Frege's double correlation thesis and Quine's set theories NF and ML (Q1076005)

The author here seeks to rehabilitate the logical notion of a set (i.e. set as the extension of a concept) as opposed to the mathematical notion (of a set as identified with its members). He modifies Frege's original logical notion by dropping the assumption that concepts can be differentiated only by having different extensions and so for example can introduce urelements other than the empty set. The bulk of a bulky paper is also taken up discussing modifications of Frege's ''Double Correlation Thesis'' (which correlates higher-level concepts ultimately with special ''objects'', the concept-correlates). The author argues that Russell's paradox can be avoided by either of two alternatives, which he contends generates systems with features similar to Quine's NF and ML. The gist of the paper is therefore to show that although these systems are not consistent with ZF set theory, and hence with the mathematical notion of a set, they have a motivation in representing this author's logical notion of set theory. The paper is long (40 pages) and includes a relaxed though readable discussion of historical views, including Quine's motivation and work in the subject, and also gives a blow-by-blow account of his own suggested conditions for avoiding Russell's paradox, and the systems which develop from these conditions.