The Semantics and Pragmatics of Appearance

Introduction

Pitt and Katz (2000) propose an account of the semantics of expressions like plastic flower and kosher bacon (which I will here dub oxymoroids), maintaining that a correct analysis requires reference to senses and thus enlisting them in an attack on extensionalism. Without claiming to make a case for extensionalism in general, I nonetheless submit that there is a different and more plausible analysis of oxymoroids on the basis of which their extensions are compositionally determined in the same prosaic way as those of unproblematical cases like plastic spoon and kosher beef, and that reference to senses is unnecessary in accounting for them. The approach I have in mind is partly pragmatic in nature but involves no wild card elements—or at least none that do not have clear analogues in widely accepted treatments of other phenomena—and can be extended in natural ways to cover a variety of additional facts to which Pitt and Katz’s analysis does not apply.

1. Outline of the analysis

On the standard analysis of expressions like plastic spoon, the adjective is taken to denote a function which, for some set A ⊆ τ(N), takes the extension of a common noun α to A ∩ ⟦α⟧—where τ(X) represents the semantic type of expressions of syntactic category X. It is for this reason that the adjective in such an expression is termed intersective, alternatively conjunctive. On the analysis of oxymoroids to be offered here, the adjective is intersective in exactly the same way that it is in plastic spoon; consequently, no reference to senses is required to determine the denotation of such an expression. What is distinctive about oxymoroids of the kind so far considered is that the head noun α of the expression is taken to denote a certain superset ((α)) of ⟦α ⟧, which I call a generalized extension (‘g-extension’) of α. For example, in plastic flower qua oxymoroid, plastic denotes what it does in plastic spoon, whereas flower denotes a set containing not only flowers but also objects which, while not flowers, nonetheless resemble them in certain outward respects; the entire phrase, accordingly, denotes the intersection of this larger set with the set of things made of plastic. I will henceforth say that the noun flower in this phrase is (extensionally) generalized.

While the tenor of the discussion so far might suggest that it is necessarily the head noun of an oxymoroid that is generalized, there are also oxymoroids in which the modifier undergoes generalization—contrast, for example, glass eye (generalized head) with glass jaw (generalized modifier). ¹ Indeed, one of Pitt and Katz’s examples, rubber chicken, is ambiguous between two oxymoroidal readings, of which Pitt and Katz cite only one (on which the expression denotes rubber objects made to resemble chickens); but rubber chicken may also denote (actual) chicken meat with a rubbery taste or consistency (as in the phrase rubber chicken circuit).² A portion of a larger expression that undergoes extensional generalization will henceforth be called a g-term of that expression. Since what I have to say about oxymoroids whose heads are g-terms carries over, mutatis mutandis, to ones in which the g-term is the modifier, I limit myself here largely to consideration of the former.³

Model theoretically, there are two ways in which one could conceive of incorporating the notion of g-extension. The first would be to think of an extensional model as an ordered triple 〈D, A₁, A₂〉, where D is a nonempty set, A₁ associates each basic expression of the language with a value of the appropriate type, and A₂ associates each such expression with a value of the same type generalized in some way. (For example, if α is a common noun, then A₁ (α) ⊆ A₂ (α) ⊆ D.) Alternatively, we could suppose that expressions are assigned values relative to a pair of models, 〈D, A₁〉 and 〈D, A₂〉, where A₁ and A₂ are related as before. I leave open the question of whether there is a substantive difference between these alternatives, and, if so, which is to be preferred.

Two points of clarification; first...