0. Introduction
0. 引言
Namgui Chang 张南基

Abstract 摘要
The history of Korean phonology of the past five centuries provides us with rich sources of diachronic studies, and there are a considerable number of insightful studies in this century. There are yet many areas where more rigorous analyses and critical reevaluations are desirable. Some basic assumptions in the theory of diachronic phonology in spite of recent advances, also, require continuing reassessment in the light of language-specific data from wider areas. Sine the advent of the generative transformational theory of grammar in the fifties, linguistic science gained a great deal of precision in analysis, but this has also created the side-effect of assuming a language to be a tightly organized body and linguistic changes as a chain of successive events that affect its organization through a series of discrete stages.

This view of linguistic structure, however, is not universally accepted. Weinreich, Labov, and Herzog (1968), for example, argue against such a view that a language is a “homogeneous object”, and presents a framework in which a language is considered “the product of combinations, alternations, or mozaics of jointly available subsystems” (Ibid: 165). Similarly, C-J. N. Bailey (1970) argues that the linguistic competence of a speaker involves something more than the knowledge of his own idiolect that he uses in his production; he must recognize, understand, evaluate, project and control variant forms, which he encounters. Bailey calls a linguistic system that enables such varieties of linguistic functions “polysystemic” or “polylectal”. According to him (Bailey 1972), linguistic changes are due to two basic mechanisms: (1) “changes resulting from the way in which children acquire their native language, which create new sub-system or lects within a given system, and (2) changes occurring through adult borrowing from other system, which may create a new system” (Ibid: 93). What is crucial here is the fundamental concept of a synchronic state at any given time as polysystemic organization containing overlaps not only of sociological or geographical dialects, but also of styles (subsystem) representing temporal progression. In this sense, a linguistic change in time may be conceived of not as a succession of System A to System B, but rather of System A to System A/B and then to System B, and since we would not think a synchronic state to have a clear beginning or ending, any state must be represented by form System A/B, namely a “ polysystemic” organization. In dealing with historical data, an analyst often encounters two or more sets of evidence which yield two or more apparently irreconcilable interpretations on one and the same linguistic progress, and if we view a synchronic state of a language to be “monosystemic”, he would then be forced to choose one solution over the other based on the relative weight of evidence, the economy of description, or even on some sense of elegance. Such examples are not hard to find in philological data. In Middle Korean (MK), for example, word initial consonant cluster (graphically so represented) may be either interpreted as such or as tense obstruents depending on which set of data one chooses to cite or ignore. The interpretation of the MK vowel system involves two contradicting sets of supporting evidence; one representing vertical alignment of the two harmonically opposing classes of vowels, and the other suggesting altered vowel relations where all vowels appear to have rotated clockwise from the earlier arrangement. Each of these, however, may be viewed as an instance of an overlap of two subsystems in transition from one state to another. The possibility and significance of diachronic processes not in a complementary (successive) relation had not been given full consideration until W. S-Y. Wang (1969) where the formal relationships between diachronic rules in various dimensions, and their effects on a synchronic state of a language were closely examined. The implementation of a sound change may be considered along several dimensions such as chronological (a period of time within which a change may take a full or partial effect), lexical (morpheme-to-morpheme), social (speaker to speaker), and phonetic (sound X to sound Y). In his subsequent article co-authored with M. Chen (1975), a wealth of cross-dialectal data in Chinese were quantitatively analyzed to demonstrate the correlation between two variables, the phonetic and the lexical. Their studies have shown that contrary to earlier beliefs a sound change may typically spread gradually across the lexicon, and often enters into a competing relationship with another, resulting in incomplete realization, which leaves residues. In this regard, particularly noteworthy are various irregularities and gaps in otherwise symmetric systems, or asymmetric distributions of features in philological data, which may reveal effects of partial intersects of diachronic forces in time.

Fifteenth century MK abounds with data, which suggest such aspects of diachronic events “caught”, so to speak, in a transitional state. The vowel system reveals both the earlier harmonic system and altered relations between vowels, the aspiration and tenseness features of obstruents show incomplete diachronic processes, and the accentual system appears to be in its last stage of transformation immediately before its extinction as a phonemic feature, as we will closely examine them in subsequent chapters. In the first chapter, we will attempt to briefly describe the fundamental characteristics of Korean phonology by sketching Morpheme Structures Conditions (MSC) and Surface Phonetic Constraints (SPC) of modern Korean as represented in the Seoul dialect. I consider these to be crucial components of phonology which functionally determine and motivate the existence and forms of phonological rules of all types. In the second chapter, the word-initial consonant clusters and the process of their reductions, which created the modern tenseness feature for obstruents, will be examined, and in the third chapter, the origin and the process of development of the aspiration feature as revealed in obstruents of MK and its subsequent cognate system. The fourth chapter will be concerned with the MK vowel system in the process of a major change, which I call “vowel rotation”, profoundly affecting the character of the MK vowel system to bring about the kind existing in modern Korean. The fifth chapter will deal with the MK prosodic system as revealed in the so-called “side dot” markings in the 15th and 16th century texts. These remarkable historical texts consistently record “tones” or pitch characteristics enabling us today to closely examine and analyze the MK prosodic system to a considerable extent. These pitch marks will be shown here as reflecting the consequences of various interactions between two prosodic subsystems in MK; one of the lexical accentual system and another of the syntactic phrase marking. Of particular interest in these areas of MK phonology is not so much in the analysis of diachronic changes from one state to another as their effects on the synchronic state of MK phonology entailing phonological variations, structural asymmetry, gaps and apparent contradictions in historical evidences.