Theoretical implications: Against the separation of form and meaning in morphology

The findings presented have important implications for a central issue in morphological theory, the relation of form and meaning in complex words. I will argue that the properties of verb-deriving processes provide strong arguments for a sign-based model of morphology, and against theories that try to separate form and meaning in morphology (e.g. Beard 1995, Don 1993, Gussmann 1987, Szymanek 1985). It is important to note that the label 'sign-based' should not be confused with 'morpheme-based'.⁷ My use of the former term refers to the output-oriented nature of morphological processes which I have frequently pointed out, and which has recently been argued for in detail by Orgun (1996).

My central arguments against separation as they emerge from the foregoing investigation concern two points, the polysemy of individual processes and the putative synonymy of competing processes. A similar approach is taken in Booij (1986). His objections against separationism are, however, almost exclusively based on the analysis of polysemy. The implications of rival processes are only very briefly discussed in that article and summarized in the final remark that "we should ... develop a more sophisticated theory of how word-formation rules with competing affixes interact" (1986:505).

Before discussing these points in detail, I will briefly summarize the key ideas of separationist theories, referring mainly to the work of Robert Beard (e.g. 1981, 1987a, 1987b, 1988, 1990a, 1990b, 1995), who has developed the most sophisticated of these theories, so-called Lexeme-Morpheme-Base Morphology (LMBM).

¹ It may not have escaped the reader's attention that I have avoided the use of the term 'morpheme' as far as possible. The theoretical problems involved with this term are notorious and need not be repeated here.