On (Greek) Linguistics as a Science

Abstract

This paper explores some of the pros and cons of the proposition that if Greek linguistics (included of course in linguistics generally) is to be taken seriously as a science, then it is not enough to simply ‘confirm/corroborate/verify’ a given theory from a given subset of data. Since corroboration can never ‘prove’ a given theory, we also submit the relevant data to the criteria of falsifiability (Popper 1963). For linguistics, this proceeds by appeal to two kinds of empirical evidence, viz.: a) intuitional (including native speaker assessment surveys as well as corpora analyses), and b) instrumental and other ‘external’ (e.g. acquisition, pathology) studies. All this despite the concession that falsifications can be circumvented (cf. Lakatos 1970); so indeed this paper deals with both the few successful (paradigm-replacing) and the many circumvented (paradigm-internal) falsifications. We begin (in section 2) with falsification-successes that culminated in paradigm-shifts in earlier sciences as well as in contemporary linguistics. We then (in sections 3-4) survey intuitive and experimental cases of falsification circumvention from contemporary linguistics. There we clarify that, although there is reasonable support for abstractions, modifications and unifications, experiments have so far offered only partial or suggestive results. Sections 5-6 sum up and conclude.