Only a small number of the world's languages have any kind of recorded history over more than a few generations, and in no case do records go back more than a few thousand years. From some perspectives, this doesn't matter. There are plenty of grammars to write and plenty of changes to describe accurately and then to explain in these recorded histories. Explanations for structural changes may be grounded in grammatical theory, and careful examination of historical changes, where the goal is explanation for how and why they happened, sometimes leads to innovations in grammatical theory, illuminating the nature of grammatical categories or the conditions for movement operations, for example. That has been the focus of some work on language change and data from changes have been used to argue for claims not only about grammatical theory but also about language acquisition, that children learn only from simple structures (DEGREE-0 COMPLEX) and that acquisition is cue-based (Lightfoot 1991, 1999). That is not to say, of course, that these propositions could not have been based on other kinds of data, but the fact is that they were based on analyses of historical change. From analyses of historical changes, we have learned things about the nature of the language faculty and about how it develops in children, unhampered by the limited inventory of changes.