Dispositions

1. The criterion of the dispositional

Whether all properties are dispositional, only some of them, or none of them, can only be answered once the criterion for being a disposition is settled. This has become a complicated and highly controversial issue, however.

At one point it was thought that dispositions contrasted with occurrences or events. Mainly this was through the influence of Ryle’s ideas in the philosophy of mind (Ryle 1949). Ryle tried to analyse mental phenomena in terms of dispositions to behave (see Ryle, G. §2). He claimed that to have a disposition was not to be in a state or to have something happening now. Rather, it was to be bound or liable to be in a state, or to undergo a change, when some other condition was realised. A disposition was not an occurrence but something that issued in occurrences. Someone may be a smoker but that does not entail that they are smoking now. Statements that ascribe dispositions are thus hypothetical and usually testable. The hypothetical can be articulated in conditional, ‘if …, then …’ statements.

Since Ryle, however, it has become more common to speak of dispositions being contrasted with categorical properties. This is largely because of a rejection of Ryle’s claim that a disposition was not a state. Ryle’s form of behaviourism was motivated by a desire to eliminate the myth of the Cartesian inner ego. His strategy was to show that mental ascriptions were dispositional and did not therefore involve inner states of any kind. Armstrong found this an incredible theory of mind, however, and disagreed in particular with Ryle’s account of dispositions (see Armstrong, D.M. §2). Any hypothetical that followed from a disposition ascription had to be made true by some standing condition of the person or object to which the ascription was made. Armstrong referred to this standing condition as the categorical basis of the disposition. The subsequent debate has concentrated on the putative contrast between categorical and dispositional properties.

Ryle came closest to holding what is now known as the basic conditional analysis of disposition ascriptions, which means that ascribing a disposition is asserting nothing more than the truth of a conditional. The analysis is controversial because it has proved difficult to find an exact relationship to conditionals that holds for all and only disposition ascriptions. As a criterion of the dispositional, the conditional analysis has been attacked from two opposite directions. From one direction, it has been argued that although disposition ascriptions entail conditionals, so do non-disposition, or categorical, ascriptions. From the other, it has been argued that while non-disposition ascriptions do not entail conditionals, neither do disposition ascriptions. The general problem is, therefore, that whatever relation holds between dispositions and conditionals arguably holds equally between non-dispositions and conditionals.

That all property ascriptions entail conditionals is an argument advanced by Mellor (1974), though it appears earlier in Popper (1959). The argument is that even the paradigmatic categorical property ascription, triangularity, can be shown to entail the conditional ‘if the corners were counted correctly, then the result would be three’. There have been various attempts to undermine this case by showing that the conditional differs significantly from those that are entailed by ascriptions of genuine dispositions. It has been alleged that the ‘correctly’ must refer to the result of the counting process for the conditional to be true for everything that is triangular. In that case, however, the conditional will be trivial, which those conditionals entailed by genuine disposition ascriptions are not. A response to this is that the conditional could still be true even if ‘correctly’ referred to the method of counting. Although circumstances are imaginable where the correct method of counting does not produce the correct result, because of some systematic deception for instance, the correct method of testing a genuine disposition may also fail to produce the correct result in some circumstances.

The general problem is that the conditionals that are entailed by the property ascription often need to be qualified. It might be said that any such conditionals that are entailed hold not in all but only in ideal conditions for the disposition. Hence a match is disposed to light when struck but only in the presence of oxygen. Oxygen, and any other such factors, will be part of the ideal conditions for the disposition to be manifested so any conditional associated with a disposition ascription will be true only when such conditions are assumed. This shows how difficult it is to distinguish fragility and triangularity, as there could again be ideal conditions for counting the corners of a triangle.

The problem of identifying a characteristic relation holding between disposition ascriptions and conditionals leads some to reject conditional entailment as a criterion of the dispositional. If dispositions are actual properties, then they exist independently of their manifestations and may still be there even if they are never manifested. Martin (1994) produced the notion of a ‘finkish’ disposition to illustrate this. Dispositions can be gained or lost over time. Paper, for example, can become brittle with age. This suggests the theoretical possibility of a disposition that can be lost, without manifesting, whenever it is tested. A finkish disposition is one where the test itself takes away the disposition. Something could have a disposition, therefore, though it is never manifested when tested. Alternately, there could be a reverse finkish case where a disposition is absent but is gained whenever it is tested for. The finkish cases seem to destroy the connection between dispositions and conditionals because when the disposition is possessed, the associated conditional will be false. When the disposition is not possessed, the associated conditional will be true.

Lewis (1997) objected to Martin, however, that the case defeats only the basic conditional analysis. In its place he put a reformed conditional analysis. This includes the idea of there being a causal base for the disposition that is possessed at the time it is tested and retained until at least the point that the disposition is manifested. That such a causal base is possessed can be built into the antecedent of the conditional and such a conditional would thereby always truly follow from a disposition ascription. Whether this is the case, and whether the reformed analysis is immune to all further counterexamples, remains controversial.

Molnar (2003) has argued further for a divorce of dispositions from conditionals. If dispositions are taken as real properties in their own right, which is a metaphysical assumption, then it makes perfect sense to allow a disposition that is randomly manifested or another that is continuously manifested. In these cases it is hard to find any non-trivial conditional that could be said to follow from the disposition ascription. Molnar argues that it has been an error to concentrate on conditional entailment as the distinguishing mark of the dispositional. Instead, he draws attention to directedness or intentionality as the mark of the dispositional, which Place has also supported (in Armstrong, Martin and Place 1996). Dispositions are directed towards their manifestations yet exist independently of those manifestations, on this view. A power is always a power for some specific behaviour or manifestation, for example, solubility is directed towards its manifestation of dissolving. This would mean that Brentano was wrong to think of intentionality as the mark of the mental (see Brentano, F.C. §3). Rather, it is the mark of the dispositional. Mental dispositions would still be intentional but, controversially, Molnar thinks that there is also physical intentionality in the world.