«Of Ghostly and Mechanical Events* JONATHAN SCHAFFER University of Massachusetts, Amherst Two of the assumptions that drive most contemporary ...»
Philosophy and Phenomenological Research
Vol. LXVIII, No. 1, January 2004
Of Ghostly and Mechanical Events*
University of Massachusetts, Amherst
Two of the assumptions that drive most contemporary philosophy of mind are
(i) the naturahtic thesis that minds, like mountains and molecules, are
macro-objects in the natural order, involving nothing ‘spooky’, and (ii) a
three-part taxonomy that recognizes substance dualism, property dualism, and monism as the available metaphysical options. Together, these assump- tions drive most contemporary philosophers of mind to either monism of the materialistic stripe, or to a version of property dualism that treats mental properties on par with other macro-properties.
Paul Pietroski, in Causing Actions (2000), aims t o articulate a dualistic framework that ‘makes room for persons’. What is especially intriguing about Pietrosh’s framework is that it denies both of the above assumptions: (i) it is resolutely non-naturalistic, and (ii) it is a dualism o events, said to steer f between substance and property dualisms. If Pietroski is right then both naturalism and the three-part taxonomy are worse than mistaken: they are in conflict with our conception of persons.
In what follows I will ask (i) whether event dualism constitutes a coherent ontological position (Sl), (ii) whether the event dualist can offer a plausible conception of mental causation (52), and (iii) whether the event dualist can explain mind-body supervenience ($3). I will answer all three questions with some form of: yes, but not in the way Pietroski suggests. Though my discus- sion will be 1argely critical, I a m overall sympathetic t o a n on-reductionist framework, albeit one which is based on tropes, which is naturalistic, and which replaces a dualism of mind and nature with a multiplicity of levels of nature. So the overall thrust of my criticisms will be that there is a better way to make room for persons.
I should say that Causing Actions is a very rewarding book: it is techni- cally sophisticated, intellectually provocative, and exceptionally creative. I should also say that Causing Actions contains important discussions of action and of intensionality, which I lack space to discuss here.’ Thanks to Brian McLaughlin, Paul Noordhof, Paul Pietroski, and the editors of PPR.
1 The reader interested in these issues should look to Pietroski’s exchange with Paul Noordhof and Roland Stout (Noordhof 2001; Stout 2001; Pietroski 2001a and 2001b).
230 JONATHAN SCHAFFER
1. Event Dualism What exactly is event dualism, and does it constitute a coherent ontological position? Pietroski provides both a positive and a negative characterization.
Positive characterization: Event dualism is the thesis that mental events are distinct from neural events. So when Booth tries to shoot Lincoln, there is a mental event (which is a trying to shoot Lincoln), and a neural event (which is, perhaps, the fusion of all Booth’s associated CNS activity), and these are distinct events.
Negative characterization: Event dualism is not substance dualism, and it is more than just property dualism. Pietroski is explicit in denying substance dualism, preferring the Strawsonian view that persons are single, primitive entities with both a mental and a corporeal aspect (2000, ch. 5, $2). And Pietroski is explicit in saying that his view is more than just property dualism (2000, p. 100). After all, many of Pietroski’s ‘neuralist’ opponents, such as Jerry Fodor (1974), embrace property dualism.
What I find problematic about these characterizations is what is missing, namely (i) an account of the individuation of events, and (ii) an ontological framework in which to locate the category event, That is to say that Pietroski maintains that mental events are distinct from neural events, while providing no account of when events are distinct, and no account of what events are in the most basic sense. And, as I will now suggest, it is unclear whether there is any decent conception of event-individuation or of ontology that is compatible with event dualism.
Individuating events: There are three main proposals for event-individuation in the literature: (i) causal individuation (Donald Davidson 1 969), (ii) locational individuation (W. V. 0. Quine 1960, Davidson 1985), and (iii) fine-grained individuation (Jaegwon Kim 1973). Pietroski mentions causal individuation as plausible (2000, p. 3; also p. 95), but clearly wants t o b e neutral here and have event dualism work with most if not all of (i)-(iii) (2000, p. 3; also p. 100). But in fact event dualism is incompatible with all of (iJ-(iiiJ.
According to causal individuation, E1#E2 iff El and E2 have distinct causes or distinct effects. Causal individuation has well-known internal probl e m : it is circular given that causation presupposes distinctness of events, and it is unprincipled when conjoined with the Davidsonian idea that some cases of seemingly distinct effects are really cases of the same effect with a different ‘causal explanation’. But never mind those. The real problem here is that causal individuation, together with Pietroski’s overdeterminative account of mental causation, entails neuralism. For Pietroski’s account of mental causation postulates that mental and neural events have the same causes and effects, as can be gleaned from a simplified version of his Kim-style diagram (2000, p. 110; also p.
151, inter alia):
REVIEW ESSAYHere M is the mental event and N is the neural event, and they share the exact same stimulus cause S and behavioral effect B, so M=N.’ According to locational individuation, E1#E2 iff El and E2 occupy different locations. Locational individuation is problematically coarse, conflating the spinning with the heating of the sphere. But never mind that. The real problem here is that locational individuation, together with Pietroski’s Strawsonian view of persons, entails neuralism. For Pietroski’s account of
persons postulates that mental and neural events have the same locations:
both occur where the person is (2000, 169), so M=N.3 p.
According to fine-grained individuation, E1#E2 iff (i) O E ~ ~ O(ii) ~, E PEI#PEZ, (iii) T E ~ ~ T Ewhere OE, PE, and TE refer to the respective or Z, object(,), property(s), and time(s) of event E. Fine-grained individuation is problematically fine, severing Brutus’s stabbing of Caesar from Brutus’s violent stabbing of Caesar. But never mind that.
The real problem here is that ‘ Pietroski allows that there can be further structure within the mental and neural chains:
But this won’t help. First, MI can’t be distinguished from NI vis-a-vis causing M2 unless M2 is already distinguished from N2. There is no foothold here. Second, causal individuation via mental intermediaries would yield a strange quasi-dualism that countenances occasional identities in the off-cases of unstructured chains. In any case I think, and I think that Pietroski is committed to thinking, that the diagram should really
include internal diagonals:
Now MI can’t be distinguished from NI even vis-a-vis causing M2.
3 Pietroski (2000, ch. 5, $3.2) makes much of Hornsby’s (1981, 1997) argument from differential vagueness of location between neural and mental events. The argument is that (i) neural events occupy relatively precise locations while mental events occupy relatively vague locations, so that (ii) neural events are not identical to mental events. (The inference from (i) to (ii) presupposes that the vagueness is ‘ontological’-merely semantic or epistemic vagueness can have no implications for real identities: see Noordhof 2001, $2) But this won’t help. First, Pietroski himself is explicit that he doesn’t want the case for event dualism to turn on this argument (2001a, 53). Second, locational individuation via differential vagueness would yield a strange quasi-dualism that countenances occasional identities in the off-cases of vagueness-matching. In any case I think that Pietroski should not rely on this argument, since the notion of ontological vagueness it presupposes is obscure at best.
232 JONATHAN SCHAFFER fine-grained individuation, together with Pietroski’s Strawsonian view of persons, collapses event dualism into property dualism. For when I enjoy
both are me, a single Strawmental event M and neural event N, (i) OM=ON:
sonian person; (ii) TM=TN: both are at the same time, or at least Pietroski doesn’t want to presume that they definitely aren’t; so (iii) the claim that M+N has effectively collapsed into the claim that P M f P ~.
So pending some hrther proposal about how to individuate events, it would seem that event dualism (insofar as it is supposed to oppose neuralism and exceed property dualism) is incoherent. Perhaps Pietroski can conjure up some further proposal for individuation that would salvage event dualism, though I would regard the failure to cohere with any of the main proposals in the literature as deeply worrisome.
Ontological framework: The standard three-part taxonomy of views allows for substance dualism, property dualism, and monism. This standard taxonomy is based on the orthodox substance-universal ontology that many attribute to Aristotle. On the substance-universal ontology events must ultimately be constituted by substances and universals, so an event dualism must ultimately be a dualism about either substances or properties: there is nothing else to be dualistic about. So event dualism can only constitute a coherent alternative to substance and property dualisms o n a n on-orthodox ontological framework.
The obvious suggestion for a non-orthodox framework that allows for event dualism would be one that countenances substances, universals, and events as basic categories. This framework, however, offends against economy. When there is an event in which Sally somersaults, the substance-universa1 portion of the ontology already recognizes a substance, Sally, instantiating a property, that of being-a-somersaulter. It seems a s i f t he a dditional event p ortion i s gratuitous. Moreover, g iven that basic c ategories are independent (dependence would seem to indicate deeper ontological structure, when by supposition there isn’t any), gratuitous categories yield incoherent combinations. Here the gratuitousness of the event category yields (i) the combination in which the substance Sally instantiates the property of being-asomersaulter without there b eing any e vent o f S ally somersaulting, and (ii) the combination in which there is an event of Sally somersaulting without the substance Sally instantiating the property of being-a-somersaulter.
So perhaps a better suggestion for a non-orthodox framework that allows for event dualism would be one that countenances events but tries to reduce substances and/or universals. At this point one wants to hear more about exactly what framework this would be, and how the alleged reductions would work. Pending the development of such a framework, I wony that event dualism is indeed ‘an obscure and panicky metaphysics’.
REVIEW ESSAYI should like to conclude this section, though, by suggesting that a trope ontology allows for a certain intermediary sort of dualism (trope dualism) that has some of the flavor of event dualism, in a way that might also resolve the problems of individuation. On a trope ontology, tropes (property tokens such a s the redness o f a rose and the roundness of the moon) are taken as primitive, substances are analyzed as compresence bundles of tropes, and property types are analyzed as resemblance classes of trope^.^ The trope ontology generates a four-part taxonomy, which includes analogues of the standard trio of substance dualism, property (-type) dualism, and monism, p lus a fourth option, trope dualism, which falls b etween s ubstance and property dualisms. The analogue of substance dualism would involve distinct mental and neural tropes in distinct compresence bundles.
The new option o f t rope dualism w ould i nvolve d istinct m ental a nd neural tropes in a single compresence bundle. The analogue of property dualism would involve a single neu-t-ral trope falling under both mental and neural resemblance classes. And the analogue of monism would involve a single neu-t-ral trope in a single resemblance class. Thus the trope ontology allows for the possibility that mentation involves one substance but two particulars (two tropes), in a way that is stronger than merely countenancing two property types.
In fact some trope theorists (such as Campbell 1981, pp. 128-30) identify events with trope sequences. Given this identification, trope dualism is very much in the spirit of event dualism.
Moreover, trope dualism might resolve the problems of individuation discussed above. Elsewhere (Schaffer 2001a) I have proposed, as a plausible principle for trope individuation, that TlzT2 iff either T1 and T2 are not exactly resembling, or not co-located. If something like that is on track then trope dualism is compatible with plausible principles for trope individuation.
Now I do not know whether Pietroski would be amenable to anything like the trope ontology sketched above. I merely wish to suggest that those skeptical of substance dualism but unsatisfied by mere property dualism have a coherent option here.
2. Mental Causation Can the event dualist offer a plausible conception of mental causation?
Pietroski provides an account of causation (or at least a sufficient condition for causation) that entails that both mental events and neural events cause the
This is the now-standard trope ontology proposed by D. C. Williams (1953) and further articulated by Keith Campbell (1981, 1990). One may also combine recognition of tropes with primitivism about substance, as suggested by C. B. Martin (1980), who attributes this view to John Locke.
And so mental events are said to be (overdetermining) causes.