J. Marshall Unger argues in this book that Korean and Japanese are cognate languages, in the technical sense used by linguists: both descend from a single linguistic parent, proto-Japanese-Korean. The book will be of interest to linguists, archaeologists, and historians concerned with the late prehistory of Northeast Asia.
While the claim of cognacy remains controversial, one aspect of Unger's position represents an emerging consensus among linguists and anthropologists: the Japanese language originated on or in the vicinity of the Korean peninsula, and it came to the Japanese archipelago as part of the Yayoi expansion, beginning (at the earliest) around three millennia ago. This view is shared with two scholars whose views Unger criticizes: Christopher Beckwith and Alexander Vovin (see Beckwith, Koguryŏ: The Language of Japan's Continental Relatives [Leiden: Brill, 2004]; and Vovin, Koreo-Japonica: A Re-evaluation of a Common Genetic Origin [Honolulu: University of Hawai‘i Press, 2008]). While neither of these scholars believes that Japanese is cognate with Korean, both concur with Unger in tracing Japanese to the Korean peninsula.
The arguments presented in the book are not primarily linguistic, although Unger contributes several new Korean-Japanese etymologies (pp. 57–59). Unger cites the existence of roughly 500 Korean-Japanese etymologies uncovered in earlier work by Samuel Martin and John Whitman (pp. 39–41), the nearly 50 percent match in functional morphemes pointed out by Martin (pp. 59–61), and the syntactic similarity between the two languages. Any treatment of their historical relationship must account for these commonalities. Unger rejects “convergence” theories: accounts that explain the similarities as a result of contact between two originally unrelated languages. He specifically rejects an account based on contact on the Korean peninsula prior to the Yayoi migration (pp. 6–17). Unger argues that if “pre-Japanese” and “pre-Korean” underwent a long period of contact, we might expect more lexical borrowings between them, and that if Japanese originated in the southern part of the peninsula, we might expect its speakers to have expanded northward; finally, he asks how Japanese might have detached itself from the convergent Korean-Japanese complex to move to the archipelago.
Unger's arguments will not be accepted by skeptics of the Korean-Japanese relationship, but he is right that proponents of a convergence model have yet to present an account of the relationship between the two languages that is consistent with known cases of convergence. There is, for example, a scenario in which one language comes to share structural resemblances with another with relatively little lexical borrowing—the process of “interference through language shift,” whereby speakers of a substrate language shift to a superstrate language while contributing little of their lexicon to the superstrate. But the linguistic and geographic facts do not support such a scenario in the case of Korean and Japanese. If an earlier Japanese substratum had shifted to Korean, for example, we would expect to find evidence of a concentration of Japanese-like features in southern Korean varieties. But we do not.
Unger argues instead that the shared features of Korean and Japanese result from a common ancestry. The differences—specifically, the fact that there is not a larger store of cognate vocabulary—result from the early date of the divergence (Unger cites, but does not endorse, Hattori Shiro's lexicostatistical dating of the separation at 2000 bce, p. 18; see Hattori, Nihongo no keitō [Tokyo: Iwanami, 1959]), and from contact with additional languages or language families. In the case of earlier Korean, the contact is with Tungusic. This is plausible, given the presence in Korean of typological properties shared with Tungusic, such as vowel harmony based on tongue root position. Korean scholars attempting to compare Korean with Tungusic have generally proceeded on the assumption that Korean and Tungusic are cognate members of a larger “Altaic” family, but these attempts have been criticized (see, e.g., Juha Janhunen and Kho Songmoo, “Is Korean Related to Tungusic?” Hangŭl 177 [1982]: 179–90). Unger's proposal highlights the need for a contact-based reconsideration of the Korean-Tungusic relationship.
In the case of Japanese, Unger suggests that contact was with the language or languages of the groups referred to as Dōngyí on the eastern China coast, and that representatives of these groups brought wet rice cultivation to the southern Korean peninsula by sea (p. 161). Dōngyí is a shifting and global term; we know nothing about the languages of this area. But given the region under consideration, Unger is right to recommend a reconsideration of earlier attempts to compare Japanese and proto-Austronesian (p. 169), as the latter is the northernmost attested language family with an origin on the China coast (specifically, Formosa). Recent research, however, dismisses a southern route for the introduction of wet rice cultivation in the Korean peninsula (see Ahn Sung-Moo, “The Emergence of Rice Agriculture in Korea: Archaeobotanical Perspectives,” Archaeological and Anthropological Sciences 2, no. 2 [2010]: 89–98). Wet rice farming was already well established in the Longshan culture in Shandong by the middle of the third millennium bce. An event such as the global cooling reported for the end of the third millennium could have triggered the dispersal of wet rice farmers from Shandong around the eastern Liaodong littoral.
