A Theory of Classes Presupposing no Canons of Type.

W. Quine
DOI: https://doi.org/10.1073/PNAS.22.5.320
1936-05-01
Abstract:Starting with a system whose elements are classes or attributes of lowest type, it is proposed to develop the higher types by definition. Expressions purporting to denote higher-type classes emerge as properly nondenotative segments of formulae about classes of lowest type; they are incomplete symbols, defined in context. Only such locutions issue from these definitions as would be countenanced by the familiar theory of types. The theory of types itself, as an index expurgatorius, therefore becomes superfluous;' the forms which it brands meaningless are already meaningless here in the strict sense of not being defined in terms of the primitives. The system is so fashioned incidentally as not to involve use of variables or parentheses. Strictly, the intention is to repudiate classes entirely and to regard a term as denoting, not a class of objects, but the several objects themselves. The more familiar idiom used above will be retained for brevity in this synopsis, but translation of the whole into the stricter idiom presents no essential difficulty. Only the primitives and definitions will be presented here. Argument will be rested intuitively upon the meanings assigned to the signs. Consistent postulates can presumably be framed to fit these meanings, since the paradoxes which arise from confusion of logical type are inexpressible in these terms. Apart from those paradoxes, indeed, certain semantic paradoxes such as Grelling's are to be apprehended in any system which involves semantic ingredients as this one does ;2 but these paradoxes seem likewise to resist translation into terms of the present primitives. Imagine first a language, comprehensive like English, satisfying these conditions: (I) Each word ends in "b" and consists otherwise solely of "a"s. (The several letters in a word are integral parts of a simple sign.) (II) Each word is of just one of the following kinds: a term, i.e., a class name; a unitary operator, i.e., a word whose prefixture to a term produces a term; a binary operator, i.e., a word whose prefixture to two juxtaposed terms produces a term; and so on. (III) Each operator is a word, and each term which is not a word consists of an n-ary operator, for some n, followed by n terms. The grouping of letters into words needs no parenthesis or spacing for its indication in any context, tor the ends of words are marked by the "'b"s. Since each operator precedes its arguments and is fixed as to the number of admissible arguments, the significant grouping of words is likewise deter320 PROC. N. A. S.
What problem does this paper attempt to address?