good. Rather, it is essential to the
distinctive reason for belief that I get from assertion that it proceeds from some-
thing freely undertaken by the other person. Only as a free declaration does it
have that value for me. Evidence, by contrast, is not dependent on presentation in
this way. A phenomenon will count as evidence however it came about, whether
by natural causes or by someone’s deliberate action, or just as easily by his inad-
vertence or carelessness. But nothing can count as someone’s assurance that was
not freely presented as such, just as talking in one’s sleep cannot count as making
an assertion or a promise.¹⁰ The two views, then, oppose each other most directly
 Richard Moran 279

over this issue of the role of the speaker’s freedom, and the hearer’s dependence
on it. On the evidential view, dependence on the freedom of the other person just
saddles us with an additional set of risks; now we have to worry not only about
misleading (natural) evidence, but deliberate distortion as well. On the assurance
view, dependence on someone’s freely assuming responsibility for the truth of
P, presenting himself as a kind of guarantor, provides me with a characteristic
reason to believe, different in kind from anything provided by evidence alone.
In the remainder of this paper, I want to sketch out a defense of the alternative
picture above, and explore the case for denying that human testimony should be
thought of as providing the same sort of reason for belief that ordinary evidence
does. A guiding question will be: As hearers faced with the question of believing
what we are told, how are we to understand the nature of our dependence on the
free assertion of the speaker, and how does this dependence affect the question of
whether our epistemic relation to what is said is ultimately an evidential one?
In a ground-breaking paper on the central questions of testimony, Angus Ross
(1986) begins by raising the question of whether it makes sense in general to treat
what people say as a form of evidence, and he explicitly relates this question to
the fact that speaking is a voluntary act. I have some differences with how he
understands this relation, but the general line of thought seems to me deeply
right and worth developing. Let me begin with a moderately lengthy quotation
from the early pages of Ross’s article.¹¹
The main problem with the idea that the hearer views the speaker’s words as evidence
arises from the fact that, unlike the examples of natural signs which spring most readily to
mind, saying something is a deliberate act under the speaker’s conscious control and the
hearer is aware that this is the case. The problem is not that of whether the hearer can in
these circumstances see the speaker’s words as good evidence; it is a question of whether
the notion of evidence is appropriate here at all. There is, of course, nothing odd about
the idea of deliberately presenting an audience with evidence in order to get them to
draw a desired conclusion, as when a photograph is produced in court. But in such a case
what is presented is, or is presented as being, evidence independently of the fact of the
presenter having chosen to present it. If a speaker’s words are evidence of anything, they
have that status only because he has chosen to use them. Speaking is not like allowing
someone to see you are blushing. The problem is not, however, that the fact of our having
chosen to use certain words cannot be evidence for some further conclusion. Our choices
can certainly be revealing. The difficulty lies in supposing that the speaker himself sees his
choice of words in this light, which in turn makes it difficult to suppose that this is how
the hearer is intended to see his choice. (Ross 1986: 72)

First of all, it should be noted that Ross’s target, like mine, is not the class
of all speech acts, nor even the class of all assertoric speech acts. Not everything
done in speech, not even everything done with sentences in the declarative mood,
involves the specific relations of telling and being believed. Assertions are also
made in the context of argument and demonstration, for instance, where there
is no assumption within the discourse that the speaker is to be believed on his
280 Getting Told and Being Believed

say-so.¹² In such a situation the speaker is not expecting to be believed, but is

attempting to provide independent convincing reasons for the truth of his view,
or laying out the steps of a proof. Telling someone something is not simply giving
expression to what’s on your mind, but is making a statement with the under-
standing that here it is your word that is to be relied on. It is a common enough
understanding, and commonly justified, but it is not one in place in such con-
texts as persuasion, argumentation, or demonstration. For different reasons it is
also not the understanding of the speech of a person in the context of therapeutic
treatment, in the oral examination of a pupil, or in the police interrogation of a
suspect. Such discourses will contain statements of various kinds, but they may be
received by the interlocutor in a very different spirit, as evidence for truth of a
very different kind from the overt subject of the subject’s statement. This again is
quite different from the exchange of information through telling and being told
in everyday life. (And on the picture of speech to be developed here, these other
discourses emerge as ultimately dependent on the central discourse of telling.)
Having said that, however, how is seeing one’s own utterance as evidence sup-
posed to be incompatible with seeing my utterance as a voluntary act of mine,
in Ross’s words seeing it as up to me what I shall say? He notes that there is
nothing in the idea of evidence itself which is inconsistent with a person’s delib-
erately presenting something as such, as when a photograph is introduced as
evidence. And, it should also be noted, an item like a photograph can serve as
good evidence even when it was not only deliberately presented, but also delib-
erately produced so as to lead one to a particular conclusion. So why cannot the
speaker have essentially the same relation to his own words, as something he
deliberately produces and presents to serve as evidence for some conclusion (and
hence to bring the hearer to some desired belief )? Part of the answer Ross gives
lies in the following view. Seeing the utterance as evidence would involve seeing
it as the outcome of some general empirical law, the sort of ‘reliable correlation’
Hume has in mind, connecting the making of the statement with the obtaining
of the facts in question. Ross acknowledges that I may see the words of others or
my own past words in this light, but
What I cannot do is see the words I now choose to utter in that light, for I cannot at
one and the same time see it as up to me what I shall say and see my choice [. . .] as
determined or constrained by facts about my own nature. (Ross 1986: 73)
Such a stance toward one’s own utterances may be barely possible, he says,
but it is hardly compatible with taking responsibility for those acts. While the
emphasis on responsibility is important, I don’t think this part of Ross’s response
leads in the right direction. For, as far as ‘reliable correlations’ go, why could I
not see my own utterance as securely linked with the truth, not in virtue of my
being determined by the facts of my own nature, but in virtue of my own free but
unswerving commitment to the truth? The sort of reliability my Humean inter-
locutor wants to count on does not abolish my freedom. I can present myself to
 Richard Moran 281

myself and others as reliable in various ways, without that meaning that my reli-
ability is a constraint to which I am passively subject. My utterance is a voluntary
act of mine, something I take responsibility for, and part of what I take responsib-
ility for is its correlation with the truth. So it seems it cannot be because I see my
utterance as freely chosen that it cannot be taken either by myself or my audience
as evidence for the truth.
However, there is another strand in what Ross is saying here that clarifies the
role of the speaker’s freedom and its clash with the idea of evidence. In the first
passage quoted he says that something like a photograph will be evidence inde-
pendently of the fact of the presenter having chosen to present it; whereas by
contrast,
If a speaker’s words are evidence of anything, they have that status only
because he has chosen to use them.
Strictly speaking, this last statement is not quite right, as we’ve already briefly
seen. If we’ve agreed that in various contexts a person’s words can be treated as
evidence, then this need not be dependent on the speaker’s having chosen to use
them. If my analyst can adopt a symptomatic stance toward my more conscious
and deliberate statements, then he may make similar revealing inferences from
my botched utterances, slips of the tongue, as well as the words I may utter under
hypnosis, or while talking in my sleep. Speaking is a form of behavior, after all,
and human behavior is infinitely interpretable, infinitely revealing, in ways that
are not at the disposal of the person to determine their meaning. One’s words can
be evidence when not chosen at all, revealing like a cry of pain; or they can be evid-
ence against one’s intent, as when someone’s tone of voice reveals that he’s lying.
What is true, however, but still in need of defense here, is that a statement only
provides the kind of reason for belief that testimony does if it is understood to be
something freely and consciously undertaken by the speaker. It is with respect to
this sort of reason for belief that we, as hearers or readers, are essentially dependent
on the free disposal of the speaker or writer. Thus, if the idea is that something
is evidence, or is being treated as evidence, when it is a reason for belief inde-
pendently of whether it was intentionally produced or presented as such, we need
a fuller characterization of the kind of ‘independence’ that pertains to the cat-
egory of evidence, and defense of the idea that testimony as such provides reason
for belief that is not independent of assumptions involving the freedom of the
speaker.

PH OTO G R A PH S A N D S TAT E M E N TS

It is here that Ross’s passing contrast between our epistemic relations to photo-
graphs and speakers is worth developing in some detail. There are many ways in
which what we see and what we believe may be dependent on what others do,
say, or show to us. In my direct experience of a footprint, I may be dependent
282 Getting Told and Being Believed

not only on the person who made it, but also perhaps on someone who drew it
to my attention. And when my epistemic relation to it is mediated by another
person in these ways, I am subject to the ordinary risks of distortion, since in
principle any evidence may be tampered with. But even with these particular risks
and dependencies, my relation to the footprint is still a perceptual one, and does
not involve me in the specific relation of believing another person. And this is so
even if my perception of it is technologically mediated in ways that involve the
doings and expertise of other people. In discussing the nature of photographic
realism, Kendall Walton (1984) compares what we see in photographs with what
we see through a microscope or in a mirror, to argue for the claim that in all three
cases we actually see the thing in question, even though this seeing is mediated in
various ways, and even though photographs can be doctored in various ways.¹³
Real experience of a thing may also be mediated or subject to various epistem-
ic risks, without that abolishing the difference between being told about it and
experiencing it oneself. As Walton points out, what I see directly when someone
points out the window may also be altered in various ways to deceive, but that
doesn’t transform the situation from perception to depiction. In Walton’s terms,
a photograph can be ‘transparent’ to the scene it depicts in part because, unlike
the case of a drawing, what we see here is not essentially dependent on what the
photographer thinks is there in the photograph. As with a telescope, we may ‘see
through’ the photograph to the scene itself.
In this regard, consider the case of the photographer in Antonioni’s
movie Blow Up (1966). He takes some pictures in the park of a woman and a
man, and then later discovers that one of his shots apparently shows the man’s
corpse lying in the bushes. This is not what he saw or believed at the time, but
it is what he sees now. Still, the photograph he took is evidence, of the most
ordinary kind, for the fact that this man has been killed. And it is evidence for
this regardless of the photographer’s beliefs about the matter. That is, it would
be evidence even if he positively disbelieved what it shows, or even if he took
the photograph and showed it to someone with the deliberate intent to deceive.
Its status as evidence is wholly independent of his beliefs or intentions. And it is
for that reason that his own relation to the photograph can be an evidential one,
like a detective or other investigator. When he gets home he crops and enlarges
and studies his photograph in order to see more deeply into what it shows, to
convince himself that the corpse on the grass is really there. In this way his own
epistemic relation to the photograph he took is the same as that of the friend
he shows it to later. They can both learn from it, or doubt what it shows. The
situation would be quite different if he were to have made a sketch of what he saw
in the park, or taken some notes on what he observed there. It would be absurd
for him take his sketch home and blow up it to examine more closely what it
shows about the man in the park. And were he to show his sketch or his notes to
another person to convince him about the man in the park, he would be offering
him a very different kind of reason to believe what happened. If he shows his
 Richard Moran 283

friend a sketch of a corpse lying in the grass, and this is to be a reason for him
to believe there was such a corpse, his friend has to assume such things as that
the sketch was not made with an intent to deceive, that the person who made it
was observing things accurately and not liable to error, and even that the aim of
the sketch was an accurate picture and not an imaginary scene, etc. In short, the
beliefs and intentions of the person who made the sketch are crucial for its status
as a reason to believe anything about what was there in the park. Without those
assumptions, the sketch does not become poorer evidence; it ceases to be evidence
of any kind, or any other reason to believe. It’s just a piece of paper, and any
correlation with the facts in the park could only be by the merest chance.
So how does the issue of freedom figure in here, in a way that distinguishes the
case of verbal testimony? After all, the photographer freely takes his picture, and
then may freely present it to another person as a reason for believing a man has
been killed. How is this different from his friend’s relation to his verbal report
of what he saw? So far we have seen the following difference. The status of the
photograph as a reason to believe something does not depend on the photograph-
er’s own attitude toward it as evidence. It depends only on the camera’s ability
to record the scene, which need not involve any choice or consciousness on the
part of the photographer at all. (The exposure could have been made by a remote
timing device.) As such the photograph can serve for him as an independent cor-
rection of his impression of the scene, in a way that his drawing cannot. It is
for this reason that when he looks at his photograph with his friend, they both
stand in the same epistemic relation to it; confronting it as independent, public,
evidence, and trying to discern its import.
By contrast, the speaker’s choice enters in essentially to the fact that his utter-
ance counts at all as a reason for belief. The point is not that his utterance is
voluntarily produced, for that in itself has no epistemic significance, and does not
distinguish the case from that of the photographer. Rather the point is that the
speaker, in presenting his utterance as an assertion, one with the force of telling the
audience something, presents himself as accountable for the truth of what he says,
and in doing so he offers a kind of guarantee for this truth. This shows up in the
fact that if we are inclined to believe what the speaker says, but then learn that
he is not, in fact, presenting his utterance as an assertion whose truth he stands
behind, then what remains is just words, not a reason to believe anything. We
misunderstood the intent of Professor Higgins when we heard him say something
about the rain in Spain, and now upon realizing this, the utterance as phenomen-
on loses the epistemic import we thought it had (whatever knowledge we may
indeed take him to have about such matters). By contrast, if we learn that the
photographer is not, in fact, presenting his photograph as true record of what
occurred in the park, the photograph as document retains all the epistemic value
for us it ever had.
284 Getting Told and Being Believed

T H E I M P O RTA N C E O F B E I N G N O N - N AT U R A L

Still, one might ask, why speak of the audience’s dependence on the freedom of
the speaker, rather than simply refer to their dependence on what the speaker has
(freely) done? The reason is that the relevant speaker’s responsibility is not simply
his responsibility for the existence of some phenomenon, in the sense that he is the
one who deliberately produced these spoken words. Rather, he is more centrally
responsible for those words having any particular epistemic status. What is the
difference, then, between the speaker’s role in providing something (his utterance)
with a particular epistemic status, and the role of someone like a photographer
who produces something with a certain epistemic import?
It is here, I think, that a consideration of Paul Grice’s original 1957 paper
‘Meaning’ proves helpful. The relation of evidence, one phenomenon’s being
an indication of something else, is the central form of what Grice calls ‘nat-
ural meaning’. Natural meaning is not something at the disposal of the speak-
er to confer or revoke, but is a matter of the independent obtaining of causal
relations in the world (e.g., the way smoke means fire, or doesn’t). Nonethe-
less, persons belong to this same natural world and may thus produce or exhibit
various evidential phenomena, and employ them to get some point across (e.g.,
pointing to the smoke pouring out of the oven). But spoken words typically
bear a different relation to the facts. In his 1957 article, Grice is primarily con-
cerned to delineate the conditions for something he calls ‘‘non-natural meaning’’,
or MeaningNN. This project famously evolved into an attempt to ground the
notions of the meaning of an expression in a language in the complex intentions
had by utterers of expressions on occasions of use; and, presented as a non-
circular account of either word-meaning or sentence-meaning, it was progress-
ively refined into baroque complexity under the pressure of counter-examples.
However, the interest and importance of the original account of non-natural
meaning is not exhausted by the prospects for an intention-based semantics of
the sort he proposed. What he isolates under the title of ‘non-natural meaning’ is
a central form of intersubjective dependence, one that is indeed paradigmatically
linguistic, but not restricted to linguistic communication.
A striking thing about the essay is how the technical notion of non-natural
meaning is introduced by contrast with natural meaning, as if this were an ante-
cedently intuitive notion, one whose definition we could progressively refine by
consulting our intuitions about a series of well-crafted cases and asking ourselves
whether we should call that a case of non-natural meaning. We are given hints,
of course, by way of both similarity and contrast with more familiar notions like
that of conventional meaning, but Grice’s target notion only emerges through
the consideration of the cases devised and presented. The cases themselves all
 Richard Moran 285

have a similar form in that in all of them one person does something which either
succeeds or not in inducing another person to some belief P. This common telos
to the cases invites two related questions. Since the end-point of each of these
encounters is that one person ends up with a new belief, we might look at the
progression of cases from an epistemological point of view and ask what it is that
brings the person at the receiving end to this new belief, what reason he may
take himself to have been given for adopting it, and why the particular kind of
reason Grice’s account of non-natural meaning zeros in on should be of special
significance, either epistemologically or otherwise. It is not, of course, as if the
other ways of inducing belief, disqualified as candidates for the non-natural, are
thought to be insufficiently grounded. Salome certainly acquired justified belief
about the fate of John the Baptist by seeing his head presented on a charger, how-
ever this may fall short as a case of non-natural meaning. Rather, the target notion
of non-natural meaning is meant to capture a way of gaining a reason to believe
something that is importantly different from others and that we have special reas-
on to be concerned with, both as purveyors and receivers of such reasons. So
the first question is: what is special about the reason for belief associated with
non-natural meaning? And secondly, as the proposed definition of non-natural
meaning is progressively refined in Grice’s essay, what pre-theoretical notion is
supposed to be guiding our intuitions along the way, so that we can feel convic-
tion about a range of cases that seem to fall more or less squarely in the category?
Here Grice is more explicit, since by way of explaining the distinction that mat-
ters to him, and why something like the case of Herod’s presentation to Salome
does not count as non-natural meaning, he says, ‘‘What we want to find is the
difference between, for example, ‘deliberately and openly letting someone know’
and ‘telling’ and between ‘getting someone to think’ and ‘telling’ ’’ (Grice 1957,
in Strawson 1967: 44). So it is the ordinary notion of telling someone something,
that way of inducing belief, that is to play a guiding role in determining which
cases satisfy the philosophical notion of non-natural meaning, and Grice’s dis-
tinction between natural and non-natural meaning can be seen as motivated by
a concern with the difference between telling a person that P and other ways of
bringing him to that same knowledge, such as providing him with evidence for P
(evidence that may be accidental or contrived, openly displayed or inadvertently
revealed).
As examples of ‘deliberately and openly letting someone know’ some fact,
Grice cites such cases as that of showing someone a compromising photograph,
or leaving the china my daughter has broken lying around for my wife to see.¹⁴
In these cases, the phenomenon in question has some independent evidential
significance, even though the person may be responsible either for drawing
attention to it (the broken china), or actually producing it (the photograph).
Their independent significance shows up in the fact that the photograph or
the china would have functioned as a reason for the belief in question without
anyone’s intervention or presentation, even if only stumbled upon accidentally.
286 Getting Told and Being Believed

By contrast, in cases of ‘telling’ or ‘non-natural meaning’, the person (hereafter

the ‘speaker’) plays quite a different role in bringing his audience to believe
something. Here, as is well known, a crucial role is played in Grice’s account by
the recognition of the speaker’s intention. Examining this role will help clarify
the specific ‘dependence on the freedom of the speaker’ that I’m claiming is
characteristic of the relation of testimony, and which distinguishes it from a
relation of evidence.
Following Grice’s progression then: a handkerchief left at the scene of the
crime may throw suspicion on someone and perhaps lead to genuine belief in his
guilt. But as a piece of evidence it would induce that belief whether or not it were
left there intentionally, and non-natural meaning (and ‘telling’, surely) must at
least be the upshot of something intentionally done. Further, it should be part of
non-natural meaning that the intention to induce a particular belief is manifest to
the person on the receiving end, and not like artfully planted evidence designed
to steer him toward the desired conclusion. Further still, this belief-inducing
intention must not simply be known to the audience, something he pieces togeth-
er despite the speaker’s best efforts at concealment; rather the speaker must fully
expect and intend that his intention will be manifest to his audience. In this way,
the audience can appreciate that another person is openly playing a role in direct-
ing him to learn something, by presenting a piece of evidence for them both to
see and assess. This makes this knowledge, or at least this awareness of the evid-
ence, ‘mutual’ between them, and hence available as an object of cognitive and
communicative cooperation between them. However, this is not yet non-natural
meaning, since these conditions are fulfilled when Herod presents the head of
St John to Salome, or when the compromising photograph is flourished. Here,
although the audience has been directly and openly led to some belief, she has
not been told anything. Rather, Salome has been shown something and reliably
left to draw her own conclusions. Herod manifests a definite intention in bring-
ing her this news, and he bears an obvious responsibility for Salome’s altered state
of belief. But there is yet another responsibility he does not assume here, which
marks the difference between ‘‘ ‘deliberately and openly letting someone know’
and ‘telling’ ’’, and this is shown in the fact that, while his intention regarding
her belief is indeed manifest, it is inert as far as Salome’s belief is concerned, just
as it is when the person is shown a compromising photograph. It isn’t doing any
epistemological work of its own. Both people would draw essentially the same
conclusions whether the evidence in question were deliberately and openly dis-
played to them or not. So we might say that Herod’s epistemic responsibility for
Salome’s belief is merely contingent, like that of the person showing the photo-
graph. In these cases they play a role in making a piece of evidence available to
another person, but they are not responsible for it having the epistemic import
it has.
For Grice, however, nothing can count as a case of non-natural meaning if the
relevant belief could be expected to be produced whether or not the intention
 Richard Moran 287

behind the action were recognized. The speaker must not only intend that the
audience recognize his intention, but this recognition must itself play a role in
inducing the belief in question, and that means that the recognition of the speak-
er’s intention must not just as a matter of fact help to bring about the relevant
belief, but must be necessary to its inducement. In this way we arrive at Grice’s
original formulation of non-natural meaning in his 1957 paper:
A uttered x with the intention of inducing a belief by means of the recogni-
tion of this intention. (Grice 1957, in Strawson 1967: 45)
If the audience could not be expected to arrive at the intended belief apart
from the recognition of the speaker’s intention regarding that belief, the speaker
must take upon himself the role of providing something with a particular
epistemic import that it otherwise would not have, and in this way Grice
sharply distinguishes non-natural meaning from the presentation of evidence.
For any phenomenon with some independent evidential import will naturally
be one which might well be expected to induce belief without the recognition
of anyone’s intention. That’s just what it is for a phenomenon to be ordinary
evidence for something else. To count as an instance of telling someone
something, however, the speaker must present his action, his utterance, as
being without epistemic significance apart from his explicit assumption of
responsibility for that significance. In this way he announces that the reason
for belief offered here is of a different kind from that stemming from externally
obtaining evidential relations.
As Ross points out (1986: 75), from the point of view of the audience, con-
sidered as a reason for belief, the role of the recognition of intention is left some-
what mysterious here. The question is: just how does my recognizing that this
speaker intends that I should believe P play a role in actually getting me to believe
that P? If we compare this case with that of other things someone may want me to
do it’s clear that the mere recognition that he wants me to do X does not, in gen-
eral, provide me with much of any reason at all for complying. Why should we
be so much more compliant when we recognize that someone wishes us to believe
something? How can the mere recognition of someone’s intention be expected to
induce belief ?
When looked at in this way, recognition of the speaker’s intention may seem
inadequate to induce belief. It may also seem pointless, adding nothing of epi-
stemic value to what the audience already has. Again, compare this with the
picture of radical interpretation, according to which the epistemic significance
of speech is that of an indication of the speaker’s beliefs. Once I employ this
scheme of interpretation to learn what the speaker believes, I am then in posses-
sion of knowledge of a certain set of facts, viz. the speaker’s state of belief, which
does have straightforward evidential value for me, quite independently of how or
whether the fact of his believing is explicitly presented to me. The speaker’s state
of mind is phenomenon, which has the same independent evidential import for
288 Getting Told and Being Believed

me, regardless of how I may have learned of it, and regardless of whether it was
manifested deliberately or inadvertently. And, as we saw, this same scheme of
interpretation can provide a basis for me to infer the likely truth of these beliefs,
and so come to share them myself. I ascribe beliefs on the basis of his verbal
behavior as I would from any other behavior, and in neither case do I rely on
recognition of any intention to manifest his states of mind. And indeed, what
could be the epistemic interest for me in learning of any such intention on his
part? By hypothesis, I already know what he takes to be true, and I can now make
of this knowledge what I will, deciding for myself whether this adds up to good
reason for me to take his belief to be true. If his verbal behavior is evidence for
his beliefs, then it doesn’t add to my evidence as interpreter to learn that, in addi-
tion to his believing P, the speaker also has the intention that I should believe
P too (and come to this belief on the basis of recognition of that very intention,
etc.). From my side, either learning of his belief is, on balance, sufficient for me
to believe P too, or it is not. Nothing further about his intentions, or just how he
would like me to arrive at this belief, will be evidentially relevant for me at all.
Or else, as before with the tainted evidence, learning that his belief was deliber-
ately manifested now casts doubt on my ascription, because the evidence of his
behavior is now contaminated by its aspect of performance.
What is needed is more direct focus on the speaker’s explicit presentation of
himself as providing a reason for belief. For it is not, in fact, the audience’s mere
awareness of the speaker’s intention that is to provide a motivation for belief. If
I simply discovered on my own that this person had the intention that I believe
P, this need not count for me as a reason for belief at all. (Why cooperate with
his designs on me, however benign?) The conditions given so far still have not
accounted for any special importance to the overt act of saying, the explicit mani-
festing of one’s intention, as opposed to simply doing something that allows one’s
intention to become known. If, unlike a piece of evidence, the speaker’s words
have no independent epistemic value as a phenomenon, then how do they acquire
the status of a reason to believe something? It seems that this can only be by vir-
tue of the speaker’s there and then explicitly presenting his utterance as a reason
to believe, with this presentation being accomplished in the act of assertion itself.
The epistemic value of his words is something publicly conferred on them by the
speaker, by presenting his utterance as an assertion. And indeed, it is because the
speaker’s words have no independent status as evidence that their contribution
to the audience’s belief must proceed through the recognition of the speaker’s
intention. Further, the intention seeking recognition must not simply be that
the audience come to believe something, but must include the intention that the
audience recognize the speaker’s act of asserting as itself constituting a reason
for belief. If it seems difficult to see how anything, even someone’s words, could
acquire some epistemic value through something like conferral, perhaps because
this suggests something too arbitrary or ceremonial to constitute a genuine reason
for belief, it should be remembered that for both parties this conferral is by its
 Richard Moran 289

nature an overt assumption of specific responsibilities on the part of the speaker.

This is no more (or less) mysterious than how an explicit agreement or contract
alters one’s responsibilities, actions which are also within the capacities of ordin-
ary speakers. The speaker’s intent then, is that for the audience, the very fact that
this speaker is freely and explicitly presenting P as worthy of belief constitutes his
speech as a reason to believe that P.
Of course, as with any public assumption of responsibility, the appropriate
abilities and other background conditions must be assumed to be in place for it
to amount to anything. For the speaker to be able to do this it must be assumed
by both parties that the speaker does indeed satisfy the right conditions for such
an act (e.g., that he possesses the relevant knowledge, trustworthiness, and reliab-
ility). These background conditions can themselves be construed as evidential, or
at any rate not at the behest of the speaker to determine, but they are not them-
selves sufficient for giving any epistemic significance to the speaker’s words, for
the relevance of these conditions only comes into play once it is understood that
a particular speech act is being performed with those words (i.e., an assertion or
promise rather than something else). The speaker has to constitute his utterance
as having this or that illocutionary force before the empirical background condi-
tions can contribute anything to its epistemic significance. Hence the idea is not
that the speaker’s word’s ‘‘all by themselves’’ should count as a reason for belief,
or that the speaker’s authority over the constitution of the particular speech act
he is performing (e.g., as assertion rather than recitation), shoulders the epistem-
ic burden all by itself. As with the explicit assumption of responsibility that goes
with making a promise, its success will depend on the various conditions that
go into the speaker’s being in a position to take on any such responsibility, and
which make for his public assumption’s being anything for another person to
count on. But in considering the speaker’s words, the audience’s belief in his
knowledge and trustworthiness do not do him any epistemic good if it is still left
open just what kind of action (if any) the speaker is presenting his utterance as.
As far as relating to his words goes, the speaker’s knowledge and trustworthiness
is epistemically inert for the audience until the question of the particular speech
act or illocution is settled. Determining his utterance as an assertion is what gets
the speaker’s words into the realm of epistemic assessment in the first place (or at
least epistemic assessment of the sort that is relevant to testimony: we may indeed
make evidential use of the words or inarticulate sounds made by someone asleep).
And in this matter, the speaker and his audience are in essentially different rela-
tions to the epistemic import of the speaker’s words. The speaker does not relate
to the question whether his utterance is a committed assertion or not as some-
thing to be settled by evidence, because as a speaker of the language he plays an
essential role in making it the case that his utterance is an assertion or not.
Hence Grice’s original formulation needs some further refinement. The speak-
er intends not just that the recognition of his intention play a role in producing
belief that P, but that the particular role this recognition should play is that of
290 Getting Told and Being Believed

showing the speaker to be assuming responsibility for the status of his utterance
as a reason to believe P. This addition is necessary since in principle there are all
sorts of ways in which the recognition of intention could ‘play a role’ in produ-
cing belief, ways that would not capture what is meant by ‘telling’ or ‘non-natural
meaning’, or the correlative notion of believing the speaker. One such way would
be manifested in the familiar situations of ‘double-bluffing’ where, e.g., I tell you
I’m traveling to Minsk, knowing you’ll take me to be lying and attempting to
conceal my plans to travel to Pinsk, and hence meaning to deceive you about my
genuine plans to go to Minsk after all. Knowing all this about me, however, you
see through the ruse and conclude that I’m indeed going to Minsk, just as I told
you. Here the recognition of intention does indeed play an essential role in the
belief arrived at, and the audience comes to believe that what I say is true, but this
is not a case of believing the speaker.¹⁵ And there are other possible ways in which
the recognition of intention might play a role, even a necessary one, but of the
wrong sort.
Grice is sensitive to an incompleteness here, when he suggests toward the end
that it should somehow be built into the definition that the intended effect must
be something which in some sense is within the control of the audience, or that
in some sense of ‘reason’ the recognition of intention behind X is for the audience
a reason and not merely a cause (Grice 1957 in Strawson 1967: 46). It is not the
speaker’s aim that the belief in question be produced by the audience’s simply
being so constituted that his awareness of the speaker’s complex self-referential
intention somehow produces the belief in him. That would fail in another way to
describe the nature of the dependence on the person as such and the importance
of mutual recognition. For the audience must not simply respond with belief,
but must understand what the speaker is saying, and must understand what the
speaker is doing in saying P, which is to say, purporting to present him with a
reason for P. And the audience must believe P because he understands what the
speaker is saying and what he is doing in saying it. In addition, and crucially,
the audience must take this entire understanding to be shared by himself and the
speaker. That is, he takes himself to be responding to just the kind of reason for
belief that the speaker is presenting himself as offering (which is why cases of
‘‘double-bluffing’’ are not cases of believing the speaker).
Any of the ‘proto-Gricean’ ways of producing belief, the cases leading up to
the full definition of non-natural meaning, provide us with something mediating
between the audience and the speaker, something other than the person as such
that is being depended on. Believing the speaker, on the other hand, involves
accepting the offer to rely on him, and not something connected with him or as a
consequence of what he has done. This direct dependence on the speaker’s offer
of responsibility is what is expressed in the ‘hereby’ that is implied, and some-
times explicitly stated, in illocutions such as ‘telling’, ‘warning’, or ‘accepting’, for
it is in this very presentation of himself that the speaker assumes responsibility
for the audience’s belief.¹⁶ The implied ‘hereby’ is thus also an expression of the
 Richard Moran 291

self-referentiality of the Gricean formula, for it declares that it is dependence on

the person as such, and not on something else he might point to, that is solicited
in saying that he hereby tells his audience that P.¹⁷ In this way we can see the pro-
gressive refinements of Grice’s definition of non-natural meaning as each aimed
at laying bare the reliance of the audience on the other person as such. The belief
is to be produced not simply by the speaker’s action, or by his intention, or by the
audience’s awareness of his intention, or anything else outside their encounter.
Just as the audience could treat the handkerchief or the photograph as evidence
for P, and thus without trusting the speaker for the truth of P, so he could treat
the speaker’s action or intention as a similar kind of evidence without trusting
the speaker, without his belief that P involving dependence on the person of the
speaker as such.
When Grice says that the belief in question should be ‘‘something which in
some sense is within the control of the audience’’, or that it should function as a
reason for him and not a mere cause, this is not meant to suggest that the audi-
ence complies with the speaker’s intention as a kind of favor, adopting the belief
on request (as Ross notes, 1986: 74). But it does serve to clarify the kind of role
that is to be played by the mutual recognition of the speaker’s intention, how that
can matter epistemically to the audience in the way suggested by Grice’s progress-
ive refinements of the account of non-natural meaning. The account of this role
suggested by the Assurance view is that the mutual recognition of intention can
play the role for the audience of providing him with a reason for belief, because
he sees the speaker as presenting himself as accountable for the truth of P, and
asking, through the recognition of his intention, that this offer of his assurance be
accepted. And it is understood by both parties that this acceptance is something
which the audience is free to give or refuse. The speaker is asking that a certain
authority of his be acknowledged; the authority to invest his utterance with a par-
ticular epistemic import, and this investment occurs by his explicit assumption of
responsibility for his utterance’s being a reason for belief. This is the role for the
recognition of intention that the speaker is asking for. And, I would argue, it is
only such a role that could account for how, in the case of speech, the recognition
of intention enhances rather than detracts from the epistemic status of the phe-
nomenon (utterance), reveals it to be something other than doctored evidence.
The idea of assertion as providing reason for belief through the explicit assump-
tion of responsibility for the truth of what is said accounts for a number of con-
trasts between belief through testimony and belief through confronting evidence.
It points the way to understanding how the recognition of intention can play a
positive role, rather than see it as something that is either epistemically irrelev-
ant or undermining to the evidential value of the utterance. Further, a specific
assumption of responsibilities is essentially an expression of a person’s freedom,
something that only makes sense as consciously assumed. It is for this reason that
292 Getting Told and Being Believed

words spoken during sleep, or under hypnosis, do not have the value of testi-
mony, because they do not count as assertions, whatever expressive psychological
value they might still retain as evidence. Like a promise or an apology, some-
thing only counts as a person’s assertion when consciously presented as such by
him.¹⁸
Promises and apologies, like acts of telling someone something, can be more
or less reflective, more or less deliberate, done more or less voluntarily or under
duress. Reference to the speaker’s assumption of responsibility for the truth of
what he says is not meant to deny that much of our speech is spontaneous and
unreflective, or that much of what we acquire from the speech of others is more
or less passively absorbed. ‘Telling’ also includes telling something by mistake,
to the wrong person, or just blurting something out when we meant to keep
silent. We express our freedom not only in our considered actions, but also in
the actions that go wrong, or are forced upon us, and the outbursts that we imme-
diately regret. Blurting something out when you meant to keep silent is still a
different matter from either talking in one’s sleep, or having the utterance of
those words be produced by electrical stimulation of the cortex. And the epistem-
ic significance for the audience is entirely different in the two kinds of cases: in
relating to the words produced by electrical stimulation we may learn something,
but what we learn need not be dependent on such assumptions as, e.g., wheth-
er the person had any understanding of the words themselves, or any sense that
he was providing anyone with a reason to believe something. These assumptions,
however, are still indispensable to the understanding of the words that escape us
or are forced from us, and they express the role of the person as such in providing
a reason. This is confirmed by the fact that both speaker and audience relate to
the blurting out differently than they would to the cases of talking in one’s sleep
or through electrical stimulation. In the latter case, the speaker would not regret
what he said or try to make amends; in a sense what happened didn’t involve him
at all.¹⁹ And for that matter, a person may also lie spontaneously or out of panic
as well as tell the truth. But surely the description of a person telling a lie makes
reference to such things as the intention, whether conscious or not, to exploit
the trust of one’s audience and present oneself as providing them with a reason
for belief. At the same time, it is consistent with the Assurance view to think of
assertions and tellings as something like the default assumption for indicative sen-
tences in the declarative mood. It requires more, rather than less, sophisticated
intentions to utter ‘The rain in Spain falls mainly on the plain’ and not mean it
as an assertion. And so, barring any special reason to think otherwise, we may be
entitled to treat an utterance of an indicative sentence as an assertion of it.²⁰ But
in a given case we may be wrong about this, and it remains true that what settles
the question about the status of the utterance is whether or not the speaker is
presenting it as true.
 Richard Moran 293

T H E S PE A K E R ’ S C O N F E R R A L : H AV I N G YO U R S AY,
G I V I N G YO U R WO R D

We are now in a position to clarify the problem Ross suggested with the idea of
a speaker presenting his utterance as evidence for his audience. The problem is
not that the speaker’s words could not be taken as evidence by his audience. In
principle, anything said or done by the speaker can be given a symptomatic read-
ing. Nor is it true that the speaker could not privately intend that his words be
taken as evidence. This would be the intention for many cases of deceit or more
everyday manipulation, for instance. In a given case, my primary aim may be for
my listener to draw the conclusion that I’m being scrupulously candid or self-
revealing, and I accomplish this by confessing some minor fault of mine. Here, I
am not telling anyone of my candor, giving my word on it (whatever good that
would do), but rather doing something (in this case: saying something) that I
hope will be taken as evidence for it. When it is a question of non-natural mean-
ing, by contrast, the speaker is not relying on evidential relations alone to get his
point across, but rather is counting on the explicit presentation of his intention
to be the very thing which makes his words a reason for believing something in
the first place. The recognition of his intention could only function this way if it
was seen to be his assurance of the truth in question, his explicit assumption of
responsibility for the truth of what he says. By contrast, the presentation of his
utterance as evidence would be an implicit denial of this responsibility, breaking
the link between the proposition he is giving his backing to and the belief he is
hoping to induce, in which case there’s no question of believing him. Thus for the
speaker to present his word as evidence would be for him to present it as a reason
to believe, while suspending the guarantee that gives it the epistemic significance
of testimony in the first place. This is the problem Ross is pointing to.²¹
And as we have seen, to present something as evidence is to present it as having
its epistemic value independent of one’s own beliefs about it, or one’s presenta-
tion of it, or the conferral of some status upon it. To offer some phenomenon as
evidence is to present it as belief-worthy independent of the fact of one’s present-
ing it as belief-worthy. When we present something as evidence for someone, we
are inviting that person to ‘see for himself’, to find it convincing as we do. And we
are prepared to offer reasons why it should be convincing, reasons independent of
our simply claiming, once again, that it is belief-worthy. To present something as
evidence is to be in a position to report that it is a reason for belief, and to be in
this position one must be presenting that claim of belief-worthiness as having a
basis in fact that is independent of one’s reporting itself. A photograph has such
an independent epistemic basis, independent of anyone’s conferral. As a phe-
nomenon, it counts as a reason for belief independently of anything concerning
how the photographer may conceive it or present it. Because of this independ-
ence, its epistemic status is something the photographer himself may discover
294 Getting Told and Being Believed

about it, or speculate about. His relation to this question is in principle no dif-
ferent from anyone else’s. He may, of course, happen to know something about
how it was produced that we don’t know, and which may affect its epistemic
status. But he may not know anything of the sort, and conversely we might know
more about it than he does. He marshals the same kinds of reasons as any other
viewer in considering the question of what beliefs the photograph may provide a
basis for.
But the speaker’s relation to the epistemic status of his own assertion is dif-
ferent from anyone else’s. For the speaker, it is not a matter of observation or
speculation whether he is indeed presenting his utterance as something with the
force of a committed assertion, and were he somehow unclear about this, then
to that degree his utterance would be something less than a committed assertion.
He may inquire into his own reliability, truthfulness, and command of the facts,
but the status of his utterance as assertion is a matter of what he is then and there
prepared to invest it with. It has been noted by more than one philosopher that
the relation of ‘believing someone’ does not have a reflexive form; it is not a rela-
tion a person can bear to himself.²² The problem with this, we can now say, is
the problem with the idea of a person offering and accepting an epistemic guar-
antee from himself, which would require him to be simultaneously in command
of and at the mercy of his own freedom. This is another basic feature of testimony
not captured by an evidential perspective on it. Speaker and audience do not con-
front the utterance as a phenomenon with an independent or natural epistemic
status which they could assess in the same spirit, for the speaker does not confront
his own assertion as a phenomenon at all, but as an issue of his commitment. To
speak of ‘conferral’ of epistemic status is intended to register the fact that to count
an utterance as, e.g., an assurance or a promise just is to count it as something
presented with a particular epistemic status, the status of a reason for some belief
(as contrasted with the status, say, of recitation or ironic mimicry). To count as a
competent speaker of a language is to be recognized as having definitive ‘say’ over
which illocution one’s utterance counts as, whether as informative assertion, or
as promise or apology, whether as a mere recitation or as a claim expressing one’s
commitment. An utterance counts as an assertion or an apology just in case the
speaker presents it as such to his audience, in the appropriate context where his
audience can be expected to recognize what it is being offered. The speaker can-
not count as having promised or asserted something if he had no such intention,
or if he did not present his utterance to be seen as a promise or assertion, whereas
the evidential import of what he says and does is independent of such conditions.
The speaker’s authority to determine the illocutionary status of his utterance
is the authority he has to present himself as accountable for the performance of
some speech act. This is not a matter of discovery for the speaker, something he
could investigate or report on, as he might with respect to the evidential status
of something. When it is a question of the evidential status of something, even
something the person himself has done, he and his interlocutors are on an equal
 Richard Moran 295

footing with respect to establishing its standing as a reason for belief. A person
does not speak with any special authority about the evidential significance of his
actions, including his verbal ones. By contrast, the authority to present oneself as
‘hereby’ assuming certain responsibilities in speech makes the speaker’s epistemic
position irreducibly different from that of his audience. For him the import of his
words is not an independently obtaining fact, something he has his own opinion
about, but is directly dependent on the import he is then and there prepared to
invest them with. And it is internal to the notion of the speaker’s authority to
confer illocutionary status on his utterance that he also has the exclusive author-
ity to cancel or revoke such status. Words can be retracted, apologies or warnings
taken back, but only by the speaker himself. At the same time, he has no author-
ity to determine, much less cancel, the evidential import of anything he has said
or done, not even of his retraction itself.
When all goes well, in testimony a speaker gives his audience a reason to believe
something, but unlike other ways of influencing the beliefs of others, in this case
the reason the audience is provided with is seen by both parties as dependent
on the speaker’s making himself accountable, conferring a right of complaint on
his audience should his claim be false. Whether this counts as a good or suf-
ficient reason for belief is not a matter of the speaker’s illocutionary authority,
but will depend both on his sincerity and on his having discharged his epistemic
responsibilities with respect to the belief in question.²³ But his presentation of his
utterance as having this particular illocutionary force is what makes it candidate
for epistemic assessment in the first place, and determines what kind of reason for
just what proposition his audience is being presented with.
This way of looking at testimony makes much of the fact that in its central
instances speech is an action addressed to another person, and that in testimony
in particular the kind of reason for belief that is presented is one that functions
in part by binding speaker and audience together, and altering the normative
relationship between them. It doesn’t follow from this, however, that someone
outside that normative relationship can’t avail himself of it and thereby acquire
a reason to believe the same thing. If one person gives his word on something to
another, whether as promise or assertion, someone overhearing this may derive a
sufficient reason to believe, say that the speaker will in fact do what he promised
or that what he asserted is true. And the overhearer improves his epistemic situ-
ation in this way, without entering into the altered normative relationship of the
two parties involved in giving and accepting of words. He has not himself been
told anything, much less promised anything, and no right of complaint has been
conferred upon him.²⁴ To say this much, however, does not provide a reason to
assimilate his situation to that of someone confronting a piece of evidence, or sug-
gest that the speaker’s illocutionary and epistemic responsibilities aren’t playing a
296 Getting Told and Being Believed

genuinely epistemic role here. For even though the statement was not addressed
to him, the overhearer is still in a different position from that of someone con-
fronting a piece of evidence like a photograph or a footprint. It still makes a dif-
ference to his epistemic relation to the overheard report that he is responding to
something whose epistemic significance is not independent in the way of a photo-
graph or footprint, but is inherited from the speaker’s assuming responsibility for
the truth (and meaning) of what he says. This is so even if, as we might say, that
responsibility was undertaken with respect to another person and not himself.
Naturally there is a certain vagueness as to just what situations will count as
overhearing, and in a given case the addressee may be a group of people. None-
theless, while the overhearer may get a reason to believe without having the
right to complaint that is conferred on the addressee, the fact that the overhear-
er of the assertion acquires any reason to believe from listening to these words
is dependent on them being addressed to someone, with the force of assuming
responsibility and thereby conferring a right of complaint. The overhearer of
testimony is not in the same normative relation to the speaker as the addressee
is, but his gaining any reason to believe is dependent on such a conferral having
been given to someone. Without that, the question of what speech act, if any, is
being performed with these words would not be settled, and hence the overhearer
could not get started on assessing their epistemic significance. (Imagine overhear-
ing someone say ‘‘The rain in Spain falls mainly on the plain’’. Until you know
what speech act, if any, is being performed here, you don’t know if considera-
tions of reliability or trustworthiness are even relevant to the status of the words as
source of knowledge about the weather in Spain.) So, while in both cases (prom-
ising and telling), the overhearer can gain a reason to believe something without
entering into the normative relation of promisor-promisee or teller-believer, in
the overhearing of testimony he only gains a reason to believe something because
such a relationship has been established by the original speaker and addressee.
This, then, is how I suggest we understand Ross’s claim that the evidential view
is inconsistent with the kind of reason for belief offered in everyday human testi-
mony. In telling his audience something the speaker does not present his utter-
ance as something with the force of evidence because that would be to present
his words as having their specific epistemic import apart from his assurance, and
the responsibility he thereby assumes. And in obscuring the speaker’s responsibil-
ity, such a stance would also obscure the nature of the audience’s dependence on
him. For if it were a matter of evidence then in principle we would both be on
an equal footing with respect to establishing its epistemic import. But this equity
does not obtain with respect to someone’s words, where it is up to the speaker
alone to determine whether they are to count as an assertion or other committed
speech act.
 Richard Moran 297

EV I D E N C E A N D D I S H A R M O N Y

The two broad views about testimony which I’ve been calling the Evidential
View and the Assurance View are in no disagreement over the status of an asser-
tion or a promise as essentially the action of a free agent. Both views are clear that
speaking is a voluntary activity, and that the speech of others has that kind of sig-
nificance in our lives. Where they differ is in how that freedom is related to the
status of the utterance as a reason to believe. For on the Assurance view, it is not
just that a particular free action is seen to have some epistemic import, but rather
that the epistemic import of what he does is dependent on the speaker’s attitude
toward his utterance and presentation of it in a certain spirit, whereas by contrast,
it is in the nature of genuinely evidential relations that they are not subject to any-
one’s conferral or revocation. It might still be asked, however, whether it doesn’t
still all come down to evidential relations in the end. The following reconstruc-
tion may be offered: Yes, the speaker freely assumes responsibility for the truth of
what he asserts. But now this very act of assurance is a fact, which the audience
confronts as evidence (of some degree of strength) for the truth of what has been
asserted. Speech is acknowledged to be importantly different from other (indicat-
ively) expressive behavior, but the audience’s relation to it, as a reason to believe
something, can only be evidential.
The claim of the Assurance view, however, is not that an assertion could not
be treated purely as evidence. It is always possible to treat anything a person says
or does as constituting further evidence for one thing or another, and there is no
level at which this somehow becomes impossible. The point instead is that refus-
ing to acknowledge any epistemic stance toward the speaker’s words other than
as evidence means that speaker and audience must always be in disharmony with
each other, for in the contexts of telling, promising, and apologizing the speaker
is not presenting his utterance as evidence. And it is internal to the speech acts of,
e.g., telling or thanking that they are not presented as evidence for one’s belief or
gratitude. To present one’s utterance as evidence would be to do something other
than to tell, promise, or apologize.
This claim may seem paradoxical. On the Assurance view, the making of an
assertion can be treated as evidence, can properly be evidence for various things,
but the practice cannot coherently be described as the offering of evidence. But
how could this be? If the speaker recognizes that his asserting can be, or even
just is, evidence for the truth of the very proposition asserted, then how could
there be anything amiss with him presenting it as something (viz. evidence) that
he sees it legitimately is? But this general possibility for self-defeat should not be
surprising. To allay the sense of paradox here, compare the assurance given in
a promise with that of an assertion, and consider the incoherence or self-defeat
in saying something like, I promise; but of course I might change my mind, or
forget, or cease caring. Here as well, the speaker is only saying something that
298 Getting Told and Being Believed

both parties know to be true, about himself and about promises in general. But
to say so is, at the very least, contrary to the spirit in which a promise is made,
contrary to the very point of making a promise. And what makes for this self-
defeat is precisely the presentation of it in an evidential spirit. For notice: for
someone to say ‘I promise, but I might change my mind’ is to refer to his promise
as a fallible indication of future performance. That is, it is to present it as a kind
of defeasible evidence for what he will do. And, of course, insofar as a promise
is seen as evidence at all it can only be seen as defeasible evidence. Hence for the
speaker to offer his promise as evidence means he must be offering it as, at best,
defeasible evidence, with respect to which the promisee is on his own. And to do
so is contrary to the point of making a promise, which is assurance.
The disharmony between speaker and audience entailed by the Evidential view
comes out in the consideration of two possible responses to receiving a prom-
ise. If someone promises to mail a letter for me, one thing I might do is accept
his promise, placing myself in his hands and taking myself to now have suffi-
cient reason for believing that he will mail the letter. If it turns out he doesn’t
mail the letter, either through carelessness or because he never really intended
to, then I will feel aggrieved and let down. This is the ordinary expectation and
liability to disappointment. I might, however, opt for another kind of response
altogether. Here I don’t accept the promise; I simply don’t go in for that sort of
thing, as I may not accept promises from a small child or (for different reasons)
from someone I despise, but in another way I do take seriously the fact that he
made one to me. In this spirit I may reason: ‘He is unlikely to make a promise
he won’t fulfill, since that would discredit him as a future promisor, and there are
great and obvious advantages in remaining someone whose promises are accep-
ted. Therefore, the fact that he made this promise to me makes it probable that
he will in fact mail the letter. So I believe he will.’ If, on this second scenario,
I later discover that he did not mail the letter after all, my reaction will be dif-
ferent. I will be disappointed, of course, and I will be surprised that he would
discredit himself in this way. But I can’t confront him with my complaint or my
resentment because I never accepted the promise in the first place. My relation to
this person’s promise is similar to my relation to the person I suspect of ‘double-
bluffing’ me. I don’t believe him; there’s no question of that. But nonetheless his
statement that he’s traveling to Minsk functions as my reason for believing that
this is what he will do. In both cases the speaker has made me a free declaration
which I then make evidential use of to infer to the truth of what he says. On the
Evidential view, this second type of response to promises and assertions would
have to be the only epistemically legitimate one, and yet such a reconstruction
would yield an incoherent description of the practices of telling or promising. It
would be incoherent because on such a view the speaker would have to be in the
position of offering assurances that are never accepted, and which he knows are
never accepted, and the audience would nonetheless be relying on the continued
offering of such free assurances to serve as his evidential base.
 Richard Moran 299

The issue of harmony between speaker and audience goes deeper than this,
however, and helps to delineate the relationship between the speaker’s author-
ity to determine the illocutionary status of his utterance, and its actual epistemic
import. In asserting that P, where the context is one of ‘telling’, the speaker is not
in a position to constitute his utterance as a good or sufficient reason for P, since
that will depend on his credentials and success as a knower, as well as his hon-
esty. But in the act of telling his audience that P, he does claim definitive ‘say’ in
determining that his utterance is being presented as a reason for belief, rather than,
say as a speculation or grammatical example, as well as determining just what it
is that he is giving his word on. From the speaker’s perspective both determin-
ations matter to the alignment of speaker and audience that he sees himself as
aiming at. From a purely evidential perspective, however, it shouldn’t matter to
the audience whether the route from the speaker’s words to a true belief involves
the loop in double-bluffing or not. And just as clearly in such a case there would
be failure of correspondence between the spirit in which the statement is made
and that in which it is received. The speaker who asserts P is not indifferent to
whether he induces belief in his audience through the loop of double-bluffing.
His assertion is asking for belief in the very proposition stated and for the very sort
of reason that he is then and there presenting to him. And that sort of reason is
bound up with his presenting himself as accountable for this truth. In double-
bluffing, the reason for belief taken by the audience is different from the reason
the speaker offers. What ‘telling’ aims at, by contrast, is that there be a corres-
pondence or identity between the reason the speaker takes himself to be offering
and what the audience accepts as a reason. So we might say, in telling his audience
that P, the speaker asks that his authority be acknowledged to determine what
sort of candidate reason for what belief is up for consideration. This is the spirit in
which his statement is made, and it is this that is denied by treating his utterance
in a wholly evidential spirit, in which the question of what is being considered a
reason for what is anybody’s business, and is not tethered to the speaker’s aware-
ness or intent. Conversation may of course move into and out of this dimension
of assessment, but for purposes of either agreement or disagreement it cannot
begin there.
In the speech act of telling, the speaker commits himself to his audience with
respect to a particular proposition, and with respect to the kind of reason being
presented. This follows from the difference between doing something that has a
certain epistemic significance (as with taking or showing a photograph) and being
responsible for something’s having the epistemic significance that it has (as with
a speaker and his words). In telling his audience something, the speaker aims at
being believed, an aim which is manifest to both parties, and which binds the
speaker and audience together with respect to a norm of correspondence between
the reason offered and the reason accepted. When an act of telling completes
itself, speaker and audience are aligned in this way through their mutual recog-
nition of the speaker’s role in determining the kind of reason for belief that is up
300 Getting Told and Being Believed

for acceptance, so that when the speaker is believed there is a non-accidental rela-
tion between the reason presented and the reason accepted. The speaker says, in
effect, ‘‘The kind of reason for belief you gain from my statement is precisely the
kind of reason for belief I am hereby presenting myself as offering you. Insofar as
there is a disparity between the two, I disavow responsibility for whatever belief
you may derive from my assertion.’’ Presenting his utterance that way is a kind
of declaration of transparency to his audience: the kind of reason overtly presen-
ted is precisely the reason that is meant to count for you. When the background
of the speaker’s knowledge and sincerity can be assumed, and the speaker is in
fact believed by his audience (a common enough occurrence, after all) the two
parties are in sync with each other in a way that they would not be if the audience
were to take the utterance either as a reason for some other belief rather than the
one stated, or a different kind of reason for that belief (as with double-bluffing).
Taking the utterance as evidence detaches the reason-giving significance of the
utterance from the speaker’s authority to determine what he is thereby commit-
ting himself to.²⁵ From an evidential perspective it may function as evidence for
any number of things, for which the speaker’s competence or responsibility may
be irrelevant. This is manifestly not the speaker’s perspective on the epistemic
significance of his statement, which he sees in terms of the nexus of a specific
responsibility assumed and a specific entitlement conferred.²⁶
For the act of telling to complete itself there must be a correspondence between
the reason being presented by the speaker and the reason accepted by his audi-
ence. This is the nexus that is aimed at in the self-reflexive aspect of the Gricean
formula, wherein the speaker asks that the very reason he is thereby presenting
be the reason that the audience thereby accepts (i.e., through recognizing that
very intention). Telling aims at being believed, which proceeds, via the speak-
er’s overt assumption of responsibility, by joining together the particular belief
proposed for acceptance, the kind of reason being presented for it, and the reas-
on accepted by the audience. An evidential stance, by contrast, de-couples all of
these from each other, to be reassembled as the observer thinks best. But such a
stance is contrary to the speaker’s perspective on his action, insofar as it pictures
his presentation of himself as meaning, in effect, that as far as reason-giving force
goes, the audience is on his own; as if the meaning of his utterance were ‘‘Now I
have spoken; make of it what you will’’, rather than ‘‘Take it from me’’.
More is conveyed in our ordinary assertions that the specific proposition asser-
ted, and more is often intended to be conveyed by the speaker, and much of this
will be picked up in an evidential spirit. All of which is to say that not all, not
nearly all, speech takes the form of one person telling something to another, testi-
fying to its truth. Not everything we need to convey is best conveyed by being
told to another, in part because not everything we need to communicate is some-
thing we could sensibly ask to be relied on for, present ourselves as accountable
for, or ask to be accepted on our say so (e.g., the occasional comedy or tragedy
in someone asserting his own dignity or probity). Nonetheless, it is clear enough
 Richard Moran 301

what Anscombe means when she speaks of the insult and injury in not being
believed.²⁷ And the offence remains even when the speaker’s audience takes his
having made the statement to count as evidence for its truth, just as above he
may take the speaker’s having made the promise to make it more probable that
he will do the thing in question. The offence lies in his refusing to accept what
the speaker freely and explicitly offers him, in favor of privately attending to what
the speaker’s action passively reveals, just as someone might refuse an apology
while still taking it in this case to be a reliable indication of remorse. What makes
sense of such refusals is the fact that acceptance of an assertion or an apology
doesn’t just put one in a different epistemic position with respect to the facts, but
brings with it certain vulnerabilities and responsibilities of its own. Accepting an
apology, for instance, brings with it the responsibility to put away one’s resent-
ment, and makes one vulnerable to a particularly bruising possibility of deceit.
These risks are avoided by simply taking the apology as more or less good evid-
ence for remorse, and then making of it what one will.
The Evidential picture puts speaker and audience into disharmony with each
other in mislocating the connection between what the speaker does and the fact
that it provides a reason for belief. From the speaker’s point of view, it is not a
matter of what his behavior passively indicates, but a matter of what he then and
there presents himself as assuming responsibility for. Unlike an evidential rela-
tion, the connection between the speaker’s words and what he asserts or what he
promises is entirely at his disposal to declare or to retract. The possibility of such
retraction is central to the meaning of speech acts of assertion, promising, and the
like, and shows how different they are in meaning and consequence from oth-
er actions. The speaker alone has the authority to bestow such epistemic import
on his words, or cancel it; whereas he speaks with no such authority over the
evidential import of anything he does or says.
Speaking of course, is an action; something with consequences in the world
like other actions, and which leaves behind evidence of itself. But the exclusivity
of the speaker’s authority shows that retracting one’s words is not to be confused
with undoing the consequences of an action that went wrong. Often enough,
another person could in principle clean up after the mess I made as well as I.
But no one else can take back what I said. And, of course, taking back what I
said does not make it as if it never happened. After I’ve taken back what I said
it may still take a long time for me to undo the damage my hard or thoughtless
words have caused. But that doesn’t mean that taking them back doesn’t accom-
plish anything, that I might just as well not have returned to them. There’s still
an important difference in the situations before and after the angry words are
retracted. Indeed, taking them back was a prior condition for the more practical
(or consequential) work of starting to try to undo the damage they caused. And
that’s a different kind of task. I cannot ‘hereby’ undo the damage, the hurt feel-
ings I caused; that takes consequential work, which may fail in unforeseen ways,
like any other action. But again, that doesn’t mean that the sort of thing I can
302 Getting Told and Being Believed

accomplish ‘hereby’ is something I perform magically or effortlessly, or that it

may not be something I can only manage with great difficulty. It will, however,
be a different kind of difficulty from that involved in repairing the actual damage.
If I try and fail here, at the level of retraction or apology, it will not be for reasons
of unforeseeable accidents, or the general resistance of the world to our wills.

REFERENCES
Anscombe, G. E. M. (1979), ‘What Is It to Believe Someone?’, in C. F. Delaney (ed.),
Rationality and Religious Belief (Notre Dame, Ind.: University of Notre Dame Press).
(1981), ‘Faith’, in Ethics, Religion and Politics: Collected Philosophical Papers, Vol. III
(Minneapolis: University of Minnesota Press).
Austin, J. L., (1962), How to Do Things with Words (Cambridge, Mass.: Harvard Univer-
sity Press, 1962).
(1979), ‘Other Minds,’ in Philosophical Papers (Oxford: Oxford University Press).
Brandom, Robert (1994), Making It Explicit (Cambridge, Mass.: Harvard University
Press, 1994).
Burge, Tyler (1993), ‘Content Preservation’, Philosophical Review (Oct.).
Cavell, Stanley (1979), The Claim of Reason (Oxford: Oxford University Press).
Coady, C. A. J. (1992), Testimony: A Philosophical Study (Oxford: Oxford University
Press).
Darwall, Stephen (forthcoming), The Second-Person Standpoint: Morality and
Accountability.
Grice, H. P. (1957), ‘Meaning’, Philosophical Review, 66: 377–88; repr. in P. F. Strawson
(ed.), Philosophy of Logic (Oxford, 1967), 39–48.
Harman, Gilbert (1986), Change in View (Cambridge, Mass.: MIT).
Holton, Richard (1994), ‘Deciding to Trust, Coming to Believe’, Australasian Journal of
Philosophy, 72: 63–76.
Hume, David (1977), ‘On Miracles’, in An Enquiry Concerning Human Understanding
(Cambridge, Mass.: Hackett).
Leite, Adam (2004), ‘On Justifying and Being Justified’, Philosophical Issues (a suppl. to
Nous), 14., Epistemology, 219–53.
Locke, John (1975), Essay Concerning Human Understanding, ed. Peter Nidditch
(Oxford: Clarendon Press).
Moran, Richard, ‘Problems of Sincerity’ (2005), Proceedings of the Aristotelian Society,
105/3: 341–61.
Quine, W. V. O., and Ullian, J. S. (1970), The Web of Belief (New York: Random
House).
Ross, Angus (1986), ‘Why Do We Believe What We Are Told?’, Ratio ( June).
Stone, Martin (2001), ‘The Significance of Doing and Suffering’, in G. Postema (ed.),
Philosophy and Tort Law (Cambridge: Cambridge University Press).
(1996), ‘On the Idea of Private of Law’, Canadian Journal of Law and Jurisprudence,
9/2 ( July) (repr. in E. Weinrib (ed.), Tort Law, The International Library of Essays in
Law and Legal Theory, 2nd ser. (Aldershot: Ashgate Publishing, 2003)).
Strawson, P. F. (1974), ‘Freedom and Resentment’, in his Freedom and Resentment (Lon-
don: Methuen), 1–25.
 Richard Moran 303

Thompson, Michael (2004), ‘What Is It to Wrong Someone? A Puzzle about Justice’, in

R. J. Wallace, Philip Pettit, Samuel Scheffler, and Michael Smith (eds.), Reason and
Value: Themes from The Moral Philosophy of Joseph Raz (Oxford: Oxford University
Press).
Traiger, Saul (1993), ‘ Humean Testimony’, Pacific Philosophical Quarterly ( June).
Walton, Kendall (1984), ‘Transparent Pictures: On the Nature of Photographic Real-
ism’, Critical Inquiry (Dec.).
Watson, Gary (2004), ‘Asserting and Promising’, Philosophical Studies, 117.
Weinrib, Ernest (1995), The Idea of Private Law (Cambridge, Mass.: Harvard University
Press).
Welbourne, Michael (1986), The Community of Knowledge (Aberdeen: Aberdeen Univer-
sity Press).
Williams, Bernard (2002), Truth and Truthfulness: An Essay in Genealogy (Princeton:
Princeton University Press).
Williamson, Timothy (1996), ‘Knowing and Asserting’, Philosophical Review, 105/4.
(repr. in his Knowledge and Its Limits (Oxford: Oxford University Press, 2000)).
Wittgenstein, Ludwig (1980), Remarks on the Philosophy of Psychology, vol. i, trans.
G. E. M. Anscombe (Oxford: Blackwell).

N OT E S
1. Cf. Locke: ‘‘For, I think, we may as rationally hope to see with other Men’s Eyes,
as to know by other Men’s Understandings. So much as we ourselves consider and
comprehend of Truth and Reason, so much we possess of real and true Knowledge.
The floating of other Men’s Opinions in our brains makes us not a jot more know-
ing, though they happen to be true.’’ (Locke 1975: 1, 4, 23.)
2. Or back to observation sentences: cf. Quine and Ullian (1970: 33–5), which makes
explicit comparison of testimony with the ‘extension of the senses’ provided by tele-
scopes and radar.
3. ‘On Miracles’ (Locke 1975).
4. Testimony: A Philosophical Study (Oxford: Oxford University Press, 1992).
5. I follow other recent writers in characterizing this as a Humean position, but I don’t
argue for the attribution. For a dissenting view, see Saul Traiger (1993).
6. Coady (1992) and Burge (1993).
7. Burge’s argument is very different, and I will not be examining its details here. In
particular, unlike Coady as we will see, his account does not appeal to a principle
of charity in the situation of radical interpretation. What it shares with Coady’s
argument is the aim of providing some a priori warrant for believing what is said.
Burge argues for what he calls the Acceptance Principle, which states that: A person
is entitled to accept as true something that is presented as true and that is intelli-
gible to him, unless there are stronger reasons not to do so (Burge 1993: 467). He
states that this is not an empirical principle (1993: 469). And the general form of
justification associated with it is meant to apply equally to our epistemic dependence
on other people (‘rational sources’) and to our dependence on certain capacities, what
304 Getting Told and Being Believed

he calls ‘resources for reason’, such as memory and perception (1993: 469–70). By
contrast, for my purposes the difference between our dependence on memory and
perception and our dependence on other people is all-important for the understand-
ing of testimony.
8. Summarizing his line of criticism of one version of the reductionist thesis associated
with Hume, Coady says ‘‘The difficulty consists in the fact that the whole enterprise
of RT’ in its present form requires that we understand what testimony is inde-
pendently of knowing that it is, in any degree, a reliable form of evidence’’ (Coady
1992: 85).
9. With some obvious exceptions; for instance, if I learned of his beliefs from the brag-
ging admission of the person who deceived him. But these sorts of cases are just what
is taken to be excluded by speaking of ‘other things being equal’.
10. When in the course of a discussion of Moore’s Paradox and the idea of ‘‘two people
speaking through my mouth’’, Wittgenstein asks, ‘‘Where is it said in logic that an
assertion [Behauptung] cannot be made in a trance?’’, I understand him roughly to
be saying: Logic (on some conception of it) may well say nothing about the speaker’s
awareness of what he is doing in making an assertion, just as the same conception of
logic permits statements of the form ‘P, but I don’t believe it’, but both possibilities
are contrary to the point and hence the meaning of assertions (Wittgenstein 1980,
§818).
11. I should add that I’ll only be discussing a part of Ross’s argument, focusing on his
criticism of an evidential view of testimony, and not his positive account of how
assertion contributes to belief. Michael Welbourne’s account of testimony in his
monograph, The Community of Knowledge has various affinities with Ross’s, includ-
ing the denial that an act of telling is presented by the speaker as evidence, and
an emphasis on believing the speaker as the target notion for an understanding of
testimony (and the concept of knowledge itself on Welbourne’s view).
12. And if we take it that assertions are made in the course of following out a proof,
or in ad hominem argument, then it’s clear that their role in providing reason for
belief needn’t even depend on the assumption that the speaker believes what he says.
See Moran (2005).
13. For purposes of the account of testimony developed here, we need not of course
follow Walton in his claim that the object itself is in fact seen in the photograph,
for it is precisely the differences between photographs and assertions that concern me
here.
14. Grice (1957), in Strawson (1967: 44). Ross discusses natural vs. non-natural mean-
ing at p. 74.
15. At the end of her paper, Anscombe (1979) points out that it is a requirement of any
successful account of the phenomenon that it explain why we only speak of believing
someone when we take them to be both right about the facts and truthful in intent.
16. Austin (1962: 57).
17. Here I agree with Harman (1986: 88) in seeing as mistaken Grice’s later attempts to
eliminate the self-referentiality in his original formulation for non-natural meaning.
(‘‘Much of this complexity is artificial and due to Grice’s refusal to stick with the
original analysis and its appeal to a self-referential intention.’’)
 Richard Moran 305

18. Coady (1992: 45–6) distinguishes genuine testimony from the situation of someone
who has been hypnotized specifically to say something, perhaps even with the expect-
ation of being believed. But his reason for excluding this case is different from mine.
For Coady, this cannot count as ‘testifying’ because it fails to satisfy the condition
that the speaker has the relevant competence, authority, or credentials to state truly
that P (1992: 42), since the words have more or less been ‘planted’ in the subject.
This focus on epistemic authority seems misplaced, however. For the speaker in this
case could after all also happen to have the requisite competence and authority on
the subject. Instead, the reason this doesn’t count as testimony is that, in his present
condition, he is not presenting himself as responsible for P’s being true.
19. Holton (1994) provides an illuminating focus on the role of what Strawson refers
to as the Participant Stance and the Reactive Attitudes, especially with regard to the
distinction between belief through trust and belief through reliance. Holton says of
the latter case, ‘‘Seen in this way, a person is like a measuring device: they respond
to the environment in various ways, and we infer from their response to what the
environment is like’’ (1994: 74).
20. Something like this assumption is expressed in Bernard Williams’s idea of assertion
as the ‘‘direct expression’’ of one’s belief (Williams (2002: 74 passim), which I discuss
in Moran (2005: 347–50)).
21. ‘‘No abandonment of the agent’s perspective, no abdication of responsibility for
one’s actions, is involved in seeing those actions as generating entitlements and
obligations, either on the part of ourselves or on the part of others. (Compare the
case of promising or issuing a command.) There is on the present account no diffi-
culty in seeing the hearer as taking the speaker’s words in the spirit in which they are
honestly offered’’ (Ross 1986: 79).
See also Robert Brandom’s richly developed account of assertion in terms of the
constellation of entitlements and obligations (Brandom 1994: ch. 3).
22. After a ‘preamble’, Anscombe begins her essay with the statement ‘‘ ‘Believe’ with
personal object cannot be reflexive’’ (1979: 144). See also Cavell: ‘‘A striking excep-
tion to the thought that I can stand in any relation to myself that I can stand to
others is that of belief. Why apparently can I not, in grammar, believe myself ?’’
(Cavell 1979: 393).
23. Cf. Williamson (1996): ‘‘To make an assertion is to confer a responsibility (on
oneself ) for the truth of its content; to satisfy the rule of assertion, by having the
requisite knowledge, is to discharge that responsibility, by epistemically ensuring the
truth of the content. Our possession of such speech acts is no more surprising than
the fact that we have a use for relations of responsibility.’’
24. For further discussion of the assertions and promises, emphasizing their differences
as well as similarities in their relations to the speaker’s responsibilities, see Watson
(2004).
25. That the directness of the audience’s dependence on the person of the speaker as
such is related to the directness of the speaker’s own relation to the reasons on which
he bases his belief is something I have been helped to see by Leite (2004). Because
the speaker’s statement of his reasons is not a hypothesis he makes about the origin
of his belief, his assertion makes him (and not, e.g., something inside him) directly
accountable for the truth or believability of his claim. Leite puts it the following
306 Getting Told and Being Believed

way: ‘‘Suppose that you consider reasons for and against a claim, find that certain
reasons decisively support holding it, and sincerely declare that you believe the claim
for those reasons. In the usual case, you thereby directly determine what the reas-
ons are for which you hold the belief. Moreover, in declaring your reasons you both
open yourself to epistemic evaluation or criticism on account of those reasons’ inad-
equacy and incur certain obligations—in particular, an obligation either to give up
the belief or to seek better reasons, should those reasons prove inadequate. A minim-
al adequacy condition for an account of the epistemic basing relation is thus that it
allow (1) that the reasons for which a belief is held can be directly determined in this
way, and (2) that one sometimes directly opens oneself to epistemic criticism and
incurs further justificatory responsibilities by sincerely declaring that one holds one’s
belief for particular reasons’’ (2004: 227–8).
26. Putting it this way describes the relationship of speaker and audience in terms of an
essentially correlative or ‘‘bi-polar’’ normativity, of the sort that has recently been
explored by a number of philosophers. By contrast, the non-personal nature of evid-
ence, the independence of its epistemic force from its being presented as a reason
to another person, expresses the ‘‘monadic’’ character of its normativity. It was only
at a late stage of working on this paper, that I began to see the direct relevance to
these issues concerning testimony of the work on ‘‘correlativity’’ ‘‘bi-polar norm-
ativity’’ and private law by Ernest Weinrib (1995), Martin Stone (1996 and 2001),
and Michael Thompson (2004); as well as Stephen Darwall’s work (forthcoming) on
the second-person standpoint, all of which will repay further study in thinking about
speech and testimony.
27. Anscombe, (1979: 150). See also Austin (1979: 100).
 Index
Addis, Laird 92 n. 7 higher order 34
Adler, Jonathan 20 n. 16, 46 n. 2, 93, 98, 100, inferential 27, 35, 79
101, 106, 115 n. 14, 183 n. 3, 184 n. 11, non-inferential 28, 35, 107
227, 228, 272 strength of 202
agency 40, 44, 227 uncritical 27
Allport, Gordon 263, 271 n. 8 voluntary control of 40–1, 48 n. 22
Alston, William 25, 47 n. 12, 104, 184 n. 13, belief formation 34, 47 n. 14, 207
215 belief forming process 65
Anscombe, G.E.M 67, 301, 304 n. 15, 305 n. Ben-Ze’ev, Aaron 258
22, 306 n. 27 Bergmann, Michael 20 n. 13, 92 n. 7, 185 n.
anti-reductionism 100, 130, 141 n. 4; see also 18
non-reductionism Blackburn, Simon 248 n. 23
assertion 57–8, 71 n. 7, 77, 84, 121, 247 n. 8, Bok, Sissela 262–3
274, 276, 278–9, 304 n. 10, BonJour, Laurence 20 n. 13, 20 n. 15, 67, 74
and entitlement 205 n. 21, 96, 98, 101, 102, 184 n. 13, 185 n.
and probability 87–9 18, 186 n. 21
and responsibility 143 n. 13, 283, 305 n. Brandom, Robert 305 n. 21
21, 305 n. 23, 305 n. 24 Brewer, Bill 92 n. 4
and signs 57–8, 84 Brunvand, Jan Harold 267, 271 n. 9
speaker’s intention 288, 291–2 Burge, Tyler 20 n. 10, 20 n. 15, 21 n. 22, 21
Audi, Robert 1, 2, 7, 19 n. 6, 20 n. 10, 20 n. n. 23, 71 n. 7, 73 n. 16, 93, 96, 100, 108,
15, 21 n. 22, 21 n. 23, 46 n. 4, 47 n. 8, 114 n. 3, 114 n. 4, 115 n. 11, 128, 139,
47 n. 13, 47 n. 14, 48 n. 15, 48 n. 20, 48 141 n. 3, 142 n. 9, 184 n. 11, 185 n. 18,
n. 22, 48 n. 23, 49 n. 25, 49 n. 27, 49 n. 186 n. 19, 204–7, 220 n. 7, 222 n. 17,
28, 49 n. 29, 68, 72 n. 13, 73 n. 17, 93, 222 n. 18, 222 n. 19, 223 n. 20, 223 n.
96, 105, 114 n. 1, 115 n. 16, 160, 184 n. 21, 248 n. 16, 272, 274, 303 n. 6, 303 n.
11, 185 n. 18, 186 n. 19, 187 n. 29, 220 7, 304 n. 8
n. 7, 222 n. 16
Austin, J.L. 15, 20 n. 10, 21 n. 23, 186 n. 19, Campbell, Tom 270 n. 1
266, 304 n. 16, 306 n. 27 Cavell, Stanley 305 n. 22
authority 128, 265, 291, 295, 301, 305 n. 18; Chakrabarti, Arindam 46 n. 2
see also deference; trust Chan, Albert 50
a priori 49 n. 39, 114 n. 1 Chisholm, Roderick 20 n. 15, 67, 86, 90, 185
fallacy of appeal to 80–1 n. 18,
Reid 33–4, 60 Clement, Fabrice 245 n. 2
reliance on 14, 80, 134, 136, 139, 231, Coady, C.A.J. 1, 14, 19 n. 1, 20 n. 10, 21 n.
236, 249 n. 32, 277 23, 46 n. 2, 61, 67, 71 n. 9, 72 n. 10, 72
autonomy 14, 240, 242, 246 n. 5, 250 n. 34; n. 11, 84, 91 n. 1, 91 n. 3, 93, 100, 149,
see also self-governance 161, 186 n. 19, 220 n. 5, 222 n. 12, 222
autonomous knower 13, 225, 243–4 n. 13, 222 n. 14, 228, 248 n. 16, 254,
265, 273, 274–5, 303 n. 6, 303 n. 7, 305
basing relation 91 n. 2, 92 n. 12, 194, 202, n. 18
209, 213, 217–8, 220 n. 4, 222 n. 15, Coady, David 267
306 n. 25 Coady, Margaret 271 n. 4
Beanblossom, Ronald 47 n. 10, 145 Conee, Earl 110, 221 n. 9
belief: coherentism 28, 74 n. 21, 95, 98, 199–201,
background 27–9, 85, 150 222 n. 15
basic 59–60, 66–7 communication 3, 19 n. 7, 132, 141 n. 1, 145,
cogito 119 229, 246 n. 8, 269
grounds for 27, 29, 35, 174, 226, 228, 241 conferral 288, 293–4
308 Index
construal 27, 42; see also interpretation expository saying 25
Corlett, J. Angelo 46 n. 1 externalism 9, 82, 90, 184 n. 13; see also
credulity: see also Reid, Thomas internalism
natural 32, 35, 68–9 arguments against 152
principle of 8, 33, 50, 52–4, 57–8, 70
n. 2
Faulkner, Paul 20 n. 16, 21 n. 18, 21 n. 21,
93, 101, 114 n. 3, 140, 144 n. 21, 186 n.
Dantzig, George B. 268 19, 188 n. 36, 188 n. 44, 207, 272
Darnton, Robert 258 Feldman, Richard 98, 110, 221 n. 9
Darwall, Stephen 306 n. 26 Ferrero, Luca 272
David, Marian 160 Foley, Richard 20 n. 10, 21 n. 20, 46 n. 2, 98,
Davidson, Donald 84, 230 157, 171, 186 n. 19, 222 n. 14
de Bary, Philip 66 folk psychology 276
deception 29, 40, 48 n. 21, 84, 101, 149, 157, foundationalism 78, 85–7, 201–2, 222 n. 16
176, 277, 290, 292 types of 94–5
Reid 51, 145–6 arguments against 151
self-deception 188 n. 45 Fricker, Elizabeth 1, 2, 13–14, 20 n. 15, 20 n.
defeaters 4–5, 43, 102, 104, 111, 130, 141 n. 16, 21 n. 17, 21 n. 18, 21 n. 22, 21 n. 23,
5, 168, 189 n. 47, 221 n. 8 21 n. 26, 25, 46 n. 2, 48 n. 16, 93, 100,
normative 20 n. 14, 20 n. 15, 185 n. 18 101, 108, 114 n. 3, 115 n. 7, 115 n. 13,
psychological 20 n. 12, 20 n. 13, 185 n. 18 115 n. 14, 127, 128, 129, 131–3, 141 n.
deference 233, 236; see also authority 1, 141 n. 2, 142 n. 7, 183 n. 3, 183 n. 4,
moral and aesthetic 237–9 183 n. 6, 183 n. 7, 184 n. 11, 185 n. 18,
Testimony Deferential Thesis 13–14 186 n. 20, 188 n. 36, 228, 246 n. 6, 247
DePaul, Michael 123 n. 1 n. 11, 247 n. 13, 248 n. 22, 249 n. 27,
Descartes, Rene 225 249 n. 28, 250 n. 35
Dretske, Fred 47 n. 14 Fumerton, Richard 1, 9, 78, 90, 92 n. 5, 92 n.
dualism, epistemic 12, 170–1, 178–9, 187 n. 7, 98
28, 187 n. 30; see also non-reductionism; Fundamentalism 72 n. 11, 96; see also
reductionism anti-reductionism; non-reductionism
Dummett, Michael 20 n. 10, 21 n. 22, 93, arguments against 100–3
108, 184 n. 11, 186 n. 19 Conservative 95, 99
Liberal 9–10, 94–5, 98–9, 108–9
Moderate 10, 98–9, 107, 109
Edwards, Jim 204, 220 n. 7, 222 n. 17 types of 98
Ekman, Paul 71
Elgin, Kate 272
entitlement 205–6, 222 n. 17, 222 n. 19, 223 Gettier cases 164–5, 219 n. 3,
n. 20, 248 n. 16, 305 n. 21; see also Gibbard, Allan 157, 222 n. 14
justification; Burge, Tyler Gilbert, Daniel T. 70 n. 1
epistemic principles 9, 86, 89–90, 96 Goldberg, Sanford C. 1, 11, 20 n. 10, 20 n.
uniqueness of 11, 127, 131–2 11, 93, 128, 160, 185 n. 16, 186 n. 19,
error 156, 157, 177, 242, 277 186 n. 24, 219 n. 3, 246 n. 7
Evans, Gareth 20 n. 10, 21 n. 23, 186 n. 19 Goldman, Alvin 20 n. 10, 20 n. 13, 20 n. 15,
evidence 11, 186 n. 18, 188 n. 41, 236, 241, 46 n. 1, 65, 66, 73 n. 18, 92 n. 11, 98,
248 n. 20, 275, 277 101, 103, 105, 143 n. 16, 184 n. 13, 185
and belief 79, 146 n. 18, 186 n. 19, 186 n. 21, 221 n. 9, 247
and freedom 280–1, 283, 286 n. 14, 250 n. 35
and justification 105, 106, 148–9, 186 n. Goodkin, Franklin 258
18 gossip 14, 254–5
photography 282–3 and curiosity 259, 261
and the self-evident 47 n. 13, 49 n. 27 and morality 255–6
telling as 277, 279 and privacy 257, 260
expertise 171, 233–5, 248 n. 18, 248 n. 19, reliability of 261–2
248 n. 21 and social norms 257
and deference 236, 242 Graham, Peter 1, 9, 19 n. 1, 19 n. 2, 20 n. 10,
moral and aesthetic 237–9, 248 n. 24 20 n. 11, 21 n. 24, 25, 46 n. 2, 46 n. 6,
 Index 309
47 n. 8, 47 n. 14, 48 n. 24, 72 n. 11, 101, Cartesianism 97–8, 102
114 n. 2, 115 n. 10, 141 n. 2, 186 n. 19, connection to truth 96, 276
193, 219 n. 3, 220 n. 5 direct acquaintance theory 78
Green, Christopher 25 doxastic 220 n. 4
Grice, H.P. 16, 143 n. 13, 246 n. 8, 284–5, epistemic support 40, 138–9, 144 n. 19
289, 290–1, 304 n. 14, 304 n. 17 Intuitionism 94, 97–8, 114 n. 5, 115
Gutman, Amy 272 n. 9
loop view of 152, 153, 156–7
Haack, Susan 106 on balance 10, 104–5, 108, 111, 115 n. 15;
Hardwig, John 20 n. 10, 21 n. 22, 46 n. 2, see also justification, pro tanto
184 n. 11, 186 n. 19, 236, 249 n. 28, 249 Pragmatism 97–8, 102, 114 n. 6
n. 32 prima facie 49 n. 28, 72 n. 11, 74 n. 20,
Harman, Gilbert 304 n. 17 104–7, 138, 143 n. 18
Harris, Paul 245 n. 2, 272 propositional 220 n. 4
Hawthorne, John 20 n. 15, 186 n. 18, 225 pro tanto 10, 72 n. 11, 104–7, 111; see also
Henze, Donald F. 63, 73 n. 15 justification, on balance
Hieronymi, Pamela 272 and reasons 34, 47 n. 12, 111, 130, 135,
Hinchman, Ted 93 147
Holland, Margaret 255, 257, 259–60 Reliabilism 65, 88, 90, 92 n. 11, 97–8,
Holton, Richard 305 n. 19 101, 102, 114 n. 6
Huemer, Michael 92 n. 6, 98, 104 taxonomy of theories of 96
Hume, David 1, 9, 20 n. 16, 64, 93 testimonial 8, 43, 171, 220 n. 5; see also
miracles 60, 84, 273 knowledge through testimony;
and reductionism 61, 149, 183 n. 3 knowledge from testimony
Hurley, Paul 93
Kant, Immanuel 256
Kawall, Jason 160
inference 9, 28, 108, 132, 196, 221 n. 11, Keely, Brian 93
226, 276–7; see also belief, Kelly, Thomas 25
inferential/non-inferential Keynes, John 86–7, 90, 92 n. 10; see also
Reid 35, 38, 61, 65, 70 n. 2 probability
testimonial 80–1, 85, 90, 100–1 Knapp, Robert H. 263
and trust 155–6 knowledge:
induction 56, 58, 74 n. 21, 161, 188 n. 34, acquisition by children 32, 60, 68, 74 n. 22,
188 n. 35, 198–9 183 n. 5; see also learning
and learning 71 n. 8, 73 n. 15 animal 68–9, 74 n. 22
and reductionism 5, 132, 149 a priori 42–3
Reid 41, 55, 58, 59 causal theory of 145–6
Insole, Christopher J. 20 n. 10, 186 n. 19 discursive 147, 151
intentionality 82–3, 277–8; see also extension of (additive &
representation quantitative) 39–40
internalism 9, 49 n. 25, 78, 81, 184 n. 13, 221 foundational 78
n. 9; see also externalism grounds for 249 n. 16
inferential 78–80, 86–7 inferential 38, 48 n. 17
interpretation 27, 41–2, 48 n. 17, 229–30, instrumental 10, 116, 121
246 n. 7, 274, 276, 287–8, 303 n. 7 interpretive 10, 121
misinterpretation 27, 43 perceptual 26, 122
of signs/symbols 9, 55, 83–4 reflective 46 n. 5
by signs 57–9
Jack, Julie 141 n. 1 testimonial 3, 8, 128, 133–4, 187 n. 30,
Jones, Karen 239, 272 220 n. 4
justification: from testimony 26, 27, 42, 187 n. 30
for accepting testimony 37, 132, 138, 142 through testimony 26, 179
n. 10, 148, 198, 276 Koenig, Melissa 245 n. 2
a posteriori 274 Kripke, Saul 73 n. 14
a priori 274 Kung, Peter 93
of the attester 31 Kusch, Martin 93, 98
310 Index
Lackey, Jennifer 1, 12, 20 n. 8, 20 n. 13, 20 n. O’Neill, John 271 n. 6
14, 20 n. 15, 20 n. 16, 21 n. 18, 21 n. 24, operational dependence of testimony 31
21 n. 25, 25, 46 n. 2, 46 n. 6, 47 n. 7, 93, other minds, problem of 83
108, 123 n. 1, 127, 141 n. 2, 141 n. 4, Owens, David 20 n. 10, 21 n. 22, 21 n. 23,
141 n. 5, 143 n. 14, 143 n. 17, 183 n. 5, 184 n. 11, 186 n. 19, 215–17, 220 n. 7,
183 n. 10, 184 n. 12, 185 n. 14, 185 n. 224 n. 26, 224 n. 27, 272
18, 186 n. 23, 186 n. 27, 187 n. 29, 187
n. 31, 188 n. 43, 193, 219 n. 3, 221 n. 8,
language: perception 33, 44, 47 n. 9, 102, 235, 246 n. 4
artificial 8, 37, 54–8, 70 n. 4, acquired 8, 47 n. 11, 54
natural 8, 37, 42, 54–8, 71 n. 4, direct realism 78
private 63, 73 n. 16 justification of 105–6, 110, 176
learning 48 n. 15, 54, 225–6, 230, 237, original 8, 54
277–8, Reid’s view of 35
Lehrer, Keith 1, 11–12, 20 n. 16, 47 n. 10, reliability of 39
93, 145, 146, 150, 153, 156, 183 n. 3, Plantinga, Alvin 20 n. 10, 20 n. 13, 21 n. 22,
222 n. 14 46 n. 2, 71 n. 9, 88, 143 n. 16, 184 n. 11,
Leite, Adam 272, 305 n. 25 185 n. 18, 186 n. 19, 186 n. 21
Lewis, C.I. 72 n. 11 Pollock, John 20 n. 12, 20 n. 13, 62, 98, 185
Lipton, Peter 20 n. 16, 100, 183 n. 3 n. 18
Locke, John 225, 303 n. 1, 303 n. 3 Postman, Leo 263, 271 n. 8
lying; see deception Poston, Ted 127, 143 n. 17
Lyons, Jack 20 n. 16, 93, 100, 183 n. 3, 222 Pritchard, Duncan 93, 99, 101, 105, 110, 115
n. 12, 222 n. 13, 222 n. 14, 227 n. 12, 115 n. 13, 115 n. 15
probability 86–7, 92 n. 10, 92 n. 11, 197–8,
McDowell, John 20 n. 10, 20 n. 15, 21 n. 22, 211, 247 n. 14; see also Keynes, John
93, 108, 184 n. 11, 185 n. 18, 186 n. 19, promising 266, 296, 298, 305 n. 21
245 n. 1 Putnam, Hilary 82, 92 n. 8
McGrath, Matt 160
Magnus, P.D. 74 n. 22 Quine, W.V.O. 246 n. 3, 303 n. 2
Mann, Wolfgang 272 Quinton, Anthony 93
Marneffe, Peter de 272
Matilal, B.K. 46 n. 2
meaning 54–5, 62–3, 83–4, 230–1, 246 n. Rafalski, Anna 1
8; see also Grice, H.P. rationality 86, 168, 170, 184 n. 13, 232, 248
natural 284–5 n. 16, 250 n. 24
non-natural 16, 284–7, 291 reasons:
memory 33–4, 39–40, 44, 64, 72 n. 11, 195 non-reductive 199
justification of 13, 196–8, 210, 221 n. 8, positive 5, 160, 161, 163, 164, 165, 166,
221 n. 10 170, 172, 174, 188 n. 41; see also
reliability of 39, 63–4 reductionism
Millgram, Elijah 20 n. 10, 186 n. 19 reductive 108, 196–8, 221 n. 11, 222 n. 14
Minar, Ed 272 sensitivity to 209–10, 212, 223 n. 22
monitoring 28 and speaker’s intention 287, 291
Moran, Richard 1, 15, 304 n. 12, 305 n. 20 statistical 221 n. 12, 222 n. 12
of the testifier 210, 212
narrative 25, 253 reductionism 5, 20 n. 16, 61–3, 67, 71 n. 10,
Neubauer, Hans-Joachim 265, 271 n. 8 100, 130, 141 n. 4, 149, 187 n. 28; see
Neyman, Jerzy 268 also anti-reductionism; non-reductionism
Nichols, Peter 1 arguments against 6, 12, 100, 170, 171–2,
non-reductionism 4, 61, 63, 72 n. 11, 149, 179, 183 n. 8, 184 n. 13, 185 n. 17,
166, 186 n. 19, 187 n. 28, 189 n. 47 see 304 n. 8
also anti-reductionism counterexamples to 163
arguments against 6, 12, 170, 179 global 5, 161–2, 183 n. 4
counterexamples to 167, 169 local 5, 21 n. 17, 162, 183 n. 7
Nozick, Robert 20 n. 13, 171, 185 n. 18, 186 Reed, Baron 1, 20 n. 13, 20 n. 14, 20 n. 15,
n. 21, 188 n. 32 160, 185 n. 18, 188 n. 45
 Index 311
Reid, Thomas 1, 7, 8–9, 20 n. 10, 44, 47 n. self-governance 229, 239, 242, 246 n. 5, 250
10, 47 n. 13, 93, 186 n. 19 n. 34; see also autonomy
basic knowledge 38, 60 Sellars, Wilfred 82
belief 34, 35, 57 Senor, Thomas 222 n. 16
children 51, 60, 74 n. 22, 145 Shah, Nishi 272
justification 37 Shibutani, Tamotsu 264, 265
language 42, 54–6, 71 n. 4, 71 n. 6, 71 n. Shieber, Joe 160
8, 73 n. 17 Shoemaker, Sydney 73 n. 15
memory 33, 64 signs:
perception 47 n. 11, 54–6 artificial; see signs, natural
principles 34–5, 47 n. 12, 49 n. 26, 50–4, natural 35, 49 n. 26, 53, 70 n. 4, 82, 92 n.
59, 66 7
psychology 59, 70 n. 2, 73 n. 17 sincerity 31, 35, 47 n. 8, 101, 114 n. 3, 121,
reliability 66 167, 176, 184 n. 11, 231, 242, 247 n. 13,
trust 33, 145 250 n. 35
reliability 48 n. 19, 114 n. 3, 114 n. 3, 115 n. skill 234
9, 141 n. 2, 141 n. 4, 150, 184 n. 13 Smith, J.C. 145
of basic sources of knowledge 39, 101, 161 Sosa, Ernest 1, 2, 10–11, 20 n. 9, 20 n. 10, 20
contextual factors 173 n. 15, 46 n. 2, 46 n. 5, 68, 106, 160, 171,
of instruments 117, 120, 136, 142 n. 10, 184 n. 13, 186 n. 18, 188 n. 33
142 n. 14 Sousa, Ronald de 270 n. 2
of the testifier 164–5, 170, 171, 173, 183 Stern, William 271 n. 8
n. 10, 187 n. 31, 188 n. 43, 261, Stevenson, Leslie 20 n. 10, 93, 186 n. 19
280–1 Stone, Martin 272, 306 n. 26
testing for 39, 48 n. 18, 151 Strawson, P.F. 20 n. 10, 21 n. 20, 93, 100,
representation : 172, 186 n. 19, 290, 304 n. 14, 305 n. 19
magical accounts of 82–3, 92 n. 8 Suls, Jerry 258
naturalistic accounts of 82 Sussman, David 272
responsibility 136, 142 n. 10, 142 n. 12, 224
n. 26, 243, 250 n. 34, 286, 289 Taylor, Gabrielle 259
individual 214, 224 n. 25, 280–1 telling 15, 25, 246 n. 6, 247 n. 8, 273, 285,
social 215–17, 224 n. 27 292, 299–300
Reynolds, Steven L. 21 n. 22, 184 n. 11 testimony 2–3
Rodin, David 271 n. 3 act of 26
Root, Michael 20 n. 16, 93, 183 n. 3, 184 n. as an a priori source of knowledge 9, 62–4
11 Assurance view 16, 278, 291
Rosati, Connie 272 as a basic source of knowledge 7, 36, 38, 41,
Rosnow, Ralph 255, 257 60, 72 n. 13
Ross, Angus 16, 19 n. 1, 20 n. 10, 21 n. 22, in contrast to basic sources of knowledge 8,
93, 184 n. 11, 186 n. 19, 272, 279, 287, 31, 33–4, 37–44, 60, 176–7
291, 293, 296, 304 n. 10, 304 n. 14, 305 corroboration of 67
n. 21 credibility requirements on testimony 31
Roth, Abraham 221 n. 7 degenerate forms of 269
rumor 15, 262–6 dependence on 31, 276, 279
merit of 264 as distinctive 127, 131, 133, 134, 162
Russell, Bertrand 78, 260–1, 269 epistemic buck-passing 11, 134–5, 137,
Rysiew, Patrick 20 n. 10, 93, 186 n. 19 139, 142 n. 11, 144 n. 20
as an essential source of knowledge 36,
safety 118–9, 171 37–8, 44
Saunders, John Turk 64, 73 n. 15 Evidential view 16, 277–9, 296–8, 301,
Schieber, Joe 186 n. 21 304 n. 11
Schiffer, Stephen 225, 246 n. 8 heterogeneity of 132, 178, 183 n. 6, 189 n.
Schmitt, Frederick 1, 12, 20 n. 10, 21 n. 20, 46
21 n. 22, 46 n. 1, 150, 151, 152, 155, as inference 80
157, 158, 184 n. 11, 186 n. 19, 188 n. and memory 213, 214, 217
34, 188 n. 35, 215, 221 n. 9, 222 n. 13, misfire of 15, 266; see also Austin, J.L.
222 n. 14 overhearing 296
312 Index
testimony (cont.) Van Cleve 1, 8–9, 20 n. 16, 21 n. 19, 47 n.
and perception 47 n. 9, 110–11, 115 n. 10, 12, 59, 66, 73 n. 17, 183 n. 3
153, 176–7, 216, 226 Velleman, David 272
as a source of basic knowledge 7, 44, 61–7, veracity, principle of 8, 33, 50, 51–4, 57–8,
72 n. 13, 70 n. 3; see also Reid, Thomas
Testimony Thesis 73 n. 15 Vogel, Jonathan 272
theatrical saying 25
Thielke, Peter 93 Walton, Kendall 282, 304 n. 13
Thomas, Laurence 254 Warfield, Fritz 25
Thompson, Michael 272, 306 n. 26 warrant 142 n. 10, 182 n. 1, 184 n. 13, 241,
Traiger, Saul 303 n. 5 249 n. 31, 276; see also justification
Transindividual Thesis 12–13, 194–5, 209, Watson, Gary 305 n. 24
212, 217 Webb, Mark Owen 20 n. 10, 21 n. 20, 46 n.
transmission 6, 21 n. 22, 21 n. 23, 143 n. 13, 2, 171, 186 n. 19
144 n. 21, 145–7, 240, 254, 260, 263, Weiner, Matthew 20 n. 10, 93, 108, 115 n. 7,
265 186 n. 19, 272
creationist counterexample 29, 46–7 n. 6 Weinrib, Ernest 305 n. 26
opacity objection 11 Weinstock, Daniel 272
Transtemporal Thesis 13, 195, 209, 213, Welbourne, Michael 20 n. 10, 21 n. 22, 184
217 n. 11, 186 n. 19, 248 n. 16, 304 n. 11
trust 11, 130, 132, 245 n. 2 Williams, Bernard 305 n. 20
of others 148–9, 152, 157, 228, 240 Williams, Michael 20 n. 15, 185 n. 18
of self 151–2, 156 Williamson, Timothy 20 n. 10, 21 n. 22, 21
truth 33, 51, 66, 78, 84–5, 86–7, 90, 96, n. 23, 184 n. 11, 186 n. 19, 188 n. 33,
119, 138, 144 n. 20, 148, 206–7, 223, 305 n. 23
253, 265, 269, 280–1, 291, 305 n. 23 witnessing 254
conductivity 164, 170, 188 n. 38, 188 Wittgenstein, Ludwig 63–4, 73 n. 16, 304 n.
n. 39 10
and trust 148–50, 156–8 Wolterstorff, Nicholas 47 n. 11, 67, 70 n. 3,
71 n. 7
Ullian, J.S. 303 n. 2
understanding 229, 275–6, 290; see also Yaffe, Gideon 50
interpretation
urban myth 15, 266–9 Zagzebski, Linda 123 n. 1
Utterance Thesis 73 n. 15 Zimmerling, Ruth 271 n. 5