1

Morgan’s Canon: Animal Psychology in the Twentieth Century and Beyond

An enormously important role has been played in twentieth century animal science by the principle
known as ‘Morgan’s Canon’, first formulated in the nineteenth century by Conwy Lloyd Morgan
(1852-1936), a British ethologist widely regarded as one of the most influential founders of
comparative psychology. The principle was expressed by Morgan over the course of his lifetime in a
number of different ways, but the one which seems to be most often quoted in the literature is from
the first edition of his Introduction to Comparative Psychology: “In no case may we interpret an
action as the outcomes of the exercise of a higher psychical faculty, if it can be interpreted as the
outcome of the exercise of one which stands lower in the psychological scale.”1 The influence that
the principle has had on the conduct and methods of comparative psychology over the succeeding
hundred and twenty years cannot be over-estimated; it has been described by one writer as
“possibly the most important single sentence in the history of the study of animal behavior.”2 It was
certainly implicated in the widespread adoption of radical psychological behaviorism during the
middle decades of the twentieth century; but its impact has outlasted the demise of that doctrine.
The Canon is still regularly explicitly invoked by scientists working on animal behavior and is perhaps
even more often silently applied as part of an implicit orthodoxy concerning the appropriate
methodology for a sober psychological science.3
However, the Canon has been severely criticized in recent years by some powerful
detractors. Daniel Dennett4, Elliott Sober5 and Jerry Fodor6 have all argued, in different ways, that
Morgan’s Canon is either redundant, defensible only on a quite seriously modified interpretation of
its meaning, or simply silly. A number of other philosophers have also contributed excellent
critiques.7 Historians of science, moreover, have shown convincingly that the Canon has been
utilized by scientists of animal behavior in ways that were never intended, and moreover, would
have been deplored, in fact, by Morgan himself.8 The case against the Canon has been thoroughly
made with considerable power, in more than one way, by a range of philosophers and historians.
Why, then, write yet another paper on Morgan’s Canon?
One reason for thinking that there is room for another nail or two in the coffin of the Canon
is that, as already noted above, the principle appears as yet to be thoroughly undead in the

1
C. Lloyd Morgan, An Introduction to Comparative Psychology, 1st edition (London and Newcastle-upon-Tyne:
The Walter Scott Publishing Co. Ltd., 1894), 53.
2
B.G. Galef Jr., “Historical origins: the making of a science”, in Foundations of Animal Behavior: Classic Papers
with Commentaries, eds. L.D. Houck and L.C.Drickamer (Chicago: University of Chicago Press, 1996): 9.
3
For examples of explicit invocation, see, for example, Robert Epstein, “The Principle of Parsimony and some
Applications in Psychology”, The Journal of Mind and Behavior 5 (1984): 119-30; Stephen Budiansky, If a Lion
Could Talk: How Animals Think (London: Wiedenfeld, 1998); M. Rosalyn Karin-D’Arcy, “The Modern Role of
Morgan’s Canon in Comparative Psychology”, International Journal of Comparative Psychology 18 (2005): 179-
201; Sara J. Shettleworth, “Clever Animals and Killjoy Explanations in Comparative Psychology”, Trends in
Cognitive Sciences 14 (2010): 477-81.
4
Daniel Dennett, “Intentional Systems in Cognitive Ethology: the Panglossian Paradigm Defended”, in The
Intentional Stance (Cambridge, MA: MIT Press, 1993): 237-68.
5
Elliott Sober, “Morgan’s Canon”, in The Evolution of Mind, ed. D.D. Cummins and C.Allen (Oxford: OUP, 1998).
6
Jerry Fodor, “Not So Clever Hans” in London Review of Books, February 4, 1999: 12-13.
7
Of particular note are those by Simon Fitzpatrick, “Doing Away with Morgan’s Canon”, Mind and Language 23
(2008): 224-46; Sean Allen-Hermanson, “Morgan’s Canon Revisited”, Philosophy of Science 72 (2005): 608-31;
and Martin Montminy, “What use is Morgan’s Canon?” Philosophical Psychology 18 (2005): 399-414.
8
See, for example, Robert H. Wozniak, “Conwy Lloyd Morgan, Mental Evolution, and The Introduction to
Comparative Psychology”, 1997. http://brynmawr.edu/psychology/rwozniak/morgan/html. Accessed January
12, 2015.
 2

disciplines to which it is of direct relevance. It still stalks the corridors of psychology departments,
making its unholy presence firmly felt9; and moreover, though this is rarely noted, its influence also
continues to be discernible well beyond the academy. The anthropomorphic tendencies of the
general public, delighting to tell anecdotes supposed revelatory of the reasoning powers, emotional
sophistication and general intelligence of their pets, have long been deplored by many Canonised
practitioners of animal science but in fact, many people reveal, even as they offer up these
anecdotes, an astonishingly high level of awareness of the possible perils of over-imaginative
interpretation, often laughingly following their tales with embarrassed acknowledgements of the
foolish anthropomorphisms they take to be thereby exhibited. One reason for this high level of
awareness is that the cautious and purportedly objective descriptions of animal behavior which have
been supposed part of a respectable scientific animal psychology have thoroughly permeated the
broadcasts and publications by means of which the public gains much of its scientific understanding
of the animal world. To take one instance, the banter between Chris Packham and Michaela
Strachan on recent editions of Springwatch and Autumnwatch, BBC nature programmes which may
be familiar to British readers, regularly involves the attribution of high-level psychological states to
the creatures observed by the giddy Michaela, lover of animals, who is gently put right on the matter
by the sober and biologically better-educated Chris, urging caution, and the strenuous avoidance of
anthropomorphism.10 The message which often emanates from these and many other wildlife
programmes, and from the language in which their descriptions of animal behavior are cautiously
couched, is that anyone who knows anything about the science of animals knows also that their
activity ought not to be interpreted by means of the tools which often seem to come naturally to
human observers: the tools made available by intentional descriptions, and the assumption of such
things as agency and perhaps even consciousness, in the creatures being observed. Real scientists,
we are encouraged to think, know better. But do they know better? The extent to which Morgan’s
Canon may be responsible for a now thankfully teetering orthodoxy needs to be appreciated, before
this question can hope to receive a serious answer.
A second reason for thinking that Morgan’s Canon deserves another airing is that the debate
about the principle is far less widely known outside psychology than it ought to be. Amongst
philosophers of mind and action, for instance, the Canon is rarely mentioned. I had been working,
broadly speaking, in the Philosophy of Mind for many years when I first heard of it, my attention
directed to its influence and interest not by a fellow philosopher, but by a colleague working in the
History of Science, Greg Radick, who had himself written about the Canon.11 I suspect a poll of
philosophers of mind would reveal similar levels of general ignorance. And yet the principle is
enormously significant. The question whether it is to be endorsed or rejected is utterly central to the
pursuit of questions about animal mind, to questions about other minds in general, to questions
about humanity’s place in nature. Philosophers of mind need to know about it, and they need to
know about the influence it may have had on the psychological work from which they increasingly
frequently take their inspiration. Not to know is to risk mistaking the products of various contingent
historical factors for the non-negotiable outcome of the inevitable onward march of an increasingly
objective science.

9
For affirmations that Morgan’s Canon continues to play a useful role in psychology, see M. Rosalyn Karin-
D’Arcy, “The Modern Role of Morgan’s Canon”; Dominic M. Dwyer and Katy F. Burgess, “Rational Accounts of
Animal Behaviour: Lessons from C.Lloyd Morgan’s Canon”, International Journal of Comparative Psychology 24
(2011): 349-64; Sara J. Shettleworth, “Clever animals and killjoy explanations”.
10
See e.g. the discussion of reed warblers’ inability to detect cuckoo chicks which have hatched in their nest,
Springwatch 2014 Episode 10.
11
Greg Radick, “Morgan’s Canon, Garner’s Phonograph, and the Evolutionary Origins of Language and
Reason”, British Journal for the History of Science 33 (2000): 3-23.
 3

A third and final reason for thinking that more needs to be said about Morgan’s Canon is
that the existing literature is rather confusing. Not everyone who criticizes the Canon is criticising
the same version of it; moreover, interpretations of its crucial terms differ, so that not everyone who
appeals explicitly to the Canon really has quite the same principle in mind. Some points which are
made in opposition to the Canon ought never to have been made, for they are very easily met, and
ought to be discarded in favour of the more powerful arguments which are available. And in
addition, one needs, of course, to give proponents of the Canon a fair hearing, and to respond to
their anxiety that to lose the Canon might be to open the door to an absurd degree of
anthropomorphic licence. Simon Fitzpatrick has argued that many endorse the Canon only because
they fail to see that the work they believe it is needed to do, can be done instead by an alternative
scientific principle, a principle which Fitzpatrick calls ‘evidentialism’. Scientific evidentialism, in its
most general form, simply states that in no case should one endorse any scientific theory or
explanation if the available empirical evidence gives us no reason to prefer it to an alternative theory
or explanation. Formulating a version specifically directed to the interpretation of animal behavior,
Fitzpatrick offers the following:

<ext>“Evidentialism: in no case should we endorse an explanation of animal behaviour in terms of

cognitive process X on the basis of the available evidence if that evidence gives us no reason to
prefer it to an alternative explanation in terms of a different cognitive process Y – whether this be
lower or higher on the ‘psychical scale’.”12</ext>

Evidentialism supports the view that in circumstances in which we genuinely have no proper
empirical reason of any sort to prefer one of two theories to another, the correct response is
agnosticism, together, perhaps, with a redoubling of efforts to ascertain (by means of
experimentation or further observation) which of the two theories really does offer the better
explanation.13 But endorsement of either theory until these efforts have been made successfully
would go too far. And endorsement is what Morgan’s Canon purports to justify. Evidentialism, then,
opposes Morgan’s Canon. And I shall concur with Fitzpatrick in arguing that it is evidentialism, and
not the Canon, which ought to be the touchstone of good science.
In some ways, I confess, it just seems obvious to me that if we are looking for a quite general
principle of theory choice in the realm of comparative psychology, evidentialism embodies a much
safer scientific attitude than does Morgan’s Canon. (In case this is not utterly plain, I shall offer in
section (v) some examples which make clear that Morgan’s Canon both can encourage, and has in
fact resulted in the endorsement of unsafe verdicts on animal behavior). The interesting question is
really how, and why, Morgan’s Canon has managed to last so long. In an attempt to answer this
question, I shall look at some of the arguments which might conceivably have convinced some of its
proponents of its legitimacy, and will seek to undermine them. But the correct explanation of its
tenacity may ultimately be one that cites causes rather than reasons. Morgan’s Canon has been of

12
Fitzpatrick, “Doing Away with Morgan’s Canon”, 242.
13
This crucial point – that in the absence of empirical reasons to prefer one explanation over another,
agnosticism is the appropriate attitude, until further empirical evidence of the sort which might enable a
rational preference to be formed – is made by many critics of the Canon, including Sober: “the canon seems to
license inferences in which the proper conclusion is that no inferences can be drawn”, “Morgan’s Canon”, 228;
Fodor: “A priori, there is nothing to choose between the hypothesis that a certain behavior has a “higher” sort
of mental aetiology and the hypothesis that it doesn’t. if you want to know which is true, you just have to go
and look” (“Not so Clever Hans”, 12); and Fitzpatrick: “… if the worry about crude anthropomorphism is that it
is not sufficiently supported by the data, it is not clear that we are thereby justified in endorsing the least
cognitively sophisticated explanation available, since that explanation may not be sufficiently supported by the
data either. After all, we could instead simply be agnostic …” (“Doing Away with Morgan’s Canon”, 235).
 4

the utmost importance to the attempts of psychology to escape such evils as anecdotalism and
introspectionism, and to psychology’s struggle to establish itself as a proper scientific endeavour,
alongside the ‘hard’ sciences. To discard it may feel to the practitioners of that science like a
tremendously dangerous step in an unwanted direction. In this paper, I shall add to the attempts
that others have made to allay this concern. Like a mercenary once recruited for a good cause, but
now exhausted and capable of being induced to fight for the other side, Morgan’s Canon is now
itself the danger to good science.

(i) The Canon according to Morgan

As already noted, Morgan’s Canon states that “In no case may we interpret an action as the
outcomes of the exercise of a higher psychical faculty, if it can be interpreted as the outcome of the
exercise of one which stands lower in the psychological scale.”14 The first duty of any philosopher
tasked with arriving at a good understanding of Morgan’s Canon is evidently the interpretation of
the crucial terms ‘higher’ and ‘lower’. But before embarking on an attempt to decide how best we
are to understand the nature of the hierarchy which is implied by these terms, it is worth pausing to
take note of a number of additions and clarifications which were made by Morgan himself.
One addendum which is made almost immediately by Morgan after stating his canon
expresses a certain discomfiture at his own use of the word ‘faculty’. Suspicion of “faculty
psychology” was in the air in the late nineteenth century; Wundt had already criticized the
misleading “mythological” conception of a psychological faculty as something which itself acts, in his
Grundzüge der physiologischen Psychologie,15 first published in 1874. Morgan suggests as an
alternative, for those who are worried by the use of the term’ faculty’, a restatement of his Canon in
terms of an ontology of processes: “In no case is an animal activity to be interpreted in terms of
higher psychological processes, if it can be fairly interpreted in terms of processes which stand lower
in the scale of psychological evolution and development.”16 This restatement offers not only an
alternative ontological framework for the understanding of the Canon, one which has tended to be
found more appealing by modern practitioners both of psychology and philosophy than the faculty-
based formulation, but also a clarification of the relevant notion of a “psychological scale”. We are
told here that the “scale” according to which the processes can be ranged in a hierarchy is a “scale of
psychological evolution and development”; I shall turn presently to some possible concerns about
this suggestion.
Another important elucidation is that when Morgan speaks, in the original formulation of his
principle of “an action” (“in no case may we interpret an action …”), he is speaking not of what
contemporary philosophers would call act tokens , but rather of act types; so that the question is not
whether, on a given occasion, we should attribute a piece of animal behavior to this or that faculty,
this or that process, a decision which, as Morgan notes, would be “a purely individual matter of
comparatively little moment,”17 given the independently evidenced assumption that the creature in
question possesses both faculties (or utilizes both processes). The question is whether a type of
activity in general – a pattern of activity which one sees repeated on a number of occasions across
different individual members of the same species, or even across members of many species, in
different times and places, such as web-spinning in spiders, the broken-wing behavior in certain

14
Morgan, An Introduction to Comparative Psychology, 53.
15
Wilhelm Wundt, Grundzüge der physiologischen Psychologie (Leipzig: Engelmann, 1874).
16
Morgan, An Introduction to Comparative Psychology, 59.
17
Morgan, An Introduction to Comparative Psychology, 54.
 5

waders, plovers and doves,18 grooming in primates, and so on, and capable of being interpreted as
the “outcome” either of a “higher” or a “lower” process – should, in the absence of any further,
independent evidence, be attributed to the “higher” or to the “lower” one. Morgan is absolutely
clear, moreover, that “it should be added, lest the range of the principle be misunderstood, that the
canon by no means excludes the interpretation of a particular activity in terms of the higher
processes, if we already have independent evidence of the occurrence of these higher processes in
the animal under observation.”19 The principle applies only when one does not have such
independent evidence. One may not, therefore, on Morgan’s view, use the Canon to insist upon
“lower” interpretations of behaviors in certain animals if one already has evidence of relevant
capacities in those animals which might permit “higher” interpretations to be more plausible.
What, then, should we make of Morgan’s talk of “higher” and “lower” faculties and
processes? The first and most important point is that Morgan’s own distinction is a distinction
amongst faculties/processes all of which are “psychical”. This is easy to miss, particularly in view of
the widespread appropriation of the metaphor of hierarchically-ranged levels in contemporary
thought to understand the relation between different sciences and their theories, explanations and
vocabularies. Confusion on this score might well incline us to leap immediately to the conclusion
that Morgan’s Canon tells us to prefer the explanations of the sciences ranked lower in this
hierarchy (for example, those of neurophysiology or biochemistry), to those ranked higher (such as
those of psychology), where they are available. But this was no part of Morgan’s own view. Morgan’s
conception of psychology was organized in terms of a hierarchy of powers or faculties which might
be invoked in the explanation of behavior, at the bottom of which was to be found “mere instinct”,
and at the top, reason. The distinction between the two had already been discussed in the
introduction to Romanes’ work, Animal Intelligence20, and this discussion represents the starting
point for Morgan’s methodological deliberations in his own Introduction to Comparative Psychology.
Morgan’s distinctive advance on Romanes was to argue that between these two domains of psychic
life, we must recognize ‘a vast field of animal and human activity which … [should be] …
distinguish[ed] from both by the application of the term ’intelligence.’”21 For Morgan, it is clear that
we are going to need (at least) a threefold distinction between instinct, intelligence and reason for
the interpretation of animal behavior. One thing which the Canon clearly entails for Morgan, then, is
at least this: that what we can perfectly well interpret as the product of instinct, we should not seek
to explain by means of the attribution of intelligence; and what we can understand as an exercise of
mere intelligence, we should not consider to be a demonstration of reason.22

18
Some species of bird defend their eggs and chicks by leading potential predators away from the nest with a
display in which they mimic having an injury to a wing. Once the predator has been safely distracted, they are
usually then able to fly off to safety themselves.
19
Morgan, An Introduction to Comparative Psychology, 59.
20
George Romanes, Animal Intelligence (London: Kegan Paul, Trench & Co., 1882).
21
Conwy Lloyd Morgan, “The Limits of Animal Intelligence”, Fortnightly Review, Aug. 1893: 223-39, 225.
Though Romanes had of course also utilized the word “intelligence” (witness the title of his major work), he
appears to conceive of it rather as a rather lowly variety of reason, rather than, as Morgan has it, a set of
powers deserving of its own distinctive category.
22
To a modern ear, it perhaps seems strange to think of explanations which invoke mere instinct as
psychological explanations. The invocation of instinct has come to be seen in some contexts as something
opposed to the invocation of mentality; and often also as something opposed to the existence of real agency in
a creature. But this was not at all how the concept was understood by Romanes, from whose discussion, as
already mentioned, Morgan’s takes its point of departure. For Romanes, instinct was to be clearly
distinguished from mere reflex by “the element of mind” (Animal Intelligence, 11). Romanes is clear that
instinct “involves mental operations” (Animal Intelligence, 11) – by which he means that a certain degree of
consciousness is implied. He is sensible, of course, that in many instances it may be difficult or impossible to
decide whether or not the mind-element is present – but his considered view is that “this is altogether a
 6

It need hardly be said that the Canon’s directive has not always been interpreted in this
strictly intra-psychic way by many modern and not-so-modern psychologists. Very often, the Canon
has been thought to sanction a preference for explanations which do not involve the attribution of
consciousness over those which do; for explanations which are mechanistic rather than intentional;
for explanations of behaviors in terms of such things as fixed action patterns, stimulus-response
mechanisms, tropes and reflexes, rather than anything which might be better understood by
reference to such things as judgement, decision, or even emotion – without any essential reference,
indeed, to anything which might make the psychological grade. Of course, it is a vexed matter to say
what it is for any kind of faculty or process to make the psychological grade but most philosophical
attempts at defining the psychic have made some kind of reference or other either to consciousness
or to intentionality, or both. Morgan’s Canon has been assumed to justify the general privileging of
the non-psychic over the psychic. But that is certainly not how Morgan intended it. Indeed, Morgan
himself would certainly be turning in his grave if he knew how the Canon had been deployed within
the science of comparative psychology of which he has a good claim to be regarded as one of the
founding fathers. Following his discussion of animal memory, he makes the following request:

<ext>“I will only ask the reader … to be good enough to credit me with an unbiassed (sic) desire to
interpret the phenomena of animal psychology without exaggeration, either in the direction of
excess or defect of mental power and differentiation. As an evolutionist who believes that the whole
range of the mental faculties have been developed by natural processes, the tendency of my bias
would assuredly not be in the direction of setting a gulf between the faculties of animals and the
faculties of man”.23 </ext>

Morgan himself is in the business of defending the principle that mind is likely to be found
throughout the animal kingdom, though reaching different degrees of sophistication in different
animals. A ‘gulf’, it is worth bearing in mind, is what he never would have countenanced. And yet a
gulf is what the practical application of his Canon has often suggested.

(ii) Extending the Canon?

It might be said, of course, that the question what Morgan himself intended is really of only
secondary and purely historical interest. Even if his was, as I have suggested, intended to be an intra-
psychic scale, it might be alleged that there is no good reason to be so restrained in the application
of the Canon. Allen-Hermanson, for example, writes that he “can see nothing incoherent about
extending the Canon to include showing a preference for nonmental faculties over ones that are
psychological.”24 But it seems to me, on the contrary, that once one makes this step, certain
difficulties at once present themselves. The reason is that if we interpret the Canon as endorsing a
preference for non-psychic over psychological explanations, it will almost certainly end up by
licensing far too much. A first attempt at expressing the worry here might go as follows: it is always
possible to offer non-psychological explanations of patterns of behavior, alongside the psychological

separate matter and has nothing to do with the question of defining instinct in a manner which shall be
formally exclusive, on the one hand, of reflex action, and on the other, of reason.” (Animal Intelligence, 11-12)
There is every reason to suppose that Morgan would have adopted the same view. For him, explanations
adverting to animal instincts were psychological explanations implying the presence of consciousness – and
the hierarchy of powers to which his Canon makes implicit reference is hence an entirely intra-psychic
hierarchy.
23
Morgan, An Introduction to Comparative Psychology, 124. My italics.
24
Allen-Hermanson, “Morgan’s Canon Revisited”, 613.
 7

ones which we might also think correct. This is true even in the human case. For example, I (and
millions of other human beings) leave our homes each morning for work where we intend to earn
the money to support ourselves and our families. To mention the aim of earning money is to offer a
psychological explanation in terms of intention and purpose – making the activity into one which can
be readily be understood as a prime case of the operation of practical reason. But there will also be a
range of lower-level explanations available of this widespread variety of human behavior. There will,
for example, be a physiological account to be given of the causal origin of the walking movements
we make in leaving (and in opening doors, stepping outside, closing our doors behind us, setting off
down the road…) in terms of activity in the motor cortex, responses to incoming perceptual data,
perhaps the setting in train of certain habitual mechanisms (we know without looking where our
own door handle is, what sort of force will be needed to turn it, and so on). And yet we do not
normally suppose that the existence of the second sort of explanation excludes the relevance and
importance of the first. Yet Morgan’s Canon, extended in the way envisaged by Allen-Hermanson,
might seem to advise us to dispense with the intentional explanation entirely, in favour of the purely
physiological one.
It might be said in response – and it would be rightly said, in my view – that this argument is
too quick. For it assumes that the physiological explanation in such a case would be an explanation
of the very same thing as the intentional one – and that this is not in fact the case. The psychological
explanation explains why a certain pattern of intentional activity has occurred - whereas the
physiological explanations merely explain why such and such bodily motions took place. We have
here two explananda and not one, and that is why two distinct explanations are admissible. But so
far as animals are concerned, the very thing which is likely to be in question, when an application of
Morgan’s Canon is envisaged, is whether one can distinguish in this way or not between similarly
differentiated explananda – whether, for example, there really is, in a given case, something it might
make sense to think of as an intentional action present, in addition to a mere set of movements, or
not. Is the Australian jewel beetle trying to mate with the beer bottle which it mistakenly believes to
be a female? or should we simply say that certain features of the bottle have triggered (in an entirely
mechanistic way) the ‘release’ of certain behaviors? We face in such cases as this a question about
how the general phenomenon, an instance of which we are currently observing, is to be described
before we can begin to even approach the question how it is to be explained. It is simply not the
case, as the formulation of Morgan’s Canon rather takes for granted (in speaking of ‘an action’ which
is ‘the outcome’ of a given process), that we can offer an entirely neutral description of some single
pattern of activity which then might admit either of a psychological or of a non-psychological
explanation. As Dennett has often argued, what we face in these cases is rather the more general,
holistic question of whether there is a psychological pattern to be found across what is perhaps
quite a wide range of behavior in different kinds of context, a psychological pattern which would be
explanatorily and predictively helpful to us in our quest to interpret the world – and the answer to
this question will dictate the description of the activity or behavior we want explained in the first
place, as well as the form of the explanation itself.25 As Dennett sees it, the crucial question is this:
does the intentional stance pay significant dividends in the attempt to understand the overall
workings of a certain organism, or not? And it is not really clear how Morgan’s Canon is to be
applied, once this difficulty is recognized. If we interpret it as telling us always to favour a non-
intentional, more mechanistic stance wherever we can do so – to look upon a creature as a mere
mechanism wherever that is possible – we seem to be in danger of outlawing psychology altogether.
For the trouble is that in a certain sense, it is always possible to take a mechanistic stance to a
pattern of activity. We can always decide (even in the case of human action) that our explanandum

25
Daniel Dennett, The Intentional Stance (Cambridge, MA: MIT Press, 1989).
 8

is merely a set of bodily motions and resolve to seek merely to explain those in terms of their
proximate neurological causes. But presumably we (mostly) think we would miss an utterly
enormous amount, were we to restrict ourselves in this way, in the human case. Why, then, think
that the decision to do so is not similarly likely to lead us astray in the case of other animals?
It might be said that this conclusion would only follow if we overlook Morgan’s own
clarification that we are allowed to offer a ‘higher’ process in explanation of a phenomenon in a case
where we have independent evidence of some sort that it exists. That is what we have in abundance,
it might be said, to justify the application of psychological interpretations in our own case and
perhaps we might be happy to concede its existence (in the form of hugely increased general
explanatory and predictive power) also in the case of certain other higher animals, such as primates
and dogs. It might be doubted, I imagine (and I would myself doubt), that we have such independent
evidence in the case of the jewel beetle. Morgan’s Canon is to function, it might be insisted, merely
as a “tie-breaker”, to be brought in only where, having taken all available evidence into account,
there still seems to be no clear reason to prefer the adoption of the intentional stance to a non-
intentional or even entirely non-psychological form of explanation. It is merely a directive that takes
effect, other things being equal. But there would be two serious worries about trying to defend
Morgan’s Canon on this more limited basis. One is that one might reasonably have the suspicion that
the number of cases in which all else really is equal is likely to be vanishingly small – indeed, that it
might end up being a null class. A situation in which even when all the evidence that could possibly
be relevant to the interpretation of a piece of behavior is taken into account, there might be nothing
which tips the balance in favor of the view that one or other of the competing hypotheses is the
more probable, is hard to imagine. The other is, of course, the question already raised of what would
justify the use of the Canon in such a case, to break the tie – of why agnosticism would not be the
proper response instead. I shall turn shortly to consider some possible answers to this important
question. But first, having cast doubt on the idea that it is safe to extend the Canon beyond
Morgan’s own intra-psychic application, let us return to the question I raised earlier, namely, the
question how we might interpret the difficult terms “higher” and “lower”.

(iii) “Higher”, “Lower” and the Psychic Hierarchy

As I have already explained, Morgan was operating with a basically tripartite understanding of the
psychic “faculties” which might underlie animal behavior – instinct, intelligence and reason – though
he seems to have been perfectly aware that this division was strictly pro tem and likely ultimately to
be superseded by a more complex set of divisions, particularly within the vast category that he calls
“intelligence”. Undoubtedly, Morgan’s own suggestion that these powers can be arranged
hierarchically was underwritten, and perhaps made to seem more obvious than it really is, by
assumptions which are false, in particular, the assumption that evolution “always marches from
simple to complex,” as Sober puts it, and the vestiges of the linear conception of a scale of nature
which Darwinism refuted, according to which every species can be ranked as “higher”, “lower” or
“equal to” each of the others. In particular, we must suspect Morgan of having taken for granted the
thought that “higher” implies “more characteristic of, or closer to, distinctively human behavior” and
hence the assumption that man is the measure of all things cognitive. From a less anthropocentric
viewpoint, one might indeed be tempted to claim, as Sober suggests at one point, that “it makes no
sense to ask whether the ant’s use of pheromones is higher or lower than the plover’s use of
deception. Both have their place in life’s diversity.”26

26
Sober, “Morgan’s Canon”, 225.
 9

But even if we must beware of anthropocentrism in the very establishment of the scale in
terms of which we are to understand Morgan’s Canon, it is surely not true that we can make nothing
sensible of the idea that some interpretations of creatures’ behavior are more sophisticated than
others. It is not, of course, that the operations of instinct are any less impressive than the workings
of reasoned thought; indeed, in some ways, the reverse is true. It is astounding that a spider can spin
a web or a bird build a nest without practice, without forethought, without teaching. Nor is the
concept of complexity, I suggest, the most useful way to think about the question which powers are
‘higher’. There are just too many different ways in which to measure complexity for this to be a very
useful suggestion. The processes behind the operation of many instincts, for instance, must be
staggeringly complex, and in some ways might even seem to exceed in complexity the sorts of
simple logical operations that are involved in some of the most basic varieties of reasoning.27 What
makes us regard certain interpretations of behavior as “higher”, I suggest, is what they seem to
imply about the overall capacities of the behaving agent. Where a behavior is attributed to instinct,
comparatively little is left for the operation of judgment, decision or thought to settle at the time of
action; and few resources are presupposed which require such things as the consideration and
weighing of alternative means to given ends. A spider may have to find a reasonable place to build
her web – but web-spinning itself is an ability she will not need to have learned or have practiced,
and which we can think of, more or less, as involving simply the execution of an entirely domain-
specific programme with which the spider comes ready-equipped. Some degree of flexibility and
adaptability is doubtless required but the calls on the spider-agent, deliberating in the moment, to
come up with spur-of-the-moment solutions to unforeseen difficulties are plausibly not many. And
should the spider for some reason be unable to construct an effective and successful web, she is
unlikely to come up with many ideas about what she might do instead to get a meal. She will not, for
example, be inventing wholly new kinds of fly trap or discussing interesting alternative possibilities
with her conspecifics. If an instinctive behavior cannot be made to do the job for which it is intended
(and no ‘back-up’ mechanisms have been selected for), then that is more or less that.
Moving up into the realms characterized by what Morgan called “intelligence”, there is likely
to be more flexibility and adaptability on show. New Caledonian crows can seemingly put together
various parts of mechanisms they have learned to manipulate in order to reach food in such a way as
to hatch a successful plan when confronted with a new and more complex mechanism than any they
have hitherto seen. This cannot be explicated by means of instinct alone; the situation they face is
new and requires immediate, on-the-spot assessment by something which it is very natural to
suppose must be a kind of thinker. But we need not suppose the existence of thought which is in all
respects like our own, in order to make sense of the postulated abilities. We need not attribute, for
example, such things as episodic memory, arithmetical abilities, a sense of self, or conceptual
thought, at least, not where we have no other behaviors to go on. What we attribute is a level and
type of intelligence of the sort necessary to explain the kind of problem-solving revealed by the
crows: no more.
Reason – a category as unsatisfactorily vague in its own way as ‘intelligence’ – is at the top of
Morgan’s hierarchy, I think, for several reasons. One is obviously its traditional place as the
differentia of the human species and its traditional associations with the presence of linguistic
abilities, via the concept of logos. Another less anthropocentric justification one might offer for
thinking that it truly belongs there is that ‘reason’ is often used to refer to a set of abilities to
manipulate information which are potentially entirely domain-general; inferential abilities, for
example, can be applied, in principle, to problems which share the same logical shape, no matter

27
As Fodor points out, all you need to make a Turing machine, in a sense an excellent reasoner, is “two kinds
of pebble and an endless roll of lavatory paper” (“Not so Clever Hans”, 13).
 10

what the topic of the reasoning involved. It is true that evidence suggests that we cannot always
exercise our reasoning abilities on some kinds of problem as well as we can on others – there may be
more domain-specificity in human thinking that has hitherto been recognized.28 But still, the
applicability of general logical principles and mathematical reasoning, etc. to diverse puzzles about
the construction of buildings, the calculation of prices, the drawing of maps, the prediction of the
motions of planets, etc., is striking, and there seems to be little evidence of such very diverse
applications of inferential thinking in creatures other than ourselves. We are probably less inclined
than Morgan would have been to suppose that evidence of the application of reason is necessarily
evidence of an agent consciously thinking, for unlike Morgan, we have seen what mere machines are
able to do in the realm of calculation, inference and problem-solving and how domain-general are
the abilities they display; and perhaps less likely also to think that reason and language necessarily
go hand in hand. But the presence of the two together certainly appears exponentially to expand the
range of a creature’s abilities. We may additionally associate with reason such things as the ability
to think about things in their absence; to engage in counterfactual thinking; to think utterly freely
about future and past; to form new concepts and categories, building on those which already exist –
all things which seem to have been crucially important in human development and to have put
within the reach of their possessors a range of capacities to act on the world of which no other
animal has so far been capable. Morgan’s categories are doubtless too broad, too vague and too all-
encompassing: but it cannot be said that we have no idea at all what to make of the suggestion that
some processes offer “higher” interpretations of creatures than others. We can at least make rough-
and-ready sense of what is at issue – enough sense, anyway, to make Morgan’s Canon definite
enough in its specific directives to be practically usable.

Suppose, then, that we adopt, roughly, the suggestion offered above as to how we are to
understand the terms ‘higher’ and ‘lower’. “Lower” processes carry with them fewer commitments
to further abilities; they are less flexible in the face of unexpected failure or alteration in
circumstance; they are more domain-specific. Is there any reason to suppose that if we have no
empirical evidence to prefer an interpretation of animal behavior in terms of “higher” processes to
one in terms of possible “lower” processes, we should endorse the lower? I want now to consider
whether there is anything to be said for Morgan’s Canon, thus understood.

(iv) The Justification of Morgan’s Canon

There seem to be roughly three kinds of argument available for the possible justification of
Morgan’s Canon. One variety of argument – perhaps the most common – suggests that it is simply a
version of some very widely endorsed scientific principle of theory choice: either a principle of
parsimony, or perhaps a principle of simplicity. A second suggests that it is a principle which has a
specifically biological grounding of some sort. A third idea is that it is required in order to correct for
an inherent bias toward anthropomorphism. All these possible justifications have been thoroughly
worked over and in my view successfully demolished in the literature. I do no more here than simply
summarize and modestly supplement points that have already been very clearly made by others.29
The idea that Morgan’s Canon can be justified on grounds of parsimony or simplicity of
course raises the prior question how such generic principles of theory choice are themselves to be

28
See Lawrence A. Hirschfeld and Susan A. Gelman, Mapping the Mind: Domain Specificity in Cognition and
Culture (Cambridge: Cambridge University Press, 1994), for a discussion of the phenomenon of domain-
specificity.
29
Notably Sober, “Morgan’s Canon”; and Fitzpatrick, “Doing Away with Morgan’s Canon”.
 11

justified, but let that pass. Some of those who have invoked the idea that the Canon is simply one
application of such a principle have been willing to bite the bullet and accept that such principles
cannot be justified but must simply be assumed.30 But it is not a parsimony principle; and though it
seems to me arguable that it is a variety of simplicity principle, there are many kinds of simplicity,
and very much more argument than has been offered would be needed to support the view that the
kind of simplicity it approves must be preferred over the other varieties. Parsimony first: there is
more than one kind of parsimony one might have in mind in putting forward a parsimony principle,
for example, straight ontological parsimony, of the sort advocated by Ockham’s razor; parsimony of
theoretical assumption; cladistic parsimony (i.e. parsimony of separate evolutionary origins for a
given trait). But parsimony is always about getting by with fewer of something; it is not about getting
by with “lower” rather than “higher” faculties and processes.31 Simplicity, admittedly, is a different
virtue from parsimony; perhaps a preference for interpretations in terms of ‘lower’ processes does
indeed amount to a preference for one kind of simplicity. But as Fitzpatrick and Fodor both
convincingly argue, there are many varieties of simplicity, and we will still need to know how to
trade off the different kinds one against another. To take a nice example from Fodor32, on one
conception of simplicity, the simplest way to play tic-tac-toe is to memorize all the possible games
and play by rote. But that is not the way anybody in fact plays, nor would it be sensible to
programme a machine to play that way, since there are rule-based solutions that work just as
effectively, but use less computational power. If “rote learning” is “lower” than the application of
rules, as might seem intuitive, Morgan’s Canon looks as though it would suggest the wrong result in
a case like this. To take another case from Fitzpatrick, it is plausible that it is “simpler” in one sense
to suppose that we should use similar psychological explanations for similar behavior, where the
behavior is found in more than one species.33 But suppose we had abundant evidence that in one of
these species, the behavior was produced by way of a “higher” process? Morgan’s Canon would
seem to insist that we prefer the “lower” process as an account of what is going on in the other
species, nevertheless; but it is not in the least obvious that we ought to do so.
Morgan’s own attempt to justify the Canon does not go by way of simplicity; indeed, Morgan
assumes that simplicity would dictate utilising the most anthropomorphic explanation available. His
own justification appears to invoke the idea that the psychic powers arrayed in his hierarchy evolve
from one another, and hence looks to be intended as a specifically biological justification:

<ext> “… it is clear that any animal may be at a stage where certain higher faculties have not
yet been evolved from their lower precursors; and hence we are logically bound not to
assume the existence of these higher faculties until good reasons shall have been shown for
such existence”.34 </ext>

But as Sober argues, this is not enough to justify the Canon. There is no reason to suppose, even if it
were true that the higher faculties were evolved from the lower, that the lower ones are more likely
to explain any given piece of animal behavior than the higher. The evolution of the higher capacities

30
See e.g. Robert Epstein “No defense of the principle … [of parsimony] … is offered, for … I believe that no
definitive defense is possible … I simply assume it, as did Ockham and others, as a first principle …” (“The
Principle of Parsimony,” 119.
31
This point appears to have been first made by E. Newbury, “Current Interpretation and Significance of
Morgan’s Canon”, Psychological Bulletin 51, 1954, 70-74: 72, who is quoted approvingly by Sober, “Morgan’s
Canon”, 230.
32
“Not so Clever Hans”, 13.
33
“Doing Away with Morgan’s Canon”, 231.
34
Morgan, “An Introduction to Comparative Psychology”, 59.
 12

might well involve the atrophy of the lower: faced with a given pattern of activity, then, there is no
prior probability that it is a lower rather than a higher capacity that explains it.
Someone might think that the best justification for Morgan’s Canon is that it guards against
an opposing bias which is present in all of us – a bias toward anthropomorphic explanations and
interpretations. I think it may well be true that human beings do have such a bias. We are good at
searching for mindedness in the world and it is undeniable that we have historically found it in many
places where it turned out not to exist, inventing gods and demons, spirits and fairies, ghosts and
goblins, to explain phenomena that did not need to be thus explained. But in general, it is not true
that the best way to compensate for one bias is to introduce another. Evidentialism alone will do the
trick, and is far safer. All that is needed, when a suspiciously anthropomorphic explanation appears
to be on the cards, is the caution that other possible explanations of the behavior in question may
need to be investigated, and empirical reasons found for preferring one of the rival explanations
over the other.
Might it be objected, though, that evidentialism faces difficulties of its own? One might
worry, perhaps, that “higher-level” explanations of any given pattern of observed animal behavior
are always available in principle, and that without Morgan’s Canon, there is simply no way to discard
these explanations, despite what might be alleged to be their intuitive extravagance. Consider, for
example, a behavior of which we might suppose ourselves to have at least the basis of an entirely
satisfactory, relatively “low-level” explanation – say, the tendency of worker ants from the same
colony to respond effectively and instantly to danger to the nest, and in particular to its queen and
larvae. In cases of serious danger, the whole nest is sometimes moved in response to an attack
which cannot be successfully repelled by soldier ants, with worker ants carrying the vulnerable
larvae to a safer place. The role of instinctive, pheromone-guided behavior in eliciting appropriate
responses from individual worker ants has been intensively studied, and much scientific literature
now exists on how these processes work to ensure the protection of the colony.35 But of course it
would be possible to envisage other, “higher level” explanations of the same behavior. Perhaps, for
example, someone might suggest that the response is mediated by emotion, by love for the colony,
say, on the part of the ants, who then respond protectively in the light of their love. And without
Morgan’s Canon, it might be wondered whether there is any way decisively to rule out a suggestion
that might strike some as intuitively absurd.
One possible riposte here would be to try to argue that evidentialism could only lead to the
suggestion that we ought to suspend judgement about this case if an inappropriately restrictive
interpretation of what could constitute a ‘reason to prefer’ one theory over another is adopted. In
this particular case, perhaps, one might expect that the ‘love’ hypothesis could be rendered deeply
improbable by further experimentation. But even if no such experiment could be devised, the point
could be re-made that reasons to prefer one theory over another may relate not only to
observations which are strictly of the specific behavior to be explained, but also to far more general
considerations – such as, for example, the necessary conditions for the attribution of emotions such
as love to a species of creature. These might involve a much more holistic view on the overall active
repertoire of a given species of animal, and indeed, might perfectly well include philosophical, as
well as scientific considerations. It seems plausible, for instance, that to attribute the capacity for
love to a creature one would require evidence not only of a well-organized behavioral response to
an attack on the purportedly loved entity, but also of distress in response to the death or destruction

35
See e.g. N. Fujiwara-Tsujii,, N. Yamagata, T. Takeda, and M. Mizonami, “Behavioral responses to the alarm
pheromone of the ant Camponotus obscuripes (Hymenoptera: Formicidae)”, Zoological Science 23 (4), 2006:
353-8; Makoto Mizunami, Nobuhiro Yamagata and Hiroshi Nishino, “Alarm Pheromone Processing in the Ant
Brain: An Evolutionary Perspective”, Frontiers in Behavioral Neuroscience 4 (28), 2010: 1-9; Tristran D Wyatt,
Pheromones and Animal Behavior: Chemical Signals and Signatures (Cambridge: CUP, 2014).
 13

of that entity; of specifically emotional forms of attachment, such as comforting and touching, for
example; and an altogether more flexible repertoire of response to different kinds of threat to the
survival and wellbeing of the larvae than ants are capable of displaying. Another possibility would be
to invoke neurological evidence of a general kind relating to the supposed neural basis of the
capacity for emotional response. At any rate, it seems very unlikely to me that there are cases in
which no such general considerations would be available to provide reasons to allow one hypothesis
to prevail over another. But in cases in which there truly were no such considerations, arguing for a
reason to prefer the lower-level explanation over the higher, I would be inclined to stand by the
verdict of evidentialism. If no reason, either scientific or philosophical, can be given for supposing
that insects such as ants do not feel emotions, then we should remain agnostic for the time being,
until such time as reasons can be brought forward for preferring one explanation over the other. The
danger otherwise is that in attempting to adopt a bias thought to be opposed to our natural
anthropomorphic tendencies, we shall underestimate some of our fellow-creatures.
Moreover, when a bias is embraced by the scientific establishment in some discipline, things
are likely to go particularly badly wrong. The very experiments which might serve (in a world where
evidentialism held sway) to undermine the bias are likely to be discouraged; and if undertaken at all,
their results may be treated with undue scepticism. In 2008, Franklin D McMillan was still able to
write that “it is still politically wise to present oneself as soundly behaviorist, if not radically so.”36The
particular bias in question, moreover, in this case may interact in dangerous ways with other societal
forces. There may be a gender component to anti-anthropomorphic sentiment in science; one can
readily enough imagine it being supposed that it is foolish and oversentimental women who
attribute mindedness to animals, and only men who are able to maintain a truly objective attitude.
And human self-interest may also play a role. If animals are complex automata, we do not have to
worry so much about agricultural methods, animal experimentation, the destruction of habitat, and
so on. We have to be on our guard against the possibility that what masquerades as a sober
methodological imperative in fact simply enshrines the prejudices that are most useful to us.
Is it mere speculation to suggest that Morgan’s Canon may have been not merely unneeded
in psychology (given the availability of the evidentialist alternative) but could also prove actually to
have been damaging? I shall offer here, to conclude my discussion, two examples of cases which
shed light on this question, in different ways. The first is a case discussed by Simon Fitzpatrick in his
excellent critique of the Canon, a case in which the methodological imperatives of the Canon were
ignored, to famously productive effect. The second is an instance in which they seem, rather, to have
been followed – but where conclusions stronger than those really admissible appear to have been
drawn in consequence, arguably as a result of Canonesque-style thinking.

(v) Pernicious Effects

If one believed in Morgan’s Canon, how might this affect one’s research strategy? Plausibly, it might
affect one’s sense of what one’s working hypothesis ought to be, in any given case. As Fitzpatrick
notes, it might make plausible what he calls the “scaling up strategy” – that “[i]n any given area of
research into animal behavior, the working hypothesis should always be the least cognitively
sophisticated explanation consistent with the available data.”37 But Fitzpatrick claims that there are
many examples from the recent history of animal psychology where this was not the working
hypothesis – and where exciting results were discovered in consequence. The case he discusses in

36
Franklin D McMillan, Mental health and Well-being in Animals (Hoboken, NJ: John Wiley & Sons, 2008), 72.
37
“Doing Away with Morgan’s Canon”, 236.
 14

detail is that of the discovery by Karl von Frisch of the famous honeybee dance, and its role in the
communication of bees about the location of food sources.38 Von Frisch discovered that the bees
figure-of-eight-shaped “waggle dance” encodes three pieces of information: the direction of the
food source (encoded in the angle of the waggle movements through the centre of the figure-of-
eight relative to the vertical, which represents the angle of the direction in which the food is located,
relative to the angle of the sun); (2) the distance from the hive of the source, encoded in the
duration of the waggle runs through the centre of the figure of eight; and (3) the quality of the food
source, encoded in the vigour of the dance.

Who would have dreamt that bees possessed such a sophisticated communicative device? In
the years immediately following the publication of von Frisch’s findings, many were extremely
sceptical of von Frisch’s claims, and critics defended an alternative “lower” hypothesis in terms of
olfaction. If von Frisch had adopted similar views as his working hypothesis, it seems exceedingly
unlikely that the significance of the honeybee dance would ever have been discovered. The exciting
explanation he eventually offered of the ability of honeybees to use the waggle dance as a means of
finding food, was only discovered as a consequence of his conviction that the dance was a
communicative strategy, and his willingness to pursue this hypothesis with dedication. The lesson to
be learned here, as Fitzpatrick suggests, is that researchers should be free to investigate whatever
hypotheses suggest themselves, all things considered, untrammelled by a prioristic methodological
assumptions. I turn now, to a second case in which the perils of the Canon can be discerned in a
somewhat different way.
In a paper entitled “Distress Vocalizations in Infant Rats: What’s all the Fuss About?”, Mark
Blumberg and his co-authors report on their investigations into ultrasonic vocalizations emitted by
infant rodents on being separated from the nest. Such vocalizations had previously been interpreted
by some investigators as cries of distress; Blumberg et al wanted to investigate the alternative
hypothesis that the vocalizations were merely “acoustic by-products of a physiological manoeuver,
the abdominal compression reaction (ACR) that increases venous return to the heart when return is
compromised.”39 This had been suggested to them by the existence of two other contexts which are
known to reliably elicit ultrasound production – extreme exposure to cold and administration of
clonidine. Each of these two contexts produces a decreased cardiac rate, against which the ACR
manoeuvre provides some protection. Their experiments revealed that administration of clonidine
did indeed increase venal return to the heart (as well as ultrasound emission). These results, they
suggest “provide strong, direct support for the ACR hypothesis, and, by doing so, underscore the
potential pitfalls of anthropomorphic interpretations of the vocalizations of infant rats.”40

If the ACR hypothesis is the hypothesis that the vocalizations produced by infant rats when removed
from the nest are mere by-products of a physiological maneuver, we must ask what it means to say
that the experimental results “provide strong direct support for the ACR hypothesis.” It is legitimate
to say that relative to the state of the evidence prior to these experiments having been done, there is
now more reason than there was previously to think the hypothesis is a possible explanation of the
vocalizations – because we now know of another context in which ultrasound production and
increased venous return go together. But that is all we can say. We cannot make any further
judgements. The ACR hypothesis is somewhat more likely than it was before to be true; but we
cannot say (on the basis of these experiments alone) that it is now more likely to be true than the

38
Karl von Frisch, The Dance Language and the Orientation of Bees (Oxford: Oxford University Press, 1967).
39
Mark S. Blumberg, Greta Sokoloff, Robert F. Kirby and Kristen J Kent, “Distress Vocalizations in Infant Rats:
What’s all the Fuss About?”, Psychological Science 11 2000, 78-81: 78.
40
Blumberg et al, “Distress Vocalizations in Infant Rats”, 78.
 15

explanation in terms of emotional distress. So the results do not provide strong or direct
experimental support for the ACR hypothesis as against the distress hypothesis. We are still entirely
in the dark about which hypothesis offers the better explanation of the phenomenon observed.

In the light of this, what entitles Blumberg and his colleagues to make the claims they do about the
perils of anthropomorphic interpretation? We have no reason whatever, as yet, to suppose that the
explanation in terms of emotional distress is mistaken. We do have reason to think that we should,
for the moment, be agnostic, until further enquiry has given us reason to think one or other
explanation is definitely better; and perhaps Blumberg et al merely mean to make the point that
other explanations for the infant rat vocalizations may be possible and should have been considered
before the distress vocalization hypothesis had been endorsed. That is true enough and mere
evidentialism would support it. But it seems to me that Blumberg and his colleagues mean to go
further than this. In a follow-up paper, they suggest that their theory, which claims that “a common
physiological event triggers ultrasound production regardless of the context or manipulation used” is
more parsimonious than that offered by the distress hypothesis – a claim which in this context is
perilously close to an invocation of the Canon, a plea to favour the “lower” process which seems to
offer a good explanation in the clonidine case, over the “higher” which attributes an emotional state
to the infant rats. But considerations of parsimony cut both ways. For we know that emotional
distress is the reason why some infant animals vocalize, so there is a variety of explanatory
parsimony one can draw on to support the “higher” explanation, too. Blumberg and his co-authors
are right that “it certainly does not follow that all infant vocalizations are emitted for a
communicatory purpose.”41 But once again, the correct response is agnosticism, and further
investigation. One can draw, I suggest, no conclusions at all about the relative likeliness that ACR
explains the ultrasonic vocalizations, as opposed to the hypothesis of emotional distress, from the
research they cite.

(vi) Conclusion

It is almost impossible for an armchair philosopher criticising the practices and assumptions of
working empirical scientists to avoid courting opprobrium. So perhaps it is necessary to say, in
conclusion, that I hold no brief for anecdotalism, nor for inappropriately anthropomorphic
interpretations of animal life; and I have the utmost respect for empirical methods. But one should
not confuse the application of the Canon with the careful interpretation of empirical results. The role
it has played, indeed, amounts, in my view, to a rejection of the cautious agnosticism that ought to
be the proper response to an as yet empirically unsettled question as to what mental capacities truly
account for a particular instance of animal behavior.

References

Allen-Hermanson, Sean. “Morgan’s Canon Revisited.” Philosophy of Science 72 (2005): 608-31.

Blumberg, Mark S., Sokoloff, Greta, Kirby, Robert F., and Kent, Kristen, J. “Distress Vocalizations in
Infant Rats: What’s all the Fuss About?” Psychological Science 11 (2000): 78-81.

41
Mark S. Blumberg and Greta Sokoloff, “Do Infant Rats Cry?”, Psychological Review 108, 2001, 83-95: 86.
 16

Blumberg, Mark S, and Sokoloff, Greta. “Do Infant Rats Cry?” Psychological Review 108 (2001): 83-
95.

Budiansky, Stephen. If a Lion Could Talk: How Animals Think. London: Wiedenfeld, 1998.

Dennett, Daniel. “Intentional Systems in Cognitive Ethology: the Panglossian Paradigm Defended.” In
his The Intentional Stance: 237-68. Cambridge, MA: MIT Press, 1993. Originally published in The
Behavioral and Brain Sciences 6 (1983): 343-90.

Dennett, Daniel. The Intentional Stance. Cambridge, MA: MIT Press, 1989.

Dwyer, Dominic M., and Burgess, Katy F. “Rational Accounts of Animal Behaviour: Lessons from
C.Lloyd Morgan’s Canon.” International Journal of Comparative Psychology 24 (2011): 349-64.

Epstein, Robert. “The Principle of Parsimony and some Applications in Psychology.” The Journal of
Mind and Behavior 5 (1984): 119-30

Fitzpatrick, Simon. “Doing Away with Morgan’s Canon.” Mind and Language 23 (2008): 224-46.

Fodor, Jerry. “Not So Clever Hans.” London Review of Books, February 4, 1999.

Fujiwara-Tsujii, N., Yamagata, N., Takeda, T., and Mizonami, M. “Behavioral responses to the alarm
pheromone of the ant Camponotus obscuripes (Hymenoptera: Formicidae).” Zoological Science 23
(4) 2006: 353-8.

Galef, B.G. Jr.. “Historical origins: the making of a science.” In Foundations of Animal Behavior:
Classic Papers with Commentaries, edited by L.D. Houck and L.C.Drickamer: 5-12. Chicago: University
of Chicago Press, 1996.

Hirschfeld, Lawrence, A., and Gelman, Susan A. Mapping the Mind: Domain Specificity in Cognition
and Culture. Cambridge: Cambridge University Press, 1994.

Karin-D’Arcy, Rosalyn. "The Modern Role of Morgan’s Canon in Comparative Psychology.”

International Journal of Comparative Psychology 18 (2005): 179-201.

Lloyd Morgan, Conwy. “The Limits of Animal Intelligence.” Fortnightly Review, August 1893: 223-39.

Lloyd Morgan, Conwy. An Introduction to Comparative Psychology, 1st edition. London and
Newcastle-upon-Tyne: The Walter Scott Publishing Co. Ltd., 1894.

Mizunami, Makoto, Yamagata, Nobuhiro and Nishino, Hiroshi. “Alarm Pheromone Processing in the
Ant Brain: An Evolutionary Perspective.” Frontiers in Behavioral Neuroscience 4 (28) 2010: 1-9.

McMillan, Franklin, D. Mental Health and Well-being in Animals. Hoboken, NJ: John Wiley & Sons,
2008.

Montminy, Martin. “What Use is Morgan’s Canon?” Philosophical Psychology 18 (2005): 399-414.

Newbury, E. “Current Interpretation and Significance of Morgan’s Canon.” Psychological Bulletin 51

(1954): 70-74.

Radick, Greg. “Morgan’s Canon, Garner’s Phonograph, and the Evolutionary Origins of Language and
Reason.” British Journal for the History of Science 33 (2000): 3-23.
 17

Romanes, George. Animal Intelligence. London: Kegan Paul, Trench & Co., 1882.

Shettleworth, Sara J. “CleverAnimals and Killjoy Explanations in Comparative Psychology.” Trends in

Cognitive Sciences 14 (2010): 477-81.

Sober, Elliott. “Morgan’s Canon.” In The Evolution of Mind edited by D.D. Cummins and C. Allen: 224-
42. Oxford: Oxford University Press, 1998.

von Frisch, Karl. The Dance Language and the Orientation of Bees. Oxford: Oxford University Press,
1967.
Wozniak, Robert H. “Conwy Lloyd Morgan, Mental Evolution and The Introduction to Comparative
Psychology. http://www.brynmawr.edu/psychology/rwozniak/morgan.html. Accessed January 12
2015.

Wundt, Wilhelm. Grundzüge der physiologischen Psychologie. Leipzig: Engelmann, 1874.

Wyatt, Tristan D. Pheromones and Animal Behavior: Chemical Signals and Signatures (Cambridge:
CUP, 2014).