Linguistic meta-theory the formal and empirical conditions of acceptability of linguistic theories and descriptions

Abstract

Most linguists acknowledge, explicitly or implicitly, the relevance of epistemological questions in linguistics but relatively few have given more than a cursory, ad hoc or incomplete consideration to them. The work of one of those few, Jan Mulder, forms the starting point for much of the present discussion. Epistemological considerations arise in many contexts in linguistics and in many guises. It is an epistemological matter whenever we test the adequacy of a description or the acceptability of a theory. Epistemological considerations are latent whenever we discuss the form or the content of linguistic theories and descriptions or their interrelations. The comparison of different approaches to linguistics inevitably raises epistemological questions concerning our approach to linguistics or our presuppositions about it. These questions are of a general nature and transcend questions about particular linguistic theories and descriptions. These epistemological questions force us to consider what we take linguistics to be. In considering questions of the type mentioned we are forced, for example, to analyse what we mean by a "linguistic theory", a "linguistic description" and what phenomena we are aiming to understand. We are, furthermore, forced to analyse the constraints which a scientific attitude places upon linguistic theorising and description-building. It is these questions concerning the acceptability of linguistic theories and descriptions which we call linguistic meta-theory.

This thesis falls into five main parts. Firstly, in Chapter One, we consider the nature and scope of linguistic meta-theory. Secondly, in Chapter Two, we look at a number of previous approaches to the subject. Other important contributions are discussed as they arise in the text. Thirdly, in Chapters Three and Four, we consider in detail the major meta-theoretical distinctions in linguistics and their consequences. In particular, we distinguish linguistic theories from linguistic descriptions and discuss the nature of linguistic phenomena. The view is put forward that linguistics is a scientific subject. The meaning of this assertion is analysed and the interrelations of linguistic theories, descriptions and phenomena are considered in the light of this analysis. The main epistemological requirement that is put forward and defended is that of the empiricism of linguistics. Certain changes in our view of the philosophy of science and in our view of the form of linguistic theories and descriptions follow from the conjunction of these major meta-theoretical positions.

Fourthly, we consider the main meta-theoretical considerations concerning theories (Chapter Five) and reject a widespread view of linguistic theory as a non-empirical study (Chapter Six) and we consider the main meta-theoretical conditions relating to linguistic descriptions and some practical examples of description -building consonant with the general positions adopted in Chapter Seven. In Chapter Eight, we look at a concrete example of theory-building in the light of the meta-theoretical conditions of acceptability previously set up. We are especially concerned to show how a theory can meet the condition of being "applicable" or "indirectly scientific" through the establishment of acceptable empirical descriptions consonant with the meta-theoretical conditions on descriptions considered earlier.

The view that linguistics is a science implies that we must be concerned with the empirical testing of descriptions and, so, the fifth part of the work is devoted to methodology. In Chapter Nine, we defend the role and necessity of methodology in linguistics and set up the logical framework of relations between the methodology and theory descriptions and phenomena. In Chapter Ten, we examine two of the known types of empirical testing and their shortcomings. Finally, in Chapter Eleven, we give an example of the successful and correct application of a methodology in order to bring out the nature of empirical testing and to demonstrate its feasibility within a scientific linguistics of the sort we imagine.