1. “Conceptualism” is the label currently in use for members of a loosely related family of philosophical doctrines that resemble each other in representing human modes of awareness as conceptual all the way down. While conceptualisms are philosophically controversial, friends and foes of the doctrines tend to agree in depicting them as implying that non-human animals lack any but very primitive qualities of mind. The reasons for this are not hard to grasp. Many philosophers champion understandings of what concepts are from which it seems to follow that no non-rational animals possess them. One result is that it may appear to be a necessary consequence of a conceptualist outlook not only that no non-rational animals are rightly credited with capacities of mind exactly analogous to those characteristic of human beings but, moreover, that no non-rational animals possess the sorts of modes of awareness that would justify us in attributing to them any significant mental qualities. In this article, I challenge this line of reasoning, arguing that there is no necessary connection between a conceptualist stance and skepticism about animal minds. My larger aim in arguing along these lines is to motivate a conceptualist orientation that, far from obscuring our understanding of animal minds, equips us to study of the mental capacities of animals clear-sightedly and without prejudice.

I proceed by sketching a well-known conceptualist position and showing that it doesn't interfere with our ability to make mental ascriptions to animals. The conceptualism from which I start is the one described in the writings of John McDowell. My selection of McDowell's work for this project may seem ill advised. McDowell undertakes his conceptualist commitments in the context of addressing philosophical anxieties that arise in connection with reflections on the relation between mind and world. His first comprehensive account of this enterprise is given in his 1991 John Locke Lectures, ‘Mind and World’, lectures that were published under the same title, together with several postscripts, in 1994.Footnote ¹ What may seem to speak against using McDowell's lectures for the purposes of this article is the fact that they get criticized by many readers specifically for depriving us of the resources to credit animals with any except very primitive capacities of mind.Footnote ² In what follows, I concede that there is a rationale for this charge in the main text of the lectures, but I also point out that what is at issue is a set of passages in which McDowell misinterprets the significance of his own conceptualist views for how we think and talk about animal mindedness. I note that already in a postscript published with his lectures McDowell begins to correct his misinterpretation and, additionally, that in yet more recent writings he goes further. Now when he discusses the conceptualist claim that characteristically human modes of awareness are intrinsically conceptual, he brings out how this claim is properly understood as part of a descriptive account of what these modes of awareness are like and that as such it places no antecedent restrictions on how similar to (or different from) ours the modes of awareness of animals may be. My object in exploring these portions of McDowell's thought is to demonstrate the availability of a conceptualist orientation that situates us to accept observations suggesting that some non-human animals possess sophisticated mental capacities.

Despite my general sympathy for McDowell's project, I do not follow his lead in every respect. After sketching the philosophical outlook in which his conceptualism is at home (section 2) and considering its bearing on questions about the mental capacities of animals (section 3), I suggest that if we want to think clearly about animal minds it is helpful to speak of concepts and conceptuality in a manner different from the way in which he speaks of these things. Whereas McDowell favors a notion of a concept that excludes the possibility that a non-rational animal might operate with concepts, I defend a contrasting notion flexible enough to enable us to accommodate the prospect of finding that some non-rational animals are concept-users. Not that this terminological departure represents a philosophical disagreement. McDowell himself stresses that it is possible talk about concepts differently than he does. Yet in order to make a case for the alternative understanding of a concept I favor I need to touch on aspects of human cognitive development that – while allowed for by McDowell's philosophical commitments – are not discussed in his writings. The flexible account of what a concept is that emerges from my reflections on cognitive development is pivotal for my main line of reasoning. I refer to it in demonstrating that it is possible to talk, from within a conceptualist framework, about substantial analogies between human and animal minds (section 4). After thus concluding my argument, I append a discussion of the kinds of familiar observations that speak for attributing concepts, as I understand them, to dogs (section 5).

2. Consider how McDowell develops his conceptualist outlook. He develops it in the course of discussing the problem of what he calls a ‘minimal empiricism’.Footnote ³ The type of empiricism that interests him is ‘minimal’ because it is not a doctrine about the relation between experience and knowledge, as classic empiricisms are, but rather, less ambitiously, a doctrine about the relation between experience and thought about the world, without regard to whether the thought is, in McDowell's pariance, ‘knowledgeable’ or not. A position qualifies as a minimal empiricism, in McDowell's sense, insofar as it represents experience as mediating how thought is directed toward and answerable to the world.Footnote ⁴ McDowell defends the plausible (but by no means uncontested) thought that a minimal empiricism is necessary in that to abandon it is to abandon the very idea of thought being answerable to the world. It is as he elaborates this thought that he begins to present the particular conceptualist position he favors. He recognizes that many philosophers reject as untenable his conceptualist approach to the problem of a minimal empiricism, and, when he responds to some of the most fundamental objections to the approach, he brings out further respects in which the conceptualism he favors is distinctive.

McDowell's story about the necessary elements of a minimal empiricism begins with a description of what it is to be mentally directed toward the world, a description that he inherits, with a small alteration, from Wilfrid Sellars. The Sellarsian insight is that to characterize something as a state of knowing is to situate it in a normatively ordered ‘space of reasons’ so that there is room for questions about, for instance, what justifies it. McDowell notes that Sellars' point has a bearing on even ‘non-knowledgeable’ thought and that ‘a normative context is necessary for the [very] idea of being in touch with the world’.Footnote ⁵ This observation is the heart of McDowell's conceptualism. To say that thought about the world is at home in a normative framework is to say that questions may arise about what justifies it, and, in the case of non-inferential perceptual thought, experience plays this justificatory role. McDowell's conceptualist contention is that it is only by representing perceptual experience as conceptual that we can account for the rational support it brings to our beliefs about the world.

When McDowell describes human perceptual experience as conceptual, he is making two distinct claims. First, he is claiming that experience has a universal form. The idea here, which he leaves mostly implicit, is that in order to play a rationalizing role in our epistemic lives, experience needs to resemble empirical judgment in being about individuals and kinds of things. And in order to be about such things it must have a universal form. For example, if someone is rightly credited with a perception of a dog – a kind of thing – then her perception must have internal to it the thought of connection to other actual or possible representations of dogs and must in this sense have a universal structure. By the same token, if someone is rightly credited with a perception of a particular dog – an individual – then her perception must have internal to it the thought of connection to other actual or possible representations of the dog and must, in an analogous sense, have a universal structure.Footnote ⁶ The second claim that McDowell is making in describing our perceptual experience as conceptual is that reason is what supplies or suffuses the relevant universal structure. When McDowell speaks of ‘reason’ he means the capacity to step back from one's inclinations to believe things and to ask whether one in fact has reason to believe what one is inclined to.Footnote ⁷ It is not difficult to see what leads him to say that, in the case of characteristically human perceptual experience, reason is the source of the universality distinctive of conceptuality. McDowell wants to do justice to the role that experience plays in empirical thought. He is committed to accommodating the observation that (apart from exceptional circumstances) we only credit a rational human being with a perceptual experience when the candidate experience is integrated into her epistemic life so that she can judge that the world is at it presents it to be. This is what leads him to say not only that our perceptual experience has a universal structure but also that this structure falls within the scope of reason.Footnote ⁸

McDowell sometimes formulates the conceptualist account of experience that emerges here in Sellarsian terms by speaking of the ‘Myth of the Given’. Sellars' phrase fits because the account opposes the idea that something that is merely causally – and hence pre-conceptually – ‘given’ might have the significance for thought that experience does. Nevertheless, in order to explain more precisely what McDowell means when he describes Givenness as a myth, it is necessary to mention one respect in which he has revised his conceptualist understanding of experience since the publication of Mind and World. Within Mind and World, McDowell suggests that when he represents perceptual experience as having conceptual content he is thereby representing it as having content that is propositional. Thus in a number of passages he formulates his conceptualist claim about experience by saying that the content of experience is ‘that things are thus and so’.Footnote ⁹ In other writings before and around the same time, McDowell similarly suggests that what he describes as the conceptuality of experience is equivalent to propositionality. For instance, he presents himself as sympathizing with a view advanced by Sellars to the effect that experiences make claims.Footnote ¹⁰ The idea that the content of experience is thus propositional is a central critical target of a number of commentators on Mind and World, including some who are otherwise receptive to McDowell's conceptualism,Footnote ¹¹ and in recent writings McDowell himself rejects this idea.Footnote ¹² This gesture of rejection is relevant here, insofar as it separates the notion of conceptuality from that of propositionality, so I want to consider what speaks in favor of making it.

A good phenomenological case can be made against understanding experience as propositional. Consider the case of visual experience. In such experience, the world becomes visually present to us. And what thus comes into view is describable in propositions. But the descriptive possibilities are unlimited, and none in particular is given. If we are to accept these features of our ordinary understanding of visual experience without correction or qualification, then we can't think of experience itself as propositional. We have to adopt a view of experience that takes seriously the recognition that, as Arthur Collins put it in a commentary on Mind and World, ‘experience does not come, as though, with subtitles’.Footnote ¹³ When McDowell discusses his reasons for rejecting the idea that the content of visual (and other forms of sensory) experience is propositional, he says he wants to take at face value an observation that is closely related to the observation, just touched on, that in perceptual experience no particular propositional description of the world is simply handed down to us. He wants to take at face value that when we move from such experience to beliefs about the world we are not drawing inferences from given propositions. On the contrary, we ordinarily take ourselves to be arriving at beliefs about the world non-inferentially, by articulating, or rendering in propositional form, content that is visually revealed to us.Footnote ¹⁴

These considerations leave untouched McDowell's grounds for speaking of a ‘Myth of Givenness’. His point is that we can't account for the rational interest of experience (i.e., its ability to provide rational support for beliefs about the world) if we conceive it as the non-conceptual, Given upshot of merely causal processes. We can respect this point without construing the content of experiences as propositional. It is enough if we allow that experiences are of individuals, as well as of kinds of things, and that in this respect they have conceptual content. Suppose that I go for a walk and see an unfamiliar dog. The content of one of my visual experiences would be ‘a dog’, and this could be mentioned in specifying rational grounds for certain of my – non-inferentially arrived at – beliefs about the world (e.g., my belief that there is a dog over there). So there is no particular difficulty about arriving at an understanding of perceptual experience on which, while not problematically propositional, it is conceptual and hence immune to worries about the Myth of the Given.

Earlier I noted that, when McDowell describes human perceptual experience as conceptual, he is exclusively concerned with forms of universality that are permeated by reason.Footnote ¹⁵ At this point, it should be clear that he does not restrict conceptuality to these specific forms of universality because he thinks concepts only figure in the sorts of articulated propositions that are the prerogatives of rational beings. McDowell recognizes that conceptuality need not be propositional. But he refuses to talk about even non-propositional conceptuality in reference to creatures that don't traffic in reasons. He refuses because he thinks doing so best enables him to highlight what is distinctive about the mindedness of rational human beings. The resulting terminology has striking consequences for how we talk about human cognitive development and the minds of animals, and these consequences are my concern in sections 3, 4 and 5. Right now, having sketched McDowell's conceptualist understanding of our perceptual experience, I want to consider how he further elaborates it in the process of responding to objections.

McDowell notes that it appears to follow from certain deeply engrained philosophical assumptions that a conceptualist understanding of experience – however necessary for a minimal empiricism – is untenable. The specific philosophical assumptions that most interest him concern the demarcation of the realm of nature, where this realm is taken to be made up of things that are real as opposed to merely apparent. The question of what falls within this realm is pertinent because it seems clear that experience is constituted by transactions (viz., impingements of the world on our senses) that qualify as natural in the relevant sense. Yet the realm of nature is often delimited in a way that would prevent experience, understood along conceptualist lines, from counting as natural. Many philosophers take nature to be limited to the subject matter of the natural sciences. McDowell claims this influential view of nature obstructs efforts to represent experience as both ineluctably conceptual and natural. His point presupposes that the intelligibility of rational relations is different in kind from the intelligibility of the individual natural sciences or, as he puts it in recent work, that such relations possess their own logic and are sui generis.Footnote ¹⁶

This presupposition is by no means philosophically uncontroversial. Many philosophers regard the natural sciences as affording not just an important but a metaphysically privileged mode of access to the world. Some who do so also insist that it must be possible adequately to capture rational relations in the terms of these sciences. These philosophers effectively bring the realm of reasons within the logical realm represented by these sciences, thereby rejecting McDowell's presupposition about the sui generis character of relations among reasons. McDowell refers to such philosophers as ‘bald naturalists’. What interests me now is an observation he makes about the consequences of helping ourselves to the presupposition, foreign to bald naturalists, that relations of reason are sui generis. Against the backdrop of this presupposition, it appears to follow from the influential conception of nature touched on in the last paragraph (i.e., the conception that limits nature to the objects of the natural sciences) that experience cannot be both natural and essentially conceptual. Now it seems clear that, if he is to defend his conceptualist outlook in an attractive naturalistic form, McDowell needs to discredit this conception of nature. And this is what he sets out to do, introducing a philosophically unorthodox metaphysic on which nature is capacious enough to accommodate relations of reason and, by extension, to accommodate his conceptualist understanding of perceptual experience.Footnote ¹⁷

If we are to appreciate the significance that McDowell attaches to his preferred metaphysic, we need to note that he represents his conceptualism about perceptual experience as an instance of a larger pattern of thought. In addition to defending an account of perceptual experience on which it is both natural and conceptual, he claims that characteristically human sensory experience and action need to be understood as natural and conceptual.Footnote ¹⁸ The conceptualist posture that McDowell uses an unorthodox metaphysic to underwrite thus turns out to involve a great deal more than perceptual experience. When he sounds conceptualist themes, he is attempting nothing less than to do justice to ways in which, as he sees it, reason pervades human experience and action. He requires his distinctive metaphysic because he wants to do this without forfeiting the recognition that we are natural beings and that rationality is ‘our special way of living an animal life’.Footnote ¹⁹

McDowell's claim is not that we are born rational but that our rational capacities are capacities that get actualized in non-mysterious, natural ways during our upbringings. In defending this claim, he introduces one of his signature terms of art, describing our rational capacities as making up our second nature.Footnote ²⁰ He talks about second or acquired nature because he wants to accent his interest in eliminating philosophical obstacles that seem to speak against taking cognitive growth, as he understands it, to be a natural developmental process. This is the aspect of McDowell's work that primarily interests me here. Below I argue that the same sorts of considerations that McDowell adduces in representing human cognitive development as a natural process also enable us to represent non-rational animals as possessing a wide range of different mental capacities. In arguing along these lines, I help myself to McDowell's broad conception of nature. So I should mention that his attitude toward the metaphysical tradition he is challenging is not at all casual. A preoccupation with attacking an overly restrictive understanding of nature is an organizing theme of his published work. He returns again and again to criticizing a priori arguments for assigning metaphysical priority to the objects of the natural sciences. He traces the picture of mental directness to the world that he thinks drives these arguments back to its roots in both rationalist and empiricist traditions, attacking the form it takes in both contexts.Footnote ²¹ Given that his introduction of a broad conception of nature is anything but adventitious, we need not be intellectually irresponsible if we follow him in conceiving nature more capaciously than philosophers often do. Nor need we be intellectually irresponsible if we take his broad conception of nature as a reference point in thinking about the minds of animals.

3. What implications does McDowell's conceptualist position have for how we think about animal minds? To appreciate the answer that McDowell himself gives to this question, we need to bear in mind that he takes the conceptuality that, as he sees it, is internal to human perceptual, sensory and practical experience to be a matter of forms of universality that are suffused by reason. It follows that he is committed to declaring – as he does declare – that non-rational animals lack our forms of experience.Footnote ²² When he makes this declaration, it may sound as though he is denying that animals possess any significant capacities of mind. But, abstracting from the question of how McDowell's claims sound, they do not imply that animals are mindless brutes.

McDowell's conceptualist position is at home within what he calls a ‘naturalism of second nature’. To grasp the position's bearing on questions about animal mindedness, we need to appreciate what it comes to to say that rationality is a second natural acquisition. The idea is that our rational capacities are capacities that get actualized, in natural or non-spooky ways, in the course of our upbringings. Or, alternately, the idea is that to be rational is to have reached a certain stage in a natural developmental process. This last formulation is telling because, in introducing the notion of a natural process of development, we imply that there can be no question of antecedently restricting the extent to which different stages on the way to rationality resemble each other. Now individual stages of growth need to be understood as locations on a naturalistic continuum. So we need to be open to the possibility of discovering that developmental stages we designate as pre-rational involve capacities of mind that resemble, in significant respects, capacities of mind characteristic of rational human beings. Moreover, in allowing for significant similarities between the mental capacities of pre-rational human beings and their rational human counterparts, we effectively also allow for significant similarities between the characteristic mental capacities of non-rational animals and those of rational human beings. It would accordingly be wrong to say that a McDowell-style conceptualist position obliges us to regard animals as mindless brutes. Properly understood, the position eliminates philosophical obstacles to allowing that some animals have qualities of mind very much like our own.

The perception that McDowell is committed to a more dismissive attitude toward animal minds is, however, not without textual grounds. In Mind and World, McDowell draws conclusions about animals' mental capacities that aren't supported by his main line of argument. After claiming that non-rational animals lack the agency and forms of experience characteristic of human beings, he invites us to understand such animals as living lives that are ‘structured exclusively by immediate biological imperatives’.Footnote ²³ His thought is that features of these animals' surroundings show up for them not as conceptualized items – i.e., not as individuals and kinds of things – but rather as mere particulars on which instinctive or learned responses operate. He marks the distinction he thus envisions between the lives of rational human beings and non-rational animals with the following jargon of Hans-Georg Gadamer's. Whereas individuals who display the conceptually saturated modes of comportment that (as both Gadamer and McDowell see it) are typical of human beings are, in Gadamer's parlance, oriented toward the world, creatures whose lives are governed by immediate biological imperatives merely inhabit environments.Footnote ²⁴ The main point that McDowell uses this Gadamerian contrast between ‘world’ and ‘environment’ to make is not sound, but, the relevant error notwithstanding, McDowell is anticipating a genuine misunderstanding of his own argument. McDowell wants to underline the fact that, in denying characteristically human agency and forms of experience to animals, he is not thereby suggesting that animals are mere automata whose lives are adequately describable in merely causal, physicalistic terms.Footnote ²⁵ He has the resources to construe animals' responsiveness to their surroundings as involving primitive normative reactions and not as merely caused in a mechanical or automatic sense. Yet in the section of Mind and World that I am discussing McDowell does more than simply observe that he is not committed to depicting non-rational beings as automata. Additionally, he claims – and it is in this connection that he makes use of Gadamer's world-environment opposition – that non-rational beings are governed wholly by merely biological impulses.

The author of Mind and World is right that it doesn't follow from the argument of his lectures that non-rational animals are mere automata.Footnote ²⁶ But he is wrong that it follows that non-rational animals must therefore be adequately describable in exclusively biological terms. There is no reason, as far as his argument is concerned, to exclude the possibility of forms of life that are not rational in his sense (i.e., not capable of stepping back from impulses to believe or do things and asking whether they in fact have a reason to believe or do what they are inclined to) and yet are not products of mere control by immediately biological drives. There is, that is, no reason to think that a creature that is not in Gadamer's sense fully ‘world-oriented’ must therefore be limited to negotiating a merely biologically structured ‘environment’. Mind and World's main line of reasoning obliges us to allow for every sort of developmental stage on the way to rationality. This means, among other things, that the book's argument obliges us to allow for modes of conduct that, while not accurately characterized as mere functions of exclusively biological urges, do not rise to the level of rationality. The argument obliges us to allow for modes of conduct that, while more sophisticated than instinctive or learned forms of responsiveness to mere particulars, do not involve a capacity to distance oneself from one's inclinations to believe or do things and to ask oneself whether one should believe or do what one is inclined to. It calls on us to make room for the possibility of discovering that some non-rational animals operate with at least primitive forms of universality and are rightly described as dealing not merely in particulars but in individuals and kinds of things. Yet, while McDowell himself effectively positions us to take this possibility seriously, he also sometimes suggests that he wants to exclude it. My point here is that, to the extent that he makes this gesture of exclusion, he misrepresents the interest of his own project as a resource for thinking about animal minds.

There is a respect in which this criticism of the author of Mind and World is out of date. Already in a Postscript to the book, McDowell distances himself from the features of his view of animals that I just criticized. ‘Perhaps’, he writes, ‘[my] talk of biological imperatives already represents a harder line than I need, even without the reductive conception [of the biological] that I disown’.Footnote ²⁷ That in representing non-rational animals as structured by exclusively biological imperatives he took a harder line than necessary is, I believe, now McDowell's official view, though perhaps he has not always made this as clear as he might have.Footnote ²⁸ In any case, although there is at least one passage in his recent writings at which McDowell acknowledges the possibility of forms of life that are neither mere products of governance by immediate biological forces nor rational,Footnote ²⁹ he excuses himself from exploring this possibility. He says that the emphasis of his work is not on continuities between the lives of rational human beings and those of non-rational animals but rather on the distinctive character of the rational human case.Footnote ³⁰ But his reticence should not prevent us from asking how a conceptualist orientation of the sort he favors might be brought to bear on the sorts of continuities that, within his own writing, never take center-stage. The remainder of this article is devoted to investigating such continuities. A good way to approach what I want to say about affinities between rational human beings and non-rational, non-human animals is to say something about affinities between rational human beings and pre-rational or pre-linguistic children.

4. Children who are not yet rational in the demanding sense at issue hereFootnote ³¹ sometimes behave in ways that resist being fully captured in merely biological terms. This is a preliminary way of formulating the claim that I want to defend in this section. I need a vocabulary for developing my claim, and for this purpose I am going to depart from McDowell. Whereas McDowell restricts talk of concepts to rational individuals, I am going to express the point I want to make by saying that children who are not yet rational may possess concepts. Here it is helpful to recall the kinds of considerations that lead McDowell to adopt his more restrictive terminology. McDowell wants to bring into relief respects in which the lives of rational individuals are distinctive. His point is not that concepts only figure in discursive episodes involving full-blown or propositionally structured reasons. On the contrary, in his recent treatments of perceptual experience he makes use of a notion of non-propositional conceptuality.Footnote ³² He would presumably also allow that conceptuality apart from propositionality has a place in sensory or ‘inner’ experience. But McDowell only speaks of even non-propositional conceptuality in reference to the experiences of individuals who count as rational in the sense that they are capable of giving and asking for (propositionally structured) reasons. This is because he wants to bring out how a rational individual's experiences are integrated into her epistemic life. They are integrated in that she can judge that the world is as they present it to be. Or, to put it in the Kantian terms McDowell sometimes employs, they are integrated in that the ‘I think’ can accompany them.Footnote ³³

My motive for employing ‘concept’ more broadly than McDowell does is not a critical one. The terminology I use is in the service of a thought that is in the spirit of McDowell's work. His broad naturalism encodes an understanding of our capacities of reason as products of a natural process of development. This applies to specific capacities for dealing in universal categories, and it also applies to those capacities we have insofar as the ‘I think’ can accompany our different representations. Suppose we refer to the acquisition of the former capacities (i.e., capacities for dealing in universal categories) as learning. Suppose we then distinguish the acquisition of the latter capacities (i.e., capacities we have insofar as the ‘I think’ can accompany our different representations) from learning by referring to it as maturation. Equipped with these terms, we can say that McDowell makes the plausible assumption that learning and maturation invariably go hand in hand. We can also say that he restricts talk of ‘concepts’ to rational individuals because he is exclusively interested in learning that individuals exhibit insofar they have arrived at an advanced stage of a natural process of maturation. My interest is different. I am interested in capacities of mind, of a sort that McDowell's broad naturalism equips us to sanction, that are precursors to those McDowell himself considers. I talk about concepts in reference to pre-rational children in order to draw attention to cognitive achievements that count as learning in my sense insofar as they meet the following description. Despite not yet demonstrating rational maturity, they already involve primitive capacities for dealing in universal categories.

There is a treatment of such cognitive achievements in Stanley Cavell's Claim of Reason. Cavell wants to get us to see that, in the acquisition of a first language, ‘there is not the clear difference between learning and maturation that we sometimes suppose there is’,Footnote ³⁴ and he invites us to recognize that coming into language involves many steps suffused with elements of both. His central illustration is an anecdote about his daughter's linguistic abilities at fifteen months. It begins with a description of the achievements that led to the recording of the word ‘kitty’ in her Baby Book. There was, Cavell writes:

Cavell adds that a few weeks after the word ‘kitty’ was recorded in the book his daughter smiled at a fur piece, stroked it and said ‘kitty’. At first he was disappointed, but he reflected that it wasn't clear how to transcribe this further performance. Perhaps it should be transcribed as ‘“This is a kitty”, or “Look at the funny kitty”, or “Aren't soft things nice?”, or “See, I remember how pleased you are when I say ‘kitty’”, or “I like to be petted”'.Footnote ³⁶ Without deciding among these options, Cavell observes that ‘in each case [his daughter's] word was produced about a soft, warm, furry object of a certain size, shape, and weight’Footnote ³⁷ and that his daughter is starting to respond in ways that reveal that she is making connections among kittens and kitten-like things or, in other words, that she is dealing in universals. This brings Cavell to the task of characterizing her accomplishment. It would be premature to say, without qualification, that she has grasped what the word ‘kitty’ means. She has yet to learn what ‘giving the meaning of a word’ is.Footnote ³⁸ It would also be premature to say, without qualification, that she has matured to the point of having understood what a kitty is. To know ‘what a kitten is’ is to know many things she doesn't, for instance, that it is a young cat; that it is the sort of creature that will typically learn to climb trees and run from dogs, to purr when happy and hiss when hostile; and that it may have fur of many different colors or even none at all.

After cautioning against saying either that his daughter has learned what the word ‘kitty’ means or that she has grasped what a kitten is, Cavell insists that she has nevertheless taken a ‘leap’ toward inhabiting a world with kittens and mastering the meaning of ‘kitty’. He says that his daughter has moved in the direction of learning a word, and this seems right. But while she has started to deal in primitive universal categories, she isn't yet all that close to word-mastery. Her primary developmental distinction is a certain repertoire of doings-appropriate-when. (Cavell reports that whereas ‘all [his daughter] did with the fur piece was, smiling, to say “kitty” once and stroke it’, when she sees real kittens ‘she not only utters her allophonic version of “kitty”, she usually squeals the word over and over, squats down near it, stretches out her arm towards it and opens and closes her fingers…purses her lips and squints with pleasure’.Footnote ³⁹ Compare this case of Cavell's with Collingwood's case of a baby who pulls off its hat in a happy rebellion, shouting ‘Hattiaw!’, a trick acquired because its mother always says ‘hatty off!’ when removing its hat.Footnote ⁴⁰) Having such a repertoire is a necessary step toward language, but early on the uttering of a word such as ‘kitty’ is just one part of the performance. My point here is not about how close Cavell's daughter is to word-mastery but about how her repertoire of doings-appropriate-when involves the ability to operate with primitive universals.

This ability is a conceptual one in my sense. So I can say that Cavell's daughter is operating with a concept of kittens. In saying this, I am not overlooking the distance that separates her from a mature speaker who possesses knowledge of what a kitten is. I am simply observing that she has progressed rationally enough to deal in forms of universality that are precursors to those internal to mature rational thought about kittens.

These reflections take for granted the sort of relaxed metaphysics of nature that McDowell favors. This metaphysics equips us to represent rational development, understood as a sui generis affair, as a natural process. Speaking of naturalness here allows us to be open about the extent to which stages en route to rationality may resemble each other. It thus saves us from the absurdities of those conceptualists who, because their restrictive metaphysics prevent them from representing cognitive development as natural, find themselves compelled to insist that children lack any substantial mental capacities until the precise moment at which they are deemed rationally mature.Footnote ⁴¹ A relaxed naturalism underwrites a more appealing account of cognitive growth, and my suggestion is that the flexible notion of a concept I have been discussing is helpful for bringing out the account's appeal. Although in this section I have been exclusively concerned with the usefulness of the relevant notion for describing mental capacities of pre-linguistic children, the notion applies to other human cases. It is useful for describing mental capacities of some human beings who are congenitally cognitively impaired as well some who are sick, injured or senile.

These remarks bring me back to this article's central project. My aim in isolating a notion of a concept that is not restricted to fully rational beings was to show that a conceptualist outlook does not prevent us from attributing sophisticated qualities of mind to non-rational animals. This notion of a concept enables us to directly challenge the line of reasoning that makes it seem as though a conceptualist orientation commits us to denying non-rational animals any but very primitive mental capacities. Conceptualists are not obliged to advance a blanket claim to the effect that non-rational animals lack concepts. So they are not obliged to deny the possibility of significant analogies between, on the one hand, typically human modes of awareness and mental qualities and, on the other, those modes of awareness and mental qualities that are typical of different non-rational animals.

This article's argument is now at a close. Before ending my discussion, I want to consider some cases in which it is in fact reasonable to speak of concepts in reference to non-rational animals. The cases I am going to discuss involve dogs. It is possible that a reader sympathetic to my argument in favor of speaking of concepts in reference to some non-rational animals will nevertheless object to the things I now want to say about dogs. So I want to preface my illustrations by noting that the merits of the argument are independent of my claims about its bearing on the canine case.

5. A good way to capture what makes it seem reasonable to talk about concepts in reference to dogs is to consider how we are limited if we exclude this way of talking. The exclusion does not prevent us from describing dogs as learning. We can still talk about canine learning that takes the form of the development of natural responses – or instincts – through more or less complex feedback loops.Footnote ⁴² But in ruling out conceptuality we commit ourselves to representing any natural or learned canine responses as operating on (or triggered by) particulars. We can no longer talk about modes of conduct that involve the recognition either of kinds of things or of individuals. This is what speaks for attributing conceptual capacities to dogs. It seems impossible adequately to capture many things dogs do without characterizing them as recognizing kinds of things and individuals.

Let me start with an example of the sort of observation that speaks for claiming that dogs deal in kinds of things. For the sake of simplicity, I am going to use an example from my life with a dog. My partner Nathaniel and I once had an Australian Shepherd named Sitka, and when Sitka was approximately a year and a half old we somehow acquired a small plastic wind-up crab. One day Nathaniel wound up the crab and placed it on the floor. As it began to whirr and move, Sitka, who had been lying nearby, leapt up into the air and, after first barking sharply several times, began to circle the toy, occasionally throwing herself back into a lunge position and issuing short yips, occasionally getting close enough so that she almost touched the thing with her nose and then jumping back abruptly as if she had received an electric shock.Footnote ⁴³ The next few times Sitka saw the toy she gave roughly the same performance, and for a while she behaved similarly when first confronted with other wind up and electronic toys. But over time her conduct changed. When she saw such a toy, she would frisk around a bit, slapping it with her paws and sometimes picking it up delicately and shaking it. In these respects, her behavior with regard to wind up and electric toys differed from her behavior with regard to, for instance, cats and squirrels. We discouraged her from chasing cats and squirrels, and by the time she was two years-old she would respond to the appearance of cat or a squirrel by doing little more than gazing at it, pricking up her ears and sniffing. Against the background of a familiarity with Sitka's characteristic ways of behaving, it seems natural to say that when she first saw the toy crab she registered a form of surprise that was a function of being confronted with a thing of a new kind. Moreover, just as Sitka arguably came to recognize various other kinds of things (e.g., dogs, squirrels, cats, human beings, trees, cars, streets, houses, stairs and dog dishes), she eventually came to recognize self-moving toys.

Consider now an example of the sort of observation that speaks for saying that dogs deal in individuals. By the time Sitka was six months old, Nathaniel and I had established a routine that prevented us from leaving her at home alone for a stretch of time longer than three hours. One day I came home at the end of a three-hour interval, grabbed Sitka's leash and collar and stepped outside with her, only to hear the phone ring. I stepped back inside, bringing Sitka with me, and answered the phone. The call was from the philosopher Cora Diamond, and I, still a student at the time, was reluctant to explain to Diamond that it would be better for me to speak another time. I stayed on the phone for thirty minutes or so, at which point Sitka – who was then already perfectly housetrained – walked across the rug until she was about four feet in front of me and, staring directly into my face, half-squatted, lifted her leg and emptied her bladder on the rug. Among the things it seems right to say about this stunt of Sitka's is that it was performed specifically for my eyes. Sitka was giving me a demonstration of her frustration, and in this respect she was interacting with me as an individual.Footnote ⁴⁴

Within memoirs people write about their lives with dogs, we find story after story that, like the one I just told about Sitka, depict dogs relating to their owners as individuals. In Bashan, Thomas Mann talks about how his short-haired setter Bashan would wait in the dark lane for his return from the city and how Bashan would greet him by performing a dance ‘consisting of rapid tramplings, of prodigious waggings…which demand[ed] tribute from of his entire hindquarters up to his very ribs’;Footnote ⁴⁵ in All the Dogs of My Life, Elizabeth von Arnim discusses how her huge dog Coco would place a protective paw over her ankle whenever she sat down and how he dragged himself from his kennel in the yard to the threshold of the house to greet her one last time moments before he died;Footnote ⁴⁶ in My Dog Tulip, J.R. Ackerly tells stories about how his female Alsatian Tulip protected him, mourned his absences and how she put herself in the darkest corner of his bedroom when she once accidentally bit him;Footnote ⁴⁷ and in The Dogs Who Came to Stay, George Pitcher inserts, among the many stories he tells about his dogs Lupa and Remus, a comic tale about how Remus once proudly exhibited to him the – in Remus' view – lovely excrement in which Remus had just rolled.Footnote ⁴⁸ In these memoirs, just as in the anecdote about Sitka that I told a moment ago, dogs are represented as relating to human beings as individuals.

Human beings are not the only individuals in dogs' lives. Dogs live lives in which there may be individuals of various sorts – not only individual human beings, but also individual places, individual inanimate objects and individual non-human animals. Several of authors whose dog-memoirs I just mentioned discuss attachments that their dogs form to other dogs. Von Arnim describes how her female fox-terrier Knobbie loved a male fox-terrier named Chunkie and became less affectionate to him after she had puppies, and also how both Knobbie and Chunkie mourned the death of their pup and canine companion Winkie;Footnote ⁴⁹ Pitcher describes how Remus loved his mother and companion Lupa and mourned her death;Footnote ⁵⁰ and during two years in which Nathaniel and I were tutors at Harvard's Mather House, Sitka, who after puppyhood was for the most part quite indifferent to other dogs, became enamored of a grey male whippet named Pensey, a dog who was walked daily by one of our students and whom we ourselves found rather silly.

Taken at face value, these vignettes about how dogs operate with individuals and kinds of things speak for the ascription of concepts in my flexible sense to dogs. It follows from this article's argument there is no general philosophical obstacle to taking the vignettes at face value. To say this is not to deny that it is possible to try to explain the behavior of dogs, without attributing concepts to them, by representing them as pure stimulus-response mechanisms or as beings governed by immediate biological impulses. That this is indeed possible will be clear to anyone with even a slight familiarity with the recent history of the study of animal behavior. But we can allow for this possibility while asking whether purported explanations of dogs' behavior that do not attribute any conceptual capacities do justice to the complexity of their lives. Admittedly, the fact that it strikes many people as natural to think and talk about dogs' lives in ways that presuppose that dogs deal in concepts in my sense does not demonstrate that we should regard explanations that don't ascribe concepts to them as inadequate. It is conceivable that a mode of canine conduct that is such that it is natural to describe it as involving conceptual capacities may, upon closer examination, turn out to be explicable in terms of more primitive capacities. By the same token, it is conceivable that some or all of the memoir-writers I just cited wrongly read sophisticated qualities of mind into what are in fact quite simple modes of canine conduct and that these writers are accordingly guilty of a sort of sentimental projection. It is no less conceivable that my account of my life with Sitka betrays a tendency toward such projection. What would show that we were indeed confronted with a case of sentimental projection? We rightly take ourselves to be confronted with such a case if, for instance, we can show that some environmental stimulus triggers a mode of canine conduct that we took to be expressive of more sophisticated capacities of mind. But the prospect of showing this in particular instances is unthreatening. We can admit that we do sometimes discover that particular aspects of dogs' behavior that we take to be concept-involving are quite primitive and also say that a clear-sighted survey of the massive, millennia-old body of observation and testimony about the lives of dogs nevertheless speaks for attributing concepts to them. This is what I am saying here.Footnote ⁵¹ Dogs recognize individuals and kinds of things, and in this respect they traffic in concepts.Footnote ⁵²

6. This article's defense of conceptualism does not depend for its success on establishing that we are right to ascribe concepts to dogs. But it does depend for its success on showing that a conceptualist outlook can accommodate the possibility that animals that are not – fully – rational may possess conceptual capacities. If a conceptualist outlook can be thus accommodating, then it is wrong to say that conceptualism as such commits us to skepticism about animal minds. It would be more correct to say that a thoughtful conceptualism equips us to accurately capture convergences and divergences between human minds and the minds of non-rational animals.