EXCLUDING TROPES*

 

            An increasingly popular strategy for resolving Kim’s exclusion problem[1] is to suggest that mental and physical property instances are identical despite the non-identity of the mental and physical properties themselves.[2] And in order to implement this strategy, an account of properties and their instances that can be reconciled with this suggestion is required. There seems to be something of an emerging consensus among advocates of this strategy that serious difficulties arise if property instances are taken to be exemplifications of universals, but that these problems can be avoided if they are instead taken to be tropes – abstract particulars.[3] Noordhof, however, offers us some cautionary advice here: we need to be careful that we simply have not replaced one “bulge in the carpet” with another.[4] Metaphysical manoeuvres of this sort risk pushing theoretical difficulties to one side rather than solving them. I want to argue that this is exactly what happens for the appeal to tropes. While it appears to accommodate the identity of mental and physical property instances, it does so at the cost of either the dispositional indeterminacy of putatively mental instances or a retreat to the species-specific identity theory. And although this latter option may not in and of itself be so bad, it undercuts the motivation for the identifying mental and physical property instances in the first place.

 

I: The Exclusion Problem

 

            The exclusion problem is thought to pose serious difficulties for the causal efficacy of mentality. The formulation of the problem depends on whether one takes the causal relata to be property instances or events. If the relata are taken to be property instances, the question is whether mental instances can ever cause physical instances. And if the relata are taken to be events, while it is generally conceded that mental events can cause physical events, the question remains whether the mental properties of mental events can ever be causally relevant. For present purposes, I am going to assume the causal relata are events and leave it open whether events are best understood as Davidsonian basic particulars or as Kimiam exemplifications-of-universals-by-objects-at-times.[5]

The exclusion argument purports to show that mental properties are never causally relevant to physical events. Intuitively, the causally relevant properties of an event are those in virtue of which it causes some other event. Suppose, for example, a soprano sings the word “shatter” at a high pitch and amplitude and, thereby, causes a nearby glass to shatter. The glass broke, not because of the meaning of the word she sang but because of the pitch and amplitude of her action, and hence only the latter are causally relevant.[6] The causal irrelevance of mental properties to physical events is supposed to follow from (i) the non-identity of mental and physical properties and (ii) the causal closure of the physical world, here formulated as the thesis that the physical properties of causes are always causally relevant to their physical effects. One might, of course, argue that in any given case both a mental property and a physical property could be causally responsible for a single physical event, each of which would have been causally sufficient in its own right. But if mental causes are uniformly and without exception overdetermined, then mental properties make no causal difference in the world. Any claim that they are causally relevant nonetheless is hardly plausible.

            It is worth emphasizing that the target of the exclusion argument is any theory of mentality according to which mental properties are non-identical to physical properties. Even a view which denies this identity on so modest a basis as that of multiple realizability is at risk. For simplicity, I am henceforward simply going to assume a functionalist account of mental properties. In addition, I am going to assume strong supervenience, roughly the view that physical indiscernibility entails mental indiscernibility.[7]

 

II: The Instance Solution

 

            In order to avoid confusion, I will henceforward refer to property instances as “instances” and properties proper as “types”.[8] The term “properties” will be used to refer indifferently to types or instances. The instance solution to the exclusion problem involves, first of all, interpreting the non-identity of mental and physical properties as the non-identity of mental and physical types. In addition, it relies on the following two presuppositions: (i) instances, and not types, are the primary bearers of causal powers; and (ii) instances of distinct types can be identical. If mental instances are identical to causally relevant physical instances, then they do not compete with them for causal efficacy. As a result, causal closure does not exclude them from causal relevance. Mental types acquire their causal relevance derivatively, from the efficacy of their instances.[9]

The success of the instance solution, however, hinges on whether or not the Instance Identity Thesis – that mental instances are identical to physical instances – can be sustained. And this depends on whether an account of the nature of instances and types, and the relations between them, can be developed which can adequately navigate the difficulties posed by the following three principles:

Type Non-identity: mental types are distinct from physical types

Character Essentialism: instances have essential characters (or such-nesses)

Necessary Coexistence: For all x and y, if x is identical to y then it is not possible for x to exist and y to fail to exist.

A few comments on each of these principles are in order. First, Type Non-identity follows straightaway from the multiple realizability of the mental – distinct physical types can be co-instantiated with a single mental type. It is worth noting that the reverse is true as well – distinct mental types can be co-instantiated with a single physical type – even if strong supervenience is assumed. After all, we attribute mental properties not just to intelligent beings, but to states (or events occurring) within intelligent beings. And physically identical states within distinct intelligent beings (or a single intelligent being at different times) can occupy different functional roles within those beings.[10] Second, the guiding idea underlying Character Essentialism is the intuition that, while, for example, a blue cup is only accidentally blue, the blueness of the cup is essentially blue. Note: the term “character” is meant to be neutral as between various accounts of the nature of instances. Finally, as should be obvious, Necessary Coexistence is simply a restricted version of the Indiscernibility of Identicals. I focus on this version both because of the contentiousness of talk of properties of instances and because of the pressure it puts on instance identity claims when combined with Character Essentialism.

            Suppose one held that the essential character of an instance is given by the type of which it is an instance. So, for example, the essential character of an instance of pain is that of being pain – that is, occupying the relevant functional role, henceforward the “pain-role.” And suppose as a matter of fact a mental type and a physical type – say pain and C-fibre activation – are co-instantiated by a state or event within an intelligent being. From Type Non-identity it follows that there are possible circumstances in which the state in question instantiates the mental type but not the physical type and vice versa.[11] Given our assumption about the essential character of an instance, Character Essentialism implies that (i) in possible circumstances in which the mental type is not instantiated the (actual) mental instance does not exist and (ii) in possible circumstances in which the physical type is not instantiated, the (actual) physical instance does not exist. Finally, given Necessary Coexistence, each of (i) and (ii) independently suffice to show that the actual mental and physical instances are non-identical.[12][13] The upshot here is that in order to sustain the Instance Identity Thesis, minimally some other account of the essential character of instances needs to be defended.  

 

III: Property Metaphysics

 

            Broadly speaking, there are two central approaches to the metaphysics of properties – type-primacy and instance-primacy – each of which will be investigated in turn. By type-primacy I mean views according to which types are ontologically fundamental and instances are analyzed in terms of types. Such views include Platonism – according to which types are transcendent entities outside of space and time – and the view that properties are universals – wholly present in distinct spatio-temporal locations – as well as various versions of type-nominalism. By instance-primacy I mean views according to which instances are taken to be fundamental and types are analyzed in terms of them. Such views paradigmatically take instances to be tropes – spatio-temporally located abstract particulars.

            Following Robb and Heil, I am going assume an in re conception of instances – roughly, the view that that “when the objects in question are material, [instances] are located in space.”[14] The idea underlying this assumption is that the character of an instance is intrinsic to it and not a function of its relations to other things. A little care in formulation is needed here. I do not mean to deny that instances can have relational, and in particular dispositional, characters. I just mean to deny that the fact that an instance has the character that it does – be it categorical or relational/ dispositional – is itself a relational fact. The in re assumption is unsatisfied by Platonism – according to which the character of an instance is a matter of its relation to a transcendent form – and various versions of class nominalism – according to which instance character is wholly a matter of membership in some class. The point of the assumption is two-fold. First, arguably the only views that are candidates for securing the efficacy of mental properties are those which satisfy it.[15] As Robb and Heil put it, “[we] find it hard to understand how being related to a non-spatio-temporal form, or falling into a certain class, or falling under a predicate, could by itself make a difference to a mind.”[16] And second, focusing on a smaller class of possible views will make the task here more manageable. Similar arguments to those presented below will presumably (or at least with any luck) apply to views that fail to satisfy the in re assumption.

            Given the structure of the objection to the Instance Identity Thesis at issue, a note about “trans-world identity conditions” for instances is in order. My inclination is to simply combine a non-duplication principle for instances (i.e., synchronic identity conditions) with a Kripkean stipulationist account of modality.[17] Proceeding in this way – and avoiding issues such as that of counterpart relations between instances in different worlds – has the virtue of simplicity and, with due diligence, should not lead us badly astray. The non-duplication principle I will presuppose is the following: no two instances can both (i) have the same essential character and (ii) exist in the same spatio-temporal location.[18] There are two comments worth making here. First, what having the “same essential character” amounts to depends on whether a universal theory or a trope theory of instances is at issue. A universal theorist would understand it in terms of being occurrences of the same universal, whereas a trope theorist would unpack it in terms of a primitive relation of exact resemblance. Second, I understand this non-duplication principle to be, at bottom, equivalent to one formulated in terms of object-time pairs rather than spatio-temporal locations.

            One final methodological point before beginning in earnest. Robb[19] has charged Noordhof [20] with requiring that a defense of the Instance Identity Thesis proceed as follows: first, defend an account of (identity conditions for) instances; second, show that the Instance Identity Thesis follows from this account of instances. And Robb has argued that this get things just backwards. The proper approach is to first defend the Instance Identity Thesis on independent grounds – presumably on the grounds that it secures mental causation – and second, find an account of instances which is both compatible with this thesis and independently plausible. My project here is to determine whether or not the second stage in Robb’s procedure can be successfully carried out.

 

IIIa: Type Primacy

 

            The central (and perhaps only) account of properties which both takes types to be primary and satisfies the in re assumption is one that invokes immanent universals. According to such a view, types are universals, entities which can be fully present in distinct locations and literally shared by objects. Instances, on this view, are derivative entities consisting of universals, objects, and times. The relation between the constituents of an instance might be characterized as the exemplification of a universal by an object at a time or, perhaps, the fact that an object bears a universal at a time. Moreover, the essential character of an instance is given by its constituent universal.

            Now, as above, if the constituent universal of a mental instance is a mental universal and the constituent universal of a physical instance is a physical universal, the Instance Identity Thesis cannot be sustained. One way to avoid this result would be to assume that the constituent universals of mental instances are physical universals. The difficulty, however, is whether or not this suggestion can be reconciled with the genuine (or 1st rate) reality of mental properties. After all, the most obvious way of implementing it would be to assume that there are no mental universals and mental instances are just physical instances which satisfy mental predicates (or concepts). But if there are no mental universals then mental properties are at best second rate, if their existence is even conceded at all.[21] So the challenge, then, is to defend a view according to which mental universals exist but, nonetheless, the constituent universals of mental instances are physical universals. In my view, the only promising way to attempt to meet this challenge is via an account of the relation between mental and physical types.[22] The two most prominent models of this relation, in the mental causation literature at least, are the relation of determinable to determinate[23] and that of second order to first order.[24] I will consider each of these in turn.   

            Paradigmatic examples of types that stand in the determinable-determinate relation are colours: redness and blueness are determinates of the property of being coloured; the property of being navy blue (or Carolina blue) is a determinate of blueness; and so on. Yablo argues that the core of the determination relation is asymmetric necessitation, where a type P asymmetrically necessitates a type Q just in case (i) for all x, if x has P then x has Q and (ii) possibly, for some Q, x has Q and lacks P.[25] There are two separate questions one might ask here. Do physical types asymmetrically necessitate mental types? And even if they do, would this reconcile the existence of mental universals with the thesis that physical universals are the constituent universals of mental instances? A quick answer to the first question is, as above, if mental properties are thought of as properties of events or states within intelligent beings, then multiple realizability and the strong supervenience of the mental properties of intelligent beings on the physical properties of intelligent beings will not yield asymmetric necessitation. The second question requires a little more comment, however. The asymmetric necessitation of the mental by the physical might be cashed out in terms of a metaphysical relation between distinct universals, or it might be captured by taking corresponding mental and physical predicates to stand in different semantic relations to the same universal.[26] The trouble is that either way mental properties are rendered second rate. If mental universals are distinct from physical universals and the constituent universals of mental instances are physical universals then mental universals are uninstantiated, that is, they do not exist.[27] But the identification of mental universals – that is universals picked out by mental predicates – with physical universals, in effect, results in the elimination of mental universals. After all, given multiple realizability, mental predicates would have to be taken to pick out different physical universals in different contexts of use.[28]

            Kim has argued that mental properties are second order properties.[29] An individual bears a second order property just in case it bears a first order property which “satisfies a certain specification.” If mental properties are functional properties, as we have been assuming, the specification is a functional or causal role. An event bears a (second order) mental property just in case it bears a (first order) physical property which realizes the requisite functional role. In Kim’s view, this suffices to establish the Instance Identity Thesis. He says, “If M has two realizers, P1 and P2, each M-instance is either a P1-instance or P2-instance, and those M-instances that are P1-instances have the causal powers of P1, and similarly, the M-instances that are P2-instances have the causal powers of P2.” [30] Now in order to avoid the charge of second rate-ism, the satisfaction of a mental specification must be understood as the instantiation of a mental universal. Moreover, if the Instance Identity Thesis is to be tenable, mental universals need to be instantiated by the same events that instantiate their (1st order) physical realizers rather than by the physical instances themselves – amalgams of events, physical universals, and times. After all, mental and physical instances cannot be identical if their constituent objects are distinct.[31] There remain two possibilities: either mental universals are distinct from their (physical) realizing universals or they are identical to them. If they are distinct then, of course, the Instance Identity Theory is false. And if they are identical, multiple realizability would leave us, as above, with mental predicates which pick out different physical universals on different occasions of use rather than with genuinely mental universals. If the Instance Identity Theory is to be sustained without defanging mental properties, type primacy will have to be abandoned.

 

IIIb: Instance Primacy

 

What remains to be determined is whether the Instance Identity Theory can be salvaged by appeal to instance primary. As above, instance primacy is the view that instances are fundamental and types are derivative. Paradigmatically, instances are taken to be tropes – spatio-temporally located abstract particulars. In keeping with the in re assumption, I will assume that tropes have intrinsic essential characters. Distinct tropes stand in similarity relations to one another in virtue of their essential characters.[32] But given that these characters are particular, there are no shared respects of similarity. As a result, the similarity relation has to be taken to be primitive. Types, on this picture, are classes of resembling tropes.

Exactly how this similarity relation generates classes of tropes needs to be addressed, however. Insofar as our concern is with types whose instances are exactly similar, this is relatively straightforward. Exact similarity is an equivalence relation and, as a result, generates equivalence classes of exactly resembling tropes. But not all types have exactly similar instances; although all instances of a specific shade of blue, for example, are exactly similar, the same is not true for all instances of the type Blue. A common strategy is to suggest that the similarity relation comes in degrees; so while the type Navy Blue, for example, consists of a class of exactly similar tropes, the type Blue will consist of a class of less than exactly similar tropes. By itself, however, this manoeuvre is unsatisfactory. The trouble is that any relation of less than exact similarity will be non-transitive and so, by itself, will not generate equivalence classes of tropes.[33] There seem to me to be two options available to the trope theorist at this point. First, one might depart from the intuitive notion of similarity and invoke a type-generating relation which comes in degrees, and all degrees of which are transitive. Alternately, one might retain the intuitive notion and make the criterion class membership some degree of similarity to the members of a specific equivalence class of exactly resembling tropes. In my view neither of these suggestions is entirely satisfactory, but I will not press the point here.[34]

            Before deploying this apparatus on behalf of the Instance Identity Thesis, a couple of comments are in order. First, there are two senses in which a type could count as second rate on this picture: the criterion for membership in the class could be a very low degree of similarity; or the class could be a mongrel, not defined by any specific degree of similarity. Second, in what follows I will be assuming that mental tropes are dispositional properties – that is, properties which characterize an object in terms of how it would behave in various circumstances – and that the physical tropes which realize mental properties are categorical – wholly here-and-now – properties.[35] Moreover, I will be assuming that it is not the case that all categorical properties are reducible to dispositional properties or vice versa.[36]

            What I want to argue is that the trope theorist faces the same difficulty of reconciling the identity of mental and physical instances with the genuine existence of mental types that undermined the appeal to universals considered above. There are two reasons for thinking that the trope variant of the Instance Identity Theory yields second rate mental types. First, given multiple realizability, a mental type M will have more than one realizing physical type. And although the members of any given realizing type may be quite similar, the members of different types will in some instances be quite dissimilar.[37] As a result, the tropes that constitute M will be bound together by a rather low degree of similarity.  Second, it will not in general be true that all members of a given realizing physical type are members of M – recall Lewis’ madman.[38] This implies that not all tropes similar to members of M – even to the aforementioned rather low degree – fall within M, rendering it a mongrel class.

Robb, however, has presented rejoinders to both of these worries.[39] He has offered two suggestions regarding how he might handle Lewis’ madman cases. First, one might take the madman’s trope to fall within M despite not occupying the requisite functional role. And second, one might take the tropes in M to include the relevant surrounding circumstances, thereby excluding the madman’s trope from the physical type which realizes M. Although nothing I say here is will be decisive, there is some reason to balk at both suggestions. The second suggestion runs the risk of brokering even greater dissimilarity among the various physical types which realize M. And the first, in effect, amounts to a retreat from the functionalist account of mental properties that has been presupposed.[40]

Robb’s response to the multiple realizability problem is, in my view, more promising. He suggests that despite consisting of tropes of distinct physical types, M is not a second rate type because its members share a high degree of dispositional similarity.[41]  Now on a naïve reading, this suggestion seems untenable. After all, if (i) the essential characters of a pair of trope determines their degree of similarity to one another, and (ii) the tropes which realize M have categorical characters which ground their classification as instances of physical types, then M-realizing tropes from different physical types will lack the dispositional characters required for a high degree of similarity.

But Robb and Heil reject this naïve reading, insisting instead that the essential characters of tropes are at the same time both categorical and dispositional:

“[it] is part of out thesis that every property – every intrinsic property of a concrete object – is at once dispositional and qualitative [= categorical]. … A property’s dispositionality and qualitativity are not different aspects or properties of the property, they are rather the property itself, differently considered.”[42]

The “Dual-Character Hypothesis” (let’s call it) allows tropes of different physical types, in virtue of their essential characters, to be quite dissimilar when considered categorically while at the same time highly similar when considered dispositionally. [43] It is worth pointing out, however, that this manoeuvre requires a more complex account of similarity. Given that the degree of similarity between a pair of tropes depends on whether they are considered categorically or dispositionally, a single two-place similarity relation will not suffice for the generation of both categorical and dispositional types. The advocate of the “Dual-Character Hypothesis” will instead need to invoke either two distinct two-place similarity relations will need to be invoked – corresponding to categorical and dispositional similarity – or one three-place relation, where the extra relatum is the manner in which the tropes are being considered.

            What I want to argue, at the end of the day, is that the Dual-Character Hypothesis leads either to dispositional (and, hence, mental) indeterminacy or to the species-specific type identity theory. And given that the latter alternative can be easily accommodated by a universalist theory of types, the appeal to tropes does not at the end of the day bear new fruit. It is worth emphasizing from the get go that my target is not the Dual-Character Hypothesis per se, but the identification of mental instances with tropes so construed. My argument will be focused on cases like that of Lewis’ madman in which physical instances fail to occupy their characteristic functional roles. Consider by way of analogy a vial of opium. When ingested by a normal human, a (categorical) physical property of the valium will cause her to become drowsy. In virtue of having this physical property, (the contents of) the vial of opium has a dispositional property, namely that of having a dormative power. A trope theorist who endorsed the Instance Identity Theory would claim that the physical trope was identical to the dormative trope. Now the very same vial of opium could have been consumed instead by an organism with a different physiology, say, a Venusian, such that the physical property in question caused him to become wide awake and highly alert. In this case, the vial would have a stimulative power in virtue of having said physical property, and the Instance Identity Theory would imply that the physical trope is identical to the stimulative trope. The question that faces us now is whether or not the dormative trope is identical to the stimulative trope. There seem to be three options. First, one could argue that when the valium is ingested by the Venusian, it lacks the dormative power. The trouble with this suggestion is that if the dormative trope is identical to the physical trope and fails to exist, then the physical trope does not exist either. Second, one could argue that when ingested by the Venusian the valium retains its dormative power. By parity of reasoning, one would have to concede that when ingested by the human, the valium had both dormative and stimulative power. The trouble is that this makes the essential character of the one trope, when considered dispositionally, to be at the same time stimulative and dormative. But this seems to be a straightforward contradiction.[44] Moreover, given the innumerable variety of effects that the one physical property of the valium could have on organisms with different physiologies, the upshot would be dispositional chaos. The third and perhaps best option would be to relativize the dispositions to physiologies. After all, while it may be contradictory to suppose that one property of the valium is both a disposition to produce drowsiness simpliciter and disposition to produce alertness (i.e., non-drowsiness) simpliciter, there is no contradiction in supposing that a property is both a disposition to produce drowsiness in humans and a disposition to produce alertness in Venusians. Note: this is not to deny that there is a single core disposition which manifests itself as drowsiness when paired with a human’s system and alertness when paired with a Venusian’s; what is being denied is that any such core disposition can be identified with either the dormative or the stimulative disposition. Only its manifestations can be so identified.

            Consider now a human whose C-fibres are firing and in whom C-fibre activation occupies the pain-role. According to the view on the table, the instance of the property of being C-fibre activation (henceforward, the “C-fibre trope”) would be identical to the pain trope. Had the person’s brain been “wired” differently, the very same property instance could have occupied a different functional role, say the pleasure-role. In such circumstances, the Instance Identity Theory would imply that the C-fibre trope is identical to the pleasure trope. As above, there seem to be three options for the trope theorist. First, one could argue that in the counterfactual circumstances, the C-fibre trope lacks the pain disposition. But, as above, since the C-fibre trope just is the pain trope, this would imply that the C-fibre trope does not exist in said circumstances. Second, one could argue that the C-fibre trope does have the pain disposition in the counterfactual circumstances, and concede, by parity of reasoning, that it has the pleasure disposition in the actual circumstances. The trouble is that this would make the very same state of our subject a state of pleasure and a state of pain. Moreover, given the variety of functional roles that the C-fibre trope could occupy in differently wired brains, the upshot would again be dispositional chaos. Finally, one could relativize mental/functional properties to physiologies. The above mentioned difficulties dissipate if C-fibre activation is taken to be pain in normal humans and pleasure in abnormal humans (i.e., those with “cross-wired brains”). After all, since our subject is actually normal, the (putative) fact that the C-fibre trope is both a pain-in-normal-humans trope and pleasure-in-abnormal-humans trope does not imply that her state of C-fibre activation is a state of pleasure as well as a state of pain. The trouble is that what we have in effect ended up with is a version of the species-specific type identity theory.[45] While this in and of itself is not so bad,[46] it undercuts the original motivation for the Instance Identity Theory and with it the need to invoke tropes to secure the efficacy of the mental. If (species-specific) mental types just are physical types then mental and physical properties do not compete for causal relevance, and so the exclusion argument cannot get a foothold in the first place.  

 

IV: Concluding Remarks

 

            The instance solution to the exclusion problem is both ingenious and philosophically elegant. As we have seen, however, the attempt to reconcile the Instance Identity Thesis with the non-identity of mental and physical types and the genuine existence of mental properties runs into insuperable difficulties. And the apparent promise of the move to tropes from universals has proven to be chimerical. What, then, is the proper solution to the exclusion problem? In my view, the lesson we should take from it is that mental properties simply are not causally efficacious, but that this is utterly unsurprising. Mental properties are dispositional properties. As such, there is no more reason to suppose that the property of pain competes for causal efficacy with the property of being C-fibre activation than there is to suppose that the dormative power of a sleeping pill competes for efficacy with the physical properties of the pill. And if this renders mental properties second rate, so be it. After all, mental events retain their efficacy even if not qua mental. And mental concepts earn their keep through their predictive-explanatory power and their cognitive efficiency. Although exclusion may render mental properties obsolete, mentality itself remains intact.

 

 

Peter Alward

Department of Philosophy

University of Lethbridge

4401 University Dr.

Lethbridge, AB

Canada T1K 3 M4

peter.alward@uleth.ca

 

REFERENCES

 

Alward, Peter (manuscript-a), “Are Functional Properties Causally Potent?” unpublished manuscript (http://www.uleth.ca/~peter.alward/papers/FunctionalProperties.htm).

 

Alward, Peter (manuscript-b), “Mad, Martian, but not Mad Martian Pain,” unpublished manuscript. (http://www.uleth.ca/~peter.alward/papers/MadMartianLewis.htm).

 

Braun, David (1995), “Causally Relevant Properties,” Philosophical Perspectives 9: 447-475.

 

Campbell, Keith (1983), “Abstract Particulars and the Philosophy of Mind,” Australasian Journal of Philosophy 61: 129-141

 

Donald Davidson (1970), “Mental Events,” in Experience and Theory, L. Foster and J.W. Swanson, eds., University of Massachusetts Press and Duckworth.

 

Ehring, Douglas (1996), “Mental Causation, Determinables, and Property Instances,” Nous 30: 461-480.

 

Ehring, Douglas (1999), “Tropeless in Seattle: the cure for insomnia,” Analysis 59: 19-24.

 

Ehring, Douglas (2002), “The Causal Argument against Natural Class Trope Nominalism,” Philosophical Studies 107: 179-190.

 

Kim, Jaegwon (1976), “Events as Property Exemplifications,” in Myles Brand and Douglas Walton(eds), Action Theory, Dordrecht: D. Reidel, pp. 159-77.

 

Kim, Jaegwon (1989), “Mechanism, Purpose, and Explanatory Exclusion,” Philosophical Perspectives 3: 77-108.

 

Kim, Jaegwon, (1993a), Supervenience and Mind, Cambridge: Cambridge University Press.

 

Kim Jaegwon (1993b), “Postscripts on Mental Causation,” in Supervenience and Mind, Cambridge: Cambridge University Press.

 

Kim, Jaegwon (1997), “Does the Problem of Mental Causation Generalize?” Proceedings of the Aristotelian Society 97: 281-297.

 

Kim, Jaegwon (1998), Mind in a Physical World, Cambridge: MIT Press.

 

Kripke, Saul (1980), Naming and Necessity, Cambridge: Harvard University press.

 

Lewis, David (1980), “Mad Pain and Martian Pain,” Readings in the Philosophy of Psychology, Vol., Ned Block, ed., Harvard University Press, pp. 216-222.

 

Martin, C.B. (1997), “On the Need for Properties: The Road to Pythagoreanism and Back,” Synthese 112: 193-231.

 

Noordhof, Paul (1998), Do Tropes Resolve the Problem of Mental Causation?” The Philosophical Quarterly 48: 221-226.

 

Robb, David (1997), “The Properties of Mental Causation,” The Philosophical Quarterly 47: 178-194.

 

Robb, David (2001), “Reply to Noordhof on Mental Causation,” The Philosophical Quarterly 51: 90-94.

 

Robb, David and Heil, John (2003), “Mental Properties,” American Philosophical Quarterly 40: 175-196.

 

Schaffer, Jonathan (2001), “The Individuation of Tropes,” Australasian Journal of Philosophy 79: 247-257.

 

Yablo, Stephen (1992), “Mental Causation,” The Philosophical Review 101: 245-280.


* Thanks are due to Dave Robb for helpful comments on an earlier draft of this paper. Thanks are also due to those in attendance when I presented a version of this paper to the Dalhousie University Philosophy Colloquium Series.

[1] See Kim (1989).

[2] See Kim (1993b, 1997), Ehring (1999), Robb (1997), and Robb and Heil (2003).

[3] Ehring (1996, 1999), Robb (1997), and Robb and Heil (2003).

[4] Noordhof (1998), p. 223.

[5] Kim (1976). Of course, on Kim’s view the relata are in effect property instances.

[6] This example comes from Braun (1995). Braun (p. 458) defends the following analysis of causal relevance: a property F of an c is causally relevant to an event e just in case (i) c is a cause of e, (ii) F is an essential property of c, and (iii) F is a “natural property.”

[7] It is worth emphasizing that the supervenience relation applies to the properties of intelligent beings, not to mental or physical events occurring within those beings. Physically indiscernible events within physically discernible intelligent beings could differ in their mental properties.

[8] Robb and Heil (2003), however, argue that properties (proper) are instances.

[9] See note #2 above for reference to advocates of this solution.

[10] Lewis’s madman (1980) is such an example. I leave it as an open question here whether the madman is in a state of pain.

[11] Of course, I am assuming that not only are mental and physical types distinct, they are moreover only contingently correlated/ co-instantiated.

[12] Note: of course, I am presupposing that in (i) the actual physical instance exists and in (ii) the actual mental instance exists. The issue of trans-world instance identity will be taken up in more detail below.

[13] See Ehring (1999) and Alward (Manuscript-a) for arguments roughly along these lines.

[14] Rob and Heil (2003) p. 175.

[15] Ehring  (2002) argues, however, the natural class trope nominalism is immune to this charge.  

[16] Rob and Heil (2003) p. 176.

[17] Kripke (1980).

[18] See Schaffer (2001) for a defense of a position along these lines.

[19] Robb (2001) p. 94 and Robb and Heil (2003) p. 193, n. 27.

[20] Noordhof (1998) p. 225. 

[21] Ehring (1999) p. 23 makes a similar point. It is worth noting that Kim (1998) p. 110 concedes this point.

[22] See Ehring (1996) pp. 461-2.

[23] See Yablo (1992).

[24] See Kim (1997, 1998).

[25] Yablo (1992) p. 252. I am just going to grant Yablo’s analysis for present purposes, but see Ehring (1996).

[26] Perhaps, following Ehring (1999, p. 23), by taking mental physical predicates to be rigid and mental predicates to be non-rigid. 

[27] But if mental universals are instantiated and distinct from physical universals, then mental and physical instances are non-identical.

[28] Much of my discussion in this section draws upon Ehring (1996). Ehring argues that although determinable instances are identical to their corresponding determinate instances, albeit only if instances are tropes, mental and physical types do not stand in the determinable-determinate relation.  And he argues that instances of types that stand in a relation of asymmetric necessitation are not identical.

[29] Kim (1997) p. 290. Strictly speaking, Kim does not so much endorse this view as attribute it to functionalists. 

[30] Kim (1997) p. 295. One might think that when Kim says that an M-instance is a P1-instance he is making a predicative rather than an identity claim. This interpretation is ruled out because (i) predication involves types and not instances and (ii) the mental property would have to be predicated of the physical property and not vice versa.

[31] Ehring (1999) p. 24 makes a similar point.

[32] Robb and Heil (2003) p. 186.

[33] Moreover, suppose that the similarity classes generated relative to a certain degree of similarity are generated as follows: all tropes similar to one another to that degree are in the same class and all tropes similar to any member of the class to that degree are in the class.  The upshot would be that the only class of tropes generated by any relation of less than exact similarity would be the class of all tropes.

[34] Briefly, my main difficulty with the former is that it renders the primitive type-generating relation mysterious. The latter option seems preferable, although one might worry that in some cases the choice of class of exactly resembling tropes relative to which membership in the larger class is defined will prove arbitrary.

[35] Robb and Heil (2003) p. 178.

[36] See, e.g., Martin (1997, pp. 213-7) for a defense of this assumption.

[37] After all, a given mental property will in principle have realizers as diverse as neuro-physiological instances, electronic instances, hydraulic instances, etc.

[38] Lewis, 1980.

[39] Robb, personal correspondence.

[40] And see my manuscript-b for difficulties for Lewis’ related attempt to accommodate mad pain.

[41] Robb, personal correspondence.

[42] Robb and Heil (2003) p.185. See also Martin (1997) p. 216.

[43] It is worth noting that a similar manoeuvre is potentially available to the universal theorist. I do not claim that much can be said on behalf of the thesis that all universals are both categorical and dispositional, but difficulties that arise for it might be expected to arise for the analogous view regarding tropes.

[44] Something has a dormative disposition just in case it produces drowsiness under a certain set of conditions, C. Since a subject that is drowsy is not alert, something that has a dormative disposition cannot produce alertness under those very same conditions, C (although it could produce alertness under distinct conditions, C’). As a result, it is a contradiction to suppose that a dormative (relative to C) disposition is identical to a stimulative (relative to C) disposition.

[45] I am assuming, of course, that the same physiological states realize functional roles across all members of a species and at all times. If this is not true, the trope theorist may have to relativize dispositions to individuals and times.

[46] But see Alward (manuscript-b).