Rule ordering in Korean phonology Chin- Wu Kim UNIVERSITY OF ILLINOIS IN this paper, I will examine a portion ofsegmental phonology ofKorean in some detail. 1 My discussion will primarily be confined to the sonorant consonants in Korean, especially the liquid, nasals, and glides. It will be divided into two parts, the first dealing with liquid-nasals and the second with glides. My main purpose is to examine the rules or phonological processes that affect these sonorants and to determine the interrelationship among these rules. I will be mostly concerned with establishing a linear ordering of these rules and with motivating a particular ordering as against other alternatives. In most cases, therefore, I will not attempt to justify a particular shape of underlying representation. Also, rules affecting other parts ofphonology, e.g., vowels and obstruents, will be discussed only insofar as they relate to sonorant consonants, and even then a particular form of such rules will not be given a full justification here, but it will be assumed that any reasonable description of Korean phonology must contain some such or similar form of rules. Most Korean forms will be cited in the Yale romanization without any special notation. When I depart from this, a pair of slashes / / will denote underlying representation, and a pair of brackets [ ], the surface phonetic form. To begin our discussion of the Korean liquid and nasals, first examine the following : 1. Research for this paper was supported in part by the Graduate Research Board, and Center for Asian Studies of the University of Illinois, Urbana-Champaign, and by the Center for Korean Studies, University of Hawaii. 1 2 CHIN W. KIM (1)sip-li [simni] ?? miles' payk-li [?e???] ?00 miles' phok-lak [phonnak] 'abrupt fall' yuk-lo [yurjno] 'land-road' In these examples, when a stop and a liquid / abut across a morpheme boundary, both segments turn into nasals. What is the path of derivation between the underlying form and the final pronunciation in the above case? There are three possibilities. Taking sip-li as a representative example, (2)a. /sip-li/ -» [simni] b./sip-li/ -> sim-li -» [simni] c./sip-li/ -> sip-ni -> [simni] In (a) the surface phonetic form [simni] is directly derived from the underlying /sip-li/. In (b), the morpheme final stop ? becomes a nasal m in front of /, then this / in turn becomes a nasal n. In (c) the / becomes a homorganic nasal ? after a stop, then the stop becomes a homorganic nasal. What sort of evidence would bear on the issue to let us choose one derivation over the others? We will exclude from our discussion the derivation (a), for, although there is a theoretical possibility of a rule that turns a stop-liquid sequence into a nasal-nasal sequence, we subscribe to the framework of phonology in which a phonological rule affects only a segment at a time. Looking now at the derivations (2b) and (2c), it becomes apparent that there are two rules involved in both cases : (3)i) Stop Nasalization ii) /-Nasalization The order of (i)-(ii) will give the derivation (2b), while the reverse order (ii)-(i) will give the derivation (2c). The task before us then is to determine the correct order in which the two rules should apply and to motivate such ordering. It has been generally assumed that the naturalness condition is a possible criterion for rule ordering. If we now look at the step by which the surface phonetic form is derived from the intermediate representation in (2b,c), it appears that the step is quite natural in both cases. In (2b), sim-li becomes [simni]. It is very plausible that a liquid / becomes a nasal ? after a nasal m. In (2c), sip-ni becomes [simni]. Again, it is quite natural that a stop becomes a homorganic nasal in front of another nasal. Both processes are rather commonly found in various languages. The argument, RULE ORDERING IN KOREAN PHONOLOGY3 therefore, in terms of plausibility or naturalness of the particular process involved does not seem to favor one analysis over the other. Actually, both processes are found independently in Korean. Examinethe following: (4)sam-li [samni] 'three miles' wang-lay [warjns] 'coming...