papers

In numerous papers Jaegwon Kim argues that nonreductive materialists (i.e., those philosophers who believe that there are no irreducible non-physical objects in the universe, and yet there are irreducible psychological properties which are indispensable in intentional psychological explanations) face two problems. One is that intentional mental properties are not causally relevant; the other is that explanations appealing to these properties are excluded by explanations appealing to physical, in particular, microphysical, properties.[1] The first problem can be called the problem of epiphenomenalism. The second problem is what Kim calls the problem of explanatory/causal exclusion. Epiphenomenalism is not just a problem for nonreductive materialists, but a problem for psychology (as a science of psychological explanation), and for anyone who believes that our thought is the cause of our action. Kim argues that the exclusion problem, on the other hand, is especially a problem for nonreductive materialists. Nonreductive materialists typically buy into the principle of the causal closure of the physical domain, according to which "any physical event that has a cause at time t has a physical cause at t..... [I]f we trace the causal ancestry of a physical event, we need never go outside the physical domain." [Kim 1989b, p. 280] One cannot reject this principle unless one is willing to go back to Cartesian dualism. But the problem is, if all causation is done at the physical level, then there is no causal work left for mental kinds. If mental kinds do not play any causal role, then intentional psychological explanations that appeal to mental properties are not warranted. Therefore, Kim concludes, nonreductive materialism is an untenable position.

Some nonreductionists (Davidson, Dretske, LePore & Loewer, Jackson & Pettit, for example) have tried to argue for the causal relevancy of mental properties from an explanatory point of view; i.e., they argue that mental properties are causally relevant because they play a significant role in psychological explanations of behavior. In other words, they claim that mental properties can be causally relevant without being causally efficacious. Kim, however, argues that this approach would not work. He thinks that a causal explanation is one that gives the causal history of an event. If we want to say, for example, that our desire for water causes us to drink water, then our desire must indeed cause us to drink water if this explanation is correct. But we also say that the neurophysiological state of our having that desire is the cause of the neurophysiological state of our water-drinking, since all the causal work is done at the neurophysiological level. There is only one causal history. Hence, only one explanation can be giving us the real causal story. Since nonreductionists do not deny that causation takes place at the neurophysiological level, mental explanation cannot give a different causal history. Kim argues that mental explanation is thus excluded by neurophysiological explanation. If Kim is right, then the problem of epiphenomenalism would ensue from the problem of explanatory exclusion. Mental properties would not be causally relevant exactly because they are not explanatorily relevant. Anyone who does not want to endorse epiphenomenalism would thus have to take the problem of explanatory/causal exclusion seriously.

In this paper I shall give a close examination of Kim's argument for the exclusion problem. Kim thinks that the only solution to the exclusion problem is reductionism.[2] I shall propose another solution that does not require that mental properties be reducible to physical properties, or that psychology be reducible to neuroscience. My paper is divided into three parts. In the first part I shall explicate how the explanatory exclusion problem is generated. In the second part I shall examine Kim's argument for the exclusion of psychological explanation, namely, his supervenience argument. And finally in the third part, I shall offer my solution to this problem. In a nutshell, my solution will be that of mental/physical properties identity defined locally, holistically. My analysis is based on Davidson's mental/physical event identity. An individual's mental event is nothing but the physical event of the individual's brain. But since properties we use to specify a mental event and properties we use to specify its corresponding physical event must be defined through the whole system in which the event at issue takes place, these properties cannot be the properties dealt with in either psychology or neuroscience. I shall explain how this analysis of identity does not commit one to mental/physical properties reductionism, or intertheoretic reductionism between psychology and neuroscience.

§1. The Generation of the Explanatory Exclusion Problem

The problem of explanatory exclusion is based on a principle of explanatory exclusion, namely:

[EE] Distinct (causal) explanations of a single explanandum exclude one another, in spite of the fact that their explanatory premises are mutually consistent. [Kim 1990, p. 123]

Taken by itself, [EE] seems too strong. And indeed it has aroused some suspicion that Kim is creating a pseudo-problem.[3] However, [EE] is generated by several assumptions that Kim makes, and these assumptions all seem to have some plausibility. In this section we shall take a close look at each of these assumptions.

1. Causal Realism

[CR] Causality itself is an objective feature of reality.[4]

[CR] is posed against a Humean analysis of causation. [CR] can also be viewed as a thesis of the extensionalist's position on causal relations, according to which causality is a relation among events themselves. Davidson, for example, is an advocate of causal extensionalism. I shall leave this assumption as uncontroversial in the context of the debate.

2. Causal Exclusion

[CE] For any single event E, there cannot be two causes C and C*, each of which is a sufficient cause of E, while they are distinct and independent of each other.

Kim considers five possible cases when we identify both C and C* as the causes of E: Case 1. Identity: C = C*. In this case the two causes are not competing with each other since there is only one cause here.

Case 2. Reduction or supervenience: C is distinct from C*, but is reducible to, or supervenient on C*. According to Kim, in this case C is not an independent cause from C* and thus the causal competition does not exist.

Case 3. Neither C nor C* is a sufficient cause of E, though each is an indispensable component of a sufficient cause. This again is an acceptable situation since neither cause is taken to be sufficient by itself.

Case 4. C and C* are different links in the same causal chain leading to the occurrence of E. Again, Kim argues that this case does not constitute a violation of [CE] since one of the two causes is causally dependent on the other cause.

Case 5. C and C* are distinct and each a sufficient cause of E. This is the case that is barred by [CE]. According to Kim, this is a case of overdetermination: "E would have occurred even if either C or C* had not occurred, or had not caused it; the other would have been sufficient to bring it about." [Kim 1989a, pp. 250 - 253]

It is not clear whether there can really be any case of overdetermination. And even if there are, we don't want to view the competition between mental causation and neurophysiological causation to be a case of overdetermination. We also don't want to view the connection between the two causes to be an example of Case 3 or Case 4. The only remaining options are Case 1 or Case 2; namely, identity or reduction. Kim thus argues that if the mind causes, then it is either identical to the brain or reducible to the brain and the two analyses are compatible with each other.

[CE] is not contested by most nonreductive materialists. The question is whether it can be extended to [EE]. Both Burge and Davidson think that causation is one thing while causal explanation is another, and there is no warranty in extending the exclusion of two causes to the exclusion of two explanations. Davidson argues that Kim is committing a slippage in his argument: “It is only if we confuse causal relations, which holds only between particulars, with causal explanations, which, so far as they are ‘sufficient’ must deal with laws, and so with types of events, that we would be tempted to accept the principle of ‘causal-explanatory exclusion’.” [Davidson 1993, p. 16] Kim, on the other hand, thinks that causal exclusion and explanatory exclusion are one and the same problem. He and Davidson or Burge simply differ on the metaphysics of explanation. For Kim, a basic metaphysical assumption concerning explanation is what he calls the explanatory realism.

3. Explanatory Realism

[ER] C is an explanans for E in virtue of the fact that c bears to e some determinate objective relation R.[5]

Davidson, for one, does not endorse [ER]. Davidson separates the issue of causation and the issue of explanation. He treats the first as a metaphysical issue and the second as an epistemological issue. Davidson stresses that explanations are merely “descriptions” of events. He says: "Redescribing an event cannot change what it causes, or change the event’s causal efficacy.” [Davidson 1993, p. 7] If psychological explanation merely gives one description while neurophysiological explanation gives another, then R1 and R2 are not "determinate objective relations." R1 and R2 are merely different relations established through linguistic variation and theory construction.

Burge also does not seem to endorse [ER]. Burge views explanations as giving descriptions of the pattern exemplified in singular events. Burge argues, “The existence of a closed system reflects a pattern of causal relations and of causal explanation that needs no supplementation from the outside. There are no gaps. It does not follow from this that such a system excludes or overrides causal relations or causal explanation in terms of properties from outside the system.” [Burge 1993, p. 102]

Both Davidson's and Burge's counter claims can be met with by another implicit assumption Kim makes concerning the nature of explanation. In contrast to Burge’s pattern view of explanations, Kim argues that explanations are “represented by singular causal propositions.” [Kim 1987, p. 230] We can formulate this assumption as Kim's propositional view of explanation.

4. The Propositional View of Explanation

[PE] 'Having’ an explanation is simply a matter of knowing a certain proposition to be true. [Kim 1987, p. 231]

The main difference between the pattern view and the propositional view of explanation, in Kim’s words, is that “gaining a new explanation, on the pattern view but not on the propositional view, does not necessarily involve acquiring new information about facts of the world.” [Kim 1987, p. 231] Since for every event there is a “unique and determinate causal history," for every event there should only be one unique singular causal proposition that is true. That is to say, if there is only one event causing, then there is only one true propositional causal explanation.

How do we reconcile the two opposing views on the nature of explanation? On the one hand, both Davidson and Burge seem to be right in pointing out the epistemological and theory-laden nature of explanation. Explanation can be given in different fields and each field has its own taxonomy and its own laws. Explanations represent our understanding of the external world. An expert from a different field can give us an explanation that carries with it significant information. If we have two explanations that have different cognitive contents, there is no reason why we must claim that only one of them is true. There is also no reason why we must say that both of them must be false because neither gives the uniquely true proposition.

But on the other hand, Kim could argue that even if those non-causal, structural dependence or pattern explanations are of epistemic value, what he is interested in is only causal explanation. In other words, Kim could grant that non-causal explanations are aimed to redescribe events, and different theories have different vocabulary and laws to offer different explanations. But when it comes to causal explanation, our interest is only in finding out the true causal story. Kim does not need to embrace an explanatory monism, the view that only one explanation is the real explanation.[6] Even though he defends [ER] in many articles, he really does not need such a strong premise to argue for the exclusion problem. What he needs is a more restricted principle as follows:

5. Causal Explanatory Principle

[CEP] A causal explanation of event e in terms of event c (“e occurred because c caused it”) is correct, or true, just in case c did as a matter of objective fact cause e.[7]

According to [CEP], to give a true causal explanation, is to cite a causal relation that actually exists between two events. Any other explanation, however helpful in enhancing our understanding of relationships between events, is not a true causal explanation.

Is [CEP] still too strong? Maybe so. However, it is a reasonable principle as long as Kim also grants other non-causal explanations to be legitimate. A true causal explanation gives a true proposition of the causal relationship. It does not have to be covered under any causal law, and thus it is not restricted to the laws and vocabulary of any discipline. It may not even be available at the present stage of our scientific knowledge. What Kim demands is simply a single proposition that cites the actual cause of a given event. This proposition is what he deems the genuine causal explanation. There may be other explanations that make the connections between events more intelligible to us. They are thus of epistemic value. But on the metaphysical level there can only be one causal connection, and what Kim's [CEP] says is that there can only be one causal explanation that really describes this causal connection. Kim says, "To be in need of an explanation is to be in an epistemically incomplete and imperfect state, and to gain an explanation is to improve one's epistemic situation; it represents an epistemic gain. However, knowledge must involve the real world: to know that p requires the truth of p, and to have a causal explanation of an event requires that the event specified as its cause be, in reality, a cause of that event." [Kim 1989a, p. 256] In other words, Kim is interested in a particular kind of explanation, one that gives us the true depiction of the real causal history. Even if we can argue that other kinds of explanations are epistemically useful, we cannot deny that what he demands is a reasonable expectation of the function of causal explanation.

If we are to give a psychological explanation of action e,[8] then, under [CEP], what our explanation says is that the mental event c characterized by the mental property or mental kind C did indeed cause e. But at the same time, a neurophysiological explanation that cites neural event n gives a complete and sufficient causal explanation of action e. According to [CE], there cannot be two independent and complete causal relations R1 and R2 in any given causation. We therefore need to, asides from explaining the connection between the mental event c and the neurological event n, explain the connection between R1 and R2. This would require us to explain the connection between the mental type C and the neurological type N. If C and N are not coextensive, or one is not reducible to the other, then we are saying that there are two independent and complete causal relations R1 and R2 in the case of action causation. This is not acceptable. So one relation has to go. Since R2, the causal relation between the neural event type N and the action type E, is an objective fact, it seems that it is R1 , or the psychological explanation which specifies R1 , that is to be excluded.[9] This leads us to the problem of explanatory exclusion. That is to say, if any other independent cause is excluded by this causal connection, then any other causal explanation is thereby excluded by the explanation that cites this causal connection. Since this problem is not separable from Kim's causal exclusion principle, Kim usually combines the two terms and calls it the explanatory/causal exclusion.

We can further add another assumption that Kim accepts, which leads one from causal exclusion directly to epiphenomenalism:

6. Alexander's Dictum

[AD] To be real is to have causal powers.

Kim simply rejects Davidson's and Burge's treating events/causation as a metaphysical issue, while treating properties/explanation as an epistemological issue. He argues, "[T]urning away from metaphysics to embrace epistemology, or away from causation to embrace explanation, will not dissipate the need for an account of mental causation." [Kim 1998b, p. 67] This account that Kim demands, is not about whether mental causation is possible, but about how mental causation is possible. Under [CEP], if mental explanations are to be correct or true explanations, then mental properties cited in these explanations must also have real “causal powers”. In giving causal explanations, we simply cannot separate causal efficacy from explanatory relevancy. If mental events do not have causal powers, then, they are not explanatorily relevant. Furthermore, by [AD], they are not real. In other words, mental realism is not tenable if we cannot resolve the exclusion problem.

In brief, explanatory exclusion and causal exclusion, for Kim, are one and the same thing. Kim argues, "[W]e think of these [causal] explanations as directly tied to the actual causal histories of the events being explained, and their 'correctness' as explanations is determined by the accuracy with which they depict the causal connections as they exist. When two such explanations of one event are on hand, we need to know how they are situated in relation to each other on the causal map of the event; one thing that cannot be allowed to happen, if our explanations are to be coherent, is that they tell two different stories about the same region of the causal map." (italics mine) [Kim 1989a, p. 255] What he is pushing for is basically a recognition of the causal exclusion problem. Explanatory exclusion is derived from the causal exclusion principle plus causal realism and causal explanatory principle. Causal realism is not in contention between Kim and nonreductive materialists. And I have argued that his causal explanatory principle is a plausible principle. But even if we reject his causal explanatory principle and thereby reject his explanatory exclusion principle, we still need to deal with the problem of causal exclusion. Mental properties can still become causally inert or causally irrelevant under the principle of causal exclusion, because they would be competing with the neurophysiological properties appealed to in neurophysiological explanations. Unless we find a way to link mental properties with neurophysiological properties, we cannot dissipate the tension between the two sets of properties, each competing to be the properties that are causally relevant.

§2. Kim's Supervenience Argument and Kim's Reductionist's Solution

In Kim's most recent book, Mind in A Physical World, he comes up with another argument against the causal efficacy of mental properties. His argument is called the supervenience argument, because it is based on a mind-body supervenience thesis that both Kim and the nonreductionists endorse. The following is a (somewhat) simplified version of Kim's supervenience argument: [Kim 1998b, pp. 38 - 46]

1. [The supervenience thesis]: Mental properties supervene on physical properties in the sense that if something instantiates any mental property M at t, there is a physical base property P such that the thing has P at t, and necessarily anything with P at a time has M at that time.

2. Suppose that an instance of mental property M causes another mental property M* to be instantiated.

3. M has a physical supervenience base P, while M* has a physical supervenience base P*.

4. But according to the supervenience thesis, M* is instantiated because its physical supervenience base P* is instantiated on this occasion.

5. Therefore, as long as P* occurs, M* must occur regardless of whether or not an instance of M preceded it.

6. Therefore, P* alone is the complete and sufficient cause of M*.

7. For M to be causally related to M*, M must be a cause to P*.

8. But P alone is already causally sufficient for P*.[10]

9. If M also causes P*, it would either violate the physical causal closure principle or would be a case of overdetermination.

10. Therefore, there is no real causal link between M and M*.

In this argument, M precedes M* and P precedes P*, and thus M ≠ M* and P ≠ P*. Distinct mental/physical events have distinct properties. The connection between M and M*, or P and P* respectively, is supposedly causation. The connection between M and P, or M* and P* respectively, is supervenience. Under the supervenience thesis, the supervenience base property necessitates its supervening property. It is accepted by the nonreductionists that there is complete causation between P and P*. What remains to be established is the causal connection between M and M*. But since P causes P* and P* necessitates M*, the causal connection between M and M* becomes superfluous. Using '®' to represent causation, and 'Ý' to represent necessitation, we can further simplify the above argument into the following diagram:

M ------------ M*

Ý Ý

P ® P*

In the previous section, we have shown that the exclusion of two independent and sufficient causes and the exclusion of two complete explanations appealing to different sets of properties go hand in hand. Now according to this supervenience argument, when there is causal/explanatory exclusion, it is mental properties that must go otiose. The only way to preserve the causal efficacy and the reality of the mental, if one accepts Kim's reasoning, is thus to find a way to show that the apparent two causes (mental and physical) are not two separate causes, and the apparent two sets of properties (M and P) are not two independent sets of properties. In other words, either an identity (Case 1) or a reduction relation (Case 2)[11] between the two sets of properties will have to be established in order to solve the exclusion problem.

Kim’s own solution to the explanatory/causal exclusion problem is to dissipate the causal competition by a disjunctive identification between mental and physical properties. He calls this theory multiple-type physicalism (or MTP for short), by which he means that a mental kind M is disjunctively identified with physical kinds P1, P2, ..., where P1, P2,... are all "realizers" (or, the supervenience bases) of M. [Kim 1993c, p. 364] In Kim's analysis, mental properties are second-order functional properties, "properties specified in terms of their roles as causal intermediaries between sensory inputs and behavioral outputs." [Kim 1998b, p. 19] These causal roles are realized by physical properties since the mental realm has the physical realm as its base. Kim calls this view 'physicalist functionalism.' In such a picture, mental properties can have different physical bases. As Kim puts it, "any base properties with the right causal/nomological relations to other properties can serve as its realizers." [Kim 1998b, p. 21] Different biological species and different cognitive structures can all have different physical bases. But as long as they manifest the same causal roles, they are said to have the same mental properties. Mental properties are thus identified functionally, and they are identifiable with, and reducible to, a disjunctive set of physical bases, each of which may realize the same mental properties differently. On the basis of this disjunctive identification, Kim claims, mental properties are reducible to physical properties. This is Kim's model of reductionism. And since mental properties are reducible to physical properties, there is no real causal/explanatory competition between the two sets of properties. Mental properties are thus not causally excluded by physical properties.

Kim's reductionism is different from the standard model of intertheoretic reductionism as explicated by Nagel. According to Nagel's model, for a theory to be reducible to another theory, there must exist universal bridge laws connecting the predicates of the two theories. These bridge laws are expressed in a biconditional form (as, for example, M iff P), such that for every property of the reduced theory, there would be a nomologically coextensive property of the base theory. In the same token, for every law in the reduced theory, there would be an a priori derivation from this law to a law in the base theory. It is against this Nagelian model of reductionism that Fodor provides a celebrated proof that special sciences are not reducible to basic physics.[12] It is also against this Nagelian model of reductionism that Davidson argues that the mental realm is anomalous and thus, irreducible to the lawful physical realm. Now Kim's model of reductionism is not based on the existence of universal bridge laws. In Kim's model, as he puts it, "the reducibility of a property critically depends on its functionalizability -- whether or not it can be construed as a second-order functional property over properties in the base domain -- not on the availability of bridge laws. Bridge laws are neither necessary nor sufficient for reduction." [Kim 1998b, p. 27] That is to say, as long as we can define a mental property -- the property of having a pain, using Kim's favorite example -- in terms of its sensory inputs (having one's finger pinched) and its behavioral outputs (wincing or saying 'ouch'), we can reduce this mental property to its physical base property. Because the same mental property could be multiply realized by different physical properties in different biological species, the reduction is generally not a one-to-one reduction but a one-to-many reduction. We can perhaps summarize the difference between the two models of reductionism in this way: on the Nagelian model, between mental properties and physical properties it is expected to have biconditional laws in the form of M iff P (where P is a single physical or neurophysiological property); on Kim's model, it is acceptable to have M iff {P1, P2,...., Pn} where P1 ≠ P2 ≠ Pn.

Kim's model of reductionism is an improvement over the standard model. It is not subject to the multiple realizability argument that is typically used against the standard model of reductionism; it is actually based on multiple realizability. Kim's model also changes the focus of debate between reductionists and anti-reductionists. The issue is no longer whether mental properties are multiply realizable or whether mental properties are anomalous, both of which have been used to defend nonreductionism. Kim acknowledges that they are. This issue now is whether mental properties are functionalizable. Kim thinks that the only kind of mental properties that are (possibly) not functionalizable are qualia or what Ned Block calls phenomenal consciousness.[Kim 1999, pp. 9-11] But under Kim's view, if these properties are not functionalizable, then they are not reducible; if they are not reducible, then, by his Exclusion Principle, they are causally inert and irrelevant. Finally, if they are causally inert, then by Alexander's Dictum, they are not real. Therefore, Kim concludes, the only real mental properties are the ones that are reducible to physical properties.

However, what most nonreductionists are arguing against is only intertheoretic reductionism, not ontological reductionism between the mental and the physical.[13] That is to say, the debate between the reductionists and the nonreductionists is not on the ontology of the mental, but on the explanatory power of general physics. Intertheoretic reduction may not be reliant on the existence of singular bridge laws; it nonetheless has to imply that the reduced theory can be fully and completely explained in terms of the base theory. Thus, even if Kim is right that to establish intertheoretic reductionism, he does not need to demonstrate that there is an exact one-to-one match between psychological kind predicates and physical kind predicates, he still needs to demonstrate that all psychological kind predicates can be fully and completely analyzed in some physical kind predicates. I argue that Kim's argument does not give us this latter demonstration. In his example, the functionalizability of the property of pain is specified in terms of its sensory inputs and behavioral outputs, and it seems possible that the descriptions of sensory inputs and behavioral outputs could be based on the vocabulary of physics (or neurophysiology) alone. Kim calls the states or properties that satisfy this causal specification (of inputs and outputs) "the 'realizers' of the mental property involved." [Kim 1999, p. 10] Thus, once the mental property can be functionalized, it has found its physical realizers. But the success of this example may have given Kim the false hope that all mental properties can be thus functionalized in physical states or properties alone. Take the property of having a propositional attitude such as 'I believe that Mr. Wang is a nice guy' for example. Even though this property could be functionalizable, its functional specification has to involve other mental properties, which again need to be further analyzed. A single mental property could involve a whole web of mental properties, each of which needs to have its causal specification if it is to be functionalized. But not all of the causal specification can involve only sensory inputs and behavior outputs, or use only physical vocabulary. The functional/causal analysis may run in circles, and it is very likely that we may ultimately not be able to use only physical vocabulary to specify the function of a single mental property. Even if all mental properties are indeed physically realizable, it does not mean that we can thus specify them in purely physical terms. In other words, Kim has not established the move in his argument from the functionalizability of mental properties to the theoretical reducibility of mental properties; nor has he shown the reducibility of psychology to neuroscience. I thus conclude that Kim's model can only guarantee us ontological reduction; not intertheoretic reduction.

Kim thinks that the advantage his MTP has over Davidson’s token physicalism is that it further explains how mental types are related to physical types. If one rejects the reduction thesis, then, so argues Kim, one has to come up with an explanation that satisfies the following conditions:

(i) it does not violate the physical closure principle;

(ii) it dissipates the causal competition between physical properties and mental properties;

(iii) it explains psychophysical type-type relations.

What Davidson’s anomalous monism fails to provide, according to Kim, is an explanation that satisfies the third condition. In the remainder of this paper, I will try to delineate a relationship between mental and physical properties such that it does not deny that the physical domain is causally closed, and yet it also resolves the causal competition between mental and physical properties. As a nonreductionist, however, I will also want to show that mental properties are real, nonreducible properties and that they are also causally relevant.

§3. A Nonreductionist's Solution: Holistic Identity

To resolve the tension between two competing causal explanations, using Kim's words, "we need to know how the two explanations are related, how the two causal stories about a single phenomenon mesh with each other. Are the two stories at bottom one story couched in different languages? Do the two stories supplement one another, each being only partial? And so on." [Kim 1998b, p. 66] In my analysis, the two causal stories would each be a sufficient and complete causal story, but there is no causal competition because they would turn out to be one story couched in different languages. In a nutshell, my solution is that of identity. But I shall argue that this kind of identity does not lead one to intertheoretic reductionism.

The best way to characterize the relationship between mental properties and physical properties is to use some version of supervenience. Kim argues that weak supervenience is too weak to be a determination relationship, and strong supervenience leads to reduction. We (the nonreductionists) seem to be in a bind here: either we give up the claim that the physical determine the mental, or we embrace the claim that the mental is reducible to the physical. Here I shall sketch an approach that does not deny strong supervenience and yet also does not accept the possibility of psychophysical reduction. I think the key is to specify which mental properties and which physical properties are involved in our formulation of psychophysical supervenience.

With Kim’s psychophysical supervenience, we have this relationship:

[PS] Every internal psychological state of an organism is supervenient on its synchronous internal physical state. [Kim 1982, p. 183][14]

At the same time, with Kim’s strong supervenience, we have the following relationship:

[SS] A strongly supervenes on B with respect to domain D just in case for any x and y in D, and any worlds w and w*, if x in w is B-indiscernible from y in w*, then x in w is A-indiscernible from y in w*. [Kim 1988, p. 111]

What [PS] and [SS] state is that if two organisms x and y are “indiscernible” in their internal (and external or relational, under some interpretation) physical states, then they share the same psychological states. Now let us take a close look at the nature of these states.

The mental states Kim considers are all post-instantiation mental states; that is, they are the actual thoughts, beliefs, desires,...that some individual or another has had. For such an individual, the internal physical states that process these thoughts, beliefs are what Kim calls the “synchronous” internal (and external/relational) physical state. This supervenience relation is based on Davidson's token-identity theory. Davidson's token-identity theory states that singular mental events are identical with singular physical events, but there is no matching between mental event types and physical event types. Davidson proposes that we define 'event' as an individual happening that is spatially and temporally confined. Events are thus unrepeatable, dated individuals. Davidson thinks that what we need for asserting event-identity is a way to fill in the blank in "If x and y are events, then x = y if and only if _______". He gives the following condition: "Many events are changes in a substance. If an event a is a change in some substance, then a = b only if b is also a change in the same substance." [Davidson 1970a, p. 173] Thus, a mental event m can be identical to a neurophysiological event n only if both m and n occur in the same individual person. Having the spatial location of both events confined to the same person constitute a necessary condition, but not a sufficient condition, of the mental-physical event-identity. The second necessary condition needed is a temporal specification: "if events are identical, they constitute identical stretches of time." [Davidson 1970a, p. 177] But while spatial and temporal demarcations are useful for the individuation of events, they are insufficient in establishing the identity claim. What is needed to have an identity relation, according to Davidson, is an identity of the causal connection. Davidson's final thesis of token-identity can be summarized as this: A mental event and a physical event are identical if and only if they happen to the same person at the same time, and they have exactly the same causes and effects.

However, under Davidson's token-identity thesis, there is still a problem for the individuation of neurophysiological events: even if we could grant that the spatiality of a neurophysiological event is easily defined as the individual's brain (or the smallest part of the brain where the change takes place), how do we dissect the ongoing processes in the brain temporally? Presumably, an neurophysiological event can be specified as whatever happens in the brain within an hour, a minute, a second, or even a millionth of a second. Which is the event we want to pick out as a neurophysiological token? It would seem that when we try to locate the neurophysiological token, what we are saying is that the neurophysiological correlate of one's having the belief that P at time t1 is a relevant token, or that the neurophysiological correlate of one's having a pain experience at time t2 is another relevant token. But when neurological events are specified this way, there are two problems. One is that the demarcation is still uncertain. For example, if the belief in question is a dispositional belief state not being presently contemplated on, what is its relevant neurophysiological correlate? Secondly, the individuation of neurophysiological events is already based on a token-identity, and thus the individuation itself loses its explanatory status.

To avoid this trouble with Davidson's token-identity theory, I suggest that we abandon talk about "physical tokens" and use a holistic picture instead. Holism, to begin with, is the principle that a part is understandable only in its relation to the whole. In other words, no token can be locally, atomically defined. Therefore, if we want to understand each single token, we have to consider it not as a simple element, but as a part of the whole system. Two tokens of different wholes cannot be compared out of their respective systems since the tokens themselves are individuated only by the roles they play in each system. We can state this thesis as follows:[15]

[Holism] The individuation of x is determined by x's contribution to the whole system.

Under this thesis, a neurophysiological event is individuated by how it occurs in the whole brain of a particular individual and what effects it brings to that brain. A mental state such as a belief that P, on the other hand, is individuated by what else the person believes (in particular, her metalinguistic beliefs concerning the meaning of words involved in the specification of P). Mental holism has been widely discussed and has gained some endorsement. What I think has been neglected is the acknowledgment of "neural holism." For two individuals, x and y, to be indiscernible in their physical states, the physical state of the brain must be viewed not merely as one particular activity in the brain, but as a cross-section of the brain. For example, even when we suppose that there is an identifiable neural activity that corresponds to an individual’s desire for coffee, the same neural activity occurring in person A may lead A to drink coffee while in person B it may not have the same effect. It is not just the neuron firing considered in isolation, but the neuron firing in that particular neural network, that plays a deterministic causal role in the person’s action. We must therefore consider the whole neurological state of the brain, or we may call it the “neural map,” in our formulation of the determinate causal history.[16]

A holistic specification of events is similar to, but not restricted to, a "causal-role" specification. The difference can be presented in these two respects: Firstly, it is not restricted to analyzing the immediate cause and effect of this particular event, but is looking at the whole history of the brain. Secondly, it allows us to consider the neuronal membrane at rest, or a dispositional mental state not presently affecting the individual's behavior.

Based on mental holism and neural holism, I propose that the mind/brain relationship be analyzed in terms of what I call holistic identity:

[Holistic Identity] Let A be a psychological property holistically specified, let B be a neurophysiological property holistically specified; let x be any object,

A is identical to B if and only if

if x has A, it necessarily has B, and vice versa.

Identity, as construed in type-identity theories or token-identity theories, is not a notion about object identity, but about property identity. As Joseph Levine puts it, questions about property identity is "questions about how or why distinct properties are coinstantiated." [Levine 1998, p. 465] Type-identity theories specify the coinstantiated properties as psychological proper types and physical proper types. Token-identity theories specify the coinstantiated properties as properties of a spatially/temporally confined token/event. When properties are specified holistically, they are properties of tokens considered in relation to a whole system to which the token belongs. By Fodor's definition, "Holistic properties are properties such that, if anything has them, then lots of other things must have them too."[17] Under this definition, the mind-brain identity relation holds only between one's mental state, considered in relation to one's whole mental life, and one's neurophysiological state, considered in relation to one's whole brain. Furthermore, when we consider neurophysiological events as tokens of a brain, the brain itself is not to be taken as a static whole. Since the brain is an organic whole, we cannot just demarcate its events temporally.[18] Each synchronic picture (each time-slice) of the brain can only be viewed as a frame in the whole dynamic system. In this kind of holistic picture, a part is related to the whole not as a unit in a well-constructed, interrelated "web," but as a unit of a stage in an ongoing process. A mental state such as a belief that P, on the other hand, is individuated by what else the person believes (in particular, her metalinguistic beliefs concerning the meaning of words involved in the specification of P), and what else the person desires, etc. These psychological characteristics of the individual also reflects the past history and personality traits of the individual, just as the individual's present brain state reflects what stimuli and inputs he or she acquired in the past. It is therefore practically impossible to find psychological twins that are not neurophysiological twins, and vice versa.[19]

Once the identity is established, we can resolve the exclusion problem. Following Davidson, I argue that mental events and physical events do not compete for causal power, since they are identical events. And under holistic identity, mental properties and physical properties also do not compete for explanatory relevance, since they are coinstantiated properties. But I want to emphasize that it is only the whole mental life of a person taken at a particular time-slice that has the determining causal power over what the person would think at the next time-slice. In the same way, it is only the whole neural map of a brain at the particular time-slice that has the determining causal power over the next neural map of the brain at the next time-slice. We can formulate the two causal explanations in this way:

Supposing that neurophysiological state N1 leads to N2, we have a strict neurophysiological causal law:

[NL] ‘If x is in N1, then x will be in N2’ or ‘N1 causes N2’.

Now suppose N1 is the neurophysiological base of M1, and N2 is the neurophysiological base of M2 . I argue that the following causal law also holds:

[ML] ‘If x is in M1, then x will be in M2’ or ‘M1 causes M2’.

Both of these causal explanations give the complete causal stories and they are simply the same causal history given different descriptions. My picture of the mental/physical kind relationship is the following ('®' stands for causation, while '½½' stands for identity):

M1 ® M2

½½ ½½

N1 ® N2

Using the above identity relation, we can also formulate a strong supervenient relationship as follows:

[SS] ‘If x is in N1 and x is in M1, when M1 is a mental kind, then necessarily for all ys, if y is in N1, then y is in M1.’

In other words, the mental is determined by the physical and is thus not an "ontological dangler."

Given this holistic identity relationship, the whole mental life of an individual is locally reducible to the whole brain extrinsically, relationally considered.[20] By local reduction, I do not mean species-specific reduction, as some (including Kim himself) take it to mean. What I mean is a reduction restricted locally to one particular person. Kim thinks that we can have species-specific psychophysical identity and this identity serves as the foundation for species-specific psychophysical reductionism. But local reduction between an individual's mental life and her brain states does not warrant species-wide reduction between general mental and physical properties. I argue that such a species-wide identity is not possible; furthermore, I argue that this local reduction is actually a disproof of intertheoretic reductionism.

To begin with, neither the mental kind M1 nor the neurophysiological kind N1 involved in this supervenience can be the proper kinds of psychology and neurophysiology per se. Kim calls the description of all the neurophysiological properties of this cross-section of the brain “the N-maximal properties.” The neurophysiological kind N1 is thus the set of all the neurophysiological properties of the brain. As we know, the neural map in our brains differs from one person to the next; it also changes constantly within the same person. There are about ten billion neurons in the cerebral cortex alone, and among these neurons, there are about one million billion neural connections in the cortical sheet. "If we consider how connections might be variously combined, the number would be hyperastronomical -- on the order of ten followed by millions of zeros…. And even more remarkably is the way in which brain cells are arranged in functioning patterns. When this exquisite arrangement of cells is taken together with the number of cells in an object the size of your brain, and when one considers the chemical reactions going on inside, one is talking about the most complicated material object in the known universe." [Edelman 1992, p. 17] Therefore, no two neural states will be exactly alike in reality, even though we could find “molecule-for-molecule” identical brains in theory. This being the reality of the brain, there would be infinitely many neurophysiological kinds from N1 to Nn. It is clear that neuroscience cannot use N1 to Nn as its proper kinds. My speculation is that neurophysiology can only give a regional measurement or a coarse-grained description of each region's causal responsibility. We can call them macro-neurophysiological kinds. In contrast, we can call N1 to Nn micro-neurophysiological kinds. With micro-neurophysiological properties, there is even a problem with measurement: No neuroscientist can ever claim to give a precise measurement of all the neural activities of every neuron in the brain. If macro-neurophysiological kinds cannot be identified with micro-neurophysiological kinds, then local physical realization does not warrant a species-wide unified physical realization. Ultimately, what neurophysiology taxonomizes as kinds are not what we are describing as N1 to Nn -- the fine-grained causally determining kinds.

In the same way, the mental kind that supervenes on this micro-neurophysiological kind is not a kind that we have in the taxonomy of psychology per se, but a much more finer-grained taxonomy of what I will call 'individual-specific psychology,' or 'IPsy.' IPsy has predicates of the form ‘x has the mental state that x believes that p and x desires that q and that.... (which exhausts all of x’s mental inclinations that are relevant to the decision).’ For example, both x and y love crabs and they are both staying at a hotel which is fourteen blocks away from a famous Crab House. A general intentional psychological explanation would be giving the following generalization: 'If x loves crabs and x believes that at the Crab House x can get crabs, then x would go to the Crab House.' It is clear that such a generalization cannot be strict: there are too many exceptions. But now if we specify all the mental inclinations that come into play: such as, both x and y hate walking fourteen blocks in the evening; both x and y think about taking a taxi to go to the Crab House; however, x does not mind taking a taxi while y has a deep fear about riding in a taxi by herself, etc., then we would put x's and y's mental states prior to their decisions into different IPsy -kinds. We can also explain why x ends up going to the Crab House while y ends up not going.

Each IPsy-kind includes the complete description of the total instantaneous states of the individual -- states which are described as the total specification of the present condition of that individual. Such a mental state determines, together with learning and memory, what the next state will be. This taxonomy gives a holistic consideration of one's mental life, taking the present mental life as the end-state of previous learning experiences as well as memory. This IPsy taxonomy is of course not very useful to us, since we cannot generalize from one person to the next. But theoretically, any other individual who shares the IPsy-state with x would just be taking the same action as x would.

Now I argue that the M1 and M2 that correspond to the N1 and N2 in the neurological causal law [NL]: N1 causes N2, are IPsy -kinds. And given this understanding, we can see that both [NL] and [ML] can be strict laws, and yet neither has any generalizability in practice. Any actual neurophysiological law that is to be strict can have only one incident, involving only one neural map of a particular brain at a time. We may call it a token-specific law. Similarly, psychological laws can also be strict -- as long as they are specified in terms of IPsy-kinds and are token-specific. In other words, if the neurological state (specified by a set of N-maximal properties) prior to an individual's taking an action is causally sufficient in bringing about this action, then the mental state (specified by a set of IPsy-maximal properties) corresponding to this neurological state is equally causally sufficient in bringing about the action. If there is a strict physical law linking the neurological state to the action, so is there a strict mental law linking this individual-specific psychological state to the action. On the token-specific, post-instantiation level, the two sets of "maximal properties" are identical and are mutually reducible. With the issue of causal exclusion, I thus conclude that there is no causal competition between IPsy-maximal properties and N-maximal properties. If it is true that (under the supervenience thesis), any y that is in N1 must be in IPsy1, and any y that is in N2 must be in IPsy2, then when there is a causal strictness between N1 and N2, there must be a causal strictness between IPsy1 and IPsy2.

But acknowledging this identification would not lead us to embrace the reduction of general psychology to general neurophysiology. We shall now establish the relationship between general psychology and general neurophysiology on the basis of the relationship between individual-specific psychology and micro neurophysiology. To begin with, the connection between IPsy and general psychology is more easily established: Psychology could be viewed as a generalization of these individual-specific psychologies. Psychological taxonomy is not based on the grouping of the whole mental life of every individual, but on the grouping of what different mental lives share in common. Now let us consider psychology per se. Psychology (Psy) and IPsy are two sciences at different levels of structure and organization. Psy is a generalization based on IPsy, or we should say, enough cases of IPsy. So, for each Psy-type, we would have numerous IPsy-types. This Psy -type is the result of generalization of numerous IPsy-types, and guided by our purpose or our interest in study, we select some common aspects of the IPsy-types as relevant in our individuation of Psy-types:

Psy i = {IPsy 1 v IPsy 2 v IPsy 3 ......IPsy n}

After the generalization process, the causally determining power is lost. When the mental state at issue is formulated as a Psy-type, a psychological law using it as the causal antecedent must add a “ceteris paribus” clause:

[PL] If Psy i, then actionj (or another mental state characterizable as Psyj), ceteris paribus.

Some of the psychological factors included in the ceteris paribus clause may actually be a determinate "countervailing" factor in some people, even though in most people they may be insignificant. This explains why general psychological laws are not strict.

The connection between each Psy-type and IPsy-type is explained as above. But on the other hand, the connection between the each micro neurophysiological type and general neurophysiological type is not so clear. We can show how different IPsy-types {IPsy 1 v IPsy 2 v IPsy 3 ......IPsy n} could be generalized into a psychological type Psyi, because we can use the same mental predicates in both taxonomies. But we have no evidence that there is any generalizability between N1 and N2, ... and Nn. How neurophysiology generates types is a more complicated issue because variations in single neuron firings, in neural patterns, or in neurotransmitter systems, etc. could be infinitely great. I suspect that neuroscience can only describe regional responsibility coarsely defined. That is to say, it may be in principle possible for neuroscientists to tell us what part in the brain is responsible for language, for sensation, etc., but it may never be possible for them to give us a complete description of the N-maximal properties that underlie a person's every mental activity however minutely individuated. From how vastly different these individual neurophysiological bases underlying the same IPsy-kind can be, we simply do not see the possibility of their being generalized into one neurophysiological kind.

We can now see why the type-type identification between general mental properties and physical properties is lost. Each psychological type has an indefinite, or even infinite, number of actual and potential neurophysiological types as its physical bases. This is not just the problem of multiple realizability. The problem is rather that in our bottom-up generalization from individual-specific psychology into general psychology, many mental details are left out for the sake of grouping. But it is the detailed, fine-grained, micro properties that have one-to-one psychophysical correspondence. Furthermore, once properties are specified holistically, they are confined in the system to which they belong. On the level of mental properties, a particular mental property is specified in relation to other mental properties. On the level of neurophysiological properties, a particular neurophysiological property is specified in relation to other neurophysiological properties. Each level has its own web of interconnections and there is little evidence that we can match the psychological map and the neural map exactly. Therefore, the local reduction that I favor is at most a property-reduction between mental properties (specified individual-psychologically) and micro-neurophysiological properties (specified as a complete description of the whole neural network of one individual). It is not a theory-reduction between psychology and neurophysiology. That there are neurophysiological state does not imply that there are neurophysiologically taxomizable states, just like the fact that there are psychological states does not mean that there are psychologically taxomizable states. The mental properties that are reducible to neurophysiological properties do not constitute a (general) psychological kind. The neurophysiological bases of this mental kind also do not constitute a (general) neurophysiological kind. Intertheoretic reduction between psychology and neurophysiology is not feasible, but this does not entail that the mental realm does not have the physical realm as its base. What I deny is merely the reductionism between two special sciences, not the reduction between two metaphysical realms.

If general psychology is not reducible to general neurophysiology, then the explanatory competition between the two theories still exists. So we are back to the explanatory exclusion problem that Kim advocates. My solution is twofold. First of all, Kim's explanatory exclusion problem is based on his Causal Explanatory Principle [CEP], but the kind of causal explanation that Kim demands through this principle can only be offered either an individual-specific psychological explanation, or as a micro-neurophysiological explanation. These explanations are what I formulated as [ML] and [NL] respectively. Therefore, on the local, individual-specific level, the two explanations do not compete. Secondly, on the level of general psychological explanation and general neurophysiological explanation (applied species-wide to human beings), Kim's [CEP] does not apply since neither theory can offer strict causal laws. We therefore do not have to choose one theory over the other since one general explanation does not necessarily exclude the other. Furthermore, I suspect that neuroscience can never offer any useful explanation for human behavior without appealing to some intentional psychological explanation. To have a neurophysiological measurement of the causal history of a behavior, we need to be able to interpret the results of our measurement. And the interpretation of a neurophysiological measurement has to rely on psychological interpretation of the behavior. In order for neuroscientists to find regional responsibility, they would also have to measure (with whatever new techniques they will produce in the future) the neural action states and rest states corresponding to a known mental phenomenon. In other words, neuroscientists do not have a way to defining the relevant neural correlate if they did not already assume a psycho-physical identity. The explanatory power of neuroscience will ultimately rest on the ontological model of psycho-physical identity. Mental properties do not become causally/explanatorily irrelevant either when they are specified relative to each individual, or when they are specified as general psychological properties. If so, then there is another way out of the explanatory exclusion: Granting that psychological explanation and neurophysiological explanation excludes each other, we see that what should be excluded is the neurophysiological explanation, not the psychological explanation.

§4. Conclusion

In this paper I have tried to argue against Kim's view that the only solution to the exclusion problem is that of reductionism. I have demonstrated a way to link the mental with the physical such that the mental is not causally inept, and at the same time psychology is not made to be reducible to neurophysiology. I want to emphasize again that my position is for the ontological reduction from the mental to the physical. But to establish intertheoretic reduction between psychology and neuroscience, we need to have the incentive that the latter theory will be able to give a simpler and yet fuller explanation than the former does. Putnam says, "It seems to me that whatever the pragmatic constraints on explanation may or may not be, one constraint is surely this: That the relevant features of a situation should be brought out by an explanation and not buried in a mass of irrelevant information." [Putnam 1973, p. 132] With the massive details on the micro-neurophysical level, neuroscience alone will not be able to give us the information that we are interested in seeking on the level of intentional psychology.

Finally, I want to point out that in this paper I am not just offering a nonreductionist's alternative to Kim's reductionistic solution, but a rejection of his reductionism. Kim's argument for reduction is based on the existence of biconditional correlation between a mental property M and a disjunctive set of M’s physical bases, N1 or N2 ...or Nn, such that M « N1 Ú N2 Ú... Ú Nn. A major problem with this biconditional is that, the disjunctive set of physical bases {N1 Ú N2 Ú... Ú Nn} is too large to be made use of in offering causal explanations. The problem is not just the multiple realizability among different species or different biological/mechanical structures. As Kim sees it, we can be just talking about species-specific reduction. The problem is rather that even within the same species such as human beings, and even within the same individual, the variations of physical instantiations of a mental whole are infinitely many. As Terence Horgan puts it, “For all we now know,... the intentional mental states we attribute to one another might turn out to be radically multiply realizable at the neurobiological level of description, even in humans; ... indeed, even in an individual human given the structure of his central nervous system at a single moment of his life. A single person S at a single moment t might be capable of undergoing (at t ) any of a variety of different physico-chemical states which are quite disparate from one another at the physico-chemical level of description,... each of which would realize (in S) the same intentional mental property.” [Horgan 1993, p. 308] Since the kind of deterministic mental state is a conjunctive set of all the individual's mental attributes relevant to the causation of her action, the physical base of this mental kind must itself be a conjunctive set of all the micro-physiological properties of the individual's brain. Given the variations of each individual's brains in their minutest details, the disjunctive set of all the possible physical bases is inexhaustible. Furthermore, given that the physical bases are bases of “instantiated” mental states, they could even compose an open infinite set since the mental states at issue could still be instantiated at any time to any individual in the future. The same criticism Putnam has toward the Turing machine states theory can be used against the neurophysiological states theory: "Besides the fact that such a description would be literally infinite, the theory is now without content..." [Putnam 1973, p. 137]

References

1. Louise Antony (1989), 'Anomalous Monism and the Problem of Explanatory Force,' The Philosophical Review, Vol. XCVIII, No. 2, pp. 153-187.

2. ____________ (1995), 'I'm a Mother, I Worry,' Philosophical Issues, 6, Content (ed.) Enrique Villanueva, pp. 160-166.

3. ______ & Joseph Levine (1997), 'Reduction with Autonomy,' Philosophical Perspective 11: Mind, Causation and World, pp. 83-105.

4. Ned Block (1990), 'Can the Mind Change the World?' in Meaning and Method: Essays in honor of Hilary Putnam, George Boolos (ed.) [Cambridge: Cambridge Univ. Press] pp. 137-170.

5. __________ (1997), 'Anti-Reductionism Slaps Back,' [manuscript].

6. Tyler Burge (1993), ‘Mind-Body Causation and Explanatory Practice’, in Mental Causation, (eds.) John Heil & Alfred Mele, [Oxford: Clarendon Press], pp. 97 - 120.

7. Manuel Campos (1995), 'Kim on the Exclusion Problem,' Philosophical Issues, 6, Content (ed.) Enrique Villanueva, pp. 167-170.

8. Nancy Cartwright (1983), How the Laws of Physics Lie, [New York: Oxford University Press].

9. Donald Davidson (1967), 'Causal Relations,' reprinted in [Davidson 1980], pp. 149-162.

10. ______________ (1970a): 'Events as Particulars,' reprinted in Davidson: Actions & Events, [Oxford, Clarendon Press], pp. 181-187.

11. ______________ (1970b): 'Mental Events,' reprinted in Davidson: Actions & Events, [Oxford, Clarendon Press], pp. 207-227.

12. _____________ (1975), 'Causal Relations,' in Causation and Conditionals, (ed.) Ernest Sosa, Oxford Univ. Press, London, pp. 82-94.

13. _____________ (1976), 'Hempel on Explaining Action,' reprinted in [Davidson 1980], pp. 261-276.

14. _____________ (1980), Essays on Actions & Events, [New York: Clarendon Press].

15. _____________ (1993), 'Thinking Causes,' in [Heil & Mele, 1993], pp. 3 - 17.

16. Marian David (1997), 'Kim's Functionalism,' Philosophical Perspective 11: Mind, Causation and World, pp. 133-148.

17. Fred Dretske (1989), 'Reasons and Causes,' in Philosophical Perspectives, 3, Philosophy of Mind and Action Theory, pp. 1-15.

18. ___________ (1995), Explaining Behavior Reasons in a World of Causes, [Cambridge: The MIT Press], fourth edition.

19. Gerald M. Edelman (1992), Bright Air, Brilliant Fire: On the Matter of the Mind, [BasicBooks, A Division of Harper/Collins Publishers, Inc.].

20. Jerry Fodor (1974), 'Special Sciences,' Reprinted in Fodor, Representations: Philosophical Essays on the Foundations of Cognitive Science, [Cambridge, Massachusetts; The MIT Press], 1983, pp. 127-145.

21. __________ (1997), 'Special Sciences: Still Autonomous After All These Years,' Philosophical Perspective 11: Mind, Causation and World, pp. 149-163.

22. John Heil & Alfred Mele (eds.) (1993), Mental Causation, [Oxford: Clarendon Press].

23. Terence Horgan (1989), 'Mental Quausation,' Philosophical Perspective, 3, Philosophy of Mind and Action Theory, pp. 47-76.

24. _____________ (1993), ‘Nonreductive Materialism and the Explanatory Autonomy of Psychology’, in Naturalism: a Critical Appraisal, (eds.) Steven Wagner & Richard Warner, [Notre Dame: University of Notre Dame Press].

25. _____________ (1997), 'Kim on Mental Causation and Causal Exclusion,' Philosophical Perspective 11: Mind, Causation and World, pp. 165-184.

26. Frank Jackson & Philip Pettit (1990), 'Program Explanation: A General Perspective,' Analysis, Vol. 50, pp. 107-117.

27. ____________ (1990b) 'Causation in the Philosophy of Mind,' Philosophy and Phenomenological Research, vol. L, Supplement, pp. 195-214.

28. Jaegwon Kim (1973a), 'Causation, Nomic Subsumption, and the Concept of Event,' Journal of Philosophy, vol. LXX, No. 8, pp. 217-236.

29. ___________ (1973b), 'Causes and Counterfactuals,' Journal of Philosophy, vol. LXX, pp. 570-572.

30. ___________ (1979), 'Causality, Identity, and Supervenience in the Mind-Body Problem,' Midwest Studies in Philosophy, Vol. 4, pp. 31-49.

31. ____________ (1982), 'Psychophysical Supervenience,' in Kim (1993a), pp. 175 - 193.

32. ____________ (1983), 'Supervenience and Supervenient Causation,' Southern Journal of Phil., supp. Vol. 22, pp. 45-55.

33. ____________ (1984), 'Epiphenomenal and Supervenient Causation,' reprinted in [Kim 1993a], pp. 92-108.

34. ___________ (1987a), ‘Explanatory Realism, Causal Realism, and Explanatory Exclusion’, in Midwest Studies in Philosophy, XII, ed. by P. French, et al, [Notre Dame: University of Notre Dame Press], pp. 225-239.

35. ___________ (1987b), '"Strong" and "Global" Supervenience Revisited,' reprinted in [Kim 1993a], pp. 79-91.

36. ___________ (1988), 'Supervenience for Multiple Domains,' reprinted in [Kim 1993a], pp. 109-130.

37. ___________ (1989a), ‘Mechanism, Purpose, and Explanatory Exclusion’, reprinted in [Kim, 1993a], pp. 237-264.

38. ___________ (1989b), ‘The Myth of Nonreductive Materialism’, reprinted in [Kim, 1993a], pp. 265-284.

39. ___________ (1990), ‘Explanatory Exclusion and the Problem of Mental Causation’, reprinted in Philosophy of Psychology: Debates on Psychological Explanation, ed. by Cynthia MacDonald, [Cambridge: Blackwell], 1995, pp. 121-141.

40. ___________ (1991), 'Dretske on How Reasons Explain Behavior,' reprinted in [Kim 1993a], pp. 285-308.

41. ___________ (1992), '"Downward Causation" in Emergentism and Nonreductive Physicalism,' In Emergence or Reductionism?, ed. A. Beckermann, H. Flohr, and J. Kim, [Berlin: De Gruyter], pp. 119-138.

42. ___________ (1993a), Supervenience and Mind: Selected Philosophical Essays, [Cambridge: Cambridge University Press].

43. ___________ (1993b), ‘The Nonreductivist’s Troubles with Mental Causation’, reprinted in [Kim, 1993a], pp. 336-357.

44. ___________ (1993c), ‘Postscripts on Mental Causation’, reprinted in [Kim, 1993a], pp. 358-367.

45. ___________ (1993d), ‘Can Supervenience and “Non-strict Laws” Save Anomalous Monism?’, in Mental Causation, (eds.) John Heil & Alfred Mele, [Oxford: Clarendon Press], 1993, pp. 19-26.

46. ___________ (1994), 'Explanatory Knowledge and Metaphysical Dependence,' Philosophical Issues, Vol. 5, 1994.

47. ___________ (1995), ‘Mental Causation: What? Me Worry?’ in Content (Philosophical Issues), No. 6,(ed.) Enrique Villanueva, [Atascadero, California: Ridgeview Publishing Company], pp. 123-151.

48. ___________ (1996), Philosophy of Mind, [Boulder: Westview Press].

49. ___________ (1998a)), 'The Mind-Body Problem After Fifty Years,' in Current Issues in Philosophy of Mind, (ed.) Anthony O'Hear, [Cambridge, England; Cambridge University Press], pp. 3-21.

50. ___________ (1998b), Mind in a Physical World: An Essay on the Mind-Body Problem and Mental Causation, [Cambridge, Massachusetts; The MIT Press].

51. ___________ (1999), 'Mental Causation and Consciousness: The Two Mind-body Problems for the Physicalists,' [manuscript].

52. Joseph Levine (1998), 'Conceivability and the Metaphysics of Mind,' Nous 32:4, pp. 449-480.

53. David Lewis (1986), 'Causation,' in David Lewis, Philosophical Papers, vol. II, Oxford: Oxford University Press, 1987, pp. 159-173.

54. Andrew Melnyk (1994), 'Being A Physicalist: How and (More Importantly) Why,' Philosophical Studies 74, pp. 221-241.

55. _____________ (1995), 'Two Cheers for Reductionism: Or, the Dim Prospect for Non-Reductive Materialism,' Philosophy of Science 62, pp. 370-388.

56. Hilary Putnam (1973), 'Reductionism and the Nature of Psychology,' Cognition 2(1), pp. 131-146.

57. Marcelo H. Sabatés (1996a), 'Kim on the Metaphysics of Explanation,' Manuscrito, Vol. XIX, No. 2, pp. 93-110.

58. ___________ (1996b) 'Should A Cognitive Psychologist Worry About the Causal Inefficacy of the Mental?' (Manuscript)

59. Searle, John (1992): 'Reductionism and the Irreducibility of Consciousness,' from The Rediscovery of the Mind, Chap. 5, reprinted in The Nature of Consciousness, (eds.) Ned Block, Owen Flanagan, and Güven Güzeldere, (Cambridge, Massachusetts: The MIT Press), pp. 451-459.

60. Ernest Sosa (1984), 'Mind-Body Interaction and Supervenient Causation,' in P.A. French et al., eds., Midwest Studies in Philosophy, Vol. IX. Minneapolis: University of Minnesota Press, 1984, pp. 271-281.

61. Paul Teller (1983), 'Comments on Kim's Paper,' Southern Journal of Phil., supp. Vol. 22, pp. 57-61.

62. James Tomberlin (1995), 'Mental Causation: A Query for Kim,' in Philosophical Issues, 6, Content (ed.) Enrique Villanueva, pp. 152-159.

63. Stephen Yablo (1992), 'Mental Causation,' The Philosophical Review, Vol. 101, No. 2, pp. 245-280.

64. ____________ (1997), 'Wide Causation,' Philosophical Perspective 11: Mind, Causation and World, pp. 251-281.

Notes