Input and Evidence: the raw material of second language acquisition is an empirical and theoretical treatment of one of the essential components of SLA: the input to language learning mechanisms. It reviews and adds to the empirical studies showing that negative evidence (correction, feedback, repetitions, reformulations) play a role in language acquisition in addition to that played by ordinary conversation. At the same time, it embeds discussion of input within a framework which includes a serious treatment of language processing, including the problem of modularity and the question of how semantic representations can influence grammatical ones. It lays the foundation for the development of a truly explanatory theory of SLA in the form of the Autonomous Induction Theory which combines a model of induction with an interpretation of Universal Grammar, thereby permitting, for the the first time, a coherent approach to the problem of constraining induction in SLA.