The story of Korean begins with the invention of the Korean alphabet. Ever since it was introduced in 1446, the Korean alphabet has been the source of precise and detailed information about the phonological and morphological structure of the language. In that year, some three years after an announcement of its creation had been made in the dynastic annals, the reigning monarch, King Sejong, promulgated a handbook introducing the new script and explaining its use, and from that point on Korean has been a language structurally accessible to future generations of linguists. Before the alphabet, there is virtually nothing in the way of quality documentation; with the alphabet, Korean structure is laid out for us to see. (The invention, how it happened, and what we know as a result, will be discussed in detail in Chapter 5.) Thus, lucid and precise written records of the Korean language go back slightly more than five and a half centuries.
That length of time may seem ancient by most standards, but it is not particularly long on the time scale of East Asian history, or even of Korean history. Chinese writing is thought to have begun around the seventeenth century BC; and it was certainly a fully developed writing system by the fourteenth century BC. That means histories were being written and literature composed almost two thousand years before the Korean alphabet was invented. That was of course in China. But on the Korean peninsula as well, local scribes most certainly wrote in Chinese – at least soon after the Han commanderies established a presence there in 108 BC. In other words, Koreans were literate and creating histories and literature about a millennium before the beginning of the alphabetic period.
But what do such early writings tell us about the Korean language? The simple answer is, frustratingly little – at least not in a direct and easily accessible way. People on the Korean peninsula were writing in Chinese, after all. But quite naturally Koreans did attempt to record elements of their native language – first and foremost proper names – and they did so with the only writing system they knew, Chinese characters. There were two ways to use these logographs: either to approximate sounds or to suggest meanings, and Koreans experimented with both methods, often in combinations. Such writing of native words was apparently practiced in all the peninsular states during the Three Kingdoms period, and evidence of that usage can still sometimes be found in the transcriptions of place names. But it was in Silla (57? BC – 935 AD), the last of the three kingdoms to take up Chinese writing, where we see the most advanced adaptation of Chinese characters to transcribe Korean. There, the poems now known as hyangga, or ‘local songs,’ were written down in a complex interweaving of Chinese graphs, one hinting at meaning, the next one or two at sounds, then perhaps another one or two with by now obscure associations. (The method is described in Chapter 3.) The Silla system might best be compared to the man'yōgana writing of early Japanese verse. But whereas almost 5,000 man'yōgana poems from the eighth century alone are still extant, no more than 25 hyangga from all the centuries in which such verse was being composed in Korea have survived. What is more, Buddhist priests in Japan soon made annotated editions of the man'yōgana poems, with readings transcribed in katakana, and these texts, too, have survived. The differences are stark. People on the Korean peninsula began writing much earlier, and Koreans were almost surely recording words in their own language earlier as well, but far fewer vestiges of those early Korean texts remain. Inscriptional fragments from ancient Korea certainly exist. And, somehow, those fragments must once have been read with the sounds and words of a poem, say. But whatever those sounds may have been, they are not overtly recoverable by the modern reader. The corpus is too small, and the transcription method too opaque for the poems to be read without supplemental knowledge of the language. For this reason, what is known as “Old Korean” is largely a reconstruction.
Structural information from the fifteenth century is used to reconstruct all pre-alphabetic stages of Korean. That dependence is as true for “Early Middle Korean” (Chapter 4) as for “Old Korean” (Chapter 3). In both cases (and for whatever “Proto-Korean” form comparativists would reconstruct as well), the departure point is always the fifteenth-century system. Recovery of the earlier system proceeds by reconciling internal reconstruction with the philological hints found in the textual corpus.
An enduring problem in Korean historical linguistics is the question of genealogy and origin. Proposed relationships to Altaic and Japanese are the most seriously considered genetic hypotheses; Korean has been compared to Altaic for almost a hundred years, and considerably longer to Japanese. Some of this comparative work has been detailed and professional, even convincing in some cases, and we describe what we believe to be positive results of comparative research in Chapter 1, “Origins.” In doing so, we present two different approaches comparativists have taken in their efforts to prove a genetic affinity of Korean with Altaic. The first and more common approach is through the classic application of the comparative method; the second, a kind of methodological shortcut to proof that in many ways is more convincing, is by looking at specific morphological details that Korean and the Altaic languages have in common, in this case, the inflectional endings of verbs used to form nominals and modifiers. We also draw attention to what might well be the most promising avenue of research of all, the comparison of Korean to Tungusic, a family of languages considered by most comparativists to be a branch of Altaic. More than half a century ago, one of us (Lee) published a preliminary study comparing Korean to the best-known Tungusic language, Manchu. We believe the genetic relationship suggested in that work deserves renewed consideration.
Nevertheless, the answer to the question of where Korean came from is still incomplete. In order for a genetic hypothesis to be truly convincing, the proposed rules of correspondence must lead to additional, often unsuspected discoveries about the relationship. Concrete facts must emerge about the history of each language being compared in order to put the hypothesis beyond challenges to its validity, and that has so far not happened in the case of Korean. As a result, we cannot yet say with complete certainty what the origin of Korean was. Chapter 1, “Origins,” is really an essay about prehistory.
Chapter 2, “The formation of Korean,” brings the descriptions in this book into the realm of recorded history. The historical narratives described there, the earliest about language and ethnicity on the Korean peninsula, were drawn from Chinese histories and were based, at least in part, upon the first-hand reports of Han Chinese observers. In annals compiled by the Han, the Wei, and others, Chinese visitors to the peninsula recorded the names of states, the earliest being that of the legendary Chosŏn; towns and settlements; and peoples, such as the Suksin, the Puyŏ, and the Hán. They wrote down the names of exotic “Eastern Barbarian” groups, including the Puyŏ, Koguryŏ, Okchŏ, and Ye, and the so-called “Three Hán”: the Mahan, Chinhan, and Pyŏnhan; they described ethnic characteristics, such things as what the locals looked like, and what some of their customs were. All of these local words and names were transcribed in Chinese characters of course, and now, more than a millennium and a half later, the sounds and meanings that those characters were intended to represent have long since been lost. The romanized forms given for the names represent the modern Korean pronunciations of the characters and nothing more. Nevertheless, much has been made of those early descriptions. Historians and linguistic historians have scrutinized every word and phrase looking for any hint, any shred of information that could be used to solve the mysteries surrounding early life, language, and culture on the Korean peninsula.
A bit more light emerges with the rise of the first true states. In the third century, Wei ethnographers had found only tribal confederations, but by the fourth century, wars and political alliances had brought about a coalescence of those groups into what were undeniably nation-states. They included, among others, the powerful northern state of Puyŏ and in the south, Kaya, or Mimana, as it is usually called in Japanese annals. But the best-known states to emerge around that time were Koguryŏ, Paekche, and Silla, the “three kingdoms” of what later became known as the Three Kingdoms period. Koguryŏ, Paekche, and Silla were also the first states to arise on the Korean peninsula for which linguistic evidence still exists. Japanese annals contain a few hints as to names and terms used in those kingdoms, but most of the lexical information comes from place names recorded in the Samguk sagi, a Koryŏ-period history from 1145 compiled out of older peninsular histories and records long since lost. How linguistic information is gleaned from that source is described in some detail in Chapter 2.
Out of those lexical fragments we build a case that what was spoken in the three kingdoms were different but closely related languages. To be sure, many controversies remain, both about that issue and about the Samguk sagi place names, particularly those found on Koguryŏ territory. We discuss some of the controversies; we show that Koguryŏ place names in particular have transcriptional characteristics that distinctively mark them as Koguryŏan.
Finally, we describe why it was the Silla language that should properly be referred to as “Old Korean.” It was Silla that effected a linguistic unification of Korea, and its speech, through military conquest and political consolidation, was the language form that eventually became the lingua franca of the entire peninsula. In that way, Sillan gave rise to Middle Korean, and is thus the direct ancestor of the language spoken throughout Korea today.
Each subsequent chapter after Chapter 2 deals with a separate period in the history and development of Korean. And although those chapters, five in all, differ greatly in detail and length, all have the same narrative structure. Each begins with a description of the historical and cultural background. The literature of each period is then listed and described, along with the script(s) used to write it. Finally, the description of each language stage is organized into the details of phonology, morphology, syntax, and lexicon.
The first known stage of Korean, “Old Korean,” is described in Chapter 3. As mentioned earlier, Sillan literati wrote in Classical Chinese, but some apparently made incipient efforts to transcribe native literature as well. All we know about such literary efforts, however, comes from much later historical records mentioning compilations of hyangga, and, of course, from the twenty-five examples of such verse that are still extant. But poems are not the only sources of linguistic information from the Old Korean period. A much more common traditional method of writing Korean was the scribal technique known as idu, the use of which goes back to the Three Kingdoms period. While mostly used for annotating Chinese texts, and providing little in the way of phonological information, idu does contain some useable information about early Korean. Both transcription systems, idu as well as the “hyangch'al” method of writing hyangga, are explained in some detail in Chapter 3. Besides idu and hyangch'al, there are also phonogramic transcriptions of Korean names; Chinese transcriptions of Korean words, loanwords into Japanese; and, finally, the information that can be surmised from the traditional Sino-Korean readings of Chinese characters, which were imported into Korea during the Three Kingdoms period.
None of these Old Korean sources is sufficient to establish its phonological system in any detail, however. The best they can be used for is to determine a few general characteristics of the system. In a word, Old Korean is reconstructed by using such philological information as reference points and triangulating from Middle Korean.
For Old Korean grammar, idu and hyangch'al provide information about the use and morphology of some particles and verb endings. There are hints about first- and second-person pronouns.
Two important lexical facts emerge from Old Korean attestations. The first observation to be made is that most of the Silla words found in extant sources correspond to reflexes in the vocabulary of Middle Korean. These correspondences are significant, because they help confirm the identification of Sillan as Old Korean. The second fact to be learned is how the growing influence of Chinese civilization affected the Korean lexicon. For the most part, Sinitic importations into Silla usage were not loanwords per se, but rather vocabulary derived from the codified readings of rime tables and dictionaries. These readings were passed down without significant additional input from China to become the traditional “Eastern Sounds” used in Middle Korean texts. As a result, the Silla readings of Chinese characters were the sources of Sino-Korean readings used today.
The term “Middle Korean” (MK) usually refers to the language of the alphabetic documents of the fifteenth and sixteenth centuries, and that is how we use it as well when the reference is clear. However, the usage can also be misleading. The language itself did not abruptly change when the alphabet was invented; instead, the linguistic period that Middle Korean represents appears to have actually begun around 500 years earlier, in the tenth century, when the capital was moved from the southeast to the middle of the peninsula. For this reason, we call the earlier centuries of the Koryŏ period “Early Middle Korean”; and, when clarity demands it, we call the language of the fifteenth and sixteenth centuries “Late Middle Korean” (LMK).
The Early Middle Korean period (Chapter 4) began when the Koryŏ established a new government and moved the geographic base for the language away from the old Silla capital. From the fragmentary evidence available to us, it appears that Koguryŏ substrata still existed in local speech at that time, but such traces gradually faded over the centuries as the Sillan-based language continued to exert its influence.
In this pre-alphabetic period, attestations of the language are hard to come by and difficult to interpret, just as they are for Old Korean. There are two important sources of phonological information about Early Middle Korean, however. The first is a vocabulary list compiled by a Chinese visitor to the Koryŏ capital in the early twelfth century, the Jīlín lèishì, or, as it is known in Korea, the Kyerim yusa. The Korean words on that list are transcribed impressionistically with Chinese characters used as phonograms, and their interpretation poses many challenges to historical reconstruction. Still, combined with internal reconstruction from the fifteenth-century system, the Jīlín lèishì evidence is a valuable phonological resource. The second Koryŏ-period resource is the thirteenth-century medical treatise, Hyangyak kugŭppang. Unlike the Jīlín lèishì, that medical compilation is a native work that contains the local names for plants and other products used in herbal cures. Though these Korean words are only occasionally written phonetically using phonograms, the transcriptions reflect an older Korean convention and are systematic enough to approach a kind of rudimentary syllabary. Philologists speculate that if the corpus were larger, the Hyangyak kugŭppang might reveal a fuller picture of Early Middle Korean phonological structure.
Another resource that must be mentioned is that of loanwords. Through Yuan-dynasty China, Koreans borrowed a number of terms from Mongolian, and these words provide information about the sounds of Korean at the time.
There is also one more important resource for Early Middle Korean: interlinear annotations of Chinese texts. In the Koryŏ period, there were two different ways of elucidating texts, both of which were unobtrusive almost to the point of being invisible. The first used simplified Chinese characters known as kugyŏl that were written by hand between the lines of Chinese; these markings were discovered in the 1990s. (Kugyŏl use and structure are illustrated in Chapter 4, with comparisons to hyangch'al and Japanese katakana.) The second marking method did not involve writing at all; it consisted of making tiny dots and angled lines with a stylus. Known as kakp'il, these marks are truly bordering on invisible; they were discovered only in 2000 with the help of a strong angled light – and, of course, sharp eyes. Both kugyŏl and kakp'il are generating considerable excitement among philologists and linguists for the information they potentially reveal about the use of particles and other grammatical markers. The final story of this linguistic resource has still to be written.
Since Early Middle Korean is less distant in time from the fifteenth century, more of its phonological system is evident from internal reconstruction than that of Old Korean is. Combined with philological clues, the method reveals something of how clusters and aspirates seen in the fifteenth century had developed through vowel syncope. There was also, we believe, a “Korean Vowel Shift” that took place between the thirteenth and fifteenth centuries; the principal evidence for the timing of the shift comes from Mongolian loanwords.
The lexical sources for Early Middle Korean show evidence of native vocabulary since lost, some of which was evidently displaced by Sinitic vocabulary. Loanwords from Mongolian and Jurchen, which were surely borrowed during the Early Middle Korean period, lingered into the alphabetic period.
As we have said, Late Middle Korean (Chapter 5) was the language's most important historical period. Its texts are consistent and phonologically precise, the textual corpus rich and voluminous. Its transcriptions record segmentals and suprasegmentals; the symbols incorporate articulatory features; spellings are standardized. For both phonological and morphological information, this textual record is unsurpassed anywhere in the premodern world. Syntax and stylistics, however, are not of the same quality. Since most writings were pedagogical interpretations of Chinese texts, they were often stylized and stilted. Philologists believe the syntax of these texts did not always represent natural, idiomatic Korean.
We try to present a reasonably exhaustive list of the many texts of the period, first by century, then by the reign period and year, describing their features, what copies are extant and where they are located. Since the nature of the writing system critically affects analyses, considerable space is devoted to describing the alphabet, Hangul, its orthographic conventions, the philological issues around its early history, and the transcription of Sino-Korean.
Linguistic structure is treated in far more detail in Chapter 5 than in any other part of the book. We pay particular attention to phonology and morphology. Over the past century and more, the phonological system of Middle Korean has been the focus of intensive research; and the findings of that research are presented in Chapter 5 together with new interpretations. We bring in comparative information from modern dialect reflexes. Morphology, too, is described in detail. In treating syntax, we have focused on ways in which fifteenth-century structure differed from that of today's language.
Early Modern Korean (Chapter 6) formed a transition between Middle Korean and Contemporary Korean. That stage is reflected in texts written between the beginning of the seventeenth century and the end of the nineteenth.
Unlike the literature of the Middle Korean period (or, of course, that of the twentieth century), writings of the Early Modern period were relatively unconstrained by convention and spelling practices. The Imjin Wars at the end of the sixteenth century, followed by disease and famine, had disrupted the social order underlying writing conventions, and ongoing changes that had long been masked by standard writing practices suddenly appeared. The textual record was different from Middle Korean in other ways as well. In addition to official government publications both new and reissued, the Early Modern corpus included such genres of literature as new types of sijo poetry, literary diaries, and, most important and popular of all, vernacular novels.
During this unstandardized period, variant spellings and transcriptional mistakes were extremely common, and it is mainly from this kind of evidence that linguistic changes have been documented. Among the most salient phonological changes the language underwent were the spread of reinforcement and aspiration, palatalization (and spirantization), the loss of the vowel /o/, monophthongization, and the erosion of vowel harmony. In its grammar, the language showed a tendency toward structural simplification in both verbal and nominal morphology. A more natural syntax and style can be seen in the Early Modern period. In the lexicon, native vocabulary continued to be lost and replaced by Sinitic words and expressions, as well as by Western words making their way into Korea through China.
“Contemporary Korean” (Chapter 7) is a description of how Korean emerged from its traditional forms to become the modern world language spoken and written in South Korea today. It begins with the script reforms of the late nineteenth century during the “enlightenment period” and the establishment of orthographic standards in 1933. These early script reforms revealed changes in the language that had long since taken place. But shifts have also taken place since the nineteenth century. The most noticeable of these more recent changes have been in the lexicon, of course; after all, Korea has become integrated into virtually every aspect of modern world culture, from economics and politics to technology to pop media, and new words are very much at the heart of these changes, as they are of what is so enthusiastically called “globalization.” But phonology and morphology have also not remained static. In this last chapter we try to document the most salient of those changes, both those that the script reforms revealed, and those that resulted later from powerful social and economic forces.
In writing this volume, we have tried to summarize what is known to date about the history of Korean. It is based upon an earlier work, Kugŏ-sa kaesŏl (‘An Introduction to the History of Korean’), originally written by one of us (Lee) and published in its first edition in 1961. That book was subsequently reissued in a completely revised edition in 1972, later reworked and enlarged numerous times, and today it is widely used as a textbook in language and literature departments in many Korean universities. In 1975 the book was translated into Japanese by Fujimoto Yukio, and in 1977 into German by Bruno Lewin. The present work is different in both presentation and structure from those translations, however. Kugŏ-sa kaesŏl was written for students studying the history of their native language, and a translation involving Korea's textual philology unavoidably confronts problems of cultural transferability difficult to surmount. As a result, we set out from the beginning to produce a different kind of work, one aimed at making the history of Korean more straightforward for, and at least a little more accessible to, an English-language readership. That goal turned out not to be a simple undertaking. One of us (Ramsey) spent a number of years working on the manuscript, consulting all the while with the other (Lee). In the end, significant changes have been made in both content and expression. Some conclusions about earlier Korean have been revised as well.
We have added considerably more detail about the phonology and morphology of Late Middle Korean, as well as inferences to be made from internal reconstruction within those systems. Although the amount of print space in Lee's original book devoted to that stage of the language was nearly as great as that used for all the other stages of the language combined, still more attention was called for, we thought, especially in addressing a Western readership unfamiliar with the alphabetic documents of the fifteenth and sixteenth centuries and their unparalleled linguistic significance.
A minor difficulty with periodization was deciding what to call the two stages of the language that followed Middle Korean. In most English-language publications, “Modern Korean” refers only to what was spoken between the seventeenth and nineteenth centuries, while what's spoken today is “Contemporary Korean.” We find that convention confusing. It's difficult to get used to talking about a “modern language” when it hasn't been spoken in over a hundred years. For this reason, we decided to call that earlier stage “Early Modern Korean” instead.
The philology presented choices. In Kugŏ-sa kaesŏl a separate chapter was devoted to a summary of the various kinds of writing systems that have historically been used in Korea. In this work, however, each type of writing is described separately, together with the stage of the language when it was employed. For example, descriptions of how Chinese characters were used to transcribe Korean can be found in the chapter on Old Korean; the structure of the early alphabet appears in the chapter on Late Middle Korean; etc.
No one system of romanization fits every purpose. To write Korean names and general terms appearing in the body of the text, we have chosen the McCune-Reischauer Romanization. That system ignores the internal structure and history of the Korean form in favor of approximating how the word sounds to English speakers, but it is also usually judged by Westerners to be esthetically pleasing, with a scholarly appearance on the page. The South Korean Ministry of Education has campaigned vigorously to win acceptance for the new revised system that it introduced in 2000, but that system ignores history and structure just as much, and as yet McCune–Reischauer remains the academic standard in the Western world. On the other hand, we have retained some non-standard spellings familiar to Western readers. Most prominently, the name of the Korean alphabet is transcribed throughout as “Hangul” (we thought McCune–Reischauer's “Han'gŭl” too freighted down with diacritics, and the Ministry of Education's revised spelling “Hangeul” intuitively odd and misleading for speakers of English). Personal names are spelled according to individual preferences when known.
For transcribing Korean linguistic forms we use the romanization found in Samuel E. Martin's Reference Grammar of Korean (RGK, 1992). That system is an adaptation of Yale Romanization that Martin created to account for the extra letters and distinctions found in Middle Korean. It is the most systematic and thoughtfully constructed transcription of earlier Korean that we have found; it is also commonly used now in professional writing about the history of Korean. Nevertheless, the system has a few troublesome features. One is the graphic complexity required to reflect all the Middle Korean symbols, including those used for suprasegmentals. Another is that the sounds represented by the letters are not always intuitively obvious. There are also a few minor philological problems. One such confusing detail, for example, is how the Middle Korean letter ○ is transcribed. That particular letter is not reflected at all in Martin's transcriptions in case it represented the “zero initial,” and this choice seems unassailable. However, in words where philologists have shown the letter to stand for a weakened, syllable-initial consonant, it is transcribed with a capital G, a choice that is also usually appropriate, because the consonant that lenited was most often a velar. But in some cases the weakened consonant was a labial, and in those cases the G can be misleading. Nevertheless, these are minor quibbles. Any romanized transcription of Middle Korean encounters difficulties.
We depart from Martin's romanization practice in three principal ways. First, and most importantly, we believe that the original Korean, including Chinese characters, must always be included for each historical citation, and that is what we have done, showing the original alongside the romanization. We also show the textual source of the citation in parentheses, along with the date of the text. Second, to reduce the complexity of the transcriptions, we have omitted tone marks, except in cases where information from those tones is required for the analysis. Third, we use the same modified Yale system for both Middle Korean and Contemporary Korean. Thus, for example, Martin romanizes the particle ˙도 ‘also, even’ as italicized ˙two when it occurs in a Middle Korean text, but as bolded to when it occurs in Contemporary Korean. We write both as two.