LINGUIST List 19.2334

Wed Jul 23 2008

Review: Historical Linguistics: Beckwith (2007)

Editor for this issue: Randall Eggert <>

        1.    Picus Ding, Koguryo: The Language of Japan's Continental Relatives

Message 1: Koguryo: The Language of Japan's Continental Relatives
Date: 15-Jul-2008
From: Picus Ding <>
Subject: Koguryo: The Language of Japan's Continental Relatives
E-mail this message to a friend

Announced at AUTHOR: Beckwith, Christopher I.TITLE: Koguryo: The Language of Japan's Continental RelativesSUBTITLE: An Introduction to the Historical-Comparative Study of theJapanese-Koguryoic Languages, with a Preliminary Description of ArchaicNortheastern Middle Chinese, Second EditionPUBLISHER: BrillYEAR: 2007

Picus Sizhi Ding, University of Hong Kong

SUMMARYThis is the second edition of Christopher Beckwith's groundbreaking work on theextinct Koguryo language. Based on toponyms recorded in the Samguk Sagi'Historical Records of the Three Kingdoms (of Korea)', Beckwith reconstructedthe Koguryo language and identified it as being closely related to Japanese. Themain body of the book consists of twelve chapters, accompanied by anintroduction (pp. 1-7) and by a compilation of Koguryo lexicon (pp. 250-254). Inaddition to a detailed index, other auxiliary materials are a note oftranscription and transliteration (for romanization of Korean, Japanese, Chineseand Tibetan) and a map of the Korean Peninsula and vicinity in Koguryo times.

Chapter one: Koguryo and the origins of Japanese (pp. 8-28)This chapter starts with a cursory discussion about modern scholarship onJapanese ethnolinguistic origins, an issue linked to the study of Koguryo fromthe very beginning. Different views on the relationships between Japanese,Koguryo and Korean are presented, all of which recognize, to varying extents, aclose relationship between Koguryo and Japanese. Most theories involve a furtherhypothetic relationship of the three languages to Altaic.

Chapter two: The ethnolinguistic history of Koguryo (pp. 29-49)According to written records from ancient Chinese sources, the proximal homelandof the Puyo-Koguryoic peoples in the fourth century B.C. was situated in thewestern part of present-day Liaoning shored by Bohai Sea, east of present-dayTientsin (Tianjin). The origin myth of the Puyo-Koguryoic peoples has thefollowing main plot: a concubine of the king of Koryo got pregnant by the touchof a beam of sunlight and gave birth to a large egg. A boy eventually emergedfrom the egg, and grew up to be an excellent archer. Later in his fleeing southunder the king's threat, he crossed a river on a bridge formed by floating fishand turtles after identifying himself as the son of the Sun. A similardescription of river-crossing is found in legends in Japan and the ancient Yüehregion of southeastern China.

Chapter three: The Old Koguryo toponyms (pp. 50-92)This chapter presents the toponym corpus taken from two of the geographicalchapters of the Samguk Sagi. The typical entry consists of a Chinesetranscription (in the Wade-Giles system), a toponym written in Chinesecharacters, a literal translation of the toponym into English, references tosources in the Samguk Sagi and glosses for pronunciation of the toponym (alsorepresented with Chinese characters, but marked by square brackets). Thephonetic value of the Koguryo pronunciation is reconstructed and discussed indetail. The corpus contains predominantly nouns, but for a few verbs,adjectives, numerals and grammatical morphemes.

Chapter four: Archaic Northeastern Middle Chinese (pp. 93-105)Given the vast territory where Chinese has been spoken (even in the earlycenturies of the Current Era), Beckwith hypothesizes the existence ofNortheastern Middle Chinese spoken in present-day northeastern China. Somecharacteristics of its sound system include the retention of *k- from OldChinese, palatalization of *ti- in Old Chinese to *tśi, and the preservation ofrhotic in syllable codas, e.g. *lir.

Chapter five: Old Koguryo phonology (pp. 106-117)The phonological system reconstructed for Old Koguryo is represented. Thecorrespondences with Old Japanese are exemplified. The word structure of OldKoguryo is also sketched.

Chapter six: Toward common Japanese-Koguryoic (pp. 118-143)Detailed comparison of probable cognates, where available, between Old Koguryoand Old Japanese is pursued in this chapter. It covers over 100 lexical items aswell as six grammatical morphemes: the genitive-attributive marker, verbderivational morpheme, adjective-attributive suffix, noun derivational morpheme,and diminutive suffix.

Chapter seven: The Proto-Japanese-Koguryoic homeland (pp. 144-163)Ancestors of the Japanese people are believed to have lived in the western partof present-day Liaoning in northeastern China with the Koguryoic people untilthey started to migrate to the southern Korean Peninsula and to northern Kyushuin Japan in the fourth century B.C. Certain cultural features of the Japaneseand lexical evidence discussed in the chapter point to a possible homeland forthe Proto-Japanese-Koguryoic people much further south in an earlier time priorto their arrival in the northeast.

Chapter eight: Koguryo and the Altaic divergence theories (pp. 164-183)This chapter rejects the romantic idea of regarding Koguryo as a member ofAltaic, ''a distant relationship theory that a century of energetic effort hasfailed to demonstrate successfully'' (p. 164). A large number of reconstructedwords for Old Koguryo are examined and their erroneous etymologies pointed out.

Chapter nine: The Altaic convergence theory (pp. 184-194)The Altaic convergence theory, which groups the Turkic, Mongolic and Tungusicfamilies of languages on the basis of their lexical similarity, resulted frommutual borrowing. Five diagnostic properties are offered as a preliminaryanalytic tool to delimit Altaic: 1. No word-initial consonant-clusters; 2.Suffixing agglutinative morphology; 3. No system of overt grammatical concord;4. Obligatory verb-final syntax; and 5. Vocabulary items in common with otherAltaic languages (p. 190).

Chapter ten: Japanese and the mixed language theory (pp. 195-213)Beckwith contests the assumption that 'basic vocabulary' is resistant toborrowing. Hence, the Mischsprache theory based on this assumption must beabandoned. Words that can be retained in languages are argued to be found byfrequency counts of large corpora: ''the very highest frequency words—the topdozen or so'' (p. 198). Various such lists are provided for British English,Quebec French, Mandarin, Norwegian, German, Russian, Japanese, Korean andseveral classical languages.

Chapter eleven: Linguistic theory and Japanese-Koguryoic (pp. 214-235)In the context of investigating the Japanese-Koguryoic connection, Beckwithdevotes this chapter to address some fundamental problems that have permeatedinto the tradition of historical-comparative linguistics in eastern Eurasia.These include the religiously upheld belief in the accuracy of the HistoricSinological Reconstruction system, oversight of linguistic changes forged byboth divergence and convergence forces, and the biased favor of divergencetheories to the exclusion of convergence theories. Linguistic convergence inEast Asia is discussed in detail and five divergence theories involving Japaneseare evaluated.

Chapter twelve: The Japanese-Koguryoic family of languages (pp. 236-249)This final chapter deals with two major issues: (i) the hypothesis that thetoponyms from the Central Korean area (the territory of the Koguryo kingdom)represent a language different from that attested in Chinese sources and (ii)understanding of the Japanese-Koguryoic theory in the light of archaeologicaldata. How the Japanese-Koguryoic languages might be related to Korean is alsoexplored.

EVALUATIONBeckwith has definitely made a valuable contribution to thehistorical-comparative study of East Asian languages with an admirable attitudetowards scholarship: he wrote in the preface that the aim of the book is ''todiscover the truth'' rather than ''to disprove the many other theories discussed''(p. x). Centering around the firmly demonstrable linguistic relationship betweenKoguryo and Japanese, the book addresses a breadth of relevant issues. Readersinterested in historical linguistics in eastern Eurasia will find many of thediscussions enlightening. (This would explain the publication of the secondedition in an interval as short as three years.)

Instead of taking it as a conclusive study to the Japanese-Koguryo theory,Beckwith explicitly invites interested researchers to undertake more studies inthe new direction. This is precisely the spirit to be expected from a genuinediscovery in scholarship: a groundbreaking achievement leads to a broaderhorizon for pursuit of knowledge, not termination of research on the topic. Inthe remainder of this review, I will take note of a couple of errors in thetoponym data of Old Koguryo and issues related to word loanability andmethodology in historical linguistics. Finally, an alternative view about theantique homeland of Proto-Japanese-Koguryoic peoples will be presented.

The major problem with analyzing the toponym data concerns the use of thecharacter 省 as a phonetic sign. Its occurrence in [未乙省] for glossing thetoponym _Kuo yüan ch'eng_ 國原城 (p. 55) is regarded as representing thephonetic value of 城 (a loanword from Chinese, meaning 'fort, city'). Likewise,for the t _Kao feng hsien_ 高烽縣 glossed as [達乙省縣] (p. 64), 省 is alsoregarded as standing for 'fort, city'. Since 達乙 is an establishedrepresentation for 高 'high' in Old Koguryo, the second character 烽 'beaconfire' is left unrepresented under this analysis. However, this toponym ends withthe character 縣 'county' and contains no character meaning 'fort, city'. Itseems to me that the word for 烽 'beacon fire' could be ☆śeŋ in Old Koguryo,represented by 省. Another problem is a typographic error on the first characterin the toponym _Li shan ch'eng_ (p. 92): 犁 'plough' has been misprinted as 梨'pear'. The same typographic error recurs on p. 251.

Although lexical borrowing is common between languages with extensive contact, Ihave found the discussion of the word for 'water' as an areal culture word (pp.154-156) and the observation that ''words for 'water' and 'river' areparticularly loanable'' (p. 177) rather puzzling. Lexical loans are copious interms for products (including merchandised animals and plants), for theiroriginal name is typically introduced with the commodity together to newsocieties. On the other hand, it is difficult to imagine that something as vitalas water, being essential to all life forms on earth, would be subject to asimilar trend of lexical borrowing. If no other linguistic theory could accountfor the similarity of this word in unrelated languages, I would consider it achance of coincidence before treating it as a loan word.

Another problematic case arises from treatment of the homophonous word *tar'high; mountain', reconstructed for Common Japanese-Koguryoic. Using thephonetic value of the character for 'charcoal', Beckwith reconstructed *tar ~*dar for the word 'mountain' as its late Middle Old Chinese form, and suggestedthat the word had been borrowed into Common Japanese-Koguryoic and,consequently, “merged with the inherited word *tar 'high'” (p. 151). Accordingto Cao and Su (1999: 513), however, the phonetic element in the character for'charcoal' is not 'mountain'. Thus the hypothesis of a semantic extension from'mountain' to 'high' in Proto-Japanese-Koguryoic (which was not explored in thebook) would best the conjecture of a loan word for 'mountain' from Old Chinese.

Beckwith reports that grammatical or structural morphological elements have beenknown for their higher extent of resistance to external replacement byconvergence (p. 219). To avoid the problem of lexical borrowing, he suggeststhat functional words should be included in producing lists of frequency wordsbased on large corpora. However, there are many technical problems to be solvedbefore the usefulness of the proposed methodology can be proven, such as:difficultly in obtaining large corpora from languages without a writingtradition, determination of essential genres for the corpora, and the number ofcomparative words in the list. A dozen items from the top of the list offrequency words, as exemplified in chapter 10, is too few in number and tooheavy in weighting for each item for the purpose of historical-comparativeinvestigation. Moreover, considerable variation regarding what constitutes topfrequency words has been observed in different studies of the same language(presumably on account of different genres in the corpora and the size of thecorpora). For instance, the negative morpheme bù in Mandarin ranks at the thirdposition in one list, but at the tenth in another (p. 199). Yet more disturbingis the outcome where an item appears in one list but disappears in another. Isuspect that the topic marker wa in Japanese would be able to make its way tothe top ten in the frequency list if the corpora were sufficiently rich ingenres, cf. Fry's (2003) study of spoken Japanese corpora.

Tibeto-Burman was probably one of the earliest groups of peoples who had comeinto intensive contact with Proto-Japanese-Koguryoic (pp. 159-163). ToBeckwith's list of typological similarities between Japanese and Tibeto-Burman,the 'pitch-accent' system, found in Prinmi - a core Qiangic language ofTibeto-Burman, can be added (see Ding 2006 for a definition of this uncommontonal system and a typological comparison). Whatever historical connection theremight be between Prinmi and Japanese is currently unclear, but the unusual tonalsystem of Prinmi can be traced back to Proto-Prinmi (Ding 2007). The earlycontact with the Tibeto-Burman people, however, does not necessarily ensure thatthe Proto-Japanese-Koguryoic peoples had resided in South China, where a numberof Tibeto-Burman languages are spoken (especially in the southwest) before theirmigration to and settlement in northeastern China. On the basis of Tibeto-Burmancontact, a more likely region for the homeland of Proto-Japanese-Koguryoic wouldfall somewhere in (north)western China, where the Qiangic and Yi-Burmese(Lolo-Burmese) peoples originated (cf. LaPolla 2001). As for cultural traitsshared between peoples in Japan and coastal China since ancient times, they canbe accounted for by the influx of refugees (who need not beProto-Japanese-Koguryoic peoples) from the coastal regions of the mainland tothe Korean Peninsula and to the Japanese Archipelago between the third centuryB.C. and the third century A.D. (cf. Wang & Cheng 2006: 6-7; 23-24). This issuecould be resolved satisfactorily when independent evidence, perhaps fromarchaeological findings, were available.

REFERENCESXianzhuo and Peicheng Su. (eds.) 1999. _An Etymology Dictionary of ChineseCharacters_. Beijing: Peking University Press.

Ding, Picus S. 2006. A typological study of tonal systems of Japanese andPrinmi: Towards a definition of pitch-accent languages. _Journal of UniversalLanguage_ 7.2: 1-35.

Ding, Picus S. 2007. The use of perception tests in studying the tonal system ofPrinmi dialects: A speaker-centered approach to descriptive linguistics._Language Documentation and Conservation_ 1.2: 154-181.

Kim, Busik. 1145. _Samguk Sagi_. Available at

Fry, John. 2003. _Ellipsis and wa-marking in Japanese Conversation_. London:Routledge.

Kessler, Brett. 2001. _The Significance of Word Lists_. Stanford: CSLI.

LaPolla, Randy. 2001. The role of migration and language contact in thedevelopment of the Sino-Tibetan language family. In R. Dixon & A. Aikhenvald(eds.), _Areal Diffusion and Genetic Inheritance: Case Studies in LanguageChange_, 225-254. Oxford: Oxford University Press.

Wang, Gaoxin and Rentao Cheng. _The History of Ancient Relations between theThree Countries of East Asia_. Beijing: Beijing University Of Technology Press.

Teaching at the University of Hong Kong, Picus Sizhi Ding has general interests in languages of China, East Asia and beyond, especially those with extensive contact with Chinese.