20th World Congress of Philosophy Logo

Philosophy of Mind

Reviving Psychophysical Supervenience

Neil Campbel

bluered.gif (1041 bytes)

ABSTRACT: Many philosophers has lost their enthusiasm for the concept of supervenience in the philosophy of mind. This is largely due to the fact that, as Jaegwon Kim has shown, familiar versions of supervenience describe relations of mere property covariation without capturing the idea of dependence. Since the dependence of the mental on the physical is a necessary requirement for even the weakest version of physicalism, it would seem that existing forms of supervenience cannot achieve that for which they were designed. My aim is to revive the concept of supervenience. I argue that if we construe supervenience along Davidsonian lines — as a relation connecting predicates rather than properties — then it avoids the shortcomings of the more familiar varieties.

bluered.gif (1041 bytes)

When it first appeared on the scene in the philosophy of mind, the concept of supervenience was warmly embraced. Supervenience was thought to capture the idea of dependence without reduction and thus promised to provide a useful framework for discussions of mental causation, phenomenal experience, and, more generally, the relation between the mental and the physical. Since then a great deal has changed. Much careful work has been done to show that philosophical applications of supervenience do not, in fact, achieve what they were thought to. For example, Jaegwon Kim, whose name is closely associated with the concept, has shown convincingly that the standard formulations of supervenience in the philosophy of mind (weak, strong, and global) do not capture the idea of psychophysical dependence. (1) Many philosophers believed that supervenience could express a form of physicalism, but since the concept of dependence is a minimal requirement for physicalism, it began to appear as though this was unlikely. Given this, Kim, along with many other philosophers, has concluded that supervenience is not a solution to the mind-body problem, but instead expresses the very problem itself. In light of this, many have lost their enthusiasm for this idea.

My goal in this paper is to renew our faith in supervenience. To do so, however, will not require the development of a new formulation of the concept; instead, it involves clearing up some misconceptions about an existing version of this relation. I refer to Davidson’s original treatment of supervenience. Most have assumed that Davidson’s brand of supervenience is equivalent to Kim’s "weak supervenience," which is too weak to express dependence. I will argue that Davidson’s conception of supervenience is, despite certain formal similarities, quite different from Kim’s and clearly captures a sense of psychophysical dependence that is of use to forms of physicalism.

As I see it, the difference between Kim’s approach and Davidson’s lies in the items supervenience is thought to connect. Kim’s formulations of supervenience connect properties, which he takes to be the ontological building blocks of events. In general then, Kim treats supervenience as a metaphysical thesis about the distribution of properties in possible worlds. By contrast, Davidson has little tolerance for the idea that events should be analyzed in terms of property exemplifications. Given Davidson’s reluctance to endorse properties, he prefers to think of supervenience as a relation between predicates. If we construe Davidson’s thesis in this way, it is quite different from Kim’s. Far from a metaphysical thesis, Davidson’s is a thesis about our use of language, about the logical relations between the application of certain words. This is corroborated by the following passage where Davidson tries to clarify what his supervenience thesis should be taken to express:

The notion of supervenience, as I have used it, is best thought of as a relation between a predicate and a set of predicates in a language: a predicate p is supervenient on a set of predicates S if for every pair of objects such that p is true of one and not of the other there is a predicate of S that is true of one and not of the other. (2)

Since we are here interested in psychophysical supervenience, the mentioned relation holds between mental predicates and sets of physical predicates. Treating the relation as one between predicates means that if the same physical predicates can be applied to two agents, then we must ascribe the same mental predicates to both. If we take Davidson seriously here and resist the usual temptation to ignore the difference between properties and predicates, this opens a way of regarding supervenience as a relation of dependence. However, the dependence expressed is not metaphysical dependence; as we shall see, it is instead a kind of logical or inferential dependence.

Before I elaborate on this treatment of Davidson’s thesis, it would be useful to say something in defense of the identified distinction between properties and predicates. One might think that Davidson does not take this distinction seriously, in which case my suggestions will be an unfair distortion of his views. In "Thinking Causes," for example, Davidson at times seems oblivious to any meaningful distinction between properties and predicates since he moves quite freely throughout the article between talk of each of these in connection with his thesis of supervenience. While this is unfortunate, I believe this is merely the result of his speaking loosely. At times, Davidson is unambiguous about his attitude toward properties. For example, when he discusses the causal role of mental properties it is clear that he does not endorse the existence of such things. Davidson writes, "Given . . . [the] extensionalist view of causal relations, it makes no literal sense . . . to speak of an event causing something as mental, or by virtue of its mental properties, or as described in one way or another." (3) For Davidson there are no properties in virtue of which events cause. This is not simply due to a peculiarity in Davidson’s understanding of causation, but instead stems from Davidson’s general refusal to reify properties, which is in turn motivated by his adoption of a Tarski-style semantics of truth. In his view there is nothing "in" events that makes it true that they can be described using certain predicates (or bring about certain effects). The application of predicates is therefore always a matter of ascription and thus depends on general principles of interpretation rather than recognition-transcendent facts. So although Davidson at times speaks as though he endorses property-talk, it is clear that he does not take such talk as seriously as someone like Kim does. For Davidson there are no properties in the sense implied by that locution. When he speaks of properties, then, Davidson is best understood as referring to predicates. Thus, when Davidson explicitly defines supervenience as a relation between predicates rather than properties we ought not to conflate the two in the way that Kim does.

I described Davidson’s definition of supervenience as one expressing a form of logical rather than metaphysical dependence. The sort of dependence I have in mind is not the familiar variety of showing that certain predicates from different areas of discourse are definitionally equivalent, and hence can be reduced one to the other in the way some have thought that moral predicates are analytically definable in terms of naturalistic ones. Davidson explicitly rules out this possibility when he says in his description of supervenience that it "does not entail reducibility through law or definition. . . ." (4) What other form of dependence might there be?

The answer to this question can be found in Davidson’s interpretationalism. According to Davidson, to have beliefs and desires is to have them ascribed by an interpreter, for there are no independent facts of the matter about mental content: "If we cannot find a way to interpret the utterances and other behaviour of a creature as revealing a set of beliefs largely consistent and true by our own standards, we have no reason to count that creature as rational, as having beliefs, or as saying anything." (5) What mental predicates are ascribed to an agent depend on the procedures we follow in the task of interpretation. There are, of course, three central things we need to consider: (1) The behaviour of the agent, (2) the agent’s relation to his or her environment, and (3) in accordance with the principle of charity, the assumption that if the agent has any beliefs at all, then they will conform largely to our own.

Conditions (1) and (2) can be expressed in terms of a set of physical predicates and I want to suggest that these are the ones on which the mental predicates supervene and depend. Condition (3), however, introduces a complication. (3) includes other mental predicates, in which case it seems that the mental predicates ascribed by an interpreter do not depend logically on physical predicates alone. This is true in a sense, and is a well-known feature of Davidson’s holism regarding the mental, but we need to clarify the status of (3). The principle of charity is a general guiding principle and not an empirical resource in the way that (1) and (2) are. That is, we do not use the principle of charity as the evidential ground that tells us whether or not an agent has any beliefs, or what those beliefs might be. This is why I put it in the form of the conditional: If an agent has any beliefs at all, then they will conform largely to our own. Consequently, the starting point of interpretation deals in physical predicates, for what we can say about the physical state of an agent and the physical conditions of the environment when an utterance is made (for instance, the passing rabbit when an agent utters "gavagai") are primary in the enterprise of interpretation; it is on the basis of such evidence that we ascribe mental content. It therefore seems clear that there is a significant sense in which mental predicates depend on physical ones. Hence, when we think of supervenience as relating predicates, and when we consider the relation that holds between mental and physical predicates in radical interpretation, it is clear that mental predicates depend logically on physical predicates. Despite Kim’s claims to the contrary, then, it would appear as though Davidson’s version of supervenience does express a form of dependence.

Although one might be convinced that Davidson’s views on interpretation and mental ascription describe a relation of logical or inferential dependence between predicates, one might worry that this relation is distinct from that of supervenience, in which case supervenience might not express dependence after all. One could argue that the relations just described concern the development of an overall theory of interpretation rather than the more specific relation between the mental and the physical supervenience is supposed to describe. I suggest that such a concern is misguided because these are not separate ideas. Consider Davidson’s definition of supervenience in "Mental Events": "that there cannot be two events alike in all physical respects but differing in some mental respect, or that an object cannot alter in some mental respect without altering in some physical respect." (6) These describe exactly the relations we expect to find in radical interpretation. If two individuals are in similar environments and behave in the same ways (for instance, they both point at a rabbit and utter "gavagai"), then since we have the same physical evidence as the ground for our interpretation of the utterances, we must ascribe the belief "That’s a rabbit" to both speakers. Similarly, we require behavioural (hence physical) evidence to ascribe a change in belief state to a speaker. If we are to say that the speaker now thinks "That’s an aardvark" and not "That’s a rabbit," the speaker must change his or her behaviour in appropriate ways to warrant a difference in our mental ascriptions. Thus, the relations of dependence generated by principles of interpretation between mental and physical predicates are precisely those described in the above definition of supervenience. Given that Davidson later on formulated supervenience in terms of a relation between predicates, and given that this is consistent with his earlier definition in "Mental Events," I see no reason to think that the form of psychophysical dependence I have described is different from the one Davidson’s supervenience is supposed to capture.

Aside from connecting Davidson’s philosophy of mind to his philosophy of language (which has daunted some philosophers), my interpretation of Davidson’s version of supervenience has the further virtue of removing a difficulty his critics have with supervenience. A number of philosophers have argued that Davidson’s treatment of supervenience entails that two people who are identical in every respect except for one seemingly irrelevant physical detail must have different beliefs (e.g., if one person has one eyelash that is longer than his or her counterpart’s). (7) This, of course, is intended as a reductio ad absurdum of Davidson’s theory. I think it is clear that this objection is unfair to Davidson if we read him in the way I have suggested. First, the criticism assumes that the relation holds between properties, which is not the case. Second, even if it did not make this mistake, it proceeds from the false assumption that there is a particular set of predicates describing certain physical features of an individual (for instance, ones describing the brain rather than eyelashes) responsible for any given mental state. If we take Davidson’s interpretationalism seriously this just is not the case. The relevant physical predicates are not those describing the neural networks of speakers (though they are not irrelevant), they are instead the broader ones that figure in the process of interpretation: the physical environment of the speaker and his or her behaviour. Thus, the objection goes wrong by assuming some form of local supervenience. Since we have seen that mental predicates supervene non-locally on physical predicates (by virtue of the role of the environment and behaviour, broadly construed), the objection misses the mark completely.


The interpretation of Davidson’s account of supervenience I have proposed certainly gives us a sense of psychophysical dependence, but many will find it dissatisfying. I suspect the reason for this is that most would prefer the kind of metaphysical dependence between mental and physical properties Kim tries to develop. While I suppose Davidson’s version is, as he himself confesses, a "bland monism," (8) his refusal to endorse mental properties makes his view less mysterious than Kim’s. If we proceed from Davidson’s assumption that events are mental only when described using mental predicates, and that predicates are simply components of a language that obey the rules of language use and interpretation, then much (though not all) of the mystery of the mental is removed. Thus, instead of trying to squeeze metaphysical dependence out of psycho-physical supervenience by formulating ever stronger modal variants of the relation as some philosophers have suggested, (9) it seems more promising to follow Davidson’s lead and think of supervenience as expressing a form of logical dependence. Of course, such an approach requires that we accept Davidson’s views on the mental and on interpretation. It is another question altogether whether these are acceptable.

bluered.gif (1041 bytes)


(1) See Jaegwon Kim, "Concepts of Supervenience" and "Supervenience as a Philosophical Concept," in Kim, Supervenience and Mind (Cambridge: Cambridge University Press, 1993).

(2) Donald Davidson, "Replies to Essays X-XII," in Essays on Davidson: Actions and Events (ed.) Bruce Vermazen and Merrill B. Hintikka (Oxford: Clarendon Press, 1985) p. 242.

(3) Donald Davidson, ‘Thinking Causes’, in Mental Causation, ed. John Heil and Alfred Mele (Oxford, Clarendon Press: 1993) p. 13.

(4) Donald Davidson, "Mental Events," in Davidson, Essays on Actions and Events (Oxford: Oxford University Press, 1980) p. 214 (emphasis added).

(5) Davidson, "Radical Interpretation," in Davidson, Inquiries into Truth and Interpretation (Oxford: Clarendon Press, 1984) p. 137.

(6) Donald Davidson, ‘Mental Events’, op. cit., p. 214.

(7) For example, see Simon Evnine, Donald Davidson (Stanford: Stanford University Press, 1991) pp. 69-70.

(8) Davidson, "Mental Events," op. cit., p. 214.

(9) For example see, Terence Horgan, "From Supervenience to Superdupervenience: Meeting the Demands of a Material World," Mind 102 (1993) pp. 554-586; Thomas Grimes, "Supervenience, Determination, and Dependency," Philosophical Studies 62 (1991) pp. 81-92.

bluered.gif (1041 bytes)


Back to the Top

20th World Congress of Philosophy Logo

Paideia logo design by Janet L. Olson.
All Rights Reserved


Back to the WCP Homepage