| OCR Text |
Show Chapter III. The Concept of a Genuine Preference 124 But consider what is thereby lost. There is no syntactic expression of the property, present continuously from one preference ranking to the next, that identifies each as a pairwise comparison, namely the preference relation itself. There is nothing explicitly displayed as a relational predicate that in each pairwise comparison joins chooser with alternatives ranked by it. There are no syntactic resources for distinguishing the preference relation from other triadic relations, whether extensional or intensional - even though, as we have seen in (3. 7)-(3. 10) above, the preference relation has the epistemic opacity that makes it similar to other intentional attitudes and unlike extensional triadic relations. There are no syntactic resources for making the truth functional distinctions explored in (3. 7)-(3. 10) consistent with preserving satisfaction of the transitivity requirement. There is no way of symbolizing preferences that take intentional objects of a non-pairwise comparative type. 8 And there is no way to express syntactically the truth-functional relationship between the preferred and nonpreferred alternatives of a pairwise comparison as itself the disjunctive intentional object of a preference. This seems unfortunate, since to be offered a choice between x and y would seem at the very least to be invited to choose either x or y. One would expect an adequate notation for preference to contain some resources for expressing not only this disjunction - perhaps the suggested notation could be expanded sententially to do that; but also its nonequivalence to the case of either being invited to choose x or being invited to choose y, in which either invitation expresses a sinister note of coercion. Because this suggested notation always expresses one alternative of a pair - a different alternative for each ranking - as a relational property, its smallest notational unit is a sentential proposition. In this respect it has some of the same defects as (6) in Section 3. Now I argue elsewhere that Kant believed all semantic interpretations to be finally reducible to syntax. Kant's belief may well go too far. But we at least should be able to do better than this. It is not unrealistic to expect an adequate notation for preference to expose enough syntactic substructure to distinguish between one complex intensional object of preference and two or more extensionally distinct preferences. A preference of F to G - itself a subsentential constituent of some sentence - has a syntactical substructure susceptible to analysis in terms that standardize the distinctions - and others like them - to which (3. 7)-(3. 10) call attention. Not only are there some noncomparative preferences, such as that of gentlemen for blondes. There are even some intensional objects of preference that cannot be interpreted as the outcome of a pairwise comparison. Here is an example of one that cannot: Myrtle prefers blonds. But given a series of pairwise comparisons between any particular blond and any non-blond potential suitor, Myrtle always selects the one who has the best sense of humor. 8 © Adrian Piper Research Archive Foundation Berlin |
