Brabantic Sandhi and the Abstractness Debate

Abstract

From the late 1960’s onwards an increasing number of phonologists have become extremely critical of the orthodox transformational-generative practice of using abstract underlying representations in the phonologies of various languages. The most extreme case was perhaps Vennemann’s Natural Generative Grammar (cf. Hooper 1976). In my own Concrete Phonology (Goyvaerts 1975,1977) I tried to put down a number of guidelines that would invariably yield relatively concrete phonologies. Also, many contributors to a forthcoming anthology on recent work in phonology (Goyvaerts 1979a) are highly critical of abstractness in phonological descriptions. What most of these ‘concrete’ phonologists have shown is that it is possible to present for any abstract solution to a problem an alternative solution in terms of a concrete framework. So far those linguists have failed to show that concrete solutions are necessary to account for certain data.