| ROA: | 122 |
|---|---|
| Title: | Sign-Based Morphology: a declarative theory of phonology-morphology interleaving |
| Authors: | Orhan Orgun |
| Comment: | |
| Length: | 28 |
| Abstract: | Sign-Based Morphology: a declarative theory of phonology-morphology interleaving ROA-122 28 pages signbase.ps, --.word6, --.rtf Monotonic Cyclicity ROA-123 13 pages monocyc.ps .word6 .rtf Cemil Orhan Orgun U.C. Berkeley orgun@cogsci.berkeley.edu These two papers develop a declarative approach to the phonology- morphology interface that accounts for cyclic versus noncyclic phonological effects, and relates the contrast to independently motivated morphological structures. "Monotonic Cyclicity" is the first work I know of that has argued that phonology-morphology interleaving is not necessarily derivational (Cole and Coleman (1993) have presented a CLS paper that makes a similar point). Since then, approaches to the phonology-morphology interface that attempt to capture interleaving in a fairly direct manner have become quite common (for example, recent works by Benua, Buckley, Kenstowizc, McCarthy). The approach I develop in the two papers announced here differs from these recent works in an important respect: the works cited above attempt to derive interleaving effects from essentially paradigmatic relationships between lexical entries. Sign-Based Morphology derives interleaving effects from constituent structures and feature percolation, two tools that most current theories of linguistics use (note that a paradigmatic interpretation of Sign-Based Morphology is possible though not essential, as I discuss in the paper). I argue that Sign-Based Morphology is superior to the paradigmatic approach in at least the following ways: (1) the inside-out nature of interleaving effects must be stipulated in the paradigmatic approach ("primacy of the base" in Benua 1995 and McCarthy 1995). No such stipulation is needed in SBM. (2) SBM accounts for noncyclic as well as cyclic effects, and relates them to independent morphological evidence. The paradigmatic approach has no way of dealing with noncyclic effects. (3) The paradigmatic approach stipulates that paradigmatic correspondence constraints apply only between words. I show that bound stems that are not independent words may be "cyclic domains". SBM deals successfully with such bound stems acting as "cyclic" domains, since it is the constituent structure that determines interleaving effects, not whether a given constituent is free or bound. Postscript advice: MAC users may have some difficulty with the postscript files that I have posted. If this happens, try deleting the first character of each file (which appears as a right square bracket "]" on most screens). Fonts: SBM uses SIL Doulos IPA, available free of charge from SIL (ftp.sil.org) in MAC and WINDOWS versions. The document uses the WINDOWS version, which unfortunately is not fully compatible with the MAC version. |
| Type: | Paper/tech report |
| Area/Keywords: | |
| Article: | Version 1 |