This book represents the first volume of the Descriptive and Comparative Grammar of Western Old Japanese. The first volume covers the phonology, writing system, lexicon, and nominal parts of speech. The second volume, includes chapters on adjectives, verbs, adverbs, particles, and conjunctions, as well as a comprehensive index for both volumes. I decided to split the publication into two volumes for three major reasons. First, as the reader can see, the first volume already has 400 pages, and the second volume will be twice as long. Producing a 1,200-page grammar in one volume is not only technically challenging, but it also creates a number of inconveniences for the reader who will have to handle such a heavy brick. Second, previous scholarship concentrated heavily on the verbal system; as a result Old Japanese nominals were neglected and not sufficiently described. I hope to rectify this unbalanced situation with this volume as well as present a number of innovative ideas. Third, and most importantly, in the process of my teaching Old Japanese over the last ten years at the University of Hawaii at Manoa, I found that the lack of an up-to-date Old Japanese grammar creates a serious obstacle for graduate students who want to learn the language and start reading Old Japanese texts on their own. From this point of view, publishing the first volume at an earlier date rather than delaying it until the second volume becomes available offers at least a partial solution to the problem.
There is a wide misconception that Old Japanese of the Asuka–Nara periods and Classical Japanese of the Heian period are essentially one and the same language. This is not true, since the differences between Old and Classical (or, better: Middle) Japanese are very significant. These differences are found in all levels of the language—in its phonology, lexicon, morphology, and syntax. Quite a number of grammatical forms used in Old Japanese did not survive into the later language, and a great number of Middle Japanese forms are late innovations not found in Old Japanese. In addition, certain forms, common for both periods, frequently have different functions. All this demonstrates that a person who is going to read and analyze Old Japanese texts just armed with knowledge of Classical Japanese will fare no better than a person who tries to read Old English texts through the prism of Middle English. One must also keep in mind that Old Japanese and Middle Japanese are based on two geographically close and in some respects similar dialects: the dialect of the Yamato plain for Old Japanese and the dialect of Heian-kyō (Kyoto) for Middle Japanese, both belonging to Central Japonic. Nevertheless, these are two different dialects, and there is no direct continuity between them. Finally, Old Japanese texts offer evidence for yet another different language: Eastern Old Japanese, which belongs to a completely different dialect group.
This grammar deals predominantly with the description of Western Old Japanese from the Yamato plain. The data from Eastern Old Japanese are used only for comparative purposes, thus the reader will not find here a description of any Eastern Old Japanese facts that have no counterparts in Western Old Japanese. Besides Eastern Old Japanese, I also used comparative data from different temporal and geographical varieties of the Ryukyuan language group. A word of caution is in order: besides being a descriptive grammar of Western Old Japanese, this book is also a comparative grammar of Western Old Japanese, but it is not a comparative grammar of the Japonic language family.
As the reader will see on several occasions, there are certain differences between early and late Western Old Japanese. Roughly speaking, the tentative boundary between these two varieties is somewhere in the late 740s or early 750s, although there are exceptions. For example, MYS 15 includes poems traditionally dated between 736–741 and is better defined as a text written in late Western Old Japanese.
Besides Japonic data, I offered in some cases possible external etymologies for Western Old Japanese morphological markers. This involves data from the ‘Altaic’ languages (Turkic, Mongolic, Tungusic, and Korean), but sometimes I also used data from Austronesian and Ainu. I do not think that the Japonic language family is a member of the ‘Altaic’ language family in the genetic sense (although a remote genetic relationship cannot be denied or disproved), but I think that it is most certainly a member of the ‘Altaic’ Sprachbund, with most intimate ties to Korean. As the reader will find on more than one occasion, there are interesting morphological parallels between Western Old Japanese and Korean which are not found in Eastern Old Japanese and Ryukyuan. I believe that these parallels represent early loans from some variety of Old Korean, possibly the language of Paekche, into Western Old Japanese. There are also possible parallels between Western Old Japanese and other ‘Altaic’ languages, which, in my opinion also represent traces of ancient contacts. In addition, there are few cases when Western Old Japanese has likely parallels in Ainu or Austronesian, which I believe also presents evidence of contact.
A few words about the presentation of the examples are in order. Examples from Western and Eastern Old Japanese texts are given in four lines: the first line provides the original spelling, the second the transliteration with morpheme breaks, the third glosses, and the fourth the English translation. The same is applicable to the examples from Old Ryukyuan and Old Korean texts. Upper case letters in the second line render the semantographic spelling, and lower case letters the phonetic one. I opted for transliteration rather than transcription for the following reasons. First, although it is believed that the orthographic distinction between kō-rui 毛 /mwo/ and otsu-rui 母 /mø/ in Western Old Japanese is preserved only in the Kojiki, it has been convincingly demonstrated recently that it also at least statistically survives into some later texts (Bentley 1997, 2002). Thus, I transliterate 毛 as /mwo/ and 母 as /mø/ in all Old Japanese texts. Second, it is presumed that Eastern Old Japanese had only five phonemic vowels: /a/, /u/, /i/, /o/, and /e/. However, Eastern Old Japanese texts are written in the Western Old Japanese orthography, which differentiated seven vowels /a/, /u/, /i/, /ï/, /o/, /´/, /e/, and a diphthong /´y/. As a result, EOJ vowel /i/ is spelled as WOJ /i/ or /ï/, EOJ /o/ as WOJ /o/ or /´/, and EOJ /e/ as /e/ or /´y/. It might seem that these variations could be ignored, and transcription uniformly rendering those as /i/, /o/, and /e/ would be less confusing. However, there is a problem. While EOJ /e/ is indeed spelled as WOJ /e/ or /´y/ with almost the same frequency, there is a clear tendency for a frequent variation of other vowel spellings only in certain syllables. One has also to keep in mind that Eastern Old Japanese is a dialect continuum, which was not homogeneous and could be roughly divided into three areas. The study of spelling variation within these areas has not been done so far. Therefore, unifying the transcription of Eastern Old Japanese may be a premature enterprise: we simply cannot throw out the baby with the bathwater. Thus, transliteration, a more conservative approach, is preferable for the time being. An even more spectacular case involves Old Ryukyuan. It is generally believed that proto-Ryukyuan vowels *o and *e in Old Ryukyuan merged with *u and *i respectively, as they have done in the modern Shuri dialect. Therefore, the tendency is to transcribe Old Ryukyuan syllables spelled with kana /Co/ and /Ce/ syllables as /Cu/ and /Ci/ respectively. However, some Old Ryukyuan spellings may be very indicative and reflect the stage before the merger (Serafim, p.c.). Thus, for example, the Old Ryukyuan word for ‘dew’ is spelled predominantly as tuyo, not tuyu, and the former is supported by dialect comparative evidence, which points to PR *tuyo, not *tuyu. Therefore, before a thorough study of Old Ryukyuan spellings is done, transliteration is also preferable.
Textual examples from other languages that include various modern Ryukyuan languages and dialects, and ‘Altaic’ languages are given in three lines: the first line provides transcription with morpheme breaks, the second glosses, and the third gives the English translation. Transcription rather than transliteration seemed to be a better solution here, because the data largely came either from modern languages without a writing system, or from old languages with an established practice of transcription. Whenever possible, I tried to preserve the transcription from original sources, but in certain cases I opted for simplification and unification. For example, I used /y/ for the palatal voiced glide throughout, replacing the IPA [j] that some sources on Ryukyuan and ‘Altaic’ employed.
When discussing the Eastern Old Japanese data, I frequently employ the reference to three different regions in the Eastern Old Japanese linguistic area: Region A, Region B, and Region C (Hino 2003: 197–201). The provinces are divided as follows: Region A—Simotuke, Kamitufusa; Region B—Hitati, Simotufusa, Kamituke, Musasi, Sagami, Suruga; and Region C—Sinano, Tofotafumi. The division into three regions is due to the fact that Eastern Old Japanese represented a dialect continuum rather than a single language. Region C was the one that was most influenced by Western Old Japanese; Region A was the least, with Region B representing the intermediate stage.
Alexander Vovin
Hawaii, December 2004