Vagueness-Related Uncertainty as a Special Sort ofPsychological Attitude
The study of vagueness is traditionally taken to be a branch of the philosophy of language. This is not the approach of this book: the treatment of vagueness outlined here falls squarely within the purview of epistemology.
Propositional vagueness is characterized completely in terms of the role it plays in thought—in this case, the role it plays in a theory of rational belief and desire. The important questions about propositional vagueness are epistemological ones, and the substance of the theory is given by settling those questions.The further question of whether the view outlined should be considered as an elaboration of epistemicism or supervaluationism strikes me as an unproductive one. Like the epistemicist, I accept the conclusion that there is a cutoff point for the property of being rich, and it is unknown where that cutoff is. But, as we noted, this does not distinguish the view from the supervaluationist either. If we follow through the definitions in section 2.4, it will similarly turn out that there is exactly one accurate interpretation and many admissible ones. But this is not surprising since, as we noted there, these were consequences of minimal assumptions (modulo definitions) that are shared by most classical approaches to vagueness, including the epistemicist and supervaluationist.
However, the view does make some distinctive predictions that are in tension with the orthodox versions of epistemicism and supervaluationism. We have briefly outlined some of the difficulties with reconciling the present view with a supervaluationist formalism in section 3.5 above. Unlike the epistemicist, this view maintains that vagueness is not merely a matter of being uncertain about the locations of cutoff points.
There is a bouletic element as well: you shouldn't care intrinsically about where the cutoff points lie. Moreover, uncertainty about the vague is in some sense derivative on uncertainty about the precise, as encoded by Rational Supervenience. On this sort of view, then, rational disagreements about the vague (in credence and in value) tend to boil down to disagreements about the precise. If vague matters were truly factual, as some epistemicists maintain, it's hard to see why vague beliefs and desires should depend in this way on precise ones.In some respects the nearest theory of vagueness is neither supervaluationism nor epistemicism, but an approach to vagueness that has recently been endorsed, with differing details, by Hartry Field, Stephen Schiffer, and Crispin Wright. According to these theorists there is a distinctive psychological attitude that is characteristic of taking a proposition to be borderline, and it is in the study of this attitude that the proper nature of vagueness is to be uncovered. For example, as we saw above in Field's theory, taking A to be borderline consists in being anticertain in both A and in —A. Although Field does not offer an explicit definition of vagueness in terms of this attitude, the thesis that you should be anticertain in propositions you are certain are borderline serves as a constraint on the conceptual role of the borderlineness operator, which helps distinguish the operator from one that merely expresses ignorance (as in an epistemicist theory). Different characterizations of this state are offered by Schiffer andWright, buttheyagree that thedistinctive features of vagueness aretobeexplained in terms of it.
In contrast to the present account, these theories are formulated in a non-classical setting: Field [54] and Schiffer [126] in a certain kind of contractionless logic (see section 1.2), and Wright [167] in intuitionistic logic (although an earlier version of Field's theory [53], discussed in chapter 7, is thoroughly classical). But more importantly, the view just outlined assumes a completely classical Bayesian epistemology. According to this theory, the uncertainty that vagueness generates is not a distinctive attitude. It is exactly the same sort of attitude you have when you are uncertain about where you left your keys, about the goings-on in other galaxies, and so on, and can be characterized quantitatively in the usual way, in terms of having a middling credence in a proposition.