This paper outlines the Autonomous Induction Theory, a theory of learning which is compatible with the hypothesis that Universal Grammar constrains the nature of natural languages. Based on Jackendoff’s architecture of the language faculty and Representational Modularity Hypothesis, it defines a trigger for the acquisition mechanisms and states a number of formal and psycholinguistic constraints on how they can operate. Finally, it argues that the theory is a necessary adjunct to purely formal learnability research.