This thesis proposes an alternative approach to the philosophy of science as applied to Government Phonology (GP), one that is concerned less with determining an ideal methodology, and more with improving the clarity and explicitness of theoretical discussion. This approach means that GP appears in this thesis not as the theory assumed, but rather as the object of study itself. For the purpose of the discussion, a distinction between a representational model of GP and more fundamental

more »

... l principles is introduced. Similarly, a distinction between an early and a late form of Generative Grammar, theory closely related to GP, is presented and argued to display a different focus on the representational model in the first case, and the theoretical background in the other. Subsequently, the methodological discussion in GP is presented and responded to, and the possibility of a logical foundation for the theory of justification is rejected. Instead, methodological theory is argued to be conventialist, tentative and goal oriented. The adverse effects for the practice of theory comparison and criticism in general are discussed, and found not to threaten the status of Generative Grammar as a scientific research activity. The final chapter provides preliminary comments to a metaphysical discussion of GP and tries to show that metaphysical assumptions form beliefs about the appropriateness of methods and theoretical assumptions. For this reason, the statement of an explicit metaphysics increases the intelligibility of a theory, while the omission of this it in theory construction increases the risk of mutual misunderstanding.