Title: Pragmatic Naturalism, Second Nature, and Communicative Practices: A Deweyan reply to McDowell


Submission type: Paper


Text word count: 3482




John Dewey and John McDowell agree in rejecting 20th non-cognitivist theories of value from a common discontent with the legacies of modern philosophical thought that underpin them. But they disagree about the best way to defend the status of values and of practical reason. Dewey champions a pragmatic, naturalistic account of values as characteristics of situations that meet our needs as biological and social beings, while McDowell rejects it on the ground that it puts us on a slippery slope to reductive materialism and determinism. In what follows I  review McDowell’s and Dewey’s objections to non-cognitivism and to the roles ‘second nature’ plays in their respective reconstructions of  value and its place in the world. I then discuss McDowell’s arguments against naturalizing second nature and a possible Deweyan response.


John Dewey and John McDowell agree in rejecting 20th non-cognitivist theories of value from a common discontent with the legacies of modern philosophical thought that underpin them. But they disagree about the best way to defend the status of values and of practical reason. Dewey champions a pragmatic, naturalistic account of values as characteristics of situations that meet our needs as biological and social beings, while McDowell rejects it on the ground that it puts us on a slippery slope to reductive materialism and determinism. In what follows I  review McDowell’s and Dewey’s objections to non-cognitivism and to the roles ‘second nature’ plays in their respective reconstructions of  value and its place in the world. I then discuss McDowell’s arguments against naturalizing second nature and a possible Deweyan response.



Modern philosophy adopted a  “disenchanted” view of nature, insisting nature is as the physical sciences depict it, containing only those properties and relations susceptible to the experimental methods those sciences employ. Values and secondary qualities were treated as projections of human subjectivity, spread upon a world in which they have no real place. Both Dewey and McDowell reject this view, holding the ‘disenchanted world’ is better understood as a complex theoretical model that stands in for the real world in scientific inquiries -- not to be identified with the actual world or its contents. Both see modern philosophy’s misinterpretation  of scientific models of the world as laying the groundwork for 20th-century non-cognitivisms.  Once one accepts that view of reality, then if human values can only be ‘real’ if what they really are are sub-species of causal forces that the natural sciences recognize.  Attitudes and desires are features of human nature that fall within the purview of the natural sciences, so values are reinterpreted accordingly. But when we do, both men argue, we leave practical reason without any significant role to play in human action, since our so-called  ‘reasons’ for action are merely organic events having no inherent cognitive significance. Say, e.g., I have a certain attitude X towards cats. This gives me no reason to behave one way or another towards cats.  As Dewey remarks, “Cats have claws and teeth and fur. They do not have implications. No physical thing has implications.”[1]  Similarly if attitudes are organic events, they can have no implications for action unless or until they are interpreted in light of some conceptual scheme.   Only if they have cognitive significance can they transfer that significance onto the world.

             Now a conceptual scheme is not just an idea, it is instead a kind of interpretative practice in virtue of which things in the world and our acts, attitudes, and desires, take on significance and become reason-giving.  Practices are rule-governed, but their rules do not conform to natural laws.  Thus when we engage in practices of interpretation we operate in a Sellarsian  ‘space of reasons’ to which natural scientific laws do not directly apply. Or as McDowell puts it, it is to take on, and act from, a second nature [2]

            We acquire second natures  through initiation into cultural practices, into the shared conceptual strategies that our cultures employ for ordering and evaluating inner and outer experience. As we acquire facility with these practices, especially language, we gain entry into that  ‘space of reasons’ in which persons uniquely operate.  Both McDowell and Dewey  assign special importance to language because it is a both a tool and a conduit of enculturation,  whose possession is essential to see and to appreciate the world as a culture’s members are able to do. As McDowell puts it:

Moulding ethical character ...is a particular case of a general phenomenon: initiation into conceptual capacities, which include responsiveness to other rational demands besides those of ethics. Such initiation is a normal part of what it is for a human being to come to maturity ... If we generalize from the way Aristotle conceives the moulding of  ethical character, we arrive at the notion of having one’s eyes opened to reason at large by acquiring a second nature..[3]


Dewey would concur


Language grew out of unintelligent babblings.... But nevertheless language once called into existence is language and operates as language. It operates not to perpetuate the forces which produced it but to modify and redirect them... What is said of the institution of language holds good of every institution ... These are not mere embellishments of the forces which produced them, idle decorations of the scene. They are additional forces ...they open new avenues of endeavor and impose new labors.[4]


That is, they do so for the initiated, those who possess second natures, for whom the world is not merely sensed but laden with meaning and value            


McDowell versus pragmatic naturalism

So far, Dewey and McDowell broadly agree.  They part company when McDowell draws two further conclusions about the relations between our nature as animals beings and the values proper to our second nature as persons:   (I) that our appreciation of  moral and non-moral properties of things, events, and persons are fundamentally different and (II) that moral properties have a special reason-giving force not possessed by non-moral properties. (I) entails that moral values and value-properties are ‘shapeless’ with respect to the non-moral – that the former operates in ways neither reducible to nor consistently linkable to the latter either by reductive or ‘bridging’ principles. (II) entails that the kinds of reasons moral values give us for judging actions within the space of reasons cannot be explained by reference to non-rational motivations we may have in virtue of our first nature as animal beings.

            Dewey would balk at this point, preferring, in  McDowell’s words, “to domesticate conceptual capacities within nature conceived as the realm of law,” and to “reconstruct the structure of the space of reasons out of conceptual materials that already belong in a natural scientific depiction of nature.”[5] 

Why, Dewey would ask, must we suppose that our first and second natures operate so differently that the operations of one could be shapeless with the regard to the other -- so shapeless that bridging principles between them are inconceivable?  McDowell’s reply is that the space of reasons and the forces operating within it are different in kind from the realm of natural scientific law. This makes it impossible that rational decisions and actions could be explained directly or indirectly in terms of the casual forces that the natural sciences recognize.  

            McDowell insists this does not commit us to rampant non-naturalism.  In “Two Sorts of Naturalism,” he writes, “of course first nature matters. It matters, for one thing because the innate endowment of human beings must put limits on the shapings of second nature that are possible for them.... First nature matters not only like that in helping to shape the space in which reflection must take place, but also in that first-natural facts can be part of what reflection takes into account.”[6]  Perhaps so, still  first natural facts cannot matter very much, for  first natural facts cannot bear any non-contingent relationship to second nature valuation. If they did, a zoological study of human nature, from the what he calls the ‘external stance’ of the ‘sheer outsider,’[7] could in principle yield a set of principles about what humans need to do in order to fare well, individually and collectively. And this he wants to deny.

            In Mind and World, Mc Dowell stresses the radical discontinuity between the life of animals who possess only first natures and of persons who also possess second natures.  Animals, he argues, are sensitive but not perceptive creatures.  Perception involves conceptualization, a capacity animals appear to lack. While animals no doubt feel the world they inhabit, McDowell believes, they neither perceive that world or themselves as distinct from it.. One cannot reason about what one cannot perceive, so it also seems to follow that animals must operate outside the space of reasons as slaves of the biological imperatives that determine their behavior. By contrast, although we start life as mere animals, initiation into the space of reasons makes our inner and outer worlds perceivable. Our emancipation from determination by biological imperatives becomes possible through questioning, experimenting upon and  evaluating our experience – all capacities he locates in second nature.

            Thus with human beings, according to McDowell, genuinely new capacities have emerged: conceptualization, perception, reason, values, and freedom from outright biological determination, capacities whose cooperation in the construction of the space of reason break with the natural forces that first produced them, a process  recapitulated by each infant human being who develops its own second nature.  The problem with Dewey’s sort of ‘domesticated naturalism’ is that it tries to deny this break. No doubt we can give “ an evolutionary account of the fact that normal human maturation includes the acquisition of a second nature which involve responsiveness to meaning;” he writes, however, “it would be quite another to give a constitutive account of what responsiveness to meaning is...”  as domesticated naturalisms attempt to do. “That,” he says, “ is a misbegotten idea.”[8]


A Deweyan rejoinder

Why is it misbegotten, Dewey would ask? What does it threaten that we should want to preserve? McDowell’s beguiling thought that we are somehow our own sui generis creations, different in kind from the other creatures with whom we share this planet?  But if it comes at the cost of making explanation of how second nature can arise from mere animal first nature impossible, shouldn’t that give us pause? Moreover there are hidden costs McDowell does not consider.  Since humans start life as biological systems and become persons only through the grace of initiation into language, what must we say about the hundreds of thousands of human beings born congenitally deaf in pre-literate societies denied this the miraculous intervention of training in a natural language? Won’t we have to presume that no matter how well they functioned in their societies they were nothing more than well trained animals?  Lacking language, how could they have been persons, inhabited any part of the space of reasons, or possessed second natures?

            What is the alternative? If we were to adopt a piece meal approach, to say that initiation into some kinds of non-linguistic cultural practices through non-linguistic modes of behavior modification could have awakened these individuals to certain kinds of reasons, then they could come to inhabit a space of reasons, if narrower than that inhabited by their linguistically competent fellows. Of course, we might have to allow that same could be true for some non-human animals, that some non-human animals might also inhabit spaces of reasons, albeit narrower than our own and marked by less spontaneity in deliberation or choice. 

            This is the alternative that Dewey would urge us to adopt. How misbegotten might it turn out to be? First, of course, it means of course we will to push down the capacity for conceptualization and, so presumably also resonance to meanings, below the level of homo sapiens. Many resist this but there seems little reason to do so. We might be able to explain the behavior of very simple a sophisticated version of the  ‘hydraulic model.’ But when we get to the level of mammals, especially predators, this is no longer plausible.

              A dog that can capture squirrels must have some conception of what a squirrel is (not a woodchuck or another dog) and of the places where squirrels retreat, e.g. trees (not holes or dog houses),  if it is to catch squirrels, even if lacks the concepts necessary for further differentiation, say between a maple tree and an elm. It must also have some conception of their physical capacities and their limits.[9]  Moreover, as dogs are pack hunters, they also must be capable of forming simple conceptions of the sorts of things that their prey are and the physical resources they possess. That is, their natures must also provide them with some grasp of intentions, because pack-hunting technique typically involves the manipulation of the prey’s behavior. (One pack-member shows itself at one location, driving the prey to flee towards an escape route where others wait in cover.)  Thus pack hunters are able to distinguish the different ways different prey respond to threats in order to successfully block their avenues of escape. Further, pack hunters must coordinate their own efforts, suggesting they can readily distinguish between their con-specifics, their prey, and the background environment, and can also judge the intentions of conspecifics from their behavior. Conceptualization of this complexity argues a capacity for perception, for experience to be meaningful and not simply undergone. [10]

            The meanings to which pack hunting animals resonate are doubtless limited and largely determined by biological imperatives. Their conceptual capacities are too limited to give rise either to communal practices or to the spontaneity distinctive of McDowellian second nature.  But at least some of the structures that make second nature possible are surely already present. If we move on to animals closer to us, e.g., chimpanzees, the suggestion that we can explain their behavior without invoking conceptual capacities is laughable.  Wild chimps make a variety of tools via practices that are learned rather than instinctual.[11] They play with objects in their environment and invent novel uses for them. Their grasp of intentions is sophisticated, as they routinely practice deception upon one another.[12]   These and other behaviors suggest chimps have the conceptual where withal to learn to recognize, recall, anticipate, and respond effectively to their conspecific’s intentions. 


Language versus Communication

Some research even suggested they might be linguistically competent, for example, the successful effort to teach a young female, Washoe, to use (180) ASL signs reliably in appropriate contexts.  Further research proved that Washoe’s abilities were not atypical, but that chimpanzee  signing fell short of full-blown language. Various hypotheses were suggested to explain their linguistic failings. Some appealed to Noam Chompsky’s theory  that language acquisition requires an innate generative grammatical structure that only humans possess.  Others argued that chimps aren’t capable of using signs linguistically at all; they simply make hand gestures for the sake of rewards they have been conditioned to expect these gestures to produce. To them, these gestures are tools not signs.

            Let’s say, for the sake of argument,  that this is correct – that chimps treat ASL signs as a kind of tool.  Since these tools work by manipulating the intentions of others, let’s call them ‘intention manipulators.’ Now let’s ask, why is it that for chimps signs are only intention manipulators? What about chimps would have to be different for them to appreciate intention-manipulators as signs? The 1st problem isn’t solved by pointing to their lack of an innate generative grammar. That would explain their inability to acquire language but not their inability to acquire signs.  The solution can’t be rational deficiency, as chimps are comparable intellectually to 3-year old human beings who do acquire signs. It can’t be that chimps fail to grasp that others act intentionally – it is precisely because they grasp this that intention-manipulating tools appeal to them.

             Let’s  turn with the 2nd problem – what would have to be different about chimps for them to appreciate intention- manipulators as signs?  Happily, one chimp has turned out to be different in what may be the relevant way: Washoe. After Washoe’s glory days as the world’s first ‘talking chimp’ were over, she became part of a study into whether chimps could learn ASL “signs” from one another, a study focused upon an infant male chimp, Loulis, who had not previously been exposed to signing. The researchers weren’t expecting any of the adult chimps to actively teach Loulis singing, but they did observe sporadic incidents of this kind. Roger Fouts and his collaborators report:

Parents of deaf children often mold the infant’s hand into the configuration and then put the hand through the movement of a sign....Washoe also molded Loulis’ hands. For example, while waiting for a candy bar that a human friend was bringing her, Washoe signed FOOD repeatedly with much excitement and food-grunts. Loulis was sitting next to her, watching. Washoe stopped signing, took Loulis’ hand, molded it into the FOOD configuration, and put it through the FOOD movement several times.... In a similar context, Washoe formed the sign GUM, but with her hand on Loulis’ cheek.... During the first few months after his arrival, Washoe was also observed to place [the sign for] DRINK on Loulis’ lips and HAT on this head, the way parents of deaf children place signs on their infants[13]


The investigators were intrigued but not perturbed by this behavior as it was too sporadic to disconfirm their theories about how Loulis would acquire ‘signing.’ In retrospect, they ought to have been astonished by it. Ongoing research into chimpanzee social behavior has revealed that they are pretty thorough-going Hobbesians, [14] social in just the ways Hobbesian agents are social. They want to be around other chimps but they view one another as competitors for food, shelter, status, and sex. Although chimps cooperate, they usually do so in ways directly beneficial to themselves, rarely helping one another in ways that offer no immediate returns. They avoid risking injury for one another. And they do not teach their young how to make or manipulate the simple tools used in their communities.[15] Each new generation has to learn to make and use these tools for itself. Chimps’s sociability does not extend to cooperating with one another in ways that are apt to confer advantages on others relative to themselves in the competition for scarce resources.

            Washoe’s attempts, albeit  sporadic, to teach Loulis the use a tool  of a supreme importance in the lives of these captive animals was thus extraordinary: (i) because of the degree of altruism she displayed towards Loulis and (ii) because the result was a communicative act. For even if the ASL signs Washoe tried to teach Loulis were, in her mind, merely a kind of tool, the act of teaching the signs was communicative.[16]

            So how would chimps have to be different to understand ‘intension manipulators’ as signs? More would have to be like Washoe, sufficiently altruistic to make teaching, the communication of information, a practice. This may be what most fundamentally distinguishes us as a species – not deliberation,  tool-making, or even language – but teaching. Let a practice of transmitting tools and practices become  established, so that succeeding generations no longer have to reinvent technologies for themselves, and they can turn their playful and creative energies loose on the refinement of those tools to them. The variety and complexity of tools would steadily increase, language being one of the many products of such primitive  communicative practices.

            This is not to say language does not matter. It matters enormously. Language-users can do things non-language users cannot – they can directly question one another’s concepts and beliefs. They can play with these tools, learning to enjoy the spontaneity of that play and to value it for its own sake. They can refine with ever greater precision the cognitive tools with which rationality operates. Further, language provides the tools to investigate a species inner life. Its members can identify and question one another’s beliefs and values. Over time, they can refine the cognitive tools – the concepts and rules – they employ, until they become so adept that they can begin to freely play with their beliefs and values, treating them as objects of interest in their own right.         

            But however valuable, Washoe reminds us that language is not the only form of communicative practice open to creatures intelligent and altruistic enough to attempt communication nor the only means by which they could gain entry into a space of reasons, albeit one less rich than our own. Thus the congenitally deaf in pre-literate societies could have possessed natures.[17]          Even earlier human species, for example, homo erectus, possessed more complex technologies than chimpanzees. At least some of these probably required deliberate teaching and so suggest the existence of communicative practices within those earlier groups.    If Washoe is less atypical than she seems, perhaps some other non-human animals could do so as well. If so, we may need to reconsider how we define ‘person.’ 


Domesticated naturalism

One unavoidable cost -- if it is a cost --of going this Deweyan route is that second nature turns out to be second, not just in emergence but also in priority for practical judgment. Viewing it as an adaptive response to biological pressures, we are warranted in assessing its performance primarily in terms of its effects upon our welfare.  However, we are not committed to the implausible position that all operations of our second nature are motivated by welfare considerations. Our delight in play, novelty,  and creativity ensure that a great deal of the activity of our second natures will have little to do with promoting our welfare, directly or indirectly (after the basic necessities have been obtained.)  And as we are disposed to fall in love with the products of our own creativity, as we do our children, the products of our reasoning and reflection can effectively compete with and even override our concern for our own and others’ welfare. 

            A pragmatic naturalistic approach to second nature diminishes its mystique, I think, but not its importance in understanding personhood, values or practical reason. We have to give up McDowell’s notion that second nature can side-step biological determination or act outside broad constraints of natural law. But Dewey would emphasize that it allows us to “put an end to the impossible attempt to live in two unrelated worlds”[18]  whose differences cannot be bridged. This seems reason enough to forgo enchantment.





[1] John Dewey “Logic of Judgments of Practice,” MW 8: 14–82, 77. And see McDowell’s discussion of whether (and which if any sorts of) naturalisms can have a robust account of practical reason in “Two Sorts of Naturalism,” Mind, Value, and Reality (Cambridge: Harvard University Press, 1998)167-197.

[2] These terms are McDowell’s. Dewey uses the older language of ‘character,’ which he is careful to define as the constituted by acquired dispositions or ‘habits’, not  invariant ‘personality traits.’ As he puts it “Character is the interpenetration of habits.” (Human Nature and Conduct, in John Dewey, The Middle Works, vol. 14, (Carbondale: Southern Illinois University Press, 1983), 29)

[3] John McDowell, Mind and World (Cambridge: Harvard University Press, 1996), p. 84.

[4] John Dewey, Human Nature and Conduct, in John Dewey, The Middle Works, vol. 14, (Carbondale: Southern Illinois University Press, 1983) 57.

[5] Mind and World, 73.

[6] “Two Sorts of Naturalism,” 190.

[7] John McDowell, “Non-Cognitivism and Rule Following” in Mind, Value, and Reality, 198-218, 214. 

[8] Mind and World, 124.

[9] I have a dog that has notably failed ever to acquire such a concept, as he invariably tries to catch squirrels who run up trees by leaping up into the trees in pursuit. Wiser dogs adopt more effective strategies in line with their actual physical capacities.  I take it my own dog’s failure to acquire an adequate conception of his physical capacities proves only he is a mentally defective dog, not that dogs as a rule lack such concepts.


[10] I’m not concerned here with how the details of an account of non-linguistic thoughts might be worked out, but those who are might wish to consult José Luis Bermúdez, Thinking Without Words (Oxford: Oxford University Press, 2003.)

[11] See, e.g.,  Richard Wrangham, W.C. McGrew, Frans B.M. de Waal, and Paul Heltne, Chimpanzee Cultures (Cambridge, MA: Harvard University Press, 1996.)

[12] For example, animals ranking low in the social hierarchy of a given group who discover a resource, not detected by others, will pointedly direct their gaze away from the resource, until higher ranking animals who would naturally claim any such prize for their own, have had their attention distracted, and the low-ranking animal can retrieve it undetected. Low ranking males hoping to seduce fertile females by displaying their erections, have been observed to hide these erections when dominant males approached – clearly anticipating what the intentions of the dominant male would be should he spy them.

[13] Roger S. Fouts, Deborah H. Fouts and Thomas E. Van Cantfort, “The Infant Loulis Learns Signs from Cross-fostered Chimpanzees,” in Teaching Sign Language to Chimpanzees, edited by R.A. Gardiner, B.T Gardiner and T.E. Van Cantfort (Albany: SUNY Press, 1989) pp 286-90

[14] Frans de Waal,  Chimpanzee Politics: Power and Sex Among Apes, (Baltimore:  Johns Hopkins University Press, 1989).

[15] Satoshi Hirata and Maura L.Celli, “Role of Mothers in the Acquisition of Tool-Use Behaviors by Captive Infant Chimpanzees,” in Animal Cognition 6 (2003) 235-244.

[16] That is, she was trying to communicate information about these tools to Loulis.

[17] Presumably some pre-modern human communities were communicative communities of just this sort, who possessed second natures and  values however constrained these might seem to their modern human descendants.

            One might argue even that some of the current debate about whether or not Neanderthals possessed language is beside the point, since language is not a pre-requisite for communication or communicative practices. Neanderthals apparently possessed a hyoid bone and thus were probably physiologically capable of speech. But whether or not they spoke, they made complex tools and shelters, skinned animals, controlled fire, and buried (some of) their dead. This suggests a range of cultural practices through which their members would have acquired a form of second nature. 


[18] Human Nature and Conduct, 11.