In defence of causal bases

Abstract

C. B. Martin's finkish cases raise one of the most serious objections to conditional analyses of dispositions. David Lewis's reformed analysis is widely considered the most promising response to the objection. Despite its sophistication, however, the reformed analysis still provokes questions concerning its ability to handle finkish cases. They focus on the applicability of the analysis to ‘baseless’ dispositions. After sketching Martin's objection and the reformed analysis, I argue that all dispositions have causal bases which the analysis can unproblematically invoke.

Notes

¹Strictly speaking, Ellis and Lierse contend that a dispositional property entails a number of counterfactuals of the kind employed on the right-hand side of the simple conditional analysis.

²I have simplified Lewis's analysis, primarily by omitting some references to time. This should not be of any relevance in the context of the considerations that follow.

³Admittedly, definitions of the notion of causal basis given in the literature do not explicitly cite the second requirement. They mention instead a property which is ‘causally active in bringing about the manifestation of a disposition’[Malzkorn 2000: 460] or, ‘together with … the antecedent circumstances, is the causally operative sufficient condition for the manifestation’[Prior, Pargetter, and Jackson 1982: 251]. But the manner in which the definitions are employed makes clear that clause (ii) captures a strand of their meaning [ibid.: 251 – 2].

⁴Of course, even if some of the properties served as causal bases of others, they would not thereby have their own bases. I assume that the implausibility of the idea that any of the properties is capable of being a causal basis of another one makes it unnecessary to show that they cannot form a circle in which each link would be the basis of its successor.

⁵For a discussion of some other challenges, see McKitrick 2003a.

⁶The analysis is rough and ready because it is silent on the magnitude of the force and on spatial separations greater than .6 femtometre. Still, as no other force in nature has the feature cited in its analysans, it succeeds in capturing the disposition to interact strongly [cf. Young and Freedman 2000: 1389 – 90].

⁷As mentioned in note 3, although Prior et al.'s definition of a causal basis of a disposition does not specify at which world the property is supposed to operate, their argument indicates that it is the closest possible world at which the activating conditions of the disposition exist 1982: 251 – 2].

⁸As noted at the end of Section II, one can also argue that Prior et al.'s conception of causal basis itself is vulnerable to finks.

⁹B. Taylor and G. Molnar offer what seem to be sketchy versions of this argument [Taylor 1999: 157 – 8; Molnar 2003: 128 – 9].

¹⁰One may also make this point by focusing on the relationship between the causal profile of a property and laws of nature: a distinct causal profile of a property would entail a difference in the laws.

¹¹The argument that follows differs in structure from Prior, Pargetter, and Jackson's proof of the thesis that probabilistic dispositions have causal bases. They focus on rejecting a possible counterexample to the thesis 1982: 252 – 3].

¹²The leading competing suggestion about the fundamental traits of dispositions ascribes to them a kind of intentionality (see, for example, Place 1996). For a critical discussion, see Mumford 1999.

¹³This may be an oversimplification. Perhaps o would retain some extrinsic properties at w* even if objects different from those in the actual world had the appropriate properties and were related to o in the right way. As the argument to follow could accommodate this case without a significant conceptual difficulty, I will ignore it for the sake of brevity.

¹⁴It may be insufficient to reformulate the reformed analysis merely by dropping the requirement that a causal basis of a disposition be an intrinsic property. There may be cases in which conditions C and an extrinsic property of an object (or the anchors of the property) would produce event m, but we would not be inclined to assert that the object has the disposition to manifest m in C. One possible response to this worry would be to insist that the object would then in fact have the disposition, and to admit at the same time that the latter would be bizarre and useless. Another would be to impose some restrictions on the extrinsic property or, perhaps more likely, on the activating conditions (e.g. to require that they stand in an appropriately specified relation to the object).

¹⁵Lewis thinks that extrinsicality of dispositions would shield their simple conditional analysis from the threat posed by finks 1999: 137 – 9]. Consequently, if he embraced extrinsic dispositions, he would not regard the argument for the thesis that they have causal bases as critical to the question of whether or not their central tenets are captured by a version of the conditional analysis. Still, an acceptance of the argument would allow him to apply the same kind of the conditional analysis to both intrinsic and extrinsic dispositions.

¹⁶Lewis does not maintain that the reformed analysis itself enshrines the complete truth about the nature of dispositions: he thinks that it leaves open the question of whether or not dispositions are identical with their causal bases. He therefore makes a separate move when he adopts Prior, Pargetter, and Jackson's idea that dispositions are distinct from their categorical causal bases. In the process he rejects Armstrong's view that dispositions are the same as their categorical causal bases, accusing it of conflating ‘dispositional and non-dispositional properties’[Lewis 1999: 143 – 4; cf. Prior, Pargetter, and Jackson 1982: 253 – 55].

¹⁷I am grateful to John T. Roberts and two anonymous referees for very helpful comments and suggestions.