Mopes, Dopes, and Tropes: A Critique of the Trope Solution to the Problem of Mental Causation
peter alward, University of Lethbridge
Abstract: A popular strategy for resolving Kim’s exclusion problem is to suggest that mental and physical property tropes are identical despite the non-identity of the mental and physical properties themselves. I argue that mental and physical tropes can be identified without losing the dispositional character of mentality only if a dual-character hypothesis regarding the intrinsic characters of tropes is endorsed. But even with this assumption, the causal efficacy of the wrong dispositions is secured.
Une stratégie populaire pour résoudre le problème de l'exclusion de Kim est de suggérer que les tropes de propriété mentale et physique soient identiques en dépit de la non-identité des propriétés mentales et physiques elles-mêmes. J'argue du fait que des tropes mentaux et physiques peuvent être identifiés sans perdre le caractère de dispositional de la mentalité seulement si une hypothèse de duel-caractère concernant les caractères intrinsèques des tropes est approuvée. Mais même avec cette prétention, l'efficacité causale des dispositions fausses est fixé.
An increasingly popular strategy for resolving Kim’s exclusion problem (cf. 1989) is to suggest that mental and physical property instances are identical despite the non-identity of the mental and physical properties themselves (Kim1993, 1997; Ehring 1999; Robb 1997; Robb and Heil 2003). In order for this strategy to succeed, an account of properties and their instances that can be reconciled with it is required. There seems to be something of an emerging consensus among advocates of the “instance identity thesis” 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 (Ehring 1996, 1999; Robb 1997; Robb and Heil 2003). What I want to argue, however, is that the appeal to tropes is ultimately unsatisfactory. It yields, at best, the causal efficacy of the dispositions that bearers of mental properties have in virtue of their specific physical constitution, rather than the efficacy of the multiply realizable dispositions required by an autonomous science of psychology.
1. The Exclusion Problem
The exclusion problem is thought to pose serious difficulties for non-reductive physicalism, as well as various forms of dualism. For simplicity, I am going to restrict my attention to views combining (i) functionalist accounts of mental properties – according to which they are multiply realizable dispositional properties – with (ii) strong supervenience – roughly, the view that the physical indiscernibility of intelligent beings entails their mental indiscernibility.1 The exclusion problem purports to demonstrate that non-reductive physicalism is incompatible with the causal efficacy of mental properties. The formulation of the problem depends on whether one takes the causal relata to be property instances or events. For present purposes, I am going to assume the latter and take the challenge of the exclusion problem to be to the causal relevance of the mental properties of mental events.2
Consider a mental event – an occurrence of pain, for example – which causes a physical event – such as the removal of a limb from a source of tissue damage. This mental cause counts as mental in virtue of the possession of a mental/dispositional property. According to the non-reductive physicalist, this property is neither identical nor otherwise reducible to any of the other physical properties the event possesses. Now the causal closure of the physical world entails that the physical properties of causes are always causally sufficient for their physical effects. As a result, any irreducible mental properties an event might have are causally superfluous.
One could, of course, argue that in any given case a mental property and a physical property are both causally responsible for the physical effect, each of which would have been sufficient in its own right. But if mental causes are uniformly and without exception over-determined then mental properties make no causal difference in the world. Any claim that they are causally relevant nonetheless is hardly plausible.
The exclusion problem can be posed as a dilemma for the non-reductive physicalist: either mental properties are reducible to physical properties, or mental properties are causally inefficacious. And neither horn is even minimally palatable for such a theorist. The first involves abandoning non-reductive physicalism in favour of reductive physicalism. And the second involves abandoning mental properties altogether, at least if a causal criterion of reality is accepted. The only alternative for the non-reductive physicalist is to find a way of splitting the horns of the dilemma, that is, to find a way of reconciling the irreducibility of mental properties with their causal efficacy.
More generally, an adequate response to the exclusion problem requires either a solution to it – finding a way of splitting the horns of the dilemma – or a defense of one or other of the horns themselves. My own view, for example, is that, despite the causal efficacy of mental events and the practical utility of mental concepts, mental properties are not themselves causally efficacious; insofar as mental properties are dispositional properties, there is no more reason to suppose that they compete for efficacy with their realizing physical properties than there is to suppose that the dormative power of a sleeping pill competes for efficacy with the physical properties of the pill.
Now what I call the “instance solution” to the problem can be understood either as a (putative) solution in a strict sense, or as a defense of the reducibility of mental properties, and a procedure for reducing them.3 The criticisms raised here, however, undermine the instance solution on either understanding, and for the same reason. Understood as a solution (in a strict sense) it yields the efficacy of the wrong property; understood as a reduction, it reduces the wrong property. For expository efficiency, I will assume the former understanding in most of what follows.
2. The Instance Solution
Underlying the instance solution to the exclusion problem is a distinction between properties – types – and instances or tokens of properties. For example, one can distinguish between the property of Blueness and the blueness of a given shirt or coffee mug. (Note: an instance of a property – the blueness of a shirt, e.g. – is distinct from a bearer of a property – a shirt, for example, which happens to be blue.) This distinction gets deployed in the instance solution as follows. Although mental properties are neither identical nor reducible to physical properties, every instance of a mental property is numerically identical to an instance of a physical property. And since mental causes just are physical causes, they do not compete with physical causes for causal efficacy and, hence, are not at risk of exclusion by them.
Consider again an occurrence of pain which causes a limb to be removed from a source of tissue damage. According to the non-reductive materialist, the disposition in virtue of which this event counts as an occurrence of pain, the “pain disposition,” is neither identical to the property which realizes it on this occasion, “C-fibre activation-hood,” nor is it reducible to any other physical property. Nevertheless the particular instance of C-fibre activation-hood is identical to the particular instance of the pain disposition. And as a result, given that the former is causally relevant to the limb movement, the latter trivially is as well.
In order for the instance identity thesis to yield a genuine solution to the exclusion problem, however, the causal efficacy of mentality cannot be purchased at the cost of the reduction (or elimination) of mental properties and their instances. And what I want to argue is that mental instances cannot be identified with physical instances without foreswearing their uniquely mental characters. Let me elaborate. Property instances have intrinsic characters or such-nesses. These characters not only play a central role in the individuation of instances (Schaffer 2001), they also play a role in the unification of distinct instances as instances of a single property: two property instances are instances of the same property if (and only if) they share an intrinsic character (as the universal theorist would have it) or have exactly (or highly) similar characters (as the trope theorist would have it). Now a uniquely mental character is a dispositional character, defined in terms of a pattern of likely causes and effects. If mental and physical property instances are identified, they will have to share a single character. And unless physicalism is eschewed, this character will have to be physical.
Consider, for example, a universalist account of properties. According to this view, properties are universals, entities which can be fully present in distinct locations and literally shared by objects. And instances are derivative entities constructed out of universals, objects, and times.4 The intrinsic character of a property instance, on this view, is given by its constituent universal. Now if the intrinsic character of a mental instance is a mental universal and the intrinsic character of a physical instance is a physical universal, then mental and physical instances cannot be identified. The identity of mental and physical instances requires that there be a single property instance with a single constituent universal. But if this universal is taken to be mental, physicalism has been abandoned. And if it is taken to be physical, the uniquely mental character of mental property instances has been lost.
3. Tropes and “Dual Character”
What remains to be determined is whether the instance solution to the exclusion problem can be salvaged by appeal to tropes. According to trope theorists, property instances are tropes – spatio-temporally located abstract particulars. The intrinsic character of a trope is particular, and so cannot literally be shared.5 Nevertheless, in virtue of their characters, tropes can stand in relations of primitive similarity to one another. Properties are classes of resembling tropes.
At first glance, it looks as though no advantage is gained by the transition from universals to tropes. After all, if the intrinsic character of a trope, which is both a mental and a physical property instance, is mental/ dispositional, physicalism has been abandoned. And if it is physical/ qualitative, mentality has been lost. Robb and Heil, however, have attempted to evade this conclusion by arguing that the intrinsic characters of tropes are at the same time both qualitative and dispositional:6
“[it] is part of our thesis that every property – every intrinsic property of a concrete object – is at once dispositional and qualitative. … A property’s dispositionality and qualitativity are not different aspects or properties of the property, they are rather the property itself, differently considered.” (Robb and Heil 2003, p.185)
As should be obvious, if the “dual-character hypothesis” can be sustained, the identification of mental and physical tropes can be reconciled with both physicalism and the dispositional character of mental property instances.7 Now although this hypothesis is far from unproblematic, I do not intend to contest it here. The reason is that even if its truth is conceded, the instance solution still runs into insuperable difficulties. This is because even though a dispositional character has been secured for mental property instances, it is ultimately the wrong dispositional character.
Consider, by way of analogy, a dose of valium which is in fact ingested by a normal human and which causes her to become drowsy. And let’s suppose that had the valium been ingested by a being with a distinct physiology, say a Venusian, it would have caused him to become wide awake and highly alert. Now the natural thing to say here is that in the actual circumstances, the valium has a dormative disposition which is realized by some of its chemical properties; and in the counterfactual circumstances, the chemical properties of the valium realize a stimulative disposition. 8 Moreover, in the actual circumstances the valium lacks the stimulative disposition and in the counterfactual circumstances it lacks the dormative disposition. After all, the dispositions of a thing are a function of the (various) system(s) in which it is embedded (which in the case at hand consists of a physiology of a certain kind).9 The valium may, of course, have a core disposition, independently of its being ingested, which explains why it has a dormative disposition when ingested by humans and a stimulative disposition when ingested by Venusians. But this core disposition is distinct from the dormative and stimulative dispositions it explains.10
Now suppose someone attempted to secure the causal efficacy of the dormative disposition of the valium against exclusion by its chemical properties by claiming that the dormative trope (whose bearer, note, is the valium) is identical to the chemical trope. This would imply that in the counterfactual circumstances the valium has both the dormative disposition and the stimulative disposition. After all, if the dispositional trope just is the chemical trope and the chemical trope exists in the counterfactual circumstances then the dispositional trope exists in those circumstances as well.11 Moreover, since by parity of reasoning the chemical trope is identical to the stimulative trope in the counterfactual circumstances, it follows that the dormative trope is identical to stimulative trope in those circumstances. And, again by parity of reasoning, the same conclusion holds in the actual circumstances. But this will not do. Not only does the valium lack a stimulative disposition in the actual circumstances (and a dormative disposition in the counterfactual circumstances), the claim that a dormative disposition is numerically identical to a stimulative disposition, and to every other disposition the valium could realize when paired with other physiologies, borders on the unintelligible.
The advocate of trope identity is faced with two options: (i) she could identify the chemical trope with a trope of the aforementioned core disposition (which explains why valium has a dormative disposition when paired with a human physiology and a stimulative disposition when paired with a Venusian physiology); or (ii) she could identify the chemical trope with a trope of a physiologically-specific disposition – the disposition to produce drowsiness in humans, for example. But both of these alternative dispositions are distinct from the dormative disposition whose efficacy the goal was to secure. Consider (a) a dose of valium ingested by a human, (b) a dose of valium ingested by a Venusian, and (c) a dose of caffeine ingested by a Venusian (which, let’s suppose, makes Venusians drowsy). Now (a) and (c) share the dormative disposition (or, as the trope theorist would put it, are similar to the requisite degree when considered dispositionally), and (b) does not. In contrast, the core disposition is shared by (a) and (b) but not (c). And the disposition to produce drowsiness in humans is possessed by (a) but neither (b) nor (c). At the end of the day, what has at best been secured is the efficacy of one of these other dispositions and not that of the dormative disposition.
Now consider a human whose C-fibres are firing and in whom C-fibre activation realizes the pain disposition. In order to generate an argument a propos mental dispositions exactly parallel to the argument just considered, I need only establish that there are possible circumstances in which the actual C-fibre trope exists but in which it realizes the pleasure disposition (or the disposition corresponding to some other functionally characterized mental state). My first inclination is to adopt a Kripkean attitude towards modality and simply stipulate such circumstances.12 But given that not everyone will find this manoeuvre persuasive, consider instead counterfactual circumstances identical to actuality up until the moment our human subject’s C-fibres begin to fire. This, I take it, suffices to establish that it is the actual C-fibre trope we are talking about. Now let’s suppose that, at the moment her C-fibres begin to fire, the subject’s brain comes to be “cross-wired” in such a way so that C-fibre activation henceforward realizes the pleasure disposition rather than the pain disposition.13 This yields counterfactual circumstances in which the actual C-fibre trope realizes the pleasure disposition rather than the pain disposition.
Again, if someone were to attempt to secure the efficacy of the pain disposition by identifying the pain trope with the C-fibre trope, she would be forced to conclude that, in both the actual and counterfactual circumstances, the pain trope is identical to the pleasure trope. As above, this follows straightaway from the identification of the C-fibre trope with the pain trope in the actual circumstances, the identification of the C-fibre trope with the pleasure trope in the counterfactual circumstances, and the fact that it is the very same C-fibre trope that exists in both set of circumstances. And given that the subject simply is neither actually in a state of pleasure nor counterfactually in a state of pain, and that the identification of a pleasure disposition with a pain disposition is probably unintelligible,14 this conclusion is untenable.
As above, the advocate of the trope solution to the exclusion problem could either (i) identify the C-fibre trope with a trope of a core disposition which explains why C-fibre activation realizes a pain disposition when it occurs in a normal human brain and a pleasure disposition when it occurs in a cross-wired brain or (ii) identify the chemical C-fibre trope with a trope of a physiologically-specific disposition, such as the disposition to produce pain responses in humans with normal brains. But again neither of these dispositions can be identified with the pain disposition and so neither option yields the efficacy of pain. Consider (a) C-fibre activation in a human with a normal brain, (b) C-fibre activation in a human with a cross-wired brain, and (c) the occurrence of whatever brain event realizes the pain disposition in a human with a cross-wired brain. Now (a) and (c) share the pain disposition (or, as the trope theorist would again have it, are similar to the requisite degree when considered dispositionally), and (b) does not. In contrast, the core disposition is shared by (a) and (b) but not (c). And the disposition to produce pain responses in humans with normal brains is possessed by (a) but neither (b) nor (c). At the end of the day, the trope solution to the exclusion problem simply bears the wrong fruit.
Note: as above it makes no difference if the trope solution is understood to yield a reduction of mental properties rather than to secure the efficacy of irreducibly mental properties. At best, what have been reduced are core dispositions – which explain why physical states realize the mental dispositions they do – or physiologically-specific dispositions. No reduction of the mental dispositions themselves has been achieved.
4. Contingent Characters
There does remain one tempting line of response to the argument presented here. One could claim that tropes have their intrinsic characters only contingently. This would enable the advocate of the instance solution to concede that the actual pain trope is identical to the counterfactual pleasure trope without being forced to conclude that the subject is actually (and counterfactually) in a state of pleasure and a state of pain. She could claim instead that the trope which has a pleasure-dispositional character (when considered dispositionally) in the counterfactual circumstances has instead a pain-dispositional character in the actual circumstances.
Now it is worth noting that this suggestion is highly counterintuitive. For example, while it is presumably true of any given blue shirt that it could have been red, it is far from clear how the blueness of the shirt could have had the character of redness. Of course, it is not obvious how much weight intuitions about unfamiliar theoretical entities like tropes should be given. More serious problems also arise, however, for the “contingent-character hypothesis.” One difficulty is that it makes questions concerning the trans-world identification of tropes more or less intractable. Given that multiple tropes can co-exist in the same spatio-temporal region, nothing except their intrinsic characters can individuate them.15 But if these characters are contingent, then we have little to guide us when faced with questions of trans-world individuation and identification.
The contingent character hypothesis also complicates attempts to identify properties with classes of resembling tropes. We can distinguish between tropes – trans-world objects – and trope-manifestations – their world-bound constituents. If tropes have essential intrinsic characters, then two tropes resemble one another to the degree their characters are similar. And there is little obvious impediment to taking properties to be classes of tropes similar to various degrees.16 But if tropes have contingent characters, then two tropes can intelligibly be said to be similar to one another to a certain degree only if they have manifestations at exactly the same worlds, each of which is similar to one another to that degree. And since distinct manifestations of a single trope can be quite dissimilar from one another, properties will have to be identified with classes of trope-manifestations, tropes proper being relegated to the sidelines.
Finally, the contingent character hypothesis requires trope theorists to make ad hoc assumptions in order to get the truth-values of certain counterfactuals to come out right. Suppose, for example, that two glasses of water and one of a highly concentrated valium solution on a tray are offered to two normal humans, Albert and Betty, and a Venusian named, Victor. Let’s suppose that (i) Albert drinks the valium solution and passes out in a chair shortly thereafter, (ii) after consuming her glass of water, Betty goes rock climbing without the benefit of safety ropes, and (iii) after consuming his glass of water, Victor falls asleep just before the climax of a movie he is watching. Now the counterfactuals
(1) If Betty had consumed the valium, she would have been injured
(2) If Victor had consumed the valium, he would seen the end of his movie
are both true in these circumstances. After all, in the closest worlds (to the one described) in which Betty consumes the valium, the valium causes her to pass out while rock climbing (without safety ropes). And in the closest worlds in which Victor – with his Venusian physiology – consumes the valium, he stays conscious throughout the movie. Moreover, there is no reason to believe that worlds in which Betty consumes the valium are any more or less close to the described circumstances than are those in which Victor consumes the valium.
Now the same trope of the chemical properties of the valium which exists in the described circumstances also exists in both worlds in which Betty consumes the valium and worlds in which Victor consumes the valium. Since this chemical trope is identical to the trope of the dormative disposition possessed by the valium in the described circumstances (assuming the instance identity thesis is correct), the dormative trope exists in both sets of counterfactual circumstances. If (1) is to be true, this trope will have to have a dormative character in worlds in which Betty consumes the valium. And if (2) is to be true, this same trope will have to have a stimulative character in the equally close worlds in which Victor consumes the valium. But if there is no independent motivation to suppose that a single trope could differ in its intrinsic character in worlds equally close to the described one, the advocate of the contingent character hypothesis will have to make an ad hoc stipulation to this effect in order to get the truth-values of (1) and (2) to come out right.
None of the worries raised here decisively undermine the contingent character hypothesis. But they do show that it comes at a non-negligible price. Trope theorists will have to decide whether it is a price worth paying to secure the efficacy of the mental.
5. Concluding Remarks
At first glance, the instance solution to the exclusion problem seems ingenious and philosophically elegant. But first impressions can be deceptive. In order to identify mental and physical property instances without eliminating the uniquely dispositional character of the former, the dual character hypothesis needs to be invoked. But unless this principle is supplemented with the equally controversial contingent character hypothesis, at best only the efficacy of non-mental dispositions is secured. But insofar as the exclusion problem poses a challenge non-reductive physicalism and the efficacy of the dispositions required in an autonomous science of psychology, this challenge is not thereby met.
Nothing I have said is meant to impugn trope metaphysics per se. The central lesson to be found here is that a solution to the exclusion problem does not simply fall out of trope theory as a philosophically free lunch. Securing the causal efficacy of the mental comes with costs which may be higher than trope theorists are, or should be, willing to pay. As a result, there is unlikely to be found in the problem of mental causation a reason to prefer a trope metaphysic over the alternatives.
 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.
2 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. This example comes from Braun (1995).
3 Kim (1998) takes the instances identity thesis to facilitate a reduction, while Robb and Heil (2003), at least as I read them, invoke it as part of a strict solution. I should note that an anonymous referee disagreed with my interpretation of the latter.
4 Instances, on the universalist picture, are superficially similar to Kim’s (1976) events. In Alward (2006), however, I argue that they differ modally: a Kimian event has its constituent property only contingently, whereas the character of a property instance is essential to it.
5 I am omitting discussion here of various nominalist accounts of trope character. See, e.g., Ehring (2002) for a discussion of such views. Following Robb and Heil (2003, p. 175), I am assuming that only an “in re” account of properties/ instances is compatible with their causal efficacy.
6 For present purposes, I will simply concede Robb and Heil’s characterization of mental properties as dispositional and of at least those physical properties which realize mental properties as qualitative.
7 It is not obvious why a universal theorist could not make a similar manoeuvre and claim that universals can be both qualitative and dispositional.
8 Of course, to say that valium has a dormative disposition when embedded in a human physiology (or a stimulative disposition when embedded in a Venusian physiology) is not to say that it always causes drowsiness (or alertness) but rather that it does so under “normal” conditions.
9 A given thing can, of course, be simultaneously embedded in multiple systems and, hence, can at a single time have a number of distinct dispositions corresponding to each of these various systems.
10 I do not mean to suggest that when the valium is ingested it ceases to have the core disposition, but rather that when it is ingested by humans it has both the core disposition and the dormative disposition, and when it is ingested by Venusians it has both the core disposition and the stimulative disposition.
11 One might, of course, argue that tropes have their intrinsic characters only contingently and conclude that a trope with a dormative character in the actual circumstances can have instead a stimulative character in counterfactual circumstances. This objection will be addressed below.
12 See Kripke, 1980. Of course, there is always the further question of whether one’s stipulations correspond to genuine possibilities.
13 Of course, if pleasure is characterized in terms of causes and effects, the subject is not, strictly speaking, in a state of pleasure in the counterfactual circumstances. Nevertheless, she is not in pain either.
14 Again I am presupposing that tropes have their intrinsic characters essentially.
15 See Shaffer (2001) for more on such issues.
16 Properties can be generated out of less than exact similarity among tropes by taking the relevant class of tropes to consist of a core class of exactly similar tropes and all other tropes similar to some specified degree to the members of the core.
2006 “Are Functional Properties Causally Potent?” Sorites, 17: 48-54.
1995 “Causally Relevant Properties.” Philosophical Perspectives, 9: 447-75.
1983 “Abstract Particulars and the Philosophy of Mind.” Australasian Journal of Philosophy, 61: 129-41
1996 “Mental Causation, Determinables, and Property Instances.” Nous, 30: 461-80.
1999 “Tropeless in
2002 “The Causal Argument against Natural Class Trope Nominalism.” Philosophical Studies, 107: 179-90.
1976 “Events as Property Exemplifications.” In Action Theory. Edited by Myles Brand and
1989 “Mechanism, Purpose, and Explanatory Exclusion.” Philosophical Perspectives, 3: 77-108.
1993 “Postscripts on Mental Causation.” In Supervenience and Mind.
1997 “Does the Problem of Mental Causation Generalize?” Proceedings of the Aristotelian Society, 97: 281-97.
in a Physical World.
1997 “On the Need for Properties: The Road to Pythagoreanism and Back.” Synthese, 112: 193-231.
1998 “Do Tropes Resolve the Problem of Mental Causation?” The Philosophical Quarterly, 48: 221-6.
1997 “The Properties of Mental Causation.” The Philosophical Quarterly, 47: 178-194.
2001 “Reply to Noordhof on Mental Causation.” The Philosophical Quarterly, 51: 90-94.
Robb, David and John Heil
2003 “Mental Properties.” American Philosophical Quarterly, 40: 175-196.
2001 “The Individuation of Tropes.” Australasian Journal of Philosophy, 79: 247-257.