"Breakwater: The New Wave, Supervenience and Individualism"(1)

John Bickle has written a stimulating book, Psychoneural Reduction: The New Wave, and I have learned a great deal from it. Not the least stimulating aspect of the book is that it has a villain, as one might expect; revolutions need villains. Indeed the book appears to have an arch-villain, the nonreductive physicalist. That includes me, but even so I felt flattered to be the first nonreductive physicalist Bickle cites. Then I read further, and soon found myself recalling Mark Twain on the occasion of his being tarred and feathered: if it weren't for the honor of the thing, he'd just as soon pass it up. Not only was I being charged with dread property dualism. What Bickle calls property dualism, following Paul Churchland, is not my view. It was as though the mob had tarred and feathered the wrong party.

According to Bickle, quoting Churchland, "The basic idea [of property dualism] is that ... [there] are properties ... held to be nonphysical in the sense that they cannot ever be reduced to or explained solely in terms of the concepts of the familiar physical sciences" (pp. 6-7).(2) But this is not property dualism in the book Bickle cites. There I define property dualism as a matter not of non-explainability but of inexpressibility: "a property is [wholly] nonphysical ... iff every predicate which ... expresses it ... belongs to some domain other than physics."(3) Furthermore, it is no part of my view that the nonphysical properties can none of them ever be explained solely in physical terms; some can, some cannot. And even those that cannot are nonetheless required to be connected, if they are genuine properties, to purely physical properties by intervening chains of intertheoretic explanations, so that the genuine nonphysical property is explained by something that is explained by something that ... is explained in purely physical terms.(4) The chains of intertheoretic explanations bottom out in purely physical terms.

True, it does not follow from there being an intervening chain of intertheoretic explanations that there is a single, leap-frog intertheoretic explanation of the nonphysical property directly in purely physical terms; intertheoretic explanation is not transitive.(5) But such leap-frog explanations are not in general to be expected anyway, not even on Bickle's notion of intertheoretic reduction, which likewise is not transitive. Furthermore, leap-frog explanations are not needed for the appropriate sort of physicalist grounding of nonphysical properties ultimately in purely physical properties. The reason is that the intervening intertheoretic explanations enable us to tell a detailed story about specifically which physical properties of which things determine whether or not a given thing has a certain nonphysical property -- a story we could not tell without the explanatory and evidential resources of the particular intervening connective theories on which the intertheoretic explanations are largely dependent.(6)

Of course there may be nonreductive physicalists so incautious as to hold that there are genuine nonphysical properties for which there is no explanatory account that bottoms out in physical properties, whether by way of a single leap-frog explanation directly in terms of physical properties or by way of an intervening chain of inter-theoretic explanations. So much the worse for them. Nothing essential to nonreductive physicalism entails any such incautious view.

Here I should pause to emphasize what may be obvious: the foregoing is not a criticism of the new-wave account of theory reduction proper, but a criticism of one of the motives for advancing it, namely to undermine nonreductive physicalism. Indeed, as we'll see below, nonreductive physicalism and the new wave may well be compatible with one another, even need each other, provided we distinguish between the new-wave account of theory reduction and some potentially misleading concrete examples often used to illustrate it.

There is another charge Bickle levels against nonreductive physicalism, and it is the one I wish to concentrate on this morning -- the charge of dread downward causation. Here too my criticism is not of the new wave proper but of one of the motives for advancing it. Following Jaegwon Kim, Bickle concludes, though tentatively, that "contemporary nonreductive physicalism [is] committed to mental properties possessing novel causal powers not had by their base properties" (p. 11). But Kim's argument for this conclusion rests on some quite problematic assumptions -- assumptions to which nonreductive physicalism is not committed (nor, I suspect, is the new wave). One of them is the principle of the Causal Individuation of Kinds:

CIK. Kinds in science are individuated on the basis of causal powers; that is, objects and events fall under a kind, or share in a property, insofar as they have similar causal powers.(7)

Kim relies on CIK, as if on a bit of settled wisdom, in his discussion of mental and other downward causation and in his argument that nonreductive physicalism is committed to a mysterious sort of downward causation "that no underlying physical-biological properties can deliver."(8) Settled wisdom it may be, but the principle suffers numerous solid empirical counter-examples, especially from biology.

Consider the kinds biologists call adaptations. The heart, for example, is an adaptation kind. Furthermore, adaptations have functions; they are for this or that, as the heart is for pumping blood. As Robert Brandon puts it in his rigorous Adaptation and Environment, "Whenever we hypothesize that some trait is an adaptation, it makes sense to inquire about its function," what it's for.(9) When we say, in the intended sense, that the heart is for pumping blood, we imply that even when a token heart cannot pump blood, nonetheless pumping blood is what it is for, what it is supposed to do; we call it defective, distinguishing between what it actually does or is disposed to do and what it is for, its function.

Already we see here the considerable gap between the adaptation kind "heart" and the causal powers of token items that fall under it -- your heart, say, or mine. It is just false that an adaptation kind -- heart, eye, memory module, signal sequence of proteins, whatever -- is so individuated that its tokens fall under the kind "insofar as they have similar causal powers." For in each case a token of the kind may totally lack the relevant powers. Such examples also counter what Kim calls Alexander's dictum, namely, "To be real is to have causal powers"(10) (as in "To be a real heart is to have the causal powers of a heart").

Furthermore, given that many psychological kinds are adaptation kinds, related examples counter another of Kim's operating principles, what we might call the synchronicity thesis:

ST. All psychological kinds supervene on the contemporaneous physical states of the organism.(11)

Why such a principle should recommend itself is hard to understand, after even a moment's reflection on psychological adaptation kinds. Consider the psychological kind called imprinting, which occurs mainly in precocial species -- those in which the young are born at a relatively advanced stage of development, making them capable of wandering away from their parents. Imprinting occurs not only in chickens, geese and ducks, but in sheep, zebras, most other ungulates, many fish, Japanese quail, guinea pigs and more. Since the young of these species are dependent on their mothers for nourishment, protection and guidance, the precocial infant must form an attachment to its mother and not move too far away. Imprinting, like some other kinds of learning (bird songs, for example), involves a sensitive period during which the infant animal must be exposed to a model. After that period imprinting will not occur.

According to psychobiologists, the primary function of imprinting is to enable the young animal not only to recognize its own mother, but to discriminate her from among other adults of its species. It may first imprint visually, then associate not only visual appearance with its mother, but scent, sound or touch. Even the first imprinting involves learning. The chick learns this shape is the one to follow, and this "observational learning" (as it is called among psychobiologists) clearly affects later behavior. Once imprinting has occurred and the object's features are identified, different objects will be discriminated from it, indeed perceived as unfamiliar, and they will provoke anxiety and attempts to get as close as possible to the more familiar object. "The imprinting of the young bird on one object ... closes down the possibility of its imprinting on others" (Mackintosh); and chicks prevented from imprinting during the sensitive period do not later follow their mothers (unless they are deprived of all visual stimuli during the sensitive period and see their mothers upon emerging).

There can also be an important element of individual recognition in imprinting's effects on sexual behavior. Experiments with Japanese quail show that their sexual preferences as adults are influenced by the precise individuals to whom they are exposed at an early age. Their preferred mate is one like the individuals on whom they imprinted, but not too like (to prevent inbreeding). One psychologist (Kimble) asks of the duckling's imprinting, "Is this instinctive or learned behavior? Manifestly it is both. The instinctive tendency to be imprinted is part of the duckling's biological heritage, while the object on which it is imprinted is a matter of experience." Mackintosh says, "The primary function of [filial] imprinting ... is to enable the young animal to recognize its own mother from among the other adults of its species." Think of the problem the new-born zebra faces in learning which zebra with all those wiggly stripes is its own mother. No wonder the mother zebra keeps other zebras at some distance from her foal for several days after birth.

The point of all this elaboration of concrete material from biology is to make clear that imprinting is a psychological adaptation kind (involving as it does perception, learning, recognition and more), and that among the things it is for is getting Junior to imprint on Mom. But it should be equally clear that the adaptation kind "imprinting," like any adaptation kind, does not supervene on the contemporaneous physical states of the organism, contrary to Kim's synchronicity thesis ST. What makes it the case that a particular cognitive mechanism in Junior has the property of being an adaptation for imprinting is not Junior's contemporaneous physical states. For one thing, Junior's imprinting mechanism might be severely damaged, deformed diseased -- even all but missing -- hence radically causally unlike his species' imprinting mechanism when operating as designed and under design conditions. For another, when we ask what kinds of information would be used to show in detail that the mechanism is indeed an adaptation, we find that the information refers not to here-now physical states of the organism, but to conditions in a natural-selective past. As Brandon puts it, five kinds of information are required:

(1) Evidence that selection has occurred, that is, that some types are better adapted than others in the relevant selective environment (and that this has resulted in differential reproduction); (2) an ecological explanation of the fact that some types are better adapted than others; (3) evidence that the traits in question are heritable; (4) information about the structures of the population from both a genetic and a selective point of view, that is, information about patterns of gene flow and patterns of selective environments; and (5) phylogenetic information concerning what has evolved from what, that is, which character states are primitive and which are derived.(12)

So the synchronicity thesis ST seems to be not only wrong but wrong-headed, once we consider adaptation kinds, properties or traits. So too for any thesis which implies ST, such as Kim's Correlation Thesis, and indeed his Restricted Correlation Thesis:

RCT. If anything has a mental property M at time t, there is some physical structure type T and physical property P such that it is a system of type T at t and has Pat t, and it holds as a matter of law that all systems of type T have M at a time just in case they have P at the same time.(13)

Because CIK and such related theses as ST and RCT are presupposed by Kim's argument that nonreductive physicalism is committed to dread downward causation, the argument fails. Furthermore, if Bickle's version of the new wave is committed to the synchronicity and individualism involved in these and like theses, it too will be in trouble. For it will be inconsistent with theory reduction in biology.(14) Is it?

Synchronicity and individualism do permeate the concrete examples Bickle uses to illustrate new-wave intertheoretic reduction. Of course this may be an accident of his choice of examples, not a consequence of his new wave approach. But consider his quite general account of the intertheoretic reduction of function. According to the account, function is to be reduced to structure plus dynamics of the mechanism X that has the function (155-158). This entails the synchronicity and individualism that are inconsistent with any adequate physicalist or other account of adaptation functions. Then there his belief that "physicalism is the prediction that theories of intentional psychology ultimately will reduce to theories pitched at the level of physical mechanisms" (17), meaning, apparently, the contemporaneous physical mechanisms of the organism. But to the extent that intentional psychology involves adaptation kinds, as it certainly does, and because adaptation kinds do not reduce to physical properties of the here-now mechanisms that fall under them, the prediction is in deep yoghurt. It's a good thing that physicalism, including nonreductive physicalism, is committed to no such prediction.

True, none of this proves that the new wave entails individualism. That depends crucially on the properties of the new-wave reduction relation , via the conditions on it (67-73). Here I admit to being a bit lazy -- let Bickle explain whether the properties of entail individualism. If the relation is individualist, or is even biased in that direction, then the resulting notion of reduction conflicts with any adequate physicalist or other account of adaptation functions. If is not individualist, as I suspect, then new wavers should emphasize that it is not, and modify their diet of unrelievedly individualist concrete examples of reduction with some fiber from evolutionary biology. Further, the way would be open for nonreductive physicalists themselves to use the considerable resources of new-wave theory reduction for various purposes, including a new-wave account of adaptation kinds and their functions, the idea being that adaptation kinds/functions reduce not to physical properties and relations of the here-now tokens that fall under them, but to spatially and temporally scattered physical conditions in their natural-selective past.

If adaptation kinds, and what a given such kind is for, are not individuated by the synchronic causal powers of their tokens, by what exactly are they individuated? If they do not supervene on contemporaneous states of the individual token organism, on what exactly do they supervene? And above all, what are we to make of talk about how the heart, for example, causes various things to happen, such as circulation of the blood?

As regards the individuation and supervenience of adaptation kinds, and of what a given such kind is for, the answer clearly must be in terms of natural-selective pressures that have been at work on heritable variations over time.(15) What determines/subvenes that a given here-now token mechanism X is an adaptation A for F -- for imprinting, learning, perceiving, whatever -- is a past series of physical conditions, events and environments over time, in which X's ancestors were selected for because often enough they performed or had the causal effect F; and it is to this past series of conditions that adaptation talk might well new-wave reduce. It is not the physical properties or relations of X that determine/subvene that X falls under an adaptation kind A for F, or to which it might reduce, but a history in which X is a late arrival. It follows that, as Ruth Millikan puts it, X's having "a teleofunction is a causally impotent fact about it. Especially, it is never directly because a thing has a certain function that it performs that function or any other," or that it has this or that particular causal or other disposition.(16)

But if having a teleofunction to do F -- that is, being an adaptation A for F -- is a causally impotent fact about A, what are we to make of talk about A's causing something? For example, what are we to make of talk about Junior's imprinter module causing what it is supposed to cause, namely imprinting Junior on Mom? Clearly, we cannot say that it is in virtue of, or qua, falling under the adaptation kind "imprinter" that Junior's imprinter X causes Junior to follow Mom around; some things that fall under the kind have no such causal powers. Rather, insofar as imprinter X causes what imprinters are supposed to cause, it does so in virtue of, or qua, both falling under the imprinter adaptation kind and operating as designed and in design conditions; only then does the mechanism X have the relevant causal powers. Even though this construal makes reference to the kind, it also makes reference to the here-now physical trait or mechanism X that causes what it is supposed to cause (under design conditions). This solves the problem Bickle stresses, following Kim, "of how a mentalistic explanation of a behavioral event can coexist with a physical explanation" (10), when the mentalistic explanation is in terms of an adaptation kind.(17)

At this point, if not long before, some philosophers are likely to object that "physical relations can be arbitrarily complex, diachronic, and context-encompassing," so that if the physical properties of X's ancestors in various environments determine that X falls under an adaptation kind, that is just a physical relational property of X. But does the predicate "the physical properties of X's ancestors in various environments determine that X falls under an adaptation kind" express a physical relational property (given among other things the predicate's talk of a nonphysical kind)? Is it even a physical predicate, namely a (first order) compound of basic physical predicates -- of those atomic predicates that do the real explanatory work in the papers, treatises and textbooks of today's best physics?(18)

Even if we assume it is a physical predicate, not every physical predicate expresses a physical property, relational properties included; physical relations can be complex, diachronic, context-encompassing, but not arbitrarily so. For example, 'is an electron of' is a basic physical relational predicate, and 'is not an electron of' would therefore be a compound physical predicate. But merely not being an electron of something is not always a physical relational property if ever. Were it so, completely nonphysical spooks would have a physical property after all, that of not being an electron of something, and so too for Cartesian mental substance, numbers, and more. Nor would there be any reason not to allow the same for the complement of any other kind of relational property. The complement of a moral relational property would itself be a moral property, and we would have to say that electrons have moral traits simply because they are not morally bad for someone (and not morally good). Worse, we would have to say that everything whatever has properties of every kind, since for each kind each thing would have the complement of many a relational property of that kind.(19)

Armstrong goes so far as to deny that negations of physical universals are universals at all, let alone physical; so too for disjunctions of them. His idea, I take it, is that for something to count as a physical property, it must be among those that would actually be used by physicists, entering smoothly, informatively and with effect into their laws, counterfactual generalizations and explanations. Such a property should do the kind of focused explanatory work performed by the robust physical universals or kinds involved in connective theories like a kinetic theory of temperature, a quantum theory of the chemical bond, various physical theories of sound and color, and so on. From this more fully naturalist point of view, a physical property is hardly just whatever is projected by the recursive predicate- or sentence-forming operations of logic and set theory, as all too many philosophers suppose from their non-naturalist armchairs. We should not mindlessly chase propertyhood up the tree of syntax.

It is worth emphasizing that Kim's whole approach, like much of Bickle's, is strongly individualist, not just in CIK, RCT, Alexander's dictum, and the like. According to Kim, all a thing's properties are determined by or supervenient on its own physical properties. This is explicit in Weak Supervenience (WS), which Kim deems "highly plausible to regard ... as minimally necessary for any claim of determination or dependency between sets of properties." WS says that necessarily, for any x and any y, if x and y have the same physical properties P, then they have the same nonphysical properties N.(20) To meet various counterexamples, Kim lets the P-properties include relational P-properties. A thing's properties all supervene on, and are determined by, its own physical properties and relations.

But it is not in the least clear, for reasons given in the last four paragraphs but one, that what X's adaptation properties supervene on -- namely a past series of physical conditions, events and environments, in which X's ancestors were selected for (or the genome responsible for them) because often enough they had a certain causal effect -- amounts to a physical property of X. Furthermore, consider mechanisms that are adaptations for producing items that map onto affairs in the world. For example, honey bees contain mechanisms for producing dances that map onto nectar at a specific direction-and-distance L. Derivatively, the dance itself is for, among other things, mapping onto the nectar's being at L. Now assume one of the bees is ill with pesticide. As a result, she visits peach blossoms west of the hive, but on returning tokens a dance for a location to the east. Assume further that nectar happens to exist at this location to the east, so that the dance tokened does map onto its location; but the location to the east is a mile away on the far side of a high hill across a lake, where no bee from the hive has actually been, and indeed not only is there no physical relation between the hive and the nectar to the east which relevantly affects the bees, there is none that does any work in determining that the dance does map onto this nectar's being located to the east.(21)

In this kind of case, a nonphysical property (of mapping onto the nectar's being at L) is determined not by the dance's own physical properties and relations, but only by these together with the physical conditions that underlie the nectar's being at L (plus the past selective conditions that determine what the dance is for). Two dances can be the same as regards their physical properties and relations yet not the same as regards all their nonphysical properties, contrary to the individualist Weak Supervenience. The same is true of certain cases of true belief, which parallel the bee-dance case.(22) Physicalists need therefore to replace any principles that entail WS with principles according to which all of x's properties are determined by physical conditions, just not always by physical conditions that amount to physical properties and relations only of x. One such principle is nonreductive determination, also called "global" supervenience, which Kim rejects -- wrongly, I've elsewhere argued elsewhere at greater length,(23) but that's another story.

Meanwhile, note that not only is a token bee-dance in a causal state (it causes waiting bees to fly in a certain distance-and-direction L), it has propositional content (that nectar lies at L). Bickle asks, somewhat skeptically, "How do theories invoking causal states with propositional content comport with theories invoking brain processes and subsequent effects on the motor system?" (16). The form of a reply can be distilled from the bee-dance case. The relation between (i) the bee-dance token X with the content that nectar lies at L, and (ii) the waiting bee-brain mechanism that interprets X and effects flight by the waiting bee to the nectar at L, is this: insofar as the here-now token mechanism causes the behavior it is supposed to cause, namely flight in conformity to (as opposed to coinciding with) the content that nectar lies at L, it does so in virtue of, or qua, falling under the relevant adaptation kind and operating as designed and in design conditions; only then does the mechanism have the relevant causal powers. Even though this makes reference to the content that nectar lies at L, it also makes reference to the here-now mechanism that causes the behavior it is supposed to cause (under design conditions).

Note in addition that a given here-now token mechanism M, qua physical stuff in a certain physical state, may produce all sorts of "behaviors." Only if M is (a) an adaptation kind for interpreting X and effecting flight in the direction X maps onto, and (b) operating as designed and in design conditions, can M be said to cause the behavior it is supposed to cause, namely flight in conformity to the content that nectar lies at L. This is the difference between the behavior caused by M conforming to the content and merely coinciding with it. Much the same sort of story can be told about the contentful causal states we call beliefs.

NOTES (references below)

1. Delivered at the Symposium on Psychoneural Reduction, University of Mississippi, February 27, 1999.
2. Unless otherwise specified, all page references are to Bickle (1998).
3. Post (1987), 197. Likewise, a property is wholly physical iff every predicate which expresses it belongs to physics. Some properties are partly physical and partly nonphysical, being expressible both by physical and by nonphysical predicates.
4. Post (1987), 5.1; Post (1999), 3.
5. Post (1987), 228; Post (1999), 3.
6. Post (1987), 5.1; Post (1999), 3.
7. Kim (1992a), 17, following Fodor among others. Kim (1992b), 135, also accepts the closely related Alexander's Dictum, "To be real is to have causal powers," which likewise is in trouble.
8. Kim (1993a), V.
9. Brandon (1990), 188. In the present paper adaptation functions are essentially what Godfrey-Smith (1994) defines as functions by his (F3).
10. Kim (1993b), 348ff. So too do such examples counter the Identity of Physical Indiscernibles, according to which x and y are identical if they share all their physical properties (relational properties included).
11. Kim (1993b), 313, 178.
12. Brandon (1990), 165.
13. Kim (1993b), 313.
14. Individualism here is the view that all a thing's properties are determined by its own physical properties, relational properties, or relations, included.
15. Cf. Brandon's kinds of information (1)-(5), above.
16. Millikan (1993), 186.
17. In those cases in which a token X of the adaptation kind is not operating as designed and under design conditions, talk of what X actually causes, even though it makes reference to the kind, is best construed as talk directly about what the token mechanism causes qua its actual causal powers, not qua adaptation kind operating as designed and under design conditions.
18. Post (1987), 3.1.1.
19. Post (1987), 178.
20. Kim (1993b), 84, 58, 79. To say that x and y have the same properties P is to say that (P)(Px <--> Py).
21. The objection that there must be a physical relational property of the dance which does work in determining that it maps is met by showing how the plausible candidate relational properties all fail to do so, in Post (1995), 83-85.
22. Post (1995), 81-85.
23. Post (1987), (1995), (1999).


Bickle, John (1998). Psychoneural Reduction: The New Wave (Cambridge, MA: MIT Press).

Brandon, Robert N. (1990). Adaptation and Environment (Princeton: Princeton University Press).

Godfrey-Smith, Peter (1994). "A Modern History Theory of Functions," Nous, 28: 344-362. Reprinted in C. Allen, M. Bekoff and G. Lauder, eds., Analyses of Function and Design in Biology (Cambridge, MA: MIT Press).

Kim, Jaegwon (1992a). "Multiple Realization and the Metaphysics of Reduction," Philosophy and Phenomenological Research, 52: 1-26.

Kim, Jaegwon (1992b). "'Downward Causation' in Emergentism and Nonreductive Physicalism," in Emergence or Reduction? Prospects for Nonreductive Physicalism, ed. Ansgar Beckermann, Hans Flohr and Jaegwon Kim (Berlin: De Gruyter), 119-138.

Kim, Jaegwon (1993a). "The Nonreductivist's Troubles with Mental Causation," in Mental Causation, ed. John Heil and Alfred Mele (Oxford: Oxford Univ. Press), 189-210.

Kim, Jaegwon (1993b). Supervenience and Mind: Selected Philosophical Essays (Cambridge: Cambridge University Press).

Millikan, Ruth Garrett (1993). White Queen Psychology and Other Essays for Alice (Cambridge, MA: MIT Press).

Post, John F. (1987). The Faces of Existence: An Essay in Nonreductive Metaphysics (Ithaca: Cornell University Press).

Post, John F. (1995). " 'Global' Supervenient Determination: Too Permissive?" in Essays on Supervenience, ed. Elias Savellos and mit Yalçin (Cambridge: Cambridge University Press, 1995), 73-100.

Post, John F. (1999). "Is Supervenience Asymmetric?" In L. C. Pereira and M. Wrigley, eds., Festshcrift in Honor of Oswaldo Chateaubriand (Manuscrito, 1999).

Supervenience, Explanation, Evidence, Reduction || Home Place || john.f.post@vanderbilt.edu