Page 37

		3.13 Lemma Scientist F identifies if and only if F identifies every text for each .

		Definition 3.12 implies that Fperiods identifies {P, O, N}, and that Ffinite identifies the collection of all finite languages. This latter collection will arise frequently in our discussion, so it is well to have a name for it.

		3.14 Definition The collection of all finite languages is denoted by .

		We thus have:

		3.15 Proposition Ffinite identifies (so is identifiable).

		Ffive identifies {W5} and no more. Indeed, every singleton collection {L} of languages is identifiable by a suitable, constant scientist. In contrast, questions about the identifiability of collections of more than one language are often nontrivial, since many such questions receive negative answers (as will be seen in Section 3.6). Such is the consequence of requiring a single scientist to determine which of several languages is inscribed in a given text.

		Constant scientists serve to highlight the liberal attitude that we have adopted about empirical discovery. The function identifies {Wm} but exhibits not the slightest "intelligence" thereby (like the man who announces an imminent earthquake every morning). Within the identification paradigm it may thus be seen that successful inquiry presupposes neither rationality nor warranted belief, but merely stable and true conjectures in the sense provided by the last three definitions. Does this liberality render identification irrelevant to human learning? The answer depends on both the domain in question and the specific criterion of rationality to hand. To take a pertinent example, normal linguistic development seems not to culminate in warranted belief in any interesting sense, since natural languages exhibit a variety of syntactic regularities that are profoundly underdetermined by the linguistic evidence available to the child (see Chomsky [42, 41]). Indeed, one might extend this argument (as does Chomsky [42]) to every nontrivial example of human learning, that is, to every situation involving a rich set of deductively interconnected beliefs to be discovered by (and not simply told to) the learner. In any such case of empirical discovery, hypothesis selection is subject to drastic underdetermination by available data, and thus selected hypotheses, however true, have little warrant. It must be admitted, however, that all of this is controversial (for an