This dissertation seeks to provide an improved syntactic analysis of the phenomena known as "island constraints" and to offer a natural explanation for these and other phenomena in behavioral terms. The author approaches the problem using the notions of "nucleus" and "satellite," and shows that these concepts are a necessary part of a descriptively adequate grammatical theory. The behavioral principles seen to be at work are erroneous closure, interrupted behavior, and perceptual conflict. The author presents a reanalysis of three constraints (the complex noun-phrase constraint, the coordinate structure constraint, and a third constraint involving adverbials) and uses the behavioral principles listed above to propose explanations for the main island constraints under consideration. (Author/VM)