Locke reports, in his discussion of substance and with some amusement, on the Indian philosopher who, when asked what the earth rests on, postulated an elephant and then, when asked in turn about the elephant, decided to go with a tortoise. Locke's amusement, of course, is justified. But it is also tempered if not downright equivocal. For he sees that at some point a very special elephant or—if we stick to the Indian's story—a very special tortoise will have to be invoked. Locke's own tortoise, or elephant, substratum or substance, has some very peculiar properties or, alternatively, an equally alarming lack of them—so much so that, as he is painfully aware, it might well seem totally implausible. Enough of Locke though. Wittgenstein's problem, unlike Locke's, lies in epistemology. His elephant consists of an indefinitely large and varied set of bench-marks or paradigms—Moore's examples, in fact, along with some additions—that function as an epistemological ‘foundation’, ‘river bed’, ‘bedrock’, ‘rock bottom’, a ‘substratum’, no less. Where is the tortoise? Wittgenstein's answer to this is that there is no tortoise, none at all. A non-existent tortoise, however, is as unacceptable here as a regular common-or-garden one. Even the Indian realized that an elephant is no good without a tortoise.