The thoroughly modern Aristotle: Was he really a functionalist?

Christopher D. Green
Department of Psychology
York University
North York, Ontario M3J 1P3


Presented Sept. 1997 at Aristotle and Contemporary Science, Thessaloniki, Greece.


Aristotle's theory of the psuchê was firmly rooted in his broader metaphysics--hylomorphism--according to which all things are a combination of matter (hulê) and form (morphê). To put things very crudely, his theory of the psuchê went like this: take a human body, combine it with the "form" of life and you have a living human. In this particular case, the form in question is that of the organization of the body's normal biological functioning . Aristotle argued that the psuchê is nothing more than the "form" that was said to turn a body into a living human.

Beyond this basic principle, however, almost every aspect of the interpretation of Aristotle's theory of the psuchê has generated controversy. Perhaps not surprisingly, the debate about what sort of theory Aristotle proposed has, at least over the last few decades, pretty well tracked the debate about what sort of a theory of mind is actually true. Different people have argued that he was, variously, a dualist, a behaviorist, and a reductionistic physicalist. Because Aristotle's theory of the psuchê seems to have had so much to do with "function," there has also been a movement of late to assimilate his thought on the topic to the contemporary position known as computational functionalism.

One of the key implications of functionalism is that minds can, in principle, be "transported" into any physical system that can be arranged so as to preserve the right functional relations. Thus, a human mind could, for instance, be instantiated in an electronic computer, provided the computer was programmed properly (the famous artificial intelligence researcher Marvin Minsky has even suggested this as a solution to the problem of human mortality!). This implication of computational functionalism is known as the "transportability thesis."

In this paper I will argue that there may be a quite general sense in which Aristotle might be considered to have been a distant relative of computational functionalists, but that it courts serious error and glaring anachronism to construe this in a way that puts Aristotle at the head of a line of thought that leads directly to that school of thought (viz., artificial intelligence-inspired cognitive science). In particular, I will show that he explicitly rejected the transportability thesis.

The first hint that Aristotle might have been a functionalist closely related to the contemporary computational sort came in the early 1970s when then-arch-functionalist Hilary Putnam (1973), quipped that "what we [functionalists] are really interested in, as Aristotle saw, is form and not matter. What is our intellectual form? is the question, not what the matter is" (1973/1975, p. 302, original italics). Putnam's quip was developed by the historian of philosophy Martha Nussbaum (1978) into a full-blown interpretation of Aristotle's psychology. In a famous faux dialogue between Aristotle and the ancient materialist, Democritus, she explicitly linked hylomorphism to functionalism (though she completely ignored the question of whether computational structure might be even one of the kinds of form Aristotle would have accepted).

In a 1983 article critical of Nussbaum's interpretation of Aristotle, Howard Robinson argued that Aristotle's theory is dualistic, and therefore not functionalist, in character. The primary piece of evidence he brought forth for this view was Aristotle's account of the active intellect (De anima, III.5), which is plainly said to be immaterial and separable from the body. Robinson (p. 124) said that materialists typically try to dismiss the active intellect as the "unfortunate" effect of Aristotle's overall metaphysical position; not as central part of his psychological theory. According to Robinson (p. 127), however, the active intellect is essential to Aristotle's theory if he is to have an account of how people can think about abstract properties (such as being a triangle). Recognizing that most people would not be convinced by this argument, however, Robinson went on to argue that there is other evidence that Aristotle's theory is dualistic through and through. In particular, he pointed to an analogy Aristotle offers in Book I of De anima--that the psuchê bears the same relation to the body as does a sailor to his ship. By Robinson's lights, no plausible non-dualistic interpretation of this is possible.

Robinson was quick to point out, however, that Aristotle's dualism is not equivalent to Descartes'. It is, he said, "more sophisticated" (p. 132) than Descartes' because the hylomorphic relation he posits between body and mind is "more intimate" (p. 132) than the simple causal relation proposed by Descartes. His attempts, however, to articulate the differences between Aristotle's and Descartes' dualisms were obscure.

In a 1984 response to Robinson, Nussbaum (1984) rejected the idea that there is any similarity at all between Aristotle's and Descartes' positions. First, there is a difference in the sheer metaphysical scope of the two theories. Dualism addresses only the issue of the relation of body and mind (or soul). Hylomorphism, by contrast, addresses the much broader question of substance. The application of this framework to the mind-body problem is but a special case of a much broader metaphysical position. To put things perhaps more concretely, dualism has nothing to say about the existence of, say, buildings, whereas hylomorphism is directed at just such problems; the difference between a pile of bricks and a building is, roughly, that the building-form has been combined with brick-matter to result in a new kind of thing that can function as a shelter. Neither the bricks alone, nor the building-form alone (whatever that might be) could do so.

Another well-known attempt to explicate Aristotle's theory can be found in Christopher Shields' (1988) paper, "Soul and body in Aristotle." Shields considers four options with respect to the interpretation of Aristotle: (1) that the psuchê is identical with the body, (2) that the psuchê is an attribute of the body, (3) that the body "constitutes" the psuchê, and (4) that the psuchê is an immaterial substance. The first of these options--identity theory--is quickly and easily rejected by Shields, and the second--that the psuchê is an attribute of the body--is shown to collapse into the third, that the body somehow constitutes the psuchê.

The notion of "constitution" being used here is a tricky one. Shields defines it thus: "Let us say that the body constitutes the [psuchê] if and only if at any given point in its history, the [psuchê] has all and only the non-historical and non-modal properties the body has" (p. 113). It would take time I don't presently have to explicate these restrictions, but in brief, all the "normal" physical properties--in particular the causal properties--of the body and the psuchê must be identical for the body to be constitutive of the psuchê. It is on this ground--that of causal properties--that Shields attempts to refute the idea that the body, in Aristotle, may be said to constitute the psuchê. To do so he cites Aristotle's own words to the effect that the psuchê (1) cannot be moved, (2) cannot be generated out of pre-existing stuff, (3) cannot be divided, and (4) is not decomposable into elements. All of these, however, are causal properties of the body: viz., it can be moved, generated, divided, and decomposed. If the body constituted the psuchê, says Shields, they would be causal properties of the psuchê as well. This point figures particularly crucially in the present discussion because Shields next writes that the most plausible form of this sort of constitutionalism is computational functionalism. So it would seem that Shields believes that Aristotle was not a functionalist. He would shortly change his position on this point dramatically, however.

Having rejected the first three options for interpreting Aristotle's view of the relation between psuchê and body, Shields opts for the fourth one: that the psuchê is an immaterial substance. He is quick to point out, however, that he does not believe Aristotle's position to have been the same as Descartes'. Instead of the psuchê being an immaterial substance utterly independent of the body, as in Descartes, Shields argues that Aristotle was a "supervenient dualist." The definition Shields gives of supervenient dualism is very--one is inclined to say "overly"--complicated. In essence, it comes down to the idea that the immaterial substance of the psuchê is in some (not very well-specified) way ontologically dependent on the material substance of the body, but not necessarily this body, or even a body very much like this one. Now except for the claim that this form of dependence, whatever it may be precisely, results in an immaterial substance being produced, this position sounds very much like contemporary functionalism. But recall that functionalism is a position that Shields had rejected in his discussion of constitutionalism.

The question of Aristotle and functionalism finally took on some urgency when Shields, in his 1990 paper "The first functionalist," bluntly proclaimed that, "Aristotle and contemporary functionalists share deep theoretical commitments. So deep are these commitments that it is fair to regard Aristotle as the first functionalist" (Shields, 1990, p. 19).

This seems to be a flat contradiction of his 1988 claim that functionalism is a form a constitutionalism, and that constitutionalism is an inadequate characterization of Aristotle's position. Gone from the 1990 paper is the careful, subtle argumentation of the 1988 paper, replaced by a tendentious, quasi-historical argument that leads to a radical new view of Aristotle. In the balance of this paper I will try to show that this view goes much too far and is, in fact, false.

Shields (p. 20) argues that Aristotle's theory was a response to theoretical pressures analogous to those that gave rise to contemporary functionalism; viz., the need to remain within the confines of materialism and the recognition that physicalist versions of materialism are too constraining to be plausible. He responds to Robinson's caution that Aristotle was committed to conscious phenomena that have no place in strict functionalism with the claim that a "weak" functionalism need not restrict the use of "mentalist vocabulary." Not content to leave it at that, however, Shields continues that because Aristotle sometimes presented definitions of mental states that do not include such vocabulary (such as anger being the heating of the blood around the heart), "there is reason to suppose that Aristotle accepts strong functionalism" (viz., the view that all mental states are functionally defined) (p. 28).

Most striking about Shields' 1990 thesis, however, was the confidence with which he asserts that Aristotle wholly endorsed the contemporary "transportability" thesis (viz., that psuchai can be moved from one material base to other quite different ones). His defense of this radical interpretation is based on what must be regarded as dubious textual evidence.

First, in a passage in the Meteorologica (390a), Aristotle claims that the eye is defined by the function of seeing. Shields takes this to mean that Aristotle believed that anything that sees is an eye. In a certain sense this is true, but Aristotle gives no hint of believing that anything other than a conventional, biological eye could possibly see. The possibility is simply given no consideration. Shields has taken the comment out of its context. Aristotle's point here was that a dead eye is, in a certain sense, no longer an eye because it can no longer see, not that something else made to see would count as an eye.

Second, in a passage in De anima (408b) Aristotle speculates that if an old man had his eyes replaced with a younger man's eyes he would "see just as a young man." Shields construes this as meaning that "one could gradually replace body parts at will with others of the right sort [italics added]…and still end up with a functioning human being". (p. 21). But again, there is no indication anywhere that Aristotle believed that anything but the conventional biological parts could possibly be "of the right sort." There is no question of prosthetic eyes being used here.

Finally, in the Metaphysics, Aristotle asks the following question: "the form of man always appears in flesh and bones and these sorts of parts; are these, therefore, the parts of the form and account [of man]?" He then answers his own question, "No, they are but matter, but because [man is not seen] coming to be in other [materials], we are not able to separate them" (1037a-b). On this basis Shields argues that in Aristotle's mind, "'nothing hinders' [man's] being realized in other ways" (p. 23). There is little to warrant this inference, however. At best Aristotle is saying that we cannot tell if people could be realized in other materials. At worst it is simply inconceivable to him that it could be done in any other way. One can only be stunned at the anachronism that seems to be being perpetrated by Shields in such interpretations. There is simply no evidence that the notion of the transportability of the psuchê was ever seriously considered by Aristotle.

A much more sober defense of Aristotle's alleged functionalism can be found in Martha Nussbaum and Hilary Putnam's (1992) paper, "Changing Aristotle's mind." They correctly observe that Aristotle rejected Democritean reductive materialism and Platonic idealism in favor of a position in which the organization, or form, of things is key to understanding their nature, even when that "thing" is the psuchê. Certainly, however, the kind of organization Aristotle had in mind was not computational. Of course, this was not made very clear when Putnam made his famous quip about functionalists having the same interest in form as Aristotle in the 1970s. It all seemed much more obvious to them in 1992, however, by which time Putnam had repudiated computational functionalism (1988). This is a much weaker notion of what counts as functionalism than is found in Shields, however.

Nussbaum and Putnam have, I think, gotten things approximately right. Even so, I believe they continue to read Aristotle a little too liberally. Contemporary functionalism puts virtually no restriction on the kind of matter that might underpin a particular computational state, save that it be "complex enough" to realize the state in question. Aristotle, I think, would reject this. He is quite clear that the only material object that has the potential to be a living thing is a (biological) body. In fact he goes so far as to say that only a living body has this potential (De anima, 412b). Even dead bodies, it would seem, no longer have the potential for life, according to Aristotle.

Now, if Aristotle was not the sort of fellow who could envision bringing a dead body back to life (i.e., rejecting the possibility that after death it retains within it a potential for life that could be "re-actualized" by the "application" of psuchê), it would seem a fair bet that neither was he the sort who would attribute that potential to computers (or thermostats, or liquid running through plastic tubes, or assemblages of beer cans and string, to use a few of contemporary cognitive scientists' more fanciful suggestions). Perhaps it is fair to say that he was a functionalist in the very general sense that he believed the psuchê to be the functioning of a particular body, but it seems unlikely that he would have endorsed even a limited form of the "transportability" thesis (except perhaps in the limiting case of an identical clone down to, say, the molecular level). The kind of life a body has is, for Aristotle, too intimately connected to the kind of body it has for it to be shifted about willy-nilly. In fact, he is quite explicit about this in the oft-ignored Book I of De anima: "[Other philosophers of the psuchê] do not try to determine anything about the body which is to receive [the psuchê], as if it were possible, as in the Pythagorean myths, that any psuchê could be clothed in any body-an absurd view, since it is apparent that each body has its own particular form and shape" (407b). If Aristotle is right about this, then it is, perhaps, Pythagoras we should be looking to as a precursor to the functionalists, not Aristotle, but that idea is for another paper.

Aristotle, then, might be considered to have been a functionalist in the very general sense that he did not believe it to be possible to explain life and mind in reductive physiological terms. His definitions of vital processes, such as perception, are functional. There is no hint, however, that he believed, even in principle, that these functions might be "lifted out" of the body in which they are found, and moved or reproduced somewhere else. In this Aristotle is at odds with contemporary functionalism. That is, his position is far enough away from that advocated by contemporary functionalists that it seems more misleading to include him among them, than to regard his theory as something apart--something sui generis. The temptation to assimilate it to our current accounts, partly in an effort to gain for them and air of ancient authority, is ever-present. As far back as the 6th century John Philoponus noted that "commentators on Aristotle are inclined to try to attribute to him doctrines which they themselves think sound." (ca. 550/1991, p. 12). Ancient as the temptation is, it is one that we should resist. If we can, we may find that Aristotle's account of the psychê is a theory from which we can still learn a thing or two.


Barnes, J. (1979). Aristotle's concept of mind. In J. Barnes, M. Schofield, & R. Sorabji (Eds.), Articles on Aristotle: Vol. 4. Psychology and aesthetics (pp. 32-41). London: Duckworth. (Originally published in Proceedings of the Aristotelian Society, 1971-2)

Burnyeat, M. F. (1992). Is an Aristotelian philosophy of mind still credible? (A draft). In Nussbaum, M. C. & Rorty, A. O. (Eds.), Essays on Aristotle's De anima (pp. 15-26). Oxford: Clarendon Press.

Durrant, M. (Ed.) (1993). Aristotle's De anima in focus. London: Routledge.

Everson, S. (1995). Psychology. In J. Barnes (Ed.), The Cambridge companion to Aristotle (pp. 168-194). Cambridge, England: Cambridge University Press.

Feigl, H. (1970). Mind-body, not a pseudo-problem. In C. V. Borst (Ed.), The mind-brain identity theory (pp. 33-42). London: Macmillan. (Original work published 1960)

Fodor, J. A. (1968). Psychological explanation: An introduction to the philosophy of psycology. New York: Random House.

Fodor, J. A. (1975). The language of thought. Cambridge, MA: Harvard University Press.

Fodor, J. A. (1981). Representations: Philosophical essays on the foundations of cognitive science. Cambridge, MA: MIT Press.

Lawson-Tancred, H. (1986). Introduction. In Aristotle, De anima (H. Lawson-Tancred, Trans., pp. 11-116). London: Penguin.

Nussbaum, M. C. (1978). Aristotle's De motu animalium. Princeton, NJ: Princeton University Press.

Nussbaum, M. C. (1984). Aristotelian dualism: Reply to Howard Robinson. In J. Annas (Ed.), Oxford studies in ancient philosophy (Vol. 2, pp. 197-207). Oxford: Clarendon Press.

Nussbaum, M. C. & Putnam, H. (1992). Changing Aristotle's mind. In Nussbaum, M. C. & Rorty, A. O. (Eds.), Essays on Aristotle's De anima (pp. 27-56). Oxford: Clarendon Press.

Place, U. T. (1956). Is consciousness a brain process? British Journal of Psychology, 47, 44-50.

Putnam, H. (1975). Minds and machines. In H. Putnam, Mind, language, and reality: Philosophical Papers (Vol. 2). Cambridge: Cambridge University Press. (Original work published 1960)

Putnam, H. (1975). Philosophy and our mental life. Philosophical papers: Vol. 2. Mind, language, and reality (pp. 291-303). Cambridge: Cambridge University Press. (Paper originally presented in 1973)

Putnam, H. (1988). Representation and reality. Cambridge, MA: MIT Press.

Robinson, H. M. (1978). Mind and body in Aristotle. Classical Quarterly, N.S. 28, 105-124.

Robinson, H. M. (1983). Aristotelian dualism. In J. Annas (Ed.), Oxford studies in ancient philosophy (Vol. 1, pp. 123-144). Oxford: Clarendon Press.

Rorty, R. (1970). Incorrigibility as the mark of the mental. Journal of Philosophy, 67, 399-424.

Ryle, G. (1949). A concept of mind. New York: Barnes & Noble.

Shields, C. (1988). Soul and body in Aristotle. In J. Annas (Ed.), Oxford studies in ancient philosophy (Vol. 6, pp. 103-136). Oxford: Clarendon Press.

Shields, C. (1990). The first functionalist. In J-C. Smith (Ed.), Historical foundations of cognitive science (pp. 19-33). Dordrecht, Netherlands: Kluwer.

Smart, J. J. C. (1970). Sensation and brain processes. In C. V. Borst (Ed.), The mind-brain identity theory (pp. 52-66). London: Macmillan. (Original work published 1959)

Sorabji, R. (1974). Body and soul in Aristotle. Philosophy, 49, 63-89.