<<

chapter 7 The Role of Internal Reconstruction in Comparing the Accent Systems of Korean Dialects

S. Robert Ramsey

Around forty years ago, I wrote a description of an accenting dialect spoken in the village of Pukcheng1 in South Hamkyeng. (At the time, information about pitch distinctions found in North Korean speech had been completely absent in the literature.) With that synchronic description in hand, I then turned to the question of how Korean accent had developed historically. The task consisted of comparing the North Korean data from my research project with data from accenting dialects spoken in Kyengsang, and then comparing both to the “tones” recorded in the Middle Korean texts of the fifteenth and sixteenth centuries. The results of this study, a triangulation from those three points of reference (Middle Korean, Hamkyeng, Kyengsang), were first published in Ramsey 1974, and then in considerably more detail in Ramsey 1978. The task of reconstructing earlier stages of the Korean language depended upon that comparative foundation and was advanced from there principally through internal reconstruction. In this way, a number of us became convinced that the farther back in time we look, the fewer pitch and vowel length dis- tinctions we find. Through this method, it was possible to see that the pitch accent (or ) systems found in Middle Korean and the modern dialects were, for the most part, produced historically through syllable and the of vowels through syncope and/or . At the same time, those processes also produced the complex initial consonant clusters found in Mid- dle Korean, as well as many of the heavily aspirated consonants still heard in Korean today. The proto-Korean system that emerged through this process was thus markedly different in structure from both Middle Korean and all modern dialects.

1 Throughout this paper Korean is romanized using the Yale system, except for personal names for which different spellings are preferred. Note, though, that Middle Korean spellings follow the modified Yale system found in Martin 1992 and Lee and Ramsey 2011. The phonetic values of syllable pitches are indicated (roughly) by transcribing low-pitched syllables with lower- case letters, and high-pitched syllables in CAPS.

© koninklijke brill nv, leiden, 2018 | doi: 10.1163/9789004351134_008 the role of internal reconstruction 67

Both the reasoning and the data underlying these conclusions can be found in many of my own publications (e.g., Ramsey 1975, 1978, 1986, 1991), as well as in Lee and Ramsey 2011. But the model was advanced in the work of oth- ers as well. I am thinking first of all of John Whitman’s 1994 paper in which he extended my original arguments regarding verb classes to account for certain accent classes found among Middle Korean nouns. Nevertheless, Whitman sur- mised that some of the accent distinctions he had examined appear to have been original and could not be accounted for through internal reconstruc- tion. Perhaps as much as anyone, my mentor Samuel E. Martin (1992, 1996) worked to advance the paradigm and to sharpen the arguments. Fukui Rei has been a major contributor to the exploration of earlier Korean through internal reconstruction as well. In particular, Fukui’s summary of what has been done along these lines in his 2012 book on the history of Korean phonology is emi- nently readable and useful. Moreover, and still more recently, Ito Chiyuki (2013) is an exhaustive review and reworking of the theories. Both Fukui and Ito add new and compelling ideas about reconstructing earlier Korean, and both are fundamentally in agreement with me that proto-Korean had no pitch accent or tonal distinctions. But my purpose in writing the present essay is not to elaborate on these latter efforts at exploring the early phonological history of Korean. Rather, I wish to stress instead the importance of that first stage of reconstruction, the one arrived at through the comparative method applied to Middle Korean and the modern dialects, the work I began with all those decades ago. After all, the reconstructive work we’ve been engaged in since that time is based upon that initial foundation. From time to time a reaffirmation of the assumptions and conclusions we began with is called for, I think. Let’s summarize a few of the findings:

Pitch Accent in South Hamkyeng

Anyone who listens to Pukcheng speech will discover right away that differ- ences in pitch distinguish words. For example, the pronunciation of MAL i MANtha ‘(he) talks a lot’ differs from mal I MANtha ‘(he) has a lot of horses’ by the pitch of the first syllable. However, this distinction is not one of syllabic tone like Chinese, since the phonological difference between mal ‘speech’ and mal ‘horse’ can only be heard as a contrast with the pitch of a following syllable, in this case, an enclitic. And so, the pitch difference between these monosyllabic nouns is not apparent if they are pronounced in isolation.