«Abstract: Charlie Martin‟s famous examples of dispositional finks have worked as a driving force for the evolution of our understanding of ...»
FINKISH DISPOSITIONS AND CONTEXTUALISM
(forthcoming in Monist)
Abstract: Charlie Martin‟s famous examples of dispositional finks have worked as a
driving force for the evolution of our understanding of dispositional properties.
Although they have been thoroughly discussed by a number of philosophers time and
again, it is still not clear whether we can get them around simply by appealing to the
context-sensitivity of dispositions. In this regard, I will put to bed what I call the contextual strategy for Martin‟s examples by separating two distinct parts of the stimulus condition associated with the dispositional ascription. This will bring us to the conclusion that Martin‟s examples should be dealt with by articulating the nature of dispositionality involved in the dispositional ascription, not by making explicit the context‟s semantic contribution.
1. THE CONTEXTUAL RESPONSE TO MARTIN‟S EXAMPLES
I will first articulate what I mean by the context-sensitivity of dispositional ascriptions. It is old news that it is possible that two people disagree about whether or not x has disposition D but they both are intuitively correct. A homemaker truly asserts that a television set is not fragile, whilst a building worker truly asserts that the same television set is fragile. There thus seem to be two true dispositional ascriptions that appear to be mutually incompatible. The most common and compelling account of this phenomenon is to invoke the idea that the semantic value of a dispositional ascription at least in part depends on the context in which the ascription is made.
In this regard, Hawthorne and Manley‟s distinction between environmentdependence and ascriber-dependence is illuminating. The idea of environmentdependence is that the semantic value of a dispositional ascription to x is partly determined by x‟s current environment. Suppose, for instance, that at home a soft blow is common and at a construction site a very hard blow is common. Then when something is currently at home, what we mean by saying simply that it is fragile is that it is fragile for a soft blow. Meanwhile, when something is currently at a construction site, by saying simply that it is fragile we mean that it is fragile for a very hard blow. In this sense, the truth-value of „x is fragile‟ depends on x‟s current environment.
Let us turn to the idea of ascriber-dependence which is that it is not x‟s environment but the ascriber‟s conversational context that determines the semantic value of a dispositional ascription to x. When the homemaker says that x is fragile, what is meant is normally that x is fragile for a soft blow. For, she is most concerned with what would happen to x if it were struck with a soft blow, and so soft blows are most salient in her conversational context.1 Meanwhile, what the building worker means by saying that x is fragile is normally that it is fragile for a very hard blow. For, what he is interested in is what would happen to x if x were struck with a very hard blow, and hence it is very hard blows that are salient in his context. As such, the truth-value of „x is fragile‟ varies from one occasion of use to another.2 When it comes to the determination of the semantic value of a dispositional ascription, the ascriber-dependent context dominates over the environment-dependent context. The homemaker and building worker disagree about the fragility of one and the same television set. Their difference, therefore, cannot be explained away by employing the idea of environment-dependence. It can be explained away by referring to different
ascriber-dependent contexts, the homemaker‟s context and building worker‟s context:
different ascriber-dependent contexts fix different strengths of a striking force with which x is supposed to be struck, which in turn fix different semantic values of „x is fragile‟. Indeed, it is ascriber-dependence that yields semantic contextualism that philosophers like Mumford, Hawthorne, and Manley consume much of their energy in an attempt to work out (Hawthorne and Manley 2006, 182-183). With this in mind, the focus of the subsequent discussion will be on the idea of ascriber-dependence.
Having introduced the context-sensitivity of dispositional ascriptions, let me turn to Martin‟s examples. They are presented as counterexamples to the simple
conditional analysis of dispositions that provides the following analysis of fragility:
(SCA) x is fragile at t iff x would break if struck at t.
This analysis of fragility, though, seems to be falsified by Martin‟s examples of finkish disposition. One of their variants goes that a glass G is struck but does not break because it is protected by a sorcerer who detects when G is about to be struck and reacts by instantaneously casting a spell that renders G no longer fragile, and thereby aborts the process of breaking. Call this case „Guarded Glass’. In this case, G is clearly fragile but wouldn‟t break if struck, which entails that this case is a counterexample to (SCA).
The sorcerer is said to be a fink to G‟s fragility, meaning that the sorcerer would join with the event of being struck to remove G‟s fragility immediately3. This is why this problem is often called the problem of finks.
But there seems to be an easy way out of this problem that takes its cue from the context-sensitivity of dispositional ascriptions I discussed earlier. For convenience, call it „contextual strategy‟. The principal idea, which is briefly suggested by Tory Cross (2005, 324-325), is that an utterance of „x is fragile‟ has different semantic values, depending on whether the presence or absence of dispositional finks is assumed in the conversational context. When „x is fragile‟ is uttered in a context where the absence of dispositional finks is salient, its semantic value can be expressed by means of the sentence that x is fragile in the absence of the sorcerer. According to the contextual strategy, it is further suggested that this sentence is analyzed by (SCA) in the following
(SCAa) x is fragile in the absence of the sorcerer at t iff x would break if struck in the absence of the sorcerer at t.
A very different story is to be told when the presence of dispositional finks is assumed relative to the context of ascription, though. The contextual strategy has it that when „x is fragile‟ is uttered in a context where the presence of dispositional finks is salient, its semantic value can be expressed by means of the sentence that x is fragile in the presence of the sorcerer. And (SCA) provides the following analysis of this
(SCAb) x is fragile in the presence of the sorcerer at t iff x would break if struck in the presence of the sorcerer at t.
When we utter „x is fragile‟ in a context where the absence of the sorcerer is assumed, therefore, our utterance must be analyzed by (SCAa); when we utter the same thing in a context where the presence of the sorcerer is assumed, it should be analyzed by (SCAb) – for short, let‟s call the first context „non-sorcerer-context‟ and the second context „sorcerer-context‟. The two utterances of „x is fragile‟, it is thus claimed, are analyzed into different counterfactual conditionals by (SCA). This is the contextual strategy for the problem of finks.4 Let us now see how the contextual strategy gets around the problem of finks.
When we are inclined to affirm that the guarded glass G is fragile, we implicitly put ourselves in a non-sorcerer-context. For, we are normally interested to sort things in terms of what would happen to them if they were struck in the absence of dispositional finks. In such a context, though, „G is fragile‟ means that G is fragile in the absence of the sorcerer. If so, the ascription of fragility at issue must be analyzed by (SCAa) into „G would break if struck in the absence of the sorcerer‟. But this is true even in Guarded Glass. Once the contextual contribution to the semantic value of „G is fragile‟ is properly counted, therefore, the simple conditional analysis of dispositions has no difficulty with our inclination to affirm that G is fragile. What is false in Guarded Glass is the counterfactual conditional that G would break if struck in the presence of the sorcerer. On the simple conditional analysis of dispositions, however, this counterfactual conditional is equated with the sentence that G is fragile in the presence of the sorcerer, which expresses the semantic content of „G is fragile‟ uttered in a sorcerer-context. It is claimed, though, that this sentence is false, which is in line with the falsehood of the corresponding counterfactual conditional.
In view of this, there seems to be nothing problematic by the light of the simple conditional analysis of dispositions. It appears that Guarded Glass engenders trouble for the simple conditional analysis of fragility as formulated by (SCA) only because we have not given due consideration to the context‟s contribution to the semantic value of „x is fragile‟. Once the context of ascription is properly amalgamated into its semantic value, it might be thought, Guarded Glass loses its force as a counterexample to (SCA).
What Guarded Glass teaches us is merely that when we set out analyzing dispositional ascriptions into counterfactual conditionals, we need to take careful consideration of the contextual contributions to their semantic values (Cross 2005, 325).
The intuitive appeal of the contextual strategy can be reinforced by drawing an analogy with an example where a building worker truly utters „The television set is fragile‟ but it does not break after struck with a soft blow – for later references, call it Building Worker. Does this example present a problem for (SCA)? Not at all. Why? The building worker‟s utterance means that the television set is fragile for a very hard blow, which, according to (SCA), is equivalent to the counterfactual conditional that it would break if struck with a very hard blow. But this counterfactual conditional is true despite the fact that it does not break after struck with a soft blow. With the building worker‟s context of ascription on board, (SCA) has no difficulty in accommodating the truth of his utterance. Note that it is the counterfactual conditional that the television set would break if struck with a soft blow that is false in Building Worker. On (SCA), however, this counterfactual conditional is equivalent to its being fragile for a soft blow, which is not meant by the building worker‟s utterance. On balance, Building Worker is no threat to (SCA).
I take it that this line of reasoning is pretty compelling; and further that it seems to bear striking resemblance to the contextual strategy for the problem of finks. This appears to add weight to the contextual strategy leading to the conclusion that the problem of finks can be deflected by taking into consideration the ascriber-dependence of dispositional ascription. It will turn out later in this paper, though, that there is a radical difference between Building Worker and Guarded Glass, which dashes the attempt to support the contextual strategy by exploiting the analogy of them. In fact, it will come to light that, despite its promising appearance, the contextual strategy for the problem of finks does not work. Notwithstanding its falsehood, however, investigating the theoretical capacity of the contextual strategy is far from a waste of time and energy.
On the contrary, we will acquire invaluable insights into the context‟s role of determining the semantic value of a dispositional ascription by examining where it goes wrong.
It is worth remarking that the contextual strategy or something close by has been entertained, sometimes explicitly but often implicitly, by many philosophers. As mentioned earlier, Cross explicitly endorses the contextual strategy despite the fact that he does not fully elaborate its details. It is also markedly close to each of Mumford‟s strategy for the problem of finks and what Manley and Wasserman (2008) call „the strategy of getting specific‟. To illustrate my point, let‟s consider the strategy of getting specific in detail, which is developed as a means to deflect the problem of maskers, where a dispositional masker is a factor that would block the manifestation of a disposition not by eliminating the disposition but by cutting short the process from the stimulus to the manifestation.5 Its key idea, which Manley and Wasserman attribute to David Lewis, is that when I simply say that x is disposed to break when struck, I mean to attribute to x a disposition whose stimulus condition is very complex, x‟s being struck with a certain (range of) strength of striking force at a certain (range of) angle in the absence of maskers and so on. Both of the contextual strategy and the strategy of getting specific thus pick up on the observation that the stimulus condition of a dispositional ascription is much more complex than it appears to be.