For a while, it seemed that my dialogue with Van Quine—a dialogue partly real, partly fictional—had been carried as far as it could profitably be continued.¹ The salient points of this dialogue are worth summing up. Quine's old objections to modal logic were not all dispelled by the development of a genuine semantics (model theory) for modal logics, contrary to what the first full-fledged possible-world semanticists had hoped—and believed. The interpreta-tional problems Quine had so vigorously made us aware of merely seemed to settle down on a new location: on the problem of cross-identification.² Against the superficial contrary claims of Kripke, Montague, and others, I argued that we cannot take cross-identifications for granted. It does not suffice simply to postulate a domain of individuals which would be prior to the possible worlds they inhabit and each of which then would (or would not) make its appearance in any given world.³ There is every reason to think that Quine would approve of the purported conclusions of my arguments. Indeed, if I am not mistaken, Quine's arguments against modal logic preserve their sting even after their precise address is changed so that they now are directed against the possibility or at least the reasonableness of cross-identification.