Humean dispositionalism

Abstract

Humean metaphysics is characterized by a rejection of necessary connections between distinct existences. Dispositionalists claim that there are basic causal powers. The existence of such properties is widely held to be incompatible with the Humean rejection of necessary connections. In this paper I present a novel theory of causal powers that vindicates the dispositionalist claim that causal powers are basic, without embracing brute necessary connections. The key assumptions of the theory are that there are natural types of causal processes, and that manifestations of powers are identified with certain kinds of causal processes. From these assumptions, the modal features of powers are explained in terms of internal relations between powers themselves and the process-types in which powers are manifested.

Notes

¹Recently, Michael Fara has suggested that conditionals are not appropriate for analysis of dispositions, but he does not thereby deny the modal nature of powers and dispositions [Fara 2005.

²Stephen Mumford 2004 suggests that the dispositionalist ought to be eliminativist about laws, but he still embraces necessary, law-like truths, which appears to be enough for these purposes.

³They are therefore natural in the sense proposed by Schaffer 2004. These properties are the grounds of similarity, but they do not necessarily constitute a minimal set of properties on which all other facts supervene.

⁴Lewis's commitment to this idea is best captured in his principle of Recombination 1986b: 87 – 92].

⁵I take this account to be largely in sympathy with comments Ellis makes about natural kinds of processes 2001. For instance, when talking about chemical processes, he suggests that having a given structure ‘is both a necessary and sufficient condition for anything's being a reaction of this kind’[161]. On the other hand, Ellis sometimes manifests a wariness about reducing process-kinds to something else, while I remain open to such reduction being possible. For instance, he says that ‘dynamic universals’ (such as process-types) cannot be reduced to relations between events [75]. Prima facie, this is in tension with the thought that it is necessary and sufficient for something's being a chemical process of a given kind that it have a certain structure.

Could the external relations between the parts of a process be exclusively spatiotemporal? On Dowe's account, for instance, this appears to be the case 2000: 109]. Other accounts, such as David Fair's 1979, appear to require additional relations of energy transfer that are not themselves spatiotemporal. Here, then, is the point where the account may come into conflict with Humean Supervenience, depending upon what is required to give an adequate account of causal processes. I leave the matter open.

⁶At least, provided we do not treat structural properties as universals, these connections should not upset Humeans. See Lewis 1986a for a suitably Humean condemnation of structural universals, as opposed to other theories of structural properties. The reason Lewis objects to structural universals is that they would have to be composed in a non-mereological fashion from simpler universals.

⁷A referee has remarked that, in a sense, sodium chloride does have parts which instantiate «is hydrogen». We may adapt the point this way: Setting aside certain isotopic complications, a nucleus of an element of atomic number n has as parts nuclei of all elements with atomic number less than n. Since nuclei are arguably what strictly speaking instantiate «is hydrogen», «is sodium», etc., hydrogen is a part of sodium in a respectable sense—and there are corresponding relations between «is sodium» and «is hydrogen». The suggestion is ingenious, but it does not tell against the present analysis. I could, after all, simply give this as my example: «is NaCl» evicts «is plutonium».

⁸C. B. Martin and John Heil are sometimes characterized by critics as ‘dual-aspect theorists’ of properties, or as holding a ‘two-sided’ theory [Armstrong 1997: 83 – 4, 250]. Both appear to reject this way of formulating their view, however [Martin 1993: 184; Martin and Heil 1999: 46 – 7; Heil 2003: §11.5].

⁹I argue elsewhere [forthcoming] that basic powers are indeed extrinsic properties—in one popular sense of the term ‘extrinsic’. But while it does appear coherent to maintain that powers are both natural and extrinsic, this is a significant disadvantage of dispositionalism, in my opinion.

¹⁰This method of explaining the Meinongian character of dispositional properties has recently been championed by Aisling Crean 2005. See also Mumford 2004: 194].

¹¹It has recently been argued that there may be extrinsic dispositional properties [Fara 2005; McKitrick 2003. This is no threat to the current proposal, however, since I am merely attempting to vindicate the dispositionalist claim that there are intrinsic properties which are essentially such as to confer certain powers.

¹²Could counterfactuals analysed in a Humean-friendly fashion do the trick? No, for the Humean has counterfactuals dependent upon contingent laws of nature [Lewis 1973. And the account to be developed here aims to remain consistent with the necessitarian aspect of dispositionalism. If a property is power-conferring, it must be essentially so, not merely contingently so. So the counterfactual dependence of the effect-role upon the cause-role would have to be necessary, and it is not possible to achieve such necessary counterfactual truths on a Humean account. Thanks to Paul Noordhof for pressing me on this point.

¹³Dowe, for example, offers an answer in terms of conjunctive forks 2000: 204 – 6]. A conjunctive fork is the relationship between two events which are not probabilistically independent and a third event which screens off the probabilistic correlation between the first two events. His theory is not only implausible because of its disjunctive nature, but also unacceptable to me for the same reason that a counterfactual account is unacceptable: it requires modal facts—facts about chance in this instance—to play a primitive role.

¹⁴This process might be thought to be the manifestation of some power of the gold: the power to transmute into lead in the presence of quirky scientists with salt, perhaps. But that is not the power we are aiming to explicate, so the possibility of this process-type is not relevant to the existence of the salient power.

¹⁵For a particularly strong breed of dualist, presumably, the world contains at least two closed causal processes: the mental and the physical.

¹⁶I am hoping that this sort of conditional is not susceptible to finks [Martin 1994 and antidotes [Bird 1998, both because it is a mere might-conditional, and because it uses implicitly dispositional terms such as ‘manifest a response’ in the consequent. So it is not suitable for a reductive analysis of dispositions, but is suitable for tracking the presence or absence of a causal power. If it turns out that such a conditional fails, then that is a problem for any attempt to explicate the nature of powers; it is not a particular problem for this account.

¹⁷While working on this paper I have been assisted by comments from John Bigelow, Stephen Barker, Alan Crosier, Antony Eagle, Patrick Emerton, and two anonymous referees, as well as audiences at the University of Nottingham, the July 2005 Australasian Association of Philosophy Conference, and the Dispositions and Causes Conference at the University of Bristol in December 2005.

The research conducted for this paper has been supported by the Australian Research Council, the Australian Academy of the Humanities, and the British Academy.