CHAPTER ONE

Questions and Clues

It has been more than a century and a half since the founding of modern scientific linguistics, yet serious questions have arisen regarding the intellectual health of the discipline and its status as a science. The history of linguistics shows a succession of theoretical positions, each welcomed with hope but later abandoned in favor of the next. Today there is a growing suspicion that the leading theoretical thrust of the last fifteen or twenty years is also inadequate, and that none of the many other existing types of grammar offers an acceptable alternative.

Linguists working in what has been called the mainstream have become increasingly aware of serious theoretical difficulties in the foundations of linguistics. There is an undercurrent of uneasiness, a feeling that we may have lost our bearings. The leading body of theory has been subjected to intensive criticism, and serious disagreements have sprung up even among its defenders, impeding their efforts to establish an adequate theoretical framework. Attempts to repair perceived shortcomings have given rise to new versions of linguistic theory. These take their place among the ever-widening group of divergent points of view, and thus tend to feed the fires of controversy rather than settle the issues that stimulated their development. The resulting multiplicity of theories has been called a post-revolutionary chaos (Binnick 1981).¹

Perhaps this diversity should not be seen as disturbing. Perhaps one should argue that a choice of linguistic theory is a matter of utility, taste, opinion, or fashion like a preference for one style of clothing over another, or that it is more like religious belief or political persuasion, where people may honestly differ in their views but each covets the right to hold firm to his own credo, and in a pluralistic society freely grants to others the right to be wrong.

But this book sees linguistics rather as a scientific enterprise devoted to understanding a complex and fascinating aspect of nature: the way in which humans communicate with each other. In its search for understanding, science tries to settle controversies rather than simply tolerate them. It does this by seeking publicly available and reproducible evidence that bears on the truth or falsity of disputed points of theory, and by seeking theoretical positions that are subject to test against publicly available and reproducible evidence. In this way science tries to eliminate false conceptions and increase our secure understanding of nature. From this point of view the current trend is indeed disturbing.

With the increasing proliferation of approaches, an adequate agreed-upon body of explicit formal theory seems far beyond our grasp. One cannot have a science when no two major theoreticians can agree on the proper shape of theory. The purpose of theory is to unify the observations in a field and make them understandable by interrelating them to one another and providing insight, but this is difficult or impossible if different investigators use different theories. We are left in confusion as to where the truth might lie.

There are complaints that the leading theories do not properly address large areas of data (Weinreich, Labov, and Herzog 1968; Hymes 1974; Dingwall 1971; Derwing 1973; Gross 1979; Naro 1980; and many others). Linguists dealing with observational data face the problem that no matter which formal grammatical framework they might consider for guiding their research and presenting their results, it will be unacceptable to a part of the potential readership. Then in the end it may turn out to be inadequate for the purpose, and there is always the risk that before the fieldwork has been completed the chosen position may already have become obsolete in the eyes of its own designers. For these reasons, observational or data-oriented work is often put forward with a bare minimum of formal theory. As a result the research suffers, and with it the discipline, for observations unconnected to adequate integrating and interrelating theory are little more than a mass of unorganized facts, and thus only a feeble contribution to linguistic knowledge.

Theoretical work finds itself equally handicapped. Insights captured within one grammatical framework may be difficult or impossible to state in other frameworks which accommodate other sets of insights. When frameworks change, the insights that they have fostered may become orphaned and in danger of being lost through lack of support, or abandoned and forgotten altogether. Indeed, many valuable results of earlier eras have already met this fate.

A perceived need for theory able to confront a broader range of communicative concerns has propelled renewed exploration in the area of semiotics. But authors lament the incompatibility between semiotic theory and grammatical theory. Work in the area of pragmatics also faces difficulties in integrating its results into the rest of linguistic theory, and the same is true for work in semantics.

Perhaps the diversity of theory should be seen instead as a healthy sign—the greater the diversity, the greater the chance of someone finding a satisfactory solution. Or there may be a hope that the best from many approaches can someday be incorporated in a grand synthesis. But a quarter century has gone by since the beginning of transformational-generative grammar, which was once a great hope, and there is no sign yet of significant convergence. Instead, the diversity and eclecticism of theory are actually increasing. Every new form of grammar that is launched testifies to perceived inadequacies in earlier forms and raises the suspicion that not one of our theories is really adequate. With an abundance of inadequate theories, it is not at all clear where next to turn.

If our theories are not adequate, our results are suspect. This could have widespread consequences beyond the boundaries of linguistics itself. Consider the importance of human speech and communication in establishing and maintaining social groups, and in carrying much of our culture. Consider that linguistic concepts and results are frequently borrowed by other disciplines, such as psycholinguistics, cognitive psychology, anthropology, social psychology, sociology, human development, and studies of aphasia and communicative disorders. Consider also the importance of speech, writing, and communication in the humanities. It would seem that any question concerning the intellectual and scientific integrity of linguistics would raise issues of considerable urgency.

If one compares linguistics to physics, the contrasts between the two sciences are striking. In place of the instability of linguistic theory, physics has a large body of standard theory and results on which a large curriculum can be based that is not much different from one university to another. Unlike the situation in linguistics, where major linguists seldom agree on theory, major physicists around the world generally agree about most matters of theory.

True, there are sharp differences of opinion in certain unsettled frontier areas of research, such as the identity and structure of the elementary particles and the nature of the associated forces. However, these differences of opinion find expression in explicit hypotheses and theories, and there is frenzied activity by experimentalists to find objective evidence for testing some of the predictions of the theories so as to be able to decide among them or obtain hints leading to other more satisfactory theories.

One finds the testing of predictions in linguistics, too, but mainly the details of theory within one or another of the leading paradigms. Major differences between different paradigms are seldom the occasion for massive searches for observational evidence to be used in deciding among them. Instead, other less-conclusive arguments are brought forward.

These differences between the conduct of research in linguistics and in physics may well provide important clues to what might underlie the widely admitted problems in linguistics. If theory is unstable, if competing theories abound, if there is difficulty in choosing among different theories, there may be problems related to the criteria of truth that are used in the discipline, the methods that are employed to determine what to accept tentatively and what to reject as false or probably false. Thus it may be that we are not just having difficulties discovering the correct form of grammar. Perhaps the problems are at a more basic level common to all the approaches.

Since that is a possibility, it seems appropriate and worthwhile to step back and reexamine the foundations and basic assumptions of linguistics. If we find the foundations to be in good order, we may be reassured at least on that point, and can then move on and look elsewhere. But if they are not in good order, we might still be ahead, for our examination may already have uncovered the source of the difficulties. Perhaps then it would be clear how to repair the foundations, or how to lay new ones on which to build a stronger discipline.