Disposition ascriptions as suppositions

Dispositions involve an apparent conditionality, where this conditionality somehow relates behaviors with dispositional properties. That is, when describing dispositions, it is common to use 'if, then' type constructs. These conditional expressions usually involve the manifestation of a behavior in response to an associated stimulus. These conditional and behavioral features, while seemingly indispensable to our understanding of dispositionality have proven difficult to account for. There are well known objections to extant analyses of dispositions in terms of conditionals. What's more, any account of dispositions appealing to counterfactuals is subject to familiar objections to counterfactual conditionals and their associated metaphysical costs. The metaphysical costs associated with analyses of dispositions in terms of counterfactual conditionals are particularly unattractive to those who endorse a certain strong kind of positivist empiricism. In this thesis I explore the possibility of giving an account of dispositions in terms of Dorothy Edgington's (1986) suppositionals, rather than in terms of standard material and counterfactual conditionals which are usually involved in conditional analyses of dispositions. The result is an account that captures the apparent conditionality of dispositions and their behaviors while avoiding both the standard counterexamples to these existing analyses and their associated metaphysical costs. As such, this suppositional account of dispositions may be of interest to those who endorse a certain strong kind of positivist empiricism. The structure of the thesis is as follows: In Chapter 2, I present a selective review of the relevant literature. I outline the way in which dispositions are related to, or perhaps imply, behaviors, and describe the origin of attempts at capturing this relationship in terms of conditionals. I present this overview in a way that suggests there is an interesting analogy between attempts at understanding the conditionality of dispositions in terms of observable behaviors and behaviorist approach to mental states. I begin with Rudolf Carnap's (1936, 1956) early attempts at accounting for dispositions in observational, extensional terms and argue that while there is something obviously correct about understanding dispositions in terms of observable behaviors, this straightforward, empirically robust behaviorist approach to dispositions generates a number of serious problems, both analytic and epistemic. In Chapter 3, I set out the standard modes in which evidence of observable behavior is taken to motivate conditional accounts of dispositions. I outline the now commonly discussed Simple Conditional Analysis of dispositions; a framework that analyses dispositionality in terms of material and counterfactual conditionals. I then discuss some of the standard counterexamples to that analysis, and related analyses, in order to motivate a rejection of, or at very least a healthy suspicion of, the Simple Conditional Analysis, its variants and sophistications. I then introduce Edgington's (1986) arguments against indicative conditionals in capturing the way in which we make use of apparently conditional (if, then) structures in natural language. I demonstrate that the four assumptions that Edgington argues must be adopted by anyone who claims that indicative conditionals capture common conditional locutions are especially representative of the reasoning involved in the standard counterexamples to conditional analyses of dispositions. Given that Edgington's arguments are mostly against indicative conditionals as a means of capturing natural language conditional locutions, and that the Simple Conditional Analysis crucially departs from Carnap's analysis insofar as it incorporates a counterfactual conditional, (a semantics for which was not available to Carnap), my claim is not that Edgington's arguments constitute a good reason for rejecting the Simple Conditional Analysis. Rather, I argue that each of the four assumptions discussed by Edgington are presupposed by any account that is subject to the standard counterexamples to the Simple Conditional Analysis. If these assumptions are not made, regardless of the type of conditional involved in the analysis, the standard counterexamples will not follow. I conclude that an account not presupposing these assumptions could, by contrast, present a plausible alternative to extant analyses, such as the Simple Conditional Analysis. In Chapter 4, I advance an account of dispositions in terms of Edgington's suppositionals. In doing so I propose that a conditional analysis actually implies at least 3 distinct conditional claims. The Simple Conditional Analysis incorporates 3 elements: a stimulus condition S, a manifestation condition M, and what Bird (2009) calls a covert disposition ascription D. I argue that upon the supposition of any pairing of these three elements, an inference is licensed to the third, and that these inferences capture the bulk of our ordinary use of dispositional language. The three inferences, and so the three activities we carry out with dispositional language, are a predictive inference, an evidential inference, and an inference to the best explanation. This explains our normal reasoning involving dispositions, the intuition that dispositions involve some sort of conditionality, and the intuition that dispositions somehow involve, imply, or are implied by behaviors, without also appealing to counterfactual conditionals, and so without taking on their metaphysical costs. In Chapter 5, having presented the suppositional account, I turn from the subject of how e get from our folk practices involving dispositions, to the sorts of metaphysical commitments that beliefs in dispositions entail. I do this by presenting a variety of functionalism about dispositions. In doing so, I largely follow the sort of 'Canberra Plan' approach that is endorsed by David Lewis (1970, 1972). I argue that the ordinary platitudes, delivered by our folk understanding of physical objects and phenomena associated with a certain disposition, fix the reference of the associated dispositional concept. It is then a matter for empirical disciplines, such as the sciences or even our common interactions with the physical world, to determine whether or not the thing to which those platitudes have referred, actually exists or not. This can be understood as a sort of realizer functionalism, according to which the dispositional property is identified with its realizer, rather than its role, but differs from some other accounts insofar as the realizer is a token property, in virtue of which it plays the functional role, and that similar token properties are of the same dispositional type (whether this can actually be called functionalism is open to debate. Just as Lewis (1999, p.307) is unsure as to whether or not he is a functionalist, so too am I). In Chapter 6, I provide both a defense and an elaboration of my suppositional account, partly using Michael Fara's (2005) account of dispositions in terms of what he calls 'habituals' as something of a foil. I respond to the potential criticism that in dispensing with counterfactual conditionals, and the possible world semantics with which they are associated, my account is not able to accommodate the sorts of antecedent assumptions of 'normal conditions' which often accompany disposition ascriptions. In response, I argue that the suppositional account allows for the inclusion of whatever antecedent suppositions one wishes to include in their suppositions concerning dispositions. These antecedent suppositions will be largely determined by the beliefs that one has concerning the disposition in question, combined with relevant context and circumstances surrounding the dispositional behavior. This suppositional account has the advantage of retaining the empirical attractiveness of early, extensional, accounts of dispositions while also retaining much of the expressive power of the later counterfactual accounts. In addition, it better captures our folk understanding of and use of dispositional concepts, and avoids the standard counterexamples to extant conditional analyses of dispositions.