17.	Block (1993) defends holistic inferential role semantics by claiming that while inferences specified in terms of narrow content (i.e., inferential roles) will all be analytic, wide content will delineate analytic from non-analytic inferences.  However, this burden-shifting may amount to forfeiting the game.  First, it would follow that the sort of content that philosophers have traditionally been interested in is wide, non-inferential role content- since that is consistent with traditional notions of analyticity, deductive validity, etc. And second, it would appear to follow that to the extent that cognitive scientists are concerned with reasoning (and many are) they too are interested in wide, non-inferential role content. Thus, Block seems to be defending an inferential role theory only by making it a relatively unimportant part of a theory of meaning. Moreover, there would seem to be a burden of proof on Block to show what sort of wide relations he has in mind and how it is that they can meet the concerns over holism and analyticity, i.e., how wide relations can work with inferential roles to achieve this. No such account has been offered (or promised) to date.<p>