Lassiter Sociolinguistics SemExternalism

Semantic Externalism, Language Variation, and
Sociolinguistic Accommodation
DANIEL LASSITER
Abstract: Chomsky (1986) has claimed that the prima facie incompatibility between
descriptive linguistics and semantic externalism proves that an externalist semantics is
impossible. Although it is true that a strong form of externalism does not cohere with
descriptive linguistics, sociolinguistic theory can unify the two approaches. The resulting
two-level theory reconciles descriptivism, mentalism, and externalism by construing
community languages as a function of social identication. This approach allows a fresh
look at names and denite descriptions while also responding to Chomskys (1993,
1995) challenge to articulate an externalist theory of meaning that can be used in the
scientic investigation of language.
Nobody can be forbidden to use any arbitrarily

producible event or object as a sign for something.
G. Frege, On Sense and Reference
1. Semantic Individualism and Semantic Externalism

Semantic externalism is rejected by many linguists, as is well known, because
it seems to conict with the mentalist presuppositions of generative linguistics
in the Chomskyan tradition: see Chomsky (1986, 1995, 2000) for extensive
discussion. Even linguists who reject mentalism and generative linguistics,
however, have occasionally objected to externalism on quite different grounds.
Some of the mundane truths of philosophy of language seem to be in conict
with equally mundane truths of descriptive linguistics: in particular, descriptive
linguists typically see variation and change as a ubiquitous part of language, while
philosophers and formal semanticists, for the most part, view a language as a
static object. Indeed, many philosophers adopt positions that make substantive
predictions about how natural language should work in its everyday use, in
particular those of an externalist persuasion, as I will show. However, insufcient
Thanks to Alan Musgrave, Colin Cheyne, Charles Pigden, Heather Dyke, Txuss Martn, and an
anonymous Mind and Language reviewer for extremely helpful comments. The bulk of this essay
was written during a year of study at the University of Otago in which I was supported by the
Frank Knox Memorial Fellowship from Harvard University, which I gratefully acknowledge.
I am also grateful to the Otago Philosophy Department for a very fruitful and enjoyable time.
All mistakes are, of course, my responsibility alone.
Address for correspondence: Department of Linguistics, New York University, 726 Broadway
7th Floor, New York, NY 10003, USA.
Email: [email protected]
Mind & Language, Vol. 23 No. 5 November 2008, pp. 607633.
2008 The Author
Journal compilation 2008 Blackwell Publishing Ltd
608
D. Lassiter
attention has been paid to the issue of whether these predictions are borne out
empirically.
The Chomskyan revolution has received much well deserved attention, but
we do not need to assume his theoretical position in order to derive philosophically
interesting results from linguistics. Rather, undeniable results of descriptive
linguistics seem to militate against dominant trends in philosophy of language.
Many linguists have taken these problems to indicate that we must choose
between externalism and descriptivism, and this is generally taken to be a
sufcient reason to prefer an individualist approach to language. I will argue that
this is a false dichotomy, and that both sides have misconstrued the terms of the
debate. Recent results of sociolinguistic theory, I will show, suggest a way to
bridge the gap between individualist and externalist accounts of meaning and
language in general without abandoning the basic commitments of either
position.
The broad outlines of the debate are as follows. Individualists believe that the
proper object of the scientic study of language is the language of an individual,
his idiolect or, in Chomskyan terms, his mental grammar, knowledge of language, or
I-language. Individualists typically believe that semantic notions such as reference and
meaning are dependent on such individualistic facts. This does not necessarily
mean that social aspects of language are unimportant or that they do not admit of
a scientic description, though some individualists have made this further claim:
cf. Chomsky (1975). However, most individualists do believe that only
individualistic aspects of language can be formalized and used to make predictions
(e.g. about entailment and grammaticality).
Semantic externalists, on the other hand, hold that a language belongs to a
community of language users, and that common languages or communalects exist
above and beyond individuals. According to this conception, a language has an
ontology (e.g. words and grammatical rules, or social practices and/or conventions)
and norms (standards of correctness) that are in some sense independent of the
linguistic competence of individual speakers. In the words of Michael Dummett,
an idiolect is merely a second-order theory: a partial, and partly incorrect, theory
about what the meanings of the expressions are in the common language, that
may be represented as a partial theory of what the correct theory of meaning for
the language is (1986, p. 469). As Dummett makes clear, it is possible for
speakers to be simply wrong in their use of language because a language exists
independently of its speakers. In contrast, under the individualist view, incorrect
usage is a murky social concept, usually a simple failure of communication or a
faux pas.
The debate is important because the side we choose will determine where we
locate crucial semantic notions such as reference, meaning, and truth. Intuitively,
though, Dummetts approach makes sense: sometimes an individuals use of
language is just wrong, and individuals often acknowledge making mistakes upon
reection or correction. Individualism simply cannot account convincingly for this
fact.
2008 The Author
Externalism, Variation, and Accommodation
609
Our discussion of these issues will rely on several works that have made this
point clearly. For reasons of space I will merely summarize these well-known
arguments of Wittgenstein, Putnam, Kripke, and Burge very briey and will not
attempt to argue for their conclusions in detail. I take them to show persuasively
that we have strong intuitions of linguistic correctness that, like intuitions of
grammaticality, are an important empirical bound on the construction of a complete
theory of language. Whether or not this assessment is correct, I hope it will become
clear even to the sceptical that externalism properly construed is not in competition
with descriptivism or mentalism and should not be rejected on these grounds
alone.
Wittgenstein argues in Philosophical Investigations (1953) that rule-following
and meaning cannot be explicated by mere description of what an individual is
doing: to say that an individual is following a particular rule already presupposes
community standards about correct application of the rule (though, unlike
Wittgenstein, I see no reason why this could not be a community of one).
Kripkes Naming and Necessity (1980) undermines the claim that reference
depends on a speakers knowledge; in Kripkes example, I could believe nothing
but falsehoods about Gdel, and yet the name Gdel would still refer to Gdel
when I say it. Kripke argues instead that names are rigid designators and that the
reference of a name is xed by a causal chain leading to an initial baptism.
Putnams paper The Meaning of Meaning (1975) uses the famous twin-Earth
thought experiment to show that the reference of natural kind terms like water
is also insensitive to speakers knowledge of reference. Putnam argues that their
reference is xed partly by environmental facts, in this case the actual chemical
structure of water. Finally, Burge argues in Individualism and the Mental (1979)
that Putnams argument, suitably modied, extends to all terms. Burge shows
that, if a speaker of English were to believe that rheumatism is called arthritis
and that he has arthritis in his thigh, the latter belief would be false (since arthritis
is an ailment of the joints), rather than being true-in-his-idiolect as a descriptive
theory emphasizing knowledge of language would predict. Burge concludes that
the meaning of arthritis is xed by the words use in a community, whether or
not an individual speaker happens to know how his community uses the word.
So, it seems, individual knowledge is not sufcient to determine reference:
environmental and sociohistorical facts are also relevant.1
To repeat, individualism faces a serious problem in accounting for these
systematic intuitions about meaning.2 However, the signicance of the arguments
In what follows I am primarily interested in externalism about linguistic meaning, rather

than externalism about mental content. Burge is primarily interested in mental content,
as is much of the philosophical work that has followed his articles. However, this tradition
focuses on the analysis of language, usually taking for granted that this method
can elucidate questions about the nature of mind. I doubt that this line of research will
turn out to be very illuminating about the nature of mind, but I will not discuss my
reservations here.
2008 The Author
610
D. Lassiter
for externalism has often been misconstrued. Individualists hold that an individuals
language just is her idiolect; the arguments just mentioned show that this position
is too strong. Environmental and sociohistorical facts are relevant to the
determination of reference, but we have seen no reason to believe that these factors
are sufcient for the purpose. A common approach (perhaps common is too
strong, since it is rarely made explicit and less often argued for) is to bite the bullet
and hold that individualistic properties are in fact irrelevant to determining the
character of the language. Idiolects, then, are of interest only to psychologists.
Dummetts quote above exemplies this claim: he takes the language to be an
external entity that is an object of knowledge for speakersnot like tables and
chairs, to be sure, but like social objects such as burial customs or money, which
individuals participate in but do not normally create. A different but related
characterization is due to Lewis (1975, p. 6), who writes that the social aspects of
language are mediated by the fact that a given language L is used by, or is a (or the)
language of, a given population P. Lewis takes for granted that there is such an
object as the population P to be found (an extremely dubious assumption, as we
will see). I will call this strong form of externalism communitarianism because
it holds that a language is possessed jointly by a community and that the identity
of an individuals language is fully determined by the community to which
he belongs (note the analogy with communitarianism in political philosophy).
Communitarianism implies that speech communities pre-exist individual speakers
and are capable of determining a unique community language with or without
their cooperation.3
The upshot of semantic externalism is that there may be a difference between
an individuals language and what she thinks her language is: there must be some
place for normativity of meaning. This clearly falls short of establishing
It is not sufcient to reject this normativity as uninteresting, as Chomsky and Davidson do.
For example Davidson (2005, p. 121): I am not impressed by Michael [Dummett]s or
Burges or Putnams insistence that words have a meaning of which both speaker and hearer
are ignorant. I dont doubt that we say this, and its fairly clear what we have in mind: speaker
and hearer are ignorant of what would be found in some dictionary, or of how people with
a better or different education or a higher income use the words. This is still meaning based
on successful communication, but it imports into the theory of meaning an elitist norm by
implying that people not in the right social swim dont know what they mean. In many cases
Davidsons comments are very much to the point, but we cannot explain all the externalist
arguments as relying subtly on elitism. First, speakers often impose norms on themselves.
Second, it is easy to imagineor indeed locatesituations in which the correct usage is not
included in dictionaries or prescriptive grammars, or is associated with groups outside the
elite. Third, our intuitions about linguistic correctness display a systematicity closely akin to
the grammaticality intuitions of theoretical linguistics. Many linguists think, as I once did, that
externalism is no more than a philosophical justication for linguistic chauvinism. My task,
then, is to outline a theoretically interesting position that accounts for the subtleties of the
intuitive, pre-theoretical notion of correctness.
A communitarian can, of course, hold that individualistic properties are relevant to determining
which language is being spoken. Lewis seems to hold this position.
2008 The Author

611
communitarianism, yet many externalists rely heavily on an unanalyzed notion of

a linguistic community, effectively making this leap without argument. It is true
that communitarianism gives a simple account of the normativity of meaning
and the possibility of error, and we might prefer it as the best available explanation
of externalism for this reason. However, as the next section will show, communitarianism is fatally incompatible with basic descriptive facts of human language.
Details of this failure point toward a new form of externalism that is compatible
with the insights of semantic individualism, which I will develop in Section 3.
I argue that communalects are neither functions of individual knowledge nor
monolithic objects somehow possessed by a community, but are determined
instead by individuals sociolinguistic dispositions. If this is that case, neither the
individual nor the social level of language can be ignored in semantics (or in other
areas of language, I believe). Rather, a theory of human language that respects the
empirical facts must connect the descriptive and the normative in an illuminating
way without disposing of either.
2. Communitarianism and Descriptive Linguistics

Communitarians often downplay or ignore problems relating to language variation
and change. It is easy to see why: if languages exist apart from the linguistic
competence of individuals, only two analyses of linguistic difference seem to be
availableeither one or both of the speakers does not know how to use the
language correctly, or they are speaking different languages.4 In either case, the
interesting questions about variation are not about individuals, but about languages
as the joint property of a community. Once we have given content to the notion
of a speech community, then, variation within the community can be neatly
dismissed as incomplete knowledge of the language. But this tactic leads us quickly
into trouble, for it is extremely difcult to give a descriptively adequate account of
speech communities.
2.1 Individuating Languages

Our ordinary conception of language seems at rst to be unproblematic: usually it
seems clear whether two people speak the same language. However, the obviousness
of this fact to well-educated speakers of modern English does not apply to human
language in general. To see this, we need to distinguish two meanings of the term
language. The rst is political: a language such as English or Urdu is dened
If we want, we can call this speaking different dialects or belonging to different speech
communities. The terminological difference does not matter as long as it is clear that the only
way for communitarians to avoid analyzing variation as incorrect usage is to invoke distinct
non-individualistic standards.
2008 The Author
612
D. Lassiter
primarily as the speech of a certain community of people, often corresponding

roughly to political boundaries. The second notion of language involves mutual
intelligibility. These two concepts do not coincide, and only the second has any real
interest for philosophers and formal linguists. For instance, standard Swedish and
standard Norwegian are mostly mutually intelligible, but some dialects of Swedish
are not. Such situations are commonplace throughout the world; as a result,
politically dened languages are not useful for the externalist.
The mutual intelligibility standard is a more promising way to individuate
languages: two people speak the same language if and only if they can communicate
successfully. An immediate problem is that successful communication is graded and
relative to communicative purposes. In addition, mutual intelligibility is not
transitive. There was presumably an unbroken chain of generations between my
distant ancestors and me such that each generation could speak to its parents and
to its children, and yet it would be absurd to say that I speak the same language as
my distant ancestors. Clearly, then, there is no natural break in the history of
English that is sufcient to distinguish community languages.
Synchronic linguistic variation is often gradual as well, although the facts are less
familiar. Although sharp boundaries between languages sometimes exist due to
geographical or sociopolitical boundaries, in their absence language tends to vary
gradually over geographical and sociopolitical space, simply because people usually
talk more like people they have contact with and less like those they have little or
no contact with. In highly developed and/or recently settled countries this can
lead to a high degree of linguistic homogeneity across vast geographical areas.
However, in less developed countries with poor communication and transport
there is usually far more linguistic variation: for instance, with roughly the area of
California or New Zealand, New Guinea has perhaps 1000 languages (or more or
fewer, depending on your preferred criteria of individuation). An interesting effect
of these sociological facts is the existence of dialect continua:
[S]harp divisions are rare . They normally occur only where there is a
geographical or sociopolitical barrier of such proportions that it severely
restricts communication between subregions. Otherwise, there is likely to be
a single complex of communalects which extends over the entire language
area. Any one communalect will share a few special features with its neighbors
to one side, and a few with those on other sides. No large clearly demarcated
dialect regions will be found. Often communalects spoken at different ends of
a dialect chain may exhibit considerable differencesthey may even be
mutually unintelligiblebut there will exist intermediate transitional
communalects which are mutually intelligible with both extremes (Pawley,
1970, p. 3).
An extreme example of this phenomenon is found among the Indo-Aryan
languages of the Indian subcontinent, where [t]he speech of each village differs
slightly from the next, without loss of mutual intelligibility, all the way from
2008 The Author
613
Assam to Afghanistan (Masica, 1991, p. 25). This vast areamillions of square

kilometreshas hundreds of millions of Indo-Aryan speakers who use hundreds of
mutually unintelligible forms of speech.
The existence of dialect continua shows clearly that mutual intelligibility is not
sufcient to individuate speech communities. Mutual intelligibility is graded and
intransitive, and language variation and change do not proceed smoothly or along
well-dened boundaries. If we were to locate an individual at an arbitrary point
on the map of the Indo-Aryan, Germanic, Fijian, Scandinavian, or Romance
dialect continua, the fact that he shares some feature with certain other individuals
in no way entails that he will share any other linguistic feature with the same
individuals. All we can say is that he will probably share some linguistic features
with them, and how much is shared will be determined by complex sociological
and historical facts. It would be extremely presumptuous to divide groups of
speakers so related into speech communities capable of sustaining the normative
notions of meaning and reference.
It is worth pausing here to be clear about the precise nature of the problem.
Several people have pointed out to me that communitarians seem to be in the
same position as many investigators with regard to their subject matter: psychologists,
for example, investigate memory without a precise denition of which mental
processes constitute memory, and biologists make good use of the species concept
despite its well-known problems and exceptions.5 Why, then, is the impossibility
of precise denition of speech communities damning for communitarians, as I
claim? The basic problem is that (as Aristotle emphasizes in Nicomachean Ethics,
Book 1, ch. 3) different subjects require different levels of precision. The concept
of a language is perfectly workable for describing clusters of speakers whose speech
is roughly similar: indeed, I have freely made use of labels such as English and
Norwegian throughout this essay, and this vague conception of a community
language is indispensable for language description.
Semantics, however, requires a much greater level of precision than these areas
of inquiry. To see this, consider the discussion of the notion species by the
biologist Edward O. Wilson, a prominent defender of the use of this concept in
biology:
Searching for an anchor, willing to compromise in order to nd some process
shared by a large fraction of organisms, biologists keep returning to the
biological-species concept. In spite of its difculties, regardless of the fact that
it can never be employed as an abstract entity like the electron to make exact
quantitative calculations, the concept is likely to continue to hold center stage
for the simple reason that it works well enough in enough studies on most kins
of organisms, most of the time (Wilson, 1999, p. 49).
Thanks to Charles Pigden and an anonymous reviewer for emphasizing the need for this
clarication.
2008 The Author
614
D. Lassiter
The vague concept of a language that is possible in light of the empirical facts of
language resembles the species concept in many respects. (Actually, languages
display fewer and much less clear lines of demarcation than species, since languages
mix quite freely while species are, by denition, incapable of interbreeding.)
However, a language concept that is similar to the species concept Wilson envisions
is not useable in semantics precisely because it can never be employed as an
abstract entity like the electron to make exact quantitative calculations. The entire
point of semantics, as I understand it, is to be able to make exact quantitative
calculations. Truth, reference, and truth-preserving inference make sense only if
the subject matter, the language, is articulated in a precise fashion.
I conclude, therefore, that in light of the sociological and dialectological facts of
human language, terms such as language, dialect and speech community cannot
be dened precisely without doing violence to the empirical facts of human
language. We must, as virtually everyone in linguistics has concluded, give up any
attempt to nd objective and absolute criteria for dening speech communities
(Hudson, 2001, p. 69). We should reject communitarianism rst of all because it
predicts that speech communities should be real objects in the world, but we have
found absolutely no evidence that this empirical prediction is borne out. More to
the point, though, a communitarian semantics is doomed because the things closest
to speech communities in the real world are incapable of supporting the basic
concepts of semantics.
To be sure, communitarianism has an initial plausibility, so much so that many
thinkers assume it uncritically. The real reason for this plausibility, I think, is that
most analytic philosophers are native speakers of one of the standardized languages
of developed countries. But such highly focused languages are not natural or
representative of human language in general. Many communities are linguistically
diffuse, and in such communities speakers may have no clear idea about what
language they are speaking; and what does and does not constitute the language
will be perceived as an issue of no great importance (Trudgill, 1986, pp. 85-6). A
semantic theory must account not just for the languages of privileged groups with
writing systems, dictionaries and grammars, but for human language in general: to
do otherwise would be to embrace an unjustiable elitism. Furthermore, the fact
that we often use speech community and language in a way that seems to
suggest that these terms refer to discrete objects in the world is irrelevant, for the
mere fact that a concept occurs in ordinary thinking about a subject in no way
guarantees that it will appear in a scientic explanation. Given the chaotic nature
of variation and change in human language, I believe, any formal theory of
meaning that relies on the claim that shared languages or speech communities
are something given in the world is hopeless.
2.2 A Problem about Semantic Change

In order to construct an externalist theory of language that avoids the pitfalls of
communitarianism, we must look closely at another descriptive failure of the
2008 The Author
615
theory. Semantic change is acknowledged as a difculty for externalism already by

Kripke in Naming and Necessity (1980), who mentions that he has no account for
what he calls reference shift except to stipulate that sometimes present intention
to refer to a given entity (or to refer ctionally) overrides the original intention to
preserve reference in the historical chain of transmission (p. 163). This is supposed
to explain how Indians refers to Native Americans, for instance, and how
Madagascar refers to an island off the coast of Africa: both names originated in
geographical errors made by European explorers, so that the actual causal source of
the name is not its referent in modern English. But Kripkes account is in conict
with his thesis that it is not how the speaker thinks he got the reference, but the
actual chain of communication, which is relevant (p. 93).6
An externalist account of language change must address three separate questions.
(i) How does speech behaviour change over time and between individuals? (ii)
How do linguistic norms change? (iii) What is the connection between changes in
individual speech behaviour and changes in linguistic norms? (ii) and (iii) are
crucial in a complete philosophical account of language, but communitarianism is
unable to answer them convincingly. The problem is not that communitarianism
rules out change, but rather that, without any specic story about the relationship
between individual use and the nature of communalects, a communitarian account
of language change amounts to hand-waving.
Lets return to Burges (1979) patient who thinks that he has arthritis in his
thigh; well call him Jim. Jim does not know that arthritis is dened as an ailment
of the joints, and has formed a theory about the meaning of arthritis that does not
match his communitys usage. Now suppose Jim tells his wife that he has arthritis
in his thigh. She believes him and also forms an incorrect theory of the meaning
of arthritis. Add Jims children and neighbours to the equation, and we can easily
imagine how Jims entire neighbourhood might begin to use the term arthritis to
refer to both arthritis and rheumatism. This process could be stopped if someone
were to consult a dictionary or a doctor, but suppose for the sake of argument that
no one does.
Changes of this sortreanalyses and other sorts of mistakes diffusing through a
communityare a primary mechanism of language change, accumulating slowly
to the point of mutual unintelligibility and beyond. Language change relies crucially
on mistakes made by individuals, and just as crucially on other speakers adopting
mistaken usages. For example, the English word bead is derived from Old English
bed, which meant prayer. This change seems bizarre until we consider the social
context:
Prayers were, as now, often recited while being counted on rosary beads, and
a phrase like to count (or tell) ones beads had at least two possible interpretations
Kripke himself did not intend to divorce reference from speakers intentions, but his theory
is often invoked in an effort to do just this.
2008 The Author
616
D. Lassiter
for someone who did not already know what was meant by bead: it could
conceivably refer to the prayers that were being counted, or the beads (in the
modern sense) that were being used for the counting. Some speakers apparently
interpreted the meaning of bead as perforated ball on a string (Fortson, 2003,
p. 651).
This reanalysis, though very much a mistake when it was made, is now correct
in English; indeed, it would be incorrect to use bead to mean prayer. Somehow,
one or more speakers or learners of English made a mistake which went
uncorrected, and the mistake spread until, somehow, it became the correct
usage.7
On the communitarian theory, though, an individuals actual use of language
and her linguistic dispositions are irrelevant to the meaning of her words: only the
sociolinguistic state of the collective body matters (Putnam, 1975, p. 229). Thus
Jims use of arthritis is incorrect; if his wife thinks he has arthritis in his thigh, she
is wrong as well; add a third person and he is wrong; and so on. Yet this sort of
slow diffusion of mistakes through a population is, in fact, how language change
takes place. Only when we try to ask the obvious questionHow does incorrect
usage snowball into correct usage?does it become apparent how little
communitarianism explains about the relationship between individual speech and
community language. The usual response is that the usage of a community can
change; but the community is not a well-dened object, as we saw in the previous
section. Likewise, the slow metamorphosis of mistakes into correct usages does not
mesh nicely with the well-dened notions of meaning, reference, and truth that
many of us hold dear. We should not allow lightly that a particular use can be
partly correct, or that a sentence containing a contested term can be sort-of-true,
as we would seem to if heterogeneous usage in a poorly dened linguistic body
determines reference.
Communitarians have several options here. One is to allow that communication
is divorced from reference: speakers can use a term incorrectly yet understand it
in the same way. However, this tactic would seem to exclude the possibility of
communalect change, so that, unbeknownst to us, all of our sentences are false
or meaningless because their semantics are determined by the communalect of
our distant ancestors. Obviously, this result cannot be countenanced. The other
option seems to be to allow that a group of deviant speakers constitutes a separate
The change need not be so drastic, however: it may affect only certain dialects or speakers,
introducing new synchronic variation into a community. Nor is this type of change restricted
to semantics. An example illustrating both possibilities is the merger of f and th in some
varieties of English, so that thin is pronounced like n, three like free, and so on. The difference
between these two sounds is relatively difcult to hear, and the th sound is normally acquired
rather late by language learners and is cross-linguistically uncommon. Thus it appears that the
innovation occurred because some children simply did not acquire the th sound, and the
merger was adopted by other children learning English.
2008 The Author

617
speech community, so that, in Jims neighbourhood, arthritis actually does

mean either arthritis or rheumatism. But this response also leads to strange
results. We can hardly avoid asking: how many people must use a term incorrectly,
and how often, before a new communalect pops into existence to validate their
usage? If we allow that Jims neighbourhood can create a separate communalect,
why not Jim and his wife, or Jim alone? Does one small difference in a sufciently
large group bring a new communalect into existence, or must the speech
behaviour of two groups of speakers differ in many respects before their
communalects are different as well? These questions cry out for an answer, but
communitarians have no story to tell about how any of this happens beyond
unhelpful formulas like meaning is use. Indeed, any specic answer would be
absurd.
Again, we have not derived a contradiction from communitarianism, but
rather shown that it is fatally unable to explain basic facts about human language
except by hand-waving. We could simply take it on faith that communitarianism
is correct and that there must be some correct answer to these questions, even if
we could never nd out what it is (as Williamson (1994) does with structurally
similar but more general problems about vagueness). But this response gives far
too much leeway to our philosophical preferences to override the normal
procedure of constructing reasonable theories and integrating them with other
areas of inquiry. In any case, Williamsons is a strategy of desperation that should
be adopted only as a last resort. I think we should look for another way out, a
theory of language that is able to relate meaning and use in specic and
descriptively correct terms.
3. Underdetermination, Theory-Building, and Sociolinguistics

We have seen that individualism and communitarianism encounter serious issues
of different kinds: individualism has no way to account for systematic and often
shared intuitions about correct usage, and communitarianism makes false empirical
predictions. Our problem, then, is to construct a theory that can account for the
normativity of meaning without taking languages or speech communities as shared
objects. Communitarianisms difculties with semantic change allowed us to focus
in on three questions that will guide our search for a solution. The earlier questions
were about language change, but each has a synchronic counterpart. If we can
answer the synchronic questions, I contend, their diachronic variants will be
answered as well. The questions are:
(i*)
(ii*)
(iii*)
What explains an individuals speech behaviour?

Which norms are relevant to an individuals speech behaviour in a
particular situation?
What is the synchronic relationship between speech behaviour and
linguistic norms?
2008 The Author
618
D. Lassiter
We will try to answer these questions with minimal theoretical assumptions,

instead attempting to establish boundary conditions for any successful externalist
theory.8
3.1 Individual Speech Behaviour

As for (i*), most theorists agree that individual speech behaviour relies on theorybuilding of some sort. For example, Fortsons explanation of the history of bead
quoted above relied implicitly on this notion. A language learner obviously cannot
reach directly into other speakers heads to learn what they mean by a word. As a
result, [a]ny speaker without direct access to the intentions of the speakers around
him or her must gure out what words mean from the contexts in which he or
she encounters them (Fortson, 2003, p. 648). When the context is ambiguous,
language learners or interpreters may form a theory of the meaning of a word,
expression, etc. that does not match the intentions of other speakers. When these
mistakes are not corrected, language change results.
Fortsons observation about language learning is suggestively similar to Quines
claim that the learner has no data to work with but the overt behavior of other
speakers (1968, p. 187). In the practice of historical linguistics, Quinean reasoning
about translation and interpretation is basically taken for granted. Fortsons language
learner is basically Quines radical translator (or perhaps better Davidsons radical
interpreter), who must build up a translation manual from scratch using contextual
clues and considerations of coherence, simplicity, etc. (Quine, 1960).9 As a result,
[m]eaning and belief play interlocking and complementary roles in the interpretation
of speech (Davidson, 1984, p. 141): at a minimum, theories are formed by
attributing beliefs and linguistic forms simultaneously to the interpretee. However
interpretive theories are to be analyzed in a synchronic sense (atomistic, molecular,
or holistic), they are tested holistically against the evidence of the speech a learner
is exposed to.
Despite these similarities, though, Fortson fails to draw Quines conclusion that
there are no meanings beyond what is implicit in peoples dispositions to overt
behavior (Quine, 1968, p. 187): rather, his account presupposes the reality of
meaning and reference. The difference between Old English bed prayer and
English bead can perhaps be described in purely behavioural terms, but the process
My discussion will focus almost exclusively on word-meaning simply because this is the main
topic of the relevant philosophical literature, but my eventual conclusion is intended to be a
more general theory of language. Detailed justication of this claim is not possible here, but
note 16 below (inserted after the main proposal has been outlined, where it will hopefully
make more sense) gives several suggestions as to how the approach of this essay could be
extended beyond word-meaning to structural aspects of language.
Language learning is far more complex and constrained than this characterization might seem
to suggest. In particular, there appear to be many constraints specic to language learning
beyond the normal considerations of explanatory power, coherence, simplicity, etc. But there
do seem to be at least these standard theory-building constraints.
2008 The Author

619
of reanalysis by which the latter was created cannot be described without reference
to someones mental states. In order to account for change, we must allow that the
theories formed by learners and interpreters of language are psychologically real.
(Incidentally, this conclusion makes it possible to endorse mentalism while also
advocating an externalist theory of meaning. The reason is that the interpretive
theories being discussed are crucially involved in the determination of reference,
as I will argue in detail below, especially in section 3.4. Since they are psychologically
real it is an empirical question whether they are partially innate, what the learning
mechanisms (or theory-forming mechanisms) are, and so on for the standard
questions of linguistic theory and psychology of language. Whatever the answer to
these questions is, it will constitute a constraint on the space of possible interpretive
theories; as Chomsky has often emphasized, learning (or interpretation, from my
perspective) is much easier with a small space of available theories, and it is doubtful
that interpretation or learning would be possible if there were no constraints on
theory formation. These psychological objects, whatever they are, will gure
crucially into the determination of reference in a community language in my
eventual theory. Thus, in contrast to the assumptions of thinkers from Chomsky
to Putnam, I do not think that externalism and mentalism are incompatible: I
think mental grammars are fully real, though I do deny the further claim that a
language just is a mental grammar.)
Returning to the main point of the section, it should be clear that adopting
Quines holistic model of translation/interpretation does not commit us to rejecting
the reality of meaning. Quine reasons that, since meaning, reference, and
grammatical structure are underdetermined by observable facts, in cases of
ambiguity there is no fact of the matter about which assignment of meaning,
reference, or grammatical structure is correct. This clearly follows only on Quines
behaviourist and vericationist assumptions, which I will reject without further
comment. However, Quine is correct about underdetermination, which explains
how it is possible for an interpreter to form mistaken theories about an interpretees
communicative intentions and thus plays a crucial role in language change.
Ironically, rather than undermining the museum myth of meaning as he intended,
Quines observations turn out to explain how this myth can be true without
contravening the truism that the learner has no data to work with but the overt
behavior of other speakers (ibid., p. 187).10
Thus Quines observations about underdetermination and Fortsons theory of
semantic change show that an explanation of individual speech behavior involving
10
Actually, historical linguists do not need anything as strong as Quinean indeterminacy: all
that is needed is that the evidence that a learner or interpreter is actually exposed to, as a
matter of contingent fact, underdetermines the analysis of certain linguistic features. At the
crucial moment, there were surely ways that Old English speakers could have used bed
prayer unambiguously so that their children would have formed a correct theory of the
words meaning; but, for whatever reason, they did not. Thus we may wish to say
underdetermined by the observed data rather than observable data.
2008 The Author
620
D. Lassiter
theory-building also provides a simple explanation for change in speech behavior:

an answer to (i*) gives us an answer to (i) for free.
3.2 Individuals and Communities in Sociolinguistics

All of this addresses only descriptive issues, however, and says nothing about
semantic change in the sense of question (ii). How do linguistic norms change? By
parity with (i) and (i*), we may suppose that, if we have an answer to the synchronic
question, the diachronic question will answer itself; so we must give a general
account of what norms are relevant to a given individuals speech behavior in a
given situation. The obvious place to look for such an account is the study of
sociolinguistics. Although theories of language often focus on the propositional
content of utterances, sociolinguists have shown that the form of an utterance also
carries important social information.11
A robust nding of sociolinguistics is that everyone is in command of multiple
forms of speech that are appropriate in different social contexts. Because forms of
speech are associated with certain groups and social situations, the form of an
utterance conveys social information. Speech patterns may be marked for social
status, region, gender, etc.: for example, in the United States, the speech of
southeasterners and Bostonians is often perceived as undesirable, while certain
other dialects are heard as being more rened. These perceptions are closely
correlated with non-linguistic stereotypes and presumably derived from them. As
a result, speakers can indicate their social relationship to listeners by manipulating
linguistic form. For instance, speakers are likely to use more standard forms in a
job interview than at the pub, and to speak differently to friends than to
acquaintances. One particularly interesting nding is that, in a face-to-face
encounter between A and B, As speech style will inuence the form of Bs
utterances. This is known as accommodation (Giles, 1994). Convergence occurs when
A wishes to express solidarity with B and so adopts a speech style closer to Bs;
divergence occurs when A wishes to emphasize social differences from B and so
adopts a dissimilar speech style.
Accommodation theory suggests that the social functions of language cannot be
explained without reference to facts about the social desires of individuals and their
beliefs about the relationship between language and social identity. Individuals use
language to express a social identity, but they can also make mistakes in the process.
Speaker A may use a formal speech style that he thinks is appropriate and be
evaluated as stuffy or pretentious by B; or A may think that a particular speech
style will be evaluated as regionally marked by B, while B interprets the style as
11
For the sake of simplicity I will pretend that variation is categorical, i.e. that speakers use
variation to indicate social facts by using either form x or form y, etc. In reality the situation
is far more complicated: usually style-shifting is accomplished by manipulating percentages of
use of x and y. Weinreich, Labov and Herzog (1968) and much following work have shown
convincingly that structured variation, not categorical usage, is the crucial notion here.
2008 The Author

621
suggesting a low social class instead. Thus a speaker must guess what speech forms
will be most effective in getting his audience to evaluate him in the way that he
wants. Like other aspects of language, then, accommodation crucially involves
fallible theory-building about the social information that a speech form will convey
to a listener.12
A further implication of accommodation theory is that the social functions
of language are implemented not on the basis of actual correlations between
linguistic form and social reality, but on perceived connections relative to an
audience. As Le Page (1997, p. 28) puts it, we do not necessarily adapt to the style of
our interlocutor, but rather to the image we have of ourselves in relation to our interlocutor
(emphasis in original). Le Page argues for a model of speakers projecting an image
of themselves in relation to their universe, and getting feedback as to the extent
to which their images coincideand then either collectively focusing, or allowing
these images to remain diffuse (p. 29). Projection is the crucial notion providing a
connection between individual speech behaviour and the social functions of
language. On this approach, the social functions of language are explained by
individuals modifying their speech behaviour on the basis of rather complex
theories involving two crucial components: (a) How the speaker wants to present
herself to her audience, i.e. where she falls in her projected social world; and (b) What
speech forms the speaker thinks will be evaluated by her audience as placing her
socially where she wants to be. Again, mistakes can be made in implementing (b).
If this is the case, we cannot banish communities altogether: they are fully real,
but dened by the social identication of individuals. In particular, the composition
of the relevant community is determined by the speakers social desires: (b) can be
evaluated as correct or incorrect only in light of (a). Thus, in the sociolinguistic
realm, individual theories of language and social facts are inextricably linked.
3.3 Reference and Deference

I want to suggest that the approach of sociolinguistic accommodation theory can
be extended to semantics and other aspects of language. Not only the social
information carried by the form of an utterance, but all of the publicly available
information it carries, is determined by facts about a community which exists only as
a projection of an individual. This is, I claim, the level at which the concepts of
meaning, reference, and truth are relevant. We can see how this works by looking
at Putnams (1975) argument for the division of linguistic labor: the meanings of
natural-kind terms are determined by criteria that are known and reliably applied
only by certain specialists, and other speakers are often willing to defer to these
experts. Putnam suggests further that this sort of structured cooperation is a
12
Success in accommodation is further constrained by knowledge of or familiarity with the

relevant forms of speech. Obviously I can accommodate less successfully to an Australian or
a monolingual speaker of French than I can to a fellow American, and probably not at all
to a monolingual speaker of a language that I do not know.
2008 The Author
622
D. Lassiter
human universal. I think that Putnam is right on both counts, though natural kind
terms are only a special case: deference (or refusal to defer) is crucial to normativity
of meaning in general.
In the case of an obscure difference such as that between elm and beech, it is
not sufcient to build instructions to defer into the lexical entry for the terms as
Chomsky (1986, p. 18) has suggested. The reason is that someone with incomplete
knowledge might not know that his knowledge was incomplete, as Burges
rheumatic patient Jim did not before his doctors appointment. But since intentions
to defer with regard to specic terms are not sufcientJim might never have
gone to the doctor, but would still have been wrong about having arthritis in his
thighword-meaning must be determined by something else.
Since deference in specic cases of which speakers are aware is insufcient, I
suggest we turn to dispositions to defer. In other words, the meaning of Jims words
is determined by who he would defer to, and in what fashion, given appropriate
information. Appropriate information is not a dodge, but rather indicates that what
a term means must be resolved on a case-by-case basis: if we wish to know what
arthritis means in Jims mouth, we must ask who Jim would defer to with regard
to the meaning of this word. (In fact Jim may or may not be disposed to defer to his
doctors use of the term, just as you may or may not agree with Putnams judgment
that water in twin-English does not refer to water. This turns out to affect meaning
and play a crucial role in semantic change, as I will argue in this and the following
section.) If this is correct, then the totality of Jims dispositions to defer in a particular
communicative situation generate a linguistic system that we can call his communalect.
Of course, it is far beyond our powers to determine what Jims communalect actually
looks like; but then, we are in the same predicament with regard to his idiolect, and
for most purposes a complete knowledge is not needed.
The relationship between individuals and communities in semantics is, I claim,
essentially the same as Le Pages account of this relationship in sociolinguistics.
Jim identies himself socially with certain groups of people, and denes his
language in relation to theirs by an (un)willingness to accept correction: in
short, he projects a speech community. As in the sociolinguistic case, the correct
interpretation of Jims utterances may not be what he thinks the correct
interpretation is. Rather, Jim forms fallible theories about what forms of language
will best communicate the social and propositional information he wishes to
convey. But best communicate cannot be analyzed without reference to the
people Jim identies with in a social sense; thus the identity of his speech community
is determined by individualistic facts, and the identity of his language is determined
by facts about this community.
I have not yet made it clear whether accepting correction is supposed to
involve modifying meanings/concepts or modifying the mapping between
sounds and pre-existing meanings.13 The conception I have in mind is the latter:
13
Thanks to a referee for emphasizing this point and for suggesting mapping as a label for the
type of interpretive theories I have in mind.
2008 The Author

623
interpretive theories are models of a sound-meaning correspondence of this

type. I assume that concepts are at least partly independent of linguistic items,
and that agreement on the sound-meaning correspondence is generally needed
for successful communication. Thus if Jim accepts correction from his doctor
about the meaning of arthritis it is because he wishes to use a word that
successfully communicates the nature of his ailment to the doctor. If the relevant
concepts exist independently of the labels, as I believe, the speaker/interpreters
problem is to match up words and meanings in a way that facilitates
communication, rather than to have the right concepts (whatever that would
mean).
My theory, then, is that the meaning of a word in the mouth of a speaker S
is determined by Ss dispositions to defer to other speakers with regard to the
meaning and use of this word. Returning to Burges example, our intuition
that Jims use of arthritis was wrong before his visit to the doctor is causally
related to the fact that, when he visited the doctor, he modied his use of the
word arthritis. Pre-theoretically, the explanation for this communicative
difculty does not seem to be that Jim and the doctor are speaking different
languages. The reason for this feeling, I think, is that (in my elaboration of
Burges story) Jim is willing to modify his usage when he learns that it did not
match that of the people he wants to communicate with. Jims use would be
wrong even if he were never to go to the doctor because (ex hypothese) Jim
would be willing change his use if he were to go. As a result, the counterfactual/
dispositional element of this theory is crucial to an explanation of our quite
systematic intuitions about the correct meaning of arthritis and, I believe,
words in general.
Now lets elaborate the story a bit and suppose that Jim accepts what the
doctor says for the purposes of his encounters with the doctor, but continues
to use arthritis to refer to rheumatism when speaking to his family and
neighbours. Here he has developed a translation manual between Ours and
Theirs, as Quine would say. We should analyze the word arthritis as having
different meanings in Jims neighbourhood on the one hand and in the
doctors mouth and medical textbooks on the other: between home and the
doctors office, Jim is switching dialects. As users of language, we create
translation manuals for such situations on a regular basis, and quite
unconsciously.
The same analysis accounts neatly for the facts adduced by Putnam and
Kripke. Putnams twin-Earth scenario relied on epistemic differences between
the inhabitants of the two planets. Speakers of English visiting Twin Earth
might use water to refer to XYZ, but only if they did not know its chemical
makeup; if they did, they would presumably reserve the term for water.
Putnams example of Archimedes use of khruss is also parallel to Jims
predicament above. In these cases, though, it is environmental rather than social
information that is at issue. If Archimedes used khruss to refer to fools gold,
he was wrong because, if confronted with modern chemistry, he would modify
2008 The Author
624
D. Lassiter
his usage.14 If other Greek speakers also made the same error, they were wrong
because they would be willing to defer to experts such as Archimedes, who was
willing in turn to defer to, e.g., the results of an experiment capable of
distinguishing gold and fools gold. In all of these cases, linguistic error does not
depend on whether the relevant speaker is ever actually confronted with
information of the relevant kind, but only on how he would change his speech
if he were (as Kripke (1979, pp. 11-12) also suggests).
The latter example illustrates a sort of deference chain that is similar to the causal
chains discussed by Kripke (1980). Kripkes causal chains are, on my theory, just a
special case of this more general phenomenon. Causal chains operate only within
a background of general dispositions to defer: in normal circumstances people are
disposed to defer to someones parents regarding his name. In contrast, the historical
origin of Madagascar or Indians is a mere curiosity. The people we are interested in
communicating with do not use the term in this way, and they are the only ones that
are relevant in determining reference. The fact that causal chains so often do
correctly predict the reference of names turns out to be the result of a sociological
fact about humans: often, we are willing to defer to others about the reference of
a name, and they are willing to defer to others, and so on. But this works only if
the people in question are socially relevant to us and communication with them is
14
A reviewer notes that no evidence has been provided to support the empirical claims about
speakers willingness to defer with regard to the term water, or Archimedes supposed
reaction to the results of modern chemistry: perhaps Archimedes would simply reply that
there are two types of khruss. Indeed, Machery et al. (2004) present an intriguing experiment
which suggests that thought experiments of the twin-earth and Gdel-Schmidt type provoke
different modal intuitions of reference in American and Chinese university students. The
reviewer is correct that I have assumed Putnams empirical claims uncritically, but this is only
for expository purposes: the possibility of variation in deferential dispositions is part and
parcel of my theory. We expect different speakers to display different dispositions to defer,
and individual speakers dispositions to defer may change over time. Thus my theory predicts
that there exist worlds in which the meaning of khruss in Archimedes mouth is the same
as that of English gold, and other worlds in which the meaning is the same as that of English
gold or fools gold. A world of the latter type could be similar to the former in every respect
except for Archimedes dispositions to defer with regard to the term khruss at a particular
moment (a new Twin Earth phenomenon, I suppose).
Likewise, if Kripkes story about Gdel and Schmidt were true, then Gdel would indeed
refer to Schmidt in the mouths of Machery et al.s Chinese subjects, assuming that Machery
et al.s conclusions are sound. Of course, westerners would be liable to understand the term
as referring to Gdel. This interpretive problem between the western and Chinese subjects
would then be similar to that of the two meanings of tea, discussed in section 3.4 below. It
follows that the direct reference theory of names is not universally or necessarily true, but is
contingent on speakers sound-meaning mapping and their deferential dispositions like any
other term. A speaker or group of speakers could map a name to a description, just as a
speaker could make up a word as a shorthand for anything he likes (cf. the quote from Frege
which is at the head of this essay). As a result the semantics of names could in principle vary
from speaker to speaker or even from name to name, though it may turn out to be a
sociological fact about humans that we prefer to map names to an individual rather than a
description (or even that name is nothing but the linguistic label for a term that is mapped
to an individual).
2008 The Author

625
important to us. People who are geographically, temporally, or linguistically

remote, on the other hand, may not be judged socially relevant. Thus we would
be willing to defer to Gdels parents, but not to the Arab traders that Marco Polo
misunderstood when he heard the name Madagascar. These are the facts that
make it true that Gdel refers to Gdel and that Madagascar refers to Madagascar.
As with Gdel, an English speaker need not know anything particular about
Madagascar in order to be able to refer to itall that is necessary is that she would
be willing to defer to English-speaking cartographers regarding its use.
A reviewer suggests the following counterargument. The account of the
reference of names given here seems just as much as those I have been arguing
against to rely on the existence of some sort of superspeakers. Now suppose that
my parents, having named me Daniel shortly after birth, lost their minds several
years later and began to call me Flapadoodle. Would my name then become
Flapadoodle? It seems not; thus, the reviewer concludes, reference is more like
an economic fact. My answer to this question will become clearer in the following
section (see especially the discussion of tea), but it is already implicit in the theory
I am arguing for. Basically, I agree with the intuition behind the example and its
conclusion, and I think that my theory can account for both if we sharpen the
issue of deference. Recall that, in the case of Jim and his communitys deviant use
of arthritis, we came to the conclusion that the deviant usage could be incorrect
in certain circumstances (e.g. talking to the doctor), but it could just as well involve
dialect-switching in which both usages are correct in different social contexts. In
the latter case, some sort of translation manual would be in order.
The following section will explore in much greater detail the conditions under
which the concept of incorrect usage can be usefully employed. Briey, though,
I see no reason why my name would not, in a certain sense, become Flapadoodle
in the situation described. But it would become my name only in the specic sense
that the sequence of sounds Flapadoodle, when used by my parents in an appropriate
contextsay, while communicating with each otherwould in fact succeed in
referring to me. Nicknames and pet names, in fact, seem to work on this principle.
Flapadoodle might or might not refer to me in a broader social context. This would
depend on the knowledge and deferential dispositions of other speakers: whether
they know that this nickname is intended to refer to me, and especially whether they
are inclined to defer rst to my parents, to me, to the law (for my legal name), etc.
(Also, the facts that my parents in the story have gone crazy, and that the name
sounds ridiculous, contribute to the intuition that Flapadoodle would not refer to
me; try substituting Jimmy.) The conclusion, then, is that there is no necessary
connection between reference and any sort of superspeakers, or any other natural
or human phenomenon. Instead, reference relies crucially on sociological facts which
can exhibit considerable variation; the tendency to defer to parents or other sources
of authority is merely a manifestation of this more general phenomenon.
If this account is right, then we have a new account of the old Humpty Dumpty
problem. Can Humpty use glory to mean a nice knock-down argument? Well,
if he is willing to isolate himself socially, yes. But very few, and only very strange,
2008 The Author
626
D. Lassiter
people would behave like this. Furthermore, the afrmative answer makes little
sense in Carrolls story simply because Humpty is talking to Alice: he must place
her somewhere in his social world in order to do so, and the story is strange and
funny precisely because Humpty violates this norm of communication. Of course
almost no one acts like Humpty in the real world, but we should not take this to
imply that it is impossible to do so. People are just normally too practical or polite
to act in this way. Nothing about meaning or reference hinges on these facts, nor
should it. Shared meaning, then, is no more than a reasonable guessit is where
theorizing begins, not where it ends.
3.4 Intentions and Semantic Change

Section 3.3 proposed an answer to (ii*), the question of what norms are relevant
in a given situation. Now what about (ii), the diachronic question? Our new
approach implies that the composition of someones speech community has no
necessary connection to facts about the distribution of any particular way of talking,
and it may or may not be the same as anyone elses speech community. Yet there
is a fact of the matter about what the correct theory of meaning for a speakers
projected community is; thus it is possible for her theory to be incorrect.15
Kripke (1979), following the discussion in Donnelan (1966), discusses cases in
which a description seems to succeed in referring even though it is technically
inaccurate. For example, someone at a party might say, The man over there
drinking champagne is happy tonight. If it turns out that the man is actually
drinking water, the description seems to refer in some sense, even if there is no
one over there drinking champagne. Kripke suggests that we can explain this by
distinguishing speakers reference from semantic reference. Semantic reference is
determined by the conventions of our language together with the intentions
of the speaker and various contextual features (1979, p. 14). Speakers reference,
15
Actually, it is more likely that a speakers projected community will be linguistically

heterogeneous, so that there may be many correct theories that differ in points of detail. A
speaker is unlikely to identify in this way with a group that shows extreme linguistic diversity,
and so the correct theories will usually agree in most respects, though this is entirely
contingent. For most purposes, this is no problem: as noted above, we usually look for
correct theories on a case-by-case basis, rather than trying to isolate entire systems, and we
can note ambiguities when they arise. Anyway, we dont want to build conformity into the
theory, since we must allow for groups of speakers either collectively focusing, or allowing
[their] images to remain diffuse (Le Page, 1997, p. 29). This heterogeneity suggests that, in
contested cases, the labels correct and incorrect are too simplistic. But this is no cause for
alarm: linguistic correctness is merely a pretheoretical concept derived from our intuitions as
speakers of highly focused languages, and we should not expect a theory of meaning to
endorse them without qualication. On the other hand, our new theory may actually match
our instincts about correctness better than communitarianism precisely because of this
heterogeneity. It would be strange to ask whether it is correct or incorrect to speak with a
Cockney accent, or to use shit to mean stuff, things as in youth slang, or to write color
instead of colour. At a minimum, we would need a clarication of the question: Correct
in what sense?
2008 The Author

627
on the other hand, is that object which the speaker wishes to talk about, on a
given occasion, and believes fulls the conditions for being the semantic referent
of the designator (1979, p. 15).
Though Kripke does not discuss the issue in his paper, interpretive difculties
caused by variation can also be absorbed to this model. Suppose a New Zealander
says to me, Tea will be at 6:30. Being a foreigner in New Zealand, I interpret this
as an indication that we will be drinking a particular hot beverage at 6:30; but the
speakers reference of tea here is dinner. As Kripkes formulation suggests, the
crucial condition for something being the speakers referent is that the speaker
must believe that the object satises the conditions for being the semantic referent.
In my terms, the speaker must intend to communicate some social and/or
propositional information (roughly as described by Grice (1957/1967)). Speakers
reference and semantic reference will differ, however, if the speaker has formed an
incorrect theory about what forms of language will actually produce this effect.
There are two ways in which this mismatch can occur, as our two examples make
clear. In the example of Tea will be at 6:30, speakers reference fails to match
semantic reference because the speaker has an incorrect theory of what forms of
language will induce me to acquire the belief that dinner will be at 6:30. In the
case of The man over there drinking champagne is happy tonight, however, the
mismatch is a result of the speaker having a false belief. The two ways in which
communication can go wrong indicate again the inextricability in communication
of theories of meaning and theories of the world.16
16
In this essay my discussion has focused on word-meaning, primarily because it is the historical
focus of the debate on externalism. Several commentators have asked, do my conclusions
about (a) the relationship between individual speakers and word-meaning, and (b) the
inextricability of theories of language and theories of the world in interpretation, generalize to
grammatical aspects of language? I think that the answer is yes, because these aspects of
language gure as importantly in the communication of social and propositional information
as word-meaning. Suppose that word-meaning were partly determined by non-individualistic
factors as I have argued, but grammatical structure were given by purely individualistic
factorssay, whatever grammatical structure the speaker had in mind. Such a theory would
fail to answer the questions about meaning that externalist theorists have raised on the level of
sentence-meaning, if sentence-meaning is compositional. That is, I have claimed that a speaker
can be wrong about the reference of a word like arthritis because there is a mismatch between
his intended meaning and the generally understood meaning in his projected speech
community. But it is easy to nd cases of grammatical mismatcha string of sounds is assigned
a different grammatical structure by the speaker and members of his communitywhich will
encounter precisely the same problems that motivated my version of externalism. The same
logic should apply to these:
Phonology. A word, phrase, or sentence may be misunderstood because of a difference in the
phonologies employed by interlocutors: for example, British speakers often hear American
English latter as ladder, and I have often witnessed non-southerners misinterpreting southeastern
US English life as laugh, height as hat, and so on (because of a rule changing [ay] to [a:] in
certain contexts in many forms of southeastern English).
Morphology. To some English speakers, you all is equivalent to all of you. To others, it is simply
the 2nd person plural pronoun. We can verify this by noting that speakers of the latter type use
all you all (or all yall) to express the concept all of you. The potential for confusion is clear.
2008 The Author
628
D. Lassiter
This conclusion provides a link between the proposed theory of speech communities,
the problem of language change, and our model of individual speech behaviour. The
potential mismatch between speakers reference and semantic reference relies not only
on the speakers beliefs about what sort of utterance would get her audience to form
some belief, but also on the speakers desire to get her audience to form this belief.
Thus the theory presupposes some social connection between speaker and audience.
General dispositions to defere.g. the fact that my friend from New Zealand would
be willing to say dinner instead of tea if he realized that I am liable to misinterpret
the latterare to be explained in light of Gricean intentions. Such intentions are not,
of course, the whole story: sometimes speakers reference may be different from
semantic reference even if no communicative difculty results. Putnams twin-Earth
arguments made this point clearly. Communication through shared errors can also be
successful when a description fails to refer semantically, e.g. if we both believe that the
man over there is drinking champagne when he is not. This is because we would
both be willing to withdraw the presupposition that the man in question is drinking
champagne if we were to learn that he is not. All of this presupposes the model of
speakers both as individualistic theory-builders and as social beings that we adopted
above. (It is worth noting further that, though Kripke presents speakers reference as
Syntax. I once saw a sign in a Delhi post ofce that read, For slow and uncourteous
service please see Chief Postmaster. English is a well-established language in India with
millions of native speakers, but Indian English has numerous grammatical and lexical
differences from standard English that are inuenced by Indian languages, particularly Hindi
(and a great deal of internal variation). To an Indian speaker of English living in Delhi the
sign mentioned would presumbly communicate the meaning that I would express by In case
of slow or uncourteous service please see Chief Postmaster. (Hindi uses the equivalent of
and under negation in some cases where we would use or. Another time I saw a sign reading
Do not sit and run in the station, which I interpreted at rst as forbidding an impossibility.)
Depending on who the sign was addressed to, then, we may or may not want to call this a
mistake; it would depend on the intended audience, the sociolinguistic aspirations of the
signs author, and other factors.
In all of these cases, misunderstanding of the type we have been discussing is possible even
if all lexical items in the language are understood in the same way by all participants in a
conversation. The reason is that world knowledge is needed in order to form a correct
theory of the communicative intentions of the interpretee, for grammar as much as for wordmeaning. Consider the reasoning you would have to go through in a potentially ambiguous
situation in order to correctly interpret an utterance of you all. Relevant facts might include
your own grammatical knowledge; your knowledge of the immediate non-linguistic
situation; your knowledge of the speakers social and regional background and stereotypical
features of the associated dialect, especially differences between this dialect and your own
idiolect; special knowledge, if any, of the speakers speech habits (verbal peculiarities beyond
social or regional stereotypes), and an unlimited amount of further knowledge (say, whether
there is a person in the audience to whom the utterance could not be addressed appropriately).
Thus I think that we can conclude that, just as I have for word meaning in the main text,
that there is (relative to a communicative situation and a choice of social identication) a
correct grammatical form in each situation, viz. the one that will correctly communicate the
intended social and propositional information to the intended audience and other members
of the speakers projected speech community.
Thanks to Txuss Martn and an anonymous referee for pointing out the need for this
clarication.
2008 The Author
629
a pragmatic phenomenon, I conceive of both speakers reference and semantic

reference as being semantic in the sense of being involving a sound-meaning mapping,
although they belong to different levels of language.)
Kripke (1979) suggests further that the possible mismatch between speakers
reference and semantic reference may explain semantic change, since what was
originally a mere speakers reference may, if it becomes habitual in a community,
evolve into a semantic reference (p. 22). Admittedly, the notion of becoming
habitual in a community is obscure since speech communities are not stable or
discrete objects in the world. However, Kripkes suggestion provides a simple
account for semantic change in the normative sense once our new notion of
speech community is implemented. Lets return to the example of bead.
Imagine that Karl is learning Old English. Karl sees someone praying with rosary
beads and asks his mother, Whats he doing? His mother answers, Hes counting his
beads. Karls mother intends for beads to refer to prayers, but Karl surmises that
beads refers to beads. Thereafter, Karl always uses beads to refer to beads, while his
mother always uses beads to refer to prayers. If Karl tells his mother Im counting my
beads, his intention is to induce in her the belief that he is counting his beads, and the
speakers reference is therefore to beads. But beads refers semantically to prayers; and
the belief that Karls mother will form is that Karl is counting his prayers. In a strict
sense, then, Karls utterance is false, and communication between Karl and his mother
has failed. However, neither of them ever uses the word beads in an unambiguous
context, and thus neither ever learns that the beliefs that they induce in each other
using the word beads are systematically different from the beliefs that they intend to
induce in each other. Though communication has failed, no communicative difculty
arises because of contingent features of the environmental and social context.
Now, suppose that Karl teaches other children the word beads by ostention.
These children now use beads intending to induce beliefs in each other about beads,
not about prayers. As long as they would be willing to defer to adults who do not use
beads in this way, however, this is another case of successful communication via
mistakes. This use of beads becomes correct only when speakers who share this error
begin to identify themselves as a sub-population with a distinct social identity. Thus
the children will identify separately with each other and with the broader community:
in this case, they project two distinct speech communities. They would, it seems,
now accept correction from adults, but only as relevant in speaking to adults, and
switch between dialects in different social contexts (as I do with tea). When they
speak to other members of their sub-population, though, the historical origin of the
word bead is beside the point, as it is for Madagascar: they now defer to the people
who made the mistake, and this is enough for correct usage. Their theory of the
meaning of beads is correct in one social context and incorrect in the other.17
17
Rather than being a cause for alarm, it is descriptively a good result that speakers can project
multiple external standards. As noted above, everyone is in command of multiple forms of
speechdialects, registers, and often languages, although often they are so similar that we do
not noticeeach with its own normative standards and social functions.
2008 The Author
630
D. Lassiter
Thus an answer to (ii), the question of change in semantic norms, seems to

involve two factors. First, there may be some change in social identication, i.e. in
individuals projections of the social world; second, there may be changes in the
correct theory for the projected communities, because of changes in other speakers
understanding, misapprehensions, etc. Either of these factors is sufcient for
change, though they often occur together. We can now rene our earlier discussion
of what the correct theory would be in this light. The correct theory is the
theory that would actually cause the audience to form the belief that the speaker
intends, unless the audience would be willing to change this theory in light of
recalcitrant social or environmental facts. The second clause is important to allow
for the two types of shared error. Shared social errors occur when all parties in a
communicative situation are under the same misapprehension about a words
meaning, but would be willing to modify their behaviour if they were corrected
because they both project more or less the same community. Shared environmental
errors include both the example of The man over there drinking champagne and
Putnams cases involving khruss and water: all parties would be willing to defer
to facts about the world. In both cases, the errors result from epistemic limitations,
either in a speakers abilities to infer other speakers linguistic intentions, or in
speakers knowledge of the world.
4. Conclusion
From the preceding considerations I conclude that the choice between individualism
and externalism is a false one. Crucial semantic and social facts can be explained
only by a theory that takes both levels to be philosophically relevant and explores
the relationship between them.
These considerations also show a considerable symmetry in the two levels of
meaning. For example, an idiolect is a theory of the communalect in the same way
that an individuals non-linguistic beliefs are a theory of the actual world. Semantic
reference is jointly determined by the communalect and facts about the world, just
as speakers reference is jointly determined by a speakers idiolect and her beliefs
about the world. If an idiolect and a communalect are both word-meaning
mappings as I have suggested, it may be possible to articulate a two-level formal
semantics with individual- and community-level operators. This may turn out to
have interesting applications to issues in formal semantics. For instance, I am
exploring the possibility of using such a two-level semantics of this type in
combination with a permissive notion of epistemic possibility (i.e. one in which
there are conceptually possible worlds that are not metaphysically possible;
cf. Gendler and Hawthorne, 2002) to address problems of substitutivity in opaque
contexts. (See Muskens, 2005 for a sketch of an implementation of a somewhat
similar idea.)
To sum up, I have argued that three aspects of individual language broadly
construeddispositions to defer, idealized communicative success, and social
2008 The Author
631
identicationmap out the same community language. This community language

is systematic and makes predictions about meaning and reference that are, I think,
intuitively correct to those without a stake in the individualism-externalism debate.
In this way we have met Chomskys challenge to articulate a notion of shared
public language that comes even close to meeting the requirements of empirical
enquiry or serious philosophical reection on language and its use (1995, p. 51).
Furthermore, if I have described the interplay between individual and social
aspects of meaning accurately, we may conclude that a good deal of the philosophical
debate between individualists and externalists has been misguided. Chomsky, for
instance, insists that only the internalist aspects of language can admit of a truly
scientic description. I have attempted to provide several counter-examples to this
claim in the form of explanations of problems that cannot be addressed or even
formulated without externalist concepts.18 Putnam responds to Chomsky by
rejecting the claim that meaning has anything at all to do with facts about
individuals; Dummett relegates the study of individual behaviour to the
philosophically irrelevant realm of psychology; and so on. None of these extremes
is correct, because, as we have seen, neither the social nor the individual levels of
meaning can be explained without reference to the other. For instance, speakers
reference and semantic reference both play important roles in a complete theory
of human language: no straightforward reduction is possible, but it is possible to
devise a theory detailing the relationships between these types of meaning. If this
is the case, individualists and externalists have in many cases simply been talking
past one another, and in reality both the individual and the social aspects of
language are interesting (philosophically, linguistically, and psychologically) and
worth investigating.
In section three, we saw how to rescue speech communities from the difculties
raised by empirical results of descriptive linguistics. This was necessary to explain
the normativity of meaning. The primary role of speech communities in this
theory is to account for externalist intuitions about meaning. However, we should
not take for granted that our reconstructed speech communities can do the
philosophical work that communalects have been called upon to do in much
previous work. They are certainly not able to provide, as Dummett (1978, p. 218)
requires, a guarantee of mutual understanding. Dummett argues that communalects
must be able to do so because otherwise, for all [a speaker] knows, or can ever
know, everyone else may attach to his words or to the symbols which he employs
a meaning quite different from that which he attaches to them (ibid.). This
consequence is intended as a reductio, but attention to the empirical facts of language
18
I do not mean to suggest that Chomsky categorically rejects externalist inquiry into language.
For instance, he admits that sociolinguistics is a perfectly legitimate inquiry, externalist by
denition (1995, p. 50). However, sociolinguistics borrows from internalist inquiry into
humans, but suggests no alternative to it (ibid.); the thrust of his argument in the quoted
passage and the surrounding text is that a scientic theory of language, i.e. a theory capable
of making testable predictions, will be thoroughly individualistic.
2008 The Author
632
D. Lassiter
shows it to be a positive boon: only a theory that does not provide such a guarantee
can provide a convincing account of language variation and change.
Although communalects are real, meaning is not their exclusive provenance.
Speakers reference is equally real and, from the point of view of interpersonal
communication, sometimes more interesting: after all, we are normally more
interested in understanding than in correcting one another. Likewise, meaning
construed as semantic reference is indeed normative, but this normativity relies
crucially on individualistic facts. Externalism is correct, it seems, but the rumours
of individualisms death have been greatly exaggerated.
a
Department of Linguistics
New York University
References
Burge, T. 1979: Individualism and the mental. In P. French, T. Uehling, and H.
Wettstein (eds), Midwest Studies in Philosophy, vol. IV: Studies in Metaphysics.
Minneapolis: University of Minnesota Press.
Chomsky, N. 1975: Reections on Language. New York: Pantheon Books.
Chomsky, N. 1986: Knowledge of Language. New York: Praeger.
Chomsky, N. 1993: Mental constructions and social reality. In E. Reuland and W.
Abraham (eds), Knowledge and Language: From Orwells Problem to Platos Problem.
Dordrecht: Kluwer.
Chomsky, N. 1995: Language and nature. Mind, 104, pp. 161.
Chomsky, N. 2000: New Horizons in the Study of Language and Mind. Cambridge:
Cambridge University Press.
Davidson, D. 1984: Inquiries into Truth and Interpretation. Oxford: Oxford University
Press.
Davidson, D. 2005: Truth, Language, and History. Oxford: Clarendon Press.
Donnellan, K. 1966: Reference and denite descriptions. Philosophical Review, 75, 281
304.
Dummett, M. 1978: Truth and Other Enigmas. Cambridge, MA: Harvard University
Press.
Dummett, M. 1986: A nice derangement of epitaphs: some comments on Davidson
and Hacking. In E. Lepore (ed.), Truth and Interpretation: Perspectives on the
Philosophy of Donald Davidson. Oxford: Basil Blackwell.
Fortson, B.W.IV. 2003: An approach to semantic change. In B. Joseph and R. Janda
(eds), The Handbook of Historical Linguistics. Malden: Blackwell.
Frege, G. 1952: On sense and reference. In P.T. Geach and M. Black (eds), Translations
from the Philosophical Writings of Gottlob Frege. Oxford: Blackwell, 5678.
Gendler Szab, T. and Hawthorne, J. 2002. Introduction: Conceivability and
Possibility. In T. Gendler Szab and J. Hawthorne (eds), Conceivability and
Possibility. Oxford: Oxford University Press.
2008 The Author
633
Giles, H. 1994: Accommodation in communication. In R.E. Asher (ed.), Encyclopedia

of Language and Linguistics. Oxford: Pergamon Press.
Grice, H.P. 1967 [1957]: Meaning. In P. Strawson (ed.), Philosophical Logic. Oxford:
Oxford University Press.
Hudson, R.A. 2001: Sociolinguistics. Cambridge: Cambridge University Press.
Kripke, S. 1979: Speakers reference and semantic reference. In T. Uehling, H.
Wettstein, and P. French (eds), Contemporary Perspectives on the Philosophy of
Language. Minneapolis: University of Minnesota Press.
Kripke, S. 1980: Naming and Necessity. Oxford: Basil Blackwell.
Le Page, R.B. 1997: A sociolinguistic theory of language. In F. Coulmas (ed.), The
Handbook of Sociolinguistics. Malden: Blackwell.
Lewis, D. 1975: Languages and language. In K. Gunderson (ed.), Language, Mind and
Knowledge. Minneapolis: University of Minnesota Press.
Masica, C. 1991: The Indo-Aryan Languages. Cambridge: Cambridge University
Press.
Machery, E., Mallon, R., Nichols, S. and Stich, S.P. 2004. Semantics, cross-cultural
style. Cognition, 92, B1B12.
Muskens, R. 2005. Synonymy, common knowledge, and the social construction of
meaning. In P. Dekker and M. Franker (eds), Proceedings of the Fifth Amsterdam
Colloquium, ILLC, University of Amsterdam.
Pawley, A.K. 1970: On The Fijian Dialects. University of Auckland Working Papers in
Anthropology, Archaeology, Linguistics, Maori Studies, no. 16.
Putnam, H. 1975: Mind, Language and Reality: Philosophical Papers vol. 2. Cambridge:
Cambridge University Press.
Quine, W.V.O. 1960: Word and Object. Cambridge, MA: MIT Press.
Quine, W.V.O. 1968: Ontological relativity. Journal of Philosophy, LXV, 185212.
Trudgill, P. 1986: Dialects in Contact. Oxford: Basil Blackwell.
Weinreich, U., Labov, W., and Herzog, M. 1968: Empirical foundations for a theory
of language change. In W.P. Lehmann and Y. Malkeil (eds), Directions for historical
linguistics: A symposium. Austin: University of Texas Press.
Williamson, T. 1994: Vagueness. London: Routledge.
Wilson, E.O. 1999: The Diversity of Life. 2nd edn. New York: W.W. Norton & Co.
Wittgenstein, L. 1953: Philosophical Investigations, trans. G.E.M. Anscombe. Oxford:
Blackwell.
2008 The Author


Lassiter Sociolinguistics SemExternalism

Uploaded by

Copyright:

Available Formats

Lassiter Sociolinguistics SemExternalism

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lassiter Sociolinguistics SemExternalism

Uploaded by

Copyright:

Available Formats

Semantic Externalism, Language Variation, and

Nobody can be forbidden to use any arbitrarily

1. Semantic Individualism and Semantic Externalism

Externalism, Variation, and Accommodation

In what follows I am primarily interested in externalism about linguistic meaning, rather

2008 The Author

Externalism, Variation, and Accommodation

communitarianism, yet many externalists rely heavily on an unanalyzed notion of

2. Communitarianism and Descriptive Linguistics

2.1 Individuating Languages

primarily as the speech of a certain community of people, often corresponding

Externalism, Variation, and Accommodation

Assam to Afghanistan (Masica, 1991, p. 25). This vast areamillions of square

2.2 A Problem about Semantic Change

Externalism, Variation, and Accommodation

theory. Semantic change is acknowledged as a difculty for externalism already by

2008 The Author

Externalism, Variation, and Accommodation

speech community, so that, in Jims neighbourhood, arthritis actually does

3. Underdetermination, Theory-Building, and Sociolinguistics

What explains an individuals speech behaviour?

We will try to answer these questions with minimal theoretical assumptions,

3.1 Individual Speech Behaviour

2008 The Author

Externalism, Variation, and Accommodation

theory-building also provides a simple explanation for change in speech behavior:

3.2 Individuals and Communities in Sociolinguistics

2008 The Author

Externalism, Variation, and Accommodation

3.3 Reference and Deference

Success in accommodation is further constrained by knowledge of or familiarity with the

2008 The Author

Externalism, Variation, and Accommodation

interpretive theories are models of a sound-meaning correspondence of this

2008 The Author

Externalism, Variation, and Accommodation

important to us. People who are geographically, temporally, or linguistically

3.4 Intentions and Semantic Change

Actually, it is more likely that a speakers projected community will be linguistically

2008 The Author

Externalism, Variation, and Accommodation

Externalism, Variation, and Accommodation

a pragmatic phenomenon, I conceive of both speakers reference and semantic

Thus an answer to (ii), the question of change in semantic norms, seems to

Externalism, Variation, and Accommodation

identicationmap out the same community language. This community language

Externalism, Variation, and Accommodation

Giles, H. 1994: Accommodation in communication. In R.E. Asher (ed.), Encyclopedia

2008 The Author

You might also like