29.	Putnam (1983) offers an argument against narrow content that, as Block (1992) notes, is in effect an argument for meaning holism for functional role theories. Very roughly, the argument is that you can have functional role twins whose environments actually differ--e.g., one's contains aluminum while the other's contains silver, yet at an early age, all their experiences are identical.  But, the holist argues, each can grow up and learn nothing but "collateral information" about the referents of their terms, and yet at some later date their concepts will differ. I.e., the claim is that at an early age they share a certain concept of a metal, but by later life one's concept has become "silver" and the other's "aluminum". Since nothing but continuous change in collateral information occurs and yet the content changes, the holist argues, there are no privileged "semantic" functional relations that determine content.<p>
	My response (which is of the same type as that of Fodor (1987), p. 94) is that there's no reason, according to my account, to think that common sense should be able to tell which environmental interactions change the structural relations of representations and thus the content. True, we can imagine that only "collateral" information is acquired, but that is only because we are largely ignorant of how learning affects cognition and its representations. When the twins' structural causal relations of their respective representations change, *that's* when the meaning changes. Perhaps with a really detailed version of the story, guesses could be made about which environmental effects lead to this, but there's no reason to suppose that common sense must supply this information.<p>