Professor Bennett freely admits that his understanding of what a "transcendental argument" is, differs from the Kantian and that he uses the term in his own — rather special — way. Although I am opposed to such terminological separatism, I shall not argue against it here. Nor shall I discuss the interesting remarks which lead up to, and which follow, what I take to be the central part of Bennett's paper, namely the characterization and exemplification of what he calls "analytic transcendental arguments". In examining them I shall be mainly concerned with their tendency to leave open deductive gaps between their premises and their conclusions, and with the difficulty of bridging these gaps in a satisfactory manner — a difficulty which they share with other transcendental arguments and which seems insuperable to me.