Experimental Linguistics

International audienceExperimental linguistics is about studying theories of linguistic representations based on quantitative evidence. This evidence can be experimental in its strict sense or derived from text corpora. In any case, the validity of the hypotheses has to be tested using inferential statistics in order to draw general conclusions from a random sample of participants or linguistic expressions, or both. While an experimental approach has been more or less standard in phonetics, and a little more recently in phonology, it is now proving to be more and more useful in morphology, syntax, semantics, and pragmatics. Experimental linguistics evidently overlaps strongly with psycholinguistics in approaching linguistic phenomena with experimental methods, and a considerable amount of the research that has been published in psycholinguistics is central to the research questions expressed in experimental linguistics. There is however a difference with respect to the major research perspective: psycholinguistics focuses on general principles of processing, and general cognitive mechanisms such as working memory constraints or executive functions play an important role. Experimental linguistics is mostly concerned with linguistic representation and with the constraints, which license variations of linguistic expressions. Whether or not this is a useful distinction is the basis for some debate in the framework of performance grammars. A substantial relation to quantitative linguistics can also be observed, in that testing of mathematically precise models with large-scale corpora or psychological experiments both are central to experimental linguistics. While quantitative work has been at the heart of experimental phonetics for more than seventy years, it took at least twenty more years to start thinking of experimental confirmation of linguistic hypotheses in Experimental Syntax. Experimental linguistics finds some of its roots in research on the “psychological reality of grammar,” which started in the 1960s. Important background for experimental linguistics can also be found in the development of performance-oriented grammars based on variants of the strong competence hypothesis. Research methods specially adapted to experimental linguistics have been developed and refined in recent years, thus overcoming some of the obstacles for using data-intensive approaches to complex linguistic questions. There is evidently some debate on the question of in how far empirical evidence beyond the intuition of well-trained linguists is necessary and useful for the development of linguistic theories. In the end, the debate amounts to a question posed by Jerry Fodor in 1981: What is it for a linguistic theory to be true of the speakers of a language? Or more concretely: Do experiments or large-scale corpus analyses enhance the reliability of linguistic theories