1. The Landscape of the Discussion

The exclusion problem is the incompatibility among the following individually plausible principles.¹

• Irreducibility: Mental properties are not identical with physical properties.

• Kimian Notion of Events: An event is analyzed as <$x,P,t$>, which means that the object $x$ instantiates property $P$ at time $t$. (Kim 1973; 1976)²

• Mental Causation: Mental events sometimes cause other events.

• Physical Causal Closure: Any event which has a cause at a given time has a sufficient physical cause at that time.

• Exclusion: Besides cases of genuine over-determination, a given effect has only one sufficient cause at a given time.

• Non-Over-Determination: There is no systematic over-determination in cases of mental causation.

The reason to support each of the principles is now quite familiar, so I will not go into the details here. Put together, these principles are incompatible. Reductive physicalists (e.g., Kim) see the problem as a reductio ad absurdum against non-reductive physicalism: other principles (especially Mental Causation) are too valuable to give up, so we had better give up Irreducibility to solve the embarrassing incompatibility.

The trope identity solution claims that none of the principles needs to be abandoned since the incompatibility is just an appearance. So, the proponents of the solution believe that we can retain both Mental Causation and Irreducibility (Ehring 1999; 2003; 2011; Heil & Robb 2003; Robb 1997; 2013; 2017; Tiehen 2019). According to them, if we convert from the universalist picture to the trope picture, the incompatibility will go away. Roughly, they believe that the term “property” in the principles of the exclusion problem is ambiguous between tropes and types (classes of tropes), and the conflation of these two readings leads to the incompatibility. The right reading, they say, should be that in Irreducibility the word “property” refers to types while in all the others it refers to tropes. That is, it is tropes, rather than universals, that constitute events.³ Mental types are not identical with physical types because mental types are “inexactly similar types” in which the tropes are only similar to each other in some (e.g., sharing the same causal role), but not all, aspects; but physical types are “exactly similar types” in which tropes are exactly similar to each other. To guarantee causal powers of mental events and to avoid the problematic over-determination, the trope solution proposes that every mental trope is identical with a physical trope. That is, every trope that falls in a mental type also falls in a physical type. Then a mental event constituted by a mental trope is identical with a physical event constituted by a physical trope since the mental trope just is the physical trope (also, the object and the time are identical). So, there will be no worry of exclusion. In this way, trope theorists claim that they non-reductively solved the exclusion problem: mental causal powers are guaranteed by trope identity, while mental irreducibility is guaranteed by type non-identity.

The most discussed issue of the trope solution is the problem of mental quausation (Macdonald & Macdonald 2006; Noordhof 1998; also see Ehring 2011; Robb 1997; 2013; 2017; Tiehen 2019 for discussions of the problem). The core idea of this problem is that even though the trope solution can guarantee the causal powers of the mental event by identifying the mental trope with the physical trope, the trope’s being mental or not does not make a difference to whether the event has the powers. The trope does not confer the causal powers qua mental, but qua physical. A common way to motivate this quausation problem is to adopt a modified version of Physical Causal Closure and Exclusion, roughly as follows.⁴

• Physical Type Causal Closure: Any trope that constitutes a sufficient cause event is a physical trope (that is, the trope falls in a physical type); and being physical is sufficient for the trope to confer sufficient causal powers for the event to cause the effect.

• Type Exclusion: If being of one type is sufficient for a trope to confer the causal powers, then the trope’s being of other types is irrelevant to its conferring the powers.

These two principles are also prima facie plausible. Physical Type Causal Closure is secured by physicalism. Type Exclusion is intuitive: as long as the trope is already of a type that is sufficient for it to confer some causal powers, its being of other types or not will not change the fact that it confers those powers. Put together, these two principles lead to the conclusion that only being of a physical type is relevant for a trope’s power-conferring. Then, the irreducible mental type is never relevant. Even though a mental trope can confer causal powers, the trope does not confer those powers qua mental. There is still no distinctive mental causal power.

This quausation problem of tropes is similar to the quausation problem of events that haunts Donald Davidson’s anomalous monism (Davidson 1980). Davidson holds that token causal relations need to be grounded in strict causal laws that only exist in physics. So, for a mental event (for Davidson, an event that is truly described by a mental predicate) to be causal, Davidson suggests that it needs to be identical with a physical event. In this way, that event is truly described by both a mental predicate and a physical predicate, thus can be grounded in a physical strict law. Critics (Honderich 1982; Kim 1984; Sosa 1984) charge that although Davidson’s identifying a mental event with a physical event will secure the causal powers of the mental event, this event, both mental and physical, is not causal qua mental, as it has the causal powers only because it is grounded in a physical strict law. The mental property, these critics say, is causally irrelevant. By presenting the quausation problem of tropes, opponents of the trope solution claim that the trope solution fares no better than anomalous monism.

The proponents of the trope solution, on the other hand, want to rebut the trope quausation problem by cutting down the similarity between the trope solution and anomalous monism. Because of the similarity, the proponents of the trope solution take the trope quausation problem to be saying that a trope can have a mental property and a physical property at the same time, and these properties determine the trope’s ability to confer causal powers; but, according to the problem, the mental property is never causally relevant (Gibb 2017; Robb 1997; 2001; 2013; 2017). To rebut the quausation problem of tropes, these proponents of the trope solution argue that, unlike events, tropes are not the kind of entities that suit for having properties. Saying that tropes have properties, as they see, is making a categorical mistake. For example, Robb remarks that “[a] causally relevant property F [(the trope)] simply does not have various aspects such that one can legitimately ask whether some but not others are responsible for F’s being causally relevant” (Robb 1997: 191), and that “[m]ore simply: there are no tropes of tropes” (Robb 2013: 222). Similarly, Gibb (2017: 271) also says that “I do not take [the quausation problem of tropes] to be a troubling one because I do not consider that second-order properties [(viz., the properties of tropes)] should be admitted in one’s ontology.” They also presented arguments against the existence of properties of tropes. Robb (1997) argues that if tropes have properties, it will lead to a vicious regress: the properties of the trope will have further properties, these further properties will have even further properties, ad infinitum. Because of this, Robb believes that we need to cut the regress from the very beginning and say that tropes do not have properties. Gibb’s argument stems from the distinction between substance and property. Here is what she says: “[t]raditionally, one way of characterizing the distinction between substance and property is that substances are entities that bear properties, whereas properties are entities that are borne by substances. If properties bear properties this cannot be correct for then the former properties would have to be classified as substances” (Gibb 2017: 272). If tropes do not have properties, trope theorists say, then they do not confer causal powers qua mental or qua physical, the quausation problem does not make sense at all. In this way, trope theorists believe that they have solved the quausation problem.

2. A Dilemma for the Trope Solution: The First Horn

I agree with trope theorists that there are important differences between the trope solution and anomalous monism, and arguably tropes do not have properties. However, simply breaking down the similarity and denying the existence of properties of tropes cannot solve the trope quausation problem. That is because even if tropes do not have properties, it is still possible for tropes to confer causal powers in virtue of other things. For example, it could be that the causal powers come from the types (which are considered as classes rather than properties of tropes) into which the tropes fall. In this picture, tropes inherit powers from types. This position is plausible under certain presuppositions, and I will come back to those presuppositions a moment later. For now, it is enough to see that if this position is true, and a given trope can fall into both a physical type and a mental type, then the trope quausation problem persists even if tropes do not have properties: following Physical Type Causal Closure and Type Exclusion, the physical type, rather than the mental one, will be chosen as causally relevant. Then the trope is still not causal qua mental.⁵

The reason why denying the existence of properties of tropes cannot solve the quausation problem is that doing so cannot secure a mental source of causal powers. The motivation behind all kinds of qua questions is to find the source of the causal powers at issue. People want to know where those powers come from ultimately. Answering “qua hot” to the question “the kettle burns me qua what, qua red or qua hot?” means that the powers come from the property of being hot. Similarly, the qua question asked to anomalous monism also intends to query the source of the causal powers of the event: the event’s having causal powers cannot be a brute fact, we need an explanation of it. In the case of tropes, to make sure that tropes are causal qua mental, it needs to be the case that the ultimate ground of causal powers is something mental. Denying the existence of properties of tropes is a decent first step towards that end since doing so stops the physical properties of tropes from being candidates of sources of causal powers. But only doing that is not enough, since there can be physical entities other than physical properties of tropes that are the sources of causal powers. A firm mental ground of the powers is yet to be found.

A promising candidate for this mental ultimate ground is the trope itself. Under the trope identity theory, the trope is both mental and physical.⁶ Then, if the trope itself can be the ultimate ground for the causal powers, the trope is causal qua mental. Then the quausation problem will be solved. Trope identity theorists do make the move towards this direction. Robb (1997: 191) claims that “[t]ropes are not causally relevant qua this or that, they are causally relevant (or not), period.” Combining this with the move to deny the existence of properties of tropes, it seems that trope theorists believe if tropes do not have properties, then the ultimate ground would automatically be tropes themselves.

However, as is shown by our previous argument, this is not the case. Causal powers can have other type-like grounds even if tropes do not have properties: under certain presuppositions, the types, as classes, can be the sources of causal powers. The presuppositions are about the nature of causation. Under certain understandings of causation, causal powers will have grounds other than tropes. For example, under the Humean notion of causation, a trope’s categorical nature (or non-causal intrinsic nature) can persist across different possible worlds while the causal powers that the trope confers can vary because causal laws are different among those worlds.⁷ To give a toy example, a black trope will absorb most of the light in our world, but in another world with very different causal laws, it could be the case that black tropes reflect, rather than absorb, most of the light. Exactly similar (in the respect of categorical nature) black tropes confer different causal powers in different possible worlds. This shows that under certain theories tropes do not confer powers primitively. Causal laws determine the powers a trope can confer under the Humean theory.

The presupposition about the nature of causation lurks behind the trope quausation problem. The problem charges that only the physical type, rather than the mental type, that the trope falls in is causally relevant. It shows that this problem presupposes that types (whatever their natures are) are the sources of causal powers, rather than tropes. Then according to our analysis above, to make types sources of causal powers, the quausation problem needs to presuppose certain kinds of causal theory. Specifically, the trope quausation problem needs to presuppose the generalist notion of causation according to which causal laws are the grounds of token causations.⁸ Only with this presupposition, the qua mental or qua physical question makes sense. Under the generalist notion of causation, a trope gets its ability to confer causal powers from the types in which it falls because it is types, rather than token tropes, that are characterized by causal laws. Only when a trope falls into a type can it inherit the causal powers from the causal law which characterizes the type. In this way, tropes do not confer causal powers primitively, the powers come from laws of nature. So, under the generalist presupposition, the qua question is actually asking, which causal law is relevant here, the mental law or the physical law? Following Physical Type Causal Closure and Type Exclusion, the quausation problem of tropes concludes that only the physical law and the physical type are relevant, so the trope does not confer causal powers qua mental.

With these analyses of the quausation problem in hand, we can see that for the trope identity theory to rebut the problem and insist that tropes are the very sources of causal powers, it needs to deny the generalist notion of causation.⁹ Instead of the generalist notion, the trope theory needs to adopt the singularist notion of causation according to which token causations are grounds of causal laws. Under the singularist notion of causation, law-like type-level causal statements such as “smoking causes lung cancer” are mere generalizations of token-level causation. Types do not confer causal powers, an event has causal powers only because of its own nature. In this way, a trope’s ability to confer causal powers does not come from a causal law but from itself. If the singularist notion is presupposed, we do not need to appeal to types to ground token causation, since there is no further ground of causal powers apart from the tropes. Only then can we safely claim that an event has the causal powers it has because of its constitutive trope, period. Only in the singularist picture can the trope be the very source of the causal powers. In this picture, it makes no sense to further ask the qua mental or qua physical question. The original quausation problem is finally solved.

Some trope identity theorists believe that the trope solution is compatible with a wide range of different accounts of causation (for example, Tiehen 2019: 153). Now we can see that they must be cautious at this point. They must carefully choose the notion of causation. For example, they cannot choose the regularity notion and the counterfactual notion, since those notions are generally taken as generalist notions,¹⁰ and this kind of notion directly leads to the quausation problem. Only under the singularist notion can the trope theory rebut the quausation problem. Now we have finished the first horn of our promised dilemma: the generalist notion of causation leads to the (original) quausation problem, so, trope theorists need to adopt a singularist notion of causation to avoid it.

3. A Dilemma for the Trope Solution: The Second Horn

The second horn of the dilemma says that adopting the singularist notion will lead the trope solution to a new quausation problem: under the singularist notion of causation, the event does not have causal powers qua non-reductively mental, but only qua reductively mental. That is, the causal powers only come from some mental properties (that is, tropes) that can be reduced to physical properties, rather than some mental properties that are irreducible to physical properties. Indeed, all the causal powers come from a mental trope, but this mental trope is identical with, that is, reducible to, a physical trope. Some reducible mental entities, rather than irreducible ones, are the sources of the causal powers. The irreducible mental stuff, the mental types, do not confer causal powers under the singularist notions of causation. We do not have something which is both irreducible and causal here. Since the existence of that kind of thing is crucial for non-reductive physicalism, then the trope identity theory cannot be non-reductive in the way we care about (trope theorists may contend that their position is non-reductive in other ways. More on this later).

Why is it important that irreducible mental stuff confer causal powers? Why cannot we be content with the situation that there are some irreducible mental things, viz. the mental types, only that those things are epiphenomenal? That is because the kind of non-reductive physicalism we care about in the discussion of the exclusion problem is a metaphysical position whose aim is to grant that non-reductive mental properties are real building blocks of our world (Kim 1993). We can call this sort of non-reductionism “property non-reductionism.” A commonly accepted criterion of being real is what Kim (1993: 348) calls “Alexander’s dictum”: to be real is to have causal powers. Then, for the mental properties to be both real and irreducible, they need to have (or confer) causal powers that are distinct from the powers which are had by physical properties. For property non-reductionists, if those non-reductive mental properties are absent in a description of the world, then this description is incomplete: it fails to capture some distinctive power-conferrers. Now adopting a singularist notion of causation makes it the case that types of tropes never confer causal powers. Then the mental type, although irreducible to the physical type, is just something epiphenomenal and irreal.¹¹ On the other hand, under the current picture, although a mental trope confers powers and is real, it is something reducible. If the trope identity theory adopts a singularist notion of causation, we will not have something that fulfills both the irreducibility requirement and the reality requirement of property non-reductionism. Trope identity theory, then, collapses into a property reductionism in this situation.¹²

Facing such criticism, trope theorists may contend that property non-reductionism is not so valuable after all—we do not need something which both confers causal powers and is irreducible, what we want is just that mental events are causal and that mental types are multiply realizable, thus irreducible (see, for example, Robb 2013). Also, they will contend that such non-causal irreducible mental types are enough for psychology since these types are explanatorily indispensable.¹³ Their idea is similar to Putnam’s famous peg-hole example (Putnam 1975). To explain why the peg cannot go through the hole, an explanation specifying all the micro-physical details of the peg and the hole can be a sufficient explanation, but not a good one. Rather, the simple explanation that the side of the peg is longer than the diameter of the hole is a good one since it adequately provides the needed illumination in an accessible way. Trope theorists may say that mental explanations, the explanations which take irreducible mental types as explanantia, are just like the simple explanation: they provide explanations that are different from, and better than, the physical explanations which take physical types as explanantia. For example, they may say that for the explanandum that Tom is crying, the pain is a better explanans than the complex neural realizer of the pain since it provides the needed information in a simple way.¹⁴

I believe the analogy with the peg-hole case cannot help the trope theory. Under the singularist picture, mental types cannot provide us with any explanations which are different from physical explanations, even though the mental types are irreducible. Indeed, explanation is an epistemic activity, which aims to improve one’s epistemic status (Kim 1989: 94). However, few take explanation to be purely epistemic, in the sense that an explanation is achieved as long as it gives people a feeling of “Aha.” (If explanation is in that sense purely epistemic, then to the question, say, “why the car is running,” the explanation saying that the driver is casting a spell would be as good as the explanation appealing to the physical mechanics of the car since both explanations would give someone a feeling of “Aha.”) Many philosophers (for example, Kim 1988; 1989; Salmon 1984) believe that true explanations are backed by objective relations between the explanans and the explanandum. This is what Kim calls “explanatory realism.” Specifically, the position is as follows.

Explanatory Realism: $C$ is an explanans for E in virtue of the fact that $c$ bears to $e$ some determinate objective relation R. (Kim 1988: 226, original italic)

Here $C$ and $E$ stand for propositions, while $c$ and $e$ are events. So, according to this position, the truth of an explanation is grounded in the objective relation between two events corresponding respectively to the explanans and the explanandum. Under explanatory realism, the explanatory work is primarily done by the objective relation. In the case of causal explanations, the objective relation is the causal relation between the two events at issue. Now under the singularist picture, all the causal work is done by the event constituted by the reducible trope which falls in both the physical type and the mental type. Then, for an explanation that contains an irreducible mental type as the explanans to be true, it needs to be backed by the causal process from the event constituted by the reducible trope to the effect. Given explanatory realism, even though we are having the irreducible mental type as the explanans, the explanatory work is primarily done by the trope-event and the causal relation between it and the effect. As a result, this mental explanation will be essentially the same as the physical explanation which contains the physical realizer of the irreducible mental type as the explanans. This is because in this physical explanation, the explanatory work, according to explanatory realism, is done by the same causal process from the trope-event to the effect. We do not have two different explanations here. Rather, we have just one, which appeals to the same causal process. The reduction of causal powers leads to the reduction of explanations. So, given explanatory realism and trope identity under the singularist picture, there is no explanatory advantage of irreducible mental types—mental explanations are not better or simpler than physical explanations, they are just the same.

The above claim is, of course, controversial. The claim adopts the external individuation criterion of explanations. According to this criterion, an explanation is individuated by the causal process it appeals to, and two explanations are the same as long as they appeal to the same causal process (Kim 1988). Such a criterion is criticized by some philosophers (Gibb 2009; Marras 1998). They contend that as long as explanation is an epistemic enterprise, explanations that contain different types as explanantia but appeal to the same causal process cannot be the same explanation because different types will have different explanatory powers in different contexts. Marras (1998) gives us the following example. Consider the following two explanations to the occurrence of the collapse of a building: (a) “The earthquake caused the building to collapse”; and (b) “The event that is on p.5 of today’s newspaper caused the building to collapse.” Marras argues that even though these two explanations are backed by the same causal process that the earthquake caused the collapse of the building, they have different explanatory powers: those who do not know which event is on the newspaper will not feel that the collapse of the building is explained. So, Marras concludes, these two explanations are different, since only one of them can change the epistemic status of those who do not know what event is in the newspaper.

However, even though it is true that in many everyday contexts like the one Marras depicts, the choice of the explanans will decide whether the explanation is successful for certain people, the external individuation criterion is still defensible in our context. The reason why in the Earthquake Case the two explanations have different explanatory powers is that knowledge backgrounds vary among different receivers of the explanations. This is what we mean by saying that the explanatory power of a certain explanation varies in different contexts. Given this feature, we can even imagine some contexts in which the receivers of the explanations do not even know what an earthquake is, then in that context explanation (a) will not be explanatory either. However, in our current context, such an influence from the difference of knowledge backgrounds and context variation will be missing. The context in which we debate whether mental types are reducible to physical types and whether mental types are explanatorily indispensable is the context of science. In this context, we want to know whether psychological explanations are necessary in addition to physical explanations, and whether psychology itself is needed in addition to physics. Here, people do not worry whether the reference of certain type terms, such as “pain” and “C-fiber firing,” is known to the scientific community since trained scientists and philosophers generally have the needed knowledge background. So the kind of concern raised in the Earthquake Case does not raise in our context. Moreover, in such a context, scientists and philosophers aim to reveal the objective structure of the world, or at least to be as close to it as possible. Then, to provide a scientific explanation, we aim to explain the occurrence of a certain effect by presenting the objective mechanism that leads to the effect. In other words, in the context of science, scientific explanations are made to reveal what is really happening out there in the world. Then, to judge whether certain explanations to a given effect are the same in the context of science under the presupposition of explanatory realism, it suffices to see whether these explanations appeal to the same structure of the world, namely, the same causal relations between the same cause and the same effect, to explain the occurrence of the effect. The influence of knowledge backgrounds of the explanation receivers removed in the context of science, we can conclude that under the singularist trope theory, a mental explanation is the same as the relevant physical explanation because both explanations appeal to the same causal process from the event constituted by the reducible trope to the effect.

That being said, people may still have the intuition that mental explanations and physical explanations are different. Given explanatory realism, this intuition, I suggest, actually comes from the implicit belief that causal-power-conferring mental properties, no matter tropes or universals, are different from physical properties. That is just the belief in property non-reductionism. If that position is true, mental explanations and physical explanations will be backed by different causal processes—one by the mental causal process and the other by the physical causal process. Then, those explanations will indeed be different. Here we spot another advantage of property non-reductionism over the trope-identity-but-type-non-identity position (if we put the exclusion problem aside for a moment): it does not only guarantee that mental causal-power-conferrers are different from physical power-conferrers but also guarantees that mental causal explanations are different from physical causal explanations, even under the presupposition of explanatory realism.¹⁵

This is the second horn of the dilemma. If the trope identity theory adopts a singularist notion of causation, it will encounter a new quausation problem which charges that the trope confers causal powers qua reductively mental, rather than qua non-reductively mental. Then, the trope identity theory cannot retain a property non-reductionist position. What is worse is that, under the singluarist picture, trope theorists cannot make sense of the claim that irreducible mental types provide different and better explanations than physical types, given the plausible presupposition of explanatory realism. The trope identity theory is both a property reductive and an explanation reductive position under the singularist notion of causation.

In sum, I have presented the following dilemma for the trope identity solution to the exclusion problem: the trope solution must either adopt a generalist notion of causation or a singularist notion of causation; but, either way, it will encounter a quausation problem. The qua problem, original or new, still haunts the trope solution and prevents it from being a successful non-reductive theory.¹⁶

4. Objections and Replies

In this section, I will consider two objections to the dilemma above, one to the first horn, the other to the second, and make my replies.

1. Ehring (2003; 2011) believes that even if types can be seen as causally relevant, the relevance of the mental type will not be excluded by the relevance of the physical type. Here Ehring can be seen as saying that even if we hold a generalist notion of causation (thus types are causally relevant), the original quausation problem can be solved. The way Ehring solves the problem is as follows. Since the mental type $M$ supervenes on physical types, on each occasion where $M$ is instantiated there is a physical type also instantiated. Then, on a given occasion, the causal powers that $M$ confers to produce a certain effect $e$ are identical with (or are a subset of) the powers that $P$, the realizing physical type, confers to produce $e$. On another occasion where $M$ is realized by a different physical type $P*$, the powers that $M$ confers to produce $e$ are identical with (or are a subset of) the powers that $P*$ confers to produce $e$. The idea is, on every occasion where an effect is produced, $M$ shares the causal relevance with its physical type realizer on that occasion. Since the causal relevance of the physical types does not exclude itself, $M$’s causal relevance is then secured. At the same time, the irreducibility of $M$ is retained because it contains different tropes from any $P$.

I believe the relevance of $M$ is still redundant in Ehring’s picture. We have an independent reason, that is, Physical Type Causal Closure, to believe that the physical types are causally relevant. But we do not have a similar independent reason to believe the mental types are relevant. Then, compared to saying that the mental type shares causal relevance with the physical type, a more plausible position is to say that the relevance belongs totally to the physical type, and the mental type occupies none. The explanatory success of psychology may appear to be a reason to say that the mental type is relevant. However, if we reason from the perspective of causal laws (since we are presupposing generalist notion of causation here), we will find that a physical law that connects the physical type with the effect type is more fundamental than the psychological law that connects the mental type to the effect type. Moreover, on each occasion of mental causation, it is highly probable that the application of the physical law can ground the application of the psychological law. This shows that the ultimate source of causal powers is the physical law and the physical type, not the mental ones. Then, the mental types will be excluded by Type Exclusion. It will not be the case that $M$ will share the causal relevance of the realizing base on each occasion, as Ehring thinks. Rather, $M$ is causally impotent in every case.

• (2)

  I have argued that mere type irreducibility cannot retain the trope solution to be a robust kind of non-reductive physicalism. Some may disagree with me on this point. They say that many classic non-reductive physicalist theories hold a token-identity-while-type-non-identity position, for example, Fodor’s theory (Fodor 1974; 1989; 1991; 1997), but we believe these positions are robust non-reductive physicalism, we do not think Fodor is a reductionist.

I have two replies to this objection. First, it is important to note that it is not my claim that every token-identity-while-type-non-identity position will collapse into a reductive position. My claim is that such a position will collapse when it adopts a singularist notion of causation. It is only under the singularist picture that mental causal-power-conferrers are reductive. Also, it is only under the singularist picture that mental explanations are the same as physical explanations. If Fodor is defending a property non-reductionism, then his theory can adopt a generalist notion of causation to avoid the collapse.

Second, and more importantly, my dilemma against the trope theory can be seen as a general argument against any token-identity-while-type-non-identity physicalist positions, or at least any such position that has a similar structure as the trope identity theory. Any such position will encounter similar situations: either it adopts a generalist notion of causation and encounters the quausation problem according to which token causes, which are both mental and physical, are not causal qua mental; or it adopts a singularist notion of causation and finds itself collapsing into a kind of reductionism of both causal powers and explanations. Therefore, Fodor’s position, or any token physicalist position, is not a robust non-reductive position either, just like the trope identity theory.

5. Conclusion

I have argued that the trope solution faces the dilemma which charges that the solution will encounter a quausation problem no matter whether it adopts a generalist notion or a singularist notion of causation. Because of this, the trope identity solution cannot solve the exclusion problem non-reductively. Moreover, almost all kinds of token physicalism will encounter similar dilemmas. So, it seems that token physicalism is a shaky position.