A Descriptive Grammar of Early Old Japanese Prose.
A Reference Grammar of Classical Japanese Prose. By ALEXANDER VOVIN. London: ROUTLEDGE CURZON, 2003. Pp. 476, tables. $135.
The problems with Bentley's book begin with his title. The texts with which he is concerned are in the main not prose, at least not in the usual understanding of that term; nor can the bulk of them be described as "Old Japanese," let alone "Early"; nor is what he has published a "descriptive grammar."
Bentley's text-corpus consists of his own selection of sixteen examples from the early Japanese religious-ceremonial texts known as Norito (a word of much disputed meaning to the understanding of which Bentley is able to contribute nothing). He refers to these throughout as "liturgies"; "intercessions" or "imprecations" would be more accurately descriptive of their content. They are intricate, lapidary literary concoctions, full of repetitions and limited in vocabulary. They preserve a number of pre- (or at least, non-) Buddhist religious terms, many now not well understood; they make generous use of poetic-stylistic devices that elsewhere may be observed in full flower in the Old Japanese poetic canon; and they are transmitted in a mixed Japanese-Chinese logogram-cum-phonogram orthography that overtly records only a minimum of their Japanese lexical items and morphological elements. For the bulk of the remainder we have only a late, often self-contradictory "reading tradition" that at best is thought to be no earlier than "early Kamakura," i.e., ca. 1200; the mixed-script texts themselves are no earlier than 927. For all these reasons, it is difficult to take Bentley's categorizations of his corpus as "early," "old," or "prose," at face value.
His claim to have written a "descriptive grammar" is equally without basis. Because of the extremely limited semantic range of his texts, he has felt it necessary to go to other examples of earlier Japanese in order to "flesh out" his description; as a result many of the morphological elements upon which he elaborates are actually absent from his corpus. Even more damaging to his claim is his penchant for introducing non-descriptive diachronic speculation, seeking to discover evidence for early Japanese forms and their meanings in words cited, too often incorrectly, from other languages. There is no mention of any hypothesis concerning a possible cognate relationship between Japanese and these other languages, or between Japanese and the language families to which these other languages may or may not belong; the forms are merely cited, and apparently that explains everything. Whatever this is, it is not descriptive grammar.
Particularly in view of the extremely recondite content of the Norito texts, establishing their meaning must be an essential first step in writing any grammatical description embracing them, whether synchronic ("descriptive") or diachronic, i.e., historical. They consist in the main of inordinately extended syntactic units in which it is not always a simple matter to establish anything resembling immediate constituents; they abound in embedded discourse, where it is frequently a theological fine-point to determine who is speaking to whom (gods to men, or men to gods?); and they have preserved more than one pious hapax of the "mumpsimus for sumpsimus" variety, for the original sense of which we now have only a tradition of free-wheeling guesses.
Bentley could have found much needed help in unraveling at least some of these semantic knots by consulting more often than he does the published Norito translations of Philippi (1952) and Bock (1979). Even more useful would have been the painstakingly literal versions by Satow and Florenz published in TASJ between 1878 and 1881, which he has not consulted at all; similarly he has overlooked such important contributions to our understanding of these texts as Dumoulin, MN 12 (1956): 121-56, 269-98, who demonstrates, inter alia, that not only what we take these texts to mean, but also in many important details the very texts themselves depend upon, and are no older than, the often widely diverging exegetical opinions of this or that Tokugawa hermeneute. Even at best, so much of the grammatical apparatus, particularly in the realms of morphological formants and case-grammar suffixes, that we find today in our editions of these texts is not overtly written in the manuscripts but rather the end-product of subsequent centuries of glossing and reading-tradition, that the accurate understanding of the original sense that would be an essential first step in Bentley's goal of "writing a descriptive grammar" of these materials is all but impossible.
This would in and of itself be a serious problem; but to it he has added another, since his own translations of the syntactic segments that he extracts and translates to illustrate his "descriptive grammar" betray a less than satisfactory command even of the language that is indubitably in the texts, much less of the exegetical body that has grown up in the course of its later hermeneutical conflations. His description of the deverbal nominal formant that he calls "the nominalizer -aku" (pp. 172-74) is illustrated with three truncated examples, plus several others from outside his text-corpus. But nowhere is there a hint of the characteristic early Old Japanese syntactic employment of this formant, particularly of its role in marking off long segments of direct discourse.
Some of his glosses are so highly unlikely that one wishes for a text-citation, but none is given (e.g., yuku '[the place I] go [to]', p. 58); relative clauses appear that violate both the syntax and the sense of the text ("the jars that are lined up," p. 163); two historical figures are mistakenly expanded into four (translating a fragment of Senmyo 19 on p. 242, where the reading tradition for the plural suffix is also misprinted); an important gloss surviving in his text is entirely misunderstood in an attempt to translate it ("what are called 'words of praise' are like salutations when lifting the cup," p. 241, but the text has nothing corresponding to his "lifting"; Bock, 1972, p. 81 with n. 354 got it right); the sense of Bussokuseki 17 is grossly misrepresented by an unwarranted interpolation ("... or so they say," p. 253; this poem has been translated correctly, or at least without this particular misunderstanding, too often to mention here). Chinese too comes in for its lumps: the title of the Yu pian [TEXT NOT REPRODUCIBLE IN ASCII] by Gu Ye-wang (519-581) is misunderstood as "The Jeweled Manuscript" (p. 253). Bentley early assures us that "it is the structure of the system that interests me, not the sentence itself" (p. 8); apparently both lexical and syntactic meaning also interest him but little.
Two further problems come between Bentley and his "descriptive grammar": he attempts to refine the received view of Old Japanese phonology by substantial if oddly selected citations from SinoKorean and Sino-Vietnamese; and he ventures (pp. 264-73) to print a new text of the Norito he has selected for study ("I have ... created my own critical text...," p. 25). What he means by this last is far from clear. He does not identify which manuscript(s) he has employed. Neither of the two manuscript sources he does cite (one, a letter from his sister-in-law in Japan, p. 16, the other an undated fragment of a Kokinshu MS that a friend examined for him in St. Petersburg, pp. 18-19) can have had much to do with the case. He does not register any of the hundreds of variant readings that he must have encountered in "creating" his text; and in more than one passage (e.g., his p. 269, line 3, the original of the important gloss mistranslated on p. 241 and already noted supra), what he has created is only a further-muddied text (with [TEXT NOT REPRODUCIBLE IN ASCII] erroneously for [TEXT NOT REPRODUCIBLE IN ASCII]).
Equally obscure are most of his phonological statements. He tries to describe in detail the acoustic details of "the eight vowels of Old Japanese" (pp. 43-47), but his transcription throughout uses the socalled "Yale System" which treats Old Japanese as a language with five vowels plus or minus glides fore-and-aft; disarmingly, he argues in favor of the Yale system by mentioning its "ease in typing" (p. 44). Any meaningful distinction between the two approaches is further blurred by his elastic use of "monophthong" (describing, e.g., a sequence -ey-, p. 56) and "diphthong" (describing the same sequence, p. 57). Not much can be made of any of this. Similar terminological confusion makes the Old Japanese writing system sound even more convoluted than it was. Bentley's "etymological spellings" (p. 59) must be understood as "historical orthography"; by "orthographic style" (p. 9) he means "calligraphy"; and what is intended by "the spelling orthography" (p. 32) remains obscure.
Bentley's misuse of "etymological" carries over into several extended passages, all out of place in a "descriptive grammar," in which he suggests "external etymologies" (p. 253 and passim). These are comparisons between Japanese and other languages, i.e., etymologies in the proper sense. But his unfamiliarity with the comparative literature and his carelessness in citing forms render almost everything he writes in this connection misleading if not simply incorrect.
Two different "rules" attempting to relate Japanese to Korean and Tungusic are cited on the same page, but both as if to account for the same Japanese form (p. 57): if one is right the other must be wrong (and probably both are incorrect). In connection with OJap. Fo '100' the Tsintsius comparative Tungus dictionary is cited as registering "[a] promising etymology [that] relates this word to the Tungusic language Orok, powo 'ten'" (p. 248). But Tsintsius (TMS 2, 40a) says nothing at all about Japanese, nor does she register "Orok powo '10'." Orok '10' is joon (TMS 2, 248b), and Bentley must have misread her gloss desyatok 'a score; a fair number (of items)' as desyat' '10'. He copies out "Mo. sumel 'the highest mountain'" from the Ono dictionary, not realizing that what Ono (here a victim of the r / l confusion) intended to write was sumer, itself a loan into Mongol from Indic. His own suggestions, Mo. sume 'shrine, temple', is (like Ono's original suggestion) vocalized incorrectly (read: sume); and of his suggestion that "Manchu-Tungusic *sama-n 'shaman' ... looks like a loan from a neighboring language, maybe Persian," the less said the better.
The Norito are rich in echoes of early Japanese shamanism, which makes it all the more unfortunate that Bentley's command of the comparative-historical literature in the field of shamanistic terminology is particularly poor (viz., FaFuri, p. 247 among other striking passages); before writing about these subjects he ought first to have familiarized himself with studies such as those available in MOAG 116 (Hamburg, 1991) and MOAG 121 (Hamburg, 1994). In Buddhist contexts too he is ill prepared. Of OJap. tumi 'sin' he writes, "I know of no external etymologies" (p. 253). One hardly doubts him. But in view of K. H. Menges, Altajische Studien, II: Japanisch und Altajisch AfKM, XLI/3, Wiesbaden 1975), p. 34, with its copious citations of earlier studies, Bentley's confession is sadly misleading concerning the actual state of Japanese-Xenic comparison with respect to this etymon. Significant publications proposing etymologies for morphological elements have also been ignored. OJap. .i is frequently mentioned (pp. 88, 242, 259 et passim), but without taking notice of its treatment, both descriptive and historical-etymological, in BJzOf 12.1: 251-91 (Bochum, 1989).
This cavalier attitude toward the existing literature is particularly unfortunate in a book that, as its author puts it, aims to "permit ... students to get a better handle [sic] on this important language" (p. 5). Bentley does not think highly of citing published studies; indeed he castigates the practice as "pay[ing] lip service" (p. 51). Instead he prefers to rely on help from his friends, either in the form of "personal communications" (on almost every page, sometimes two or more to the page) or "references" that turn out to lead one only to unpublished papers by his fellow-students (pp. 95, 96, 99). He records in detail how he was astonished to learn from one such friendly source that the Old Japanese corpus actually includes "one set of poems literally engraved in stone" (p. 53, Bentley's italics), not realizing that elsewhere he cites (but mistranslates) a poem from these same lithic lyrics (pp. 252-53), and in his "Bibliography" lists (p. 278) a 1975 monograph entirely devoted to the same. This innocent level of naivete is matched only by the announcement of his independent discovery that "the Wei zhi chapter concerning the barbarians" tells us about "a place called Wa (ancient Japan)" (p. 249); is it possible that he knows nothing of the enormous literature centering upon the Gishi wajinden that has accumulated over the centuries?
Most highly favored among Bentley's tributes to the work of his friends are dozens of references to "Vovin (forthcoming)." Fortunately Vovin has now come forth, so it is possible to compare some of what Bentley writes with at least one of his sources.
Vovin's book steers clear of the comparative-historical pitfalls that are so frequently Bentley's undoing. He bases his description on three Heian monogatari texts (which he over-generously describes as "masterpieces," p. 6), the Taketori, Ise, and Hamamatsu chunagon, but he ignores the Genji. The third of Vovin's chosen texts in particular causes him considerable difficulty. He misunderstands its title throughout ("The tale of Hamamatsu and the Chunagon," p. 6; "The tale of the Chunagon and Hamamatsu," p. 463, but both are wrong), and even transcribing the long u of chunagon frequently baffles him (on p. 6 he renders it several times with a misshapen version of Chin. [TEXT NOT REPRODUCIBLE IN ASCII] yan; why?). His decision to exclude the Genji from his materials, though surprising, was probably wise. This is partly because of his evident unfamiliarity with Heian culture (sedoka is said to mean "rower's song," p. 5; the Makura no soshi is described as an "essay," p. 6; nishi no kyo becomes "the Western Capital," p. 59), but also because the Genji's exuberant syntactic structures would have posed severe problems for his analysis. Nor can much be said in favor of his account of Heian phonology ([m], [n], [[eta]] are described as "nasal stops," p. 14; the syllabic nasal is said to be found only "[i]n Chinese loanwords," loc. cit., and its syllabicity, not to mention its employment in such words as Vovin's wonna 'woman', pp. 69, 79, et passim, ignored throughout).
But all this is as nothing compared with Vovin's treatment of the Japanese script, specifically the hiragana syllabary, which on the evidence of these pages, he has yet to master. From p. 12 on he introduces a non-existent kana sign, something like Chin. [TEXT NOT REPRODUCIBLE IN ASCII] ren, which he alleges is used in Heian texts to write ya or -j-. Unlike Bentley he does not "create" entire texts; but he carries out much the same operation with this bogus kana symbol, as also with the hundreds of text-fragments which he prints to illustrate his grammatical analysis; a high percentage of these citations display shocking unfamiliarity with both kana and kanji.
The hiragana for ya [TEXT NOT REPRODUCIBLE IN ASCII] (the real thing this time) is used over and over to write hito (e.g., p. 111), or doubled, hitobito (p. 82), or in a Chinese compound, even nin (p. 79); that for ma [TEXT NOT REPRODUCIBLE IN ASCII] is said to write take (p. 54); conversely, Chin. ren is said to write the ya of yama (p. 86), while a hiragana sequence fu no ya [TEXT NOT REPRODUCIBLE IN ASCII] is said to write miyako no hito (p. 53). Vovin writes that he has relied for his citations "[i]n most cases" on the printed texts in the Nihon koten bungaku taikei series, "but on some rare occasions I deviated from them, relying on actual manuscripts of the texts" (p. 2); he goes on to list the facsimile editions he consulted. There are far too many of these blunders, literally hundreds, to check for the purpose of a review; but some seemed so very curious that I could not resist verifying what the manuscripts actually had. For the Ise monogatari I consulted the manuscript in the Imperial Household Agency collection reproduced in Suzuki Tomotaro, Kochu Ise monogatari (Tokyo, 1971), as well as three others from the Tenri University Library reproduced in Katagiri Yoichi, Ise monogatari shohonshu (1973). Nothing in any of these texts corresponds to any of Vovin's exotic "readings" (e.g., for his fu no ya = miyako no hito, p. 53, the manuscript in Suzuki, p. 148 has, as we might expect, simply [TEXT NOT REPRODUCIBLE IN ASCII] [i.e., miyako] no hito). Another level of textual irreality is created when Vovin cites the same text-fragment twice in difference places, but has it wrong both times on the evidence of the Suzuki manuscript (e.g., his p. 71, contradicted on p. 78, and neither supported by Suzuki, p. 141). His book claims to be "a vital tool ... for scholars and students ... who want to learn to read texts in the original script" (p. [i]): what can one add but, good luck to them!
Confronted with this orthographic Walpurgisnacht, one may be forgiven for not taking too seriously the bulk of the descriptive accounts that Vovin has evolved for his often misread and frequently misunderstood text-fragments. Nevertheless, some of them are of interest, particularly when read together with Bentley's accounts of the same feature in his texts, here as so frequently, again in debt to "Vovin (forthcoming)."
One such concerns two verbal stem-formant suffixes, +t- and +n- (Bentley, pp. 288-90; Vovin, pp. 305-14). In Old Japanese, secondary formations with the former suffix were employed for actions involving overt external causation, those with the latter for actions, especially natural, without outside stimulus; and this admittedly tenuous distinction was maintained, mutatis mutandis, into the Heian texts. Bentley does little apart from bad-mouthing the first (1959) edition of Bruno Lewin's estimable Abriss der japanische Grammatik auf der Grundlage der klassischen Schriftsprache with a garbled remark, "[he] gives offers [sic] little help on how to differentiate the two"; thereupon he "follows Vovin (forthcoming)," reproducing at second-hand the treatment of this problem in Kolpacki's Ocerki po istorii japonskogo jazyka (1956). Vovin ignores Lewin's treatment (though he lists the Abriss on p. 462), elaborates further on Kolpacki, and finally suggests dignifying what she learned about this problem with the canonical designation "Kolpacki's constraint."
If either Bentley or Vovin had consulted the revised and enlarged 1975 second printing of Lewin, they would have found (pp. 166-69) a discussion (and refutation) of Kolpacki, and even more importantly, a citation of the major existing descriptive treatment of Old Japanese, Yokoyama Masako's Language Dissertation Nr. 45 (1950), nowhere mentioned in either of these books; there, inter alia, the problem of these two suffixes is deftly analyzed. (Moreover, Lewin was already familiar with the Kolpacki Ocerki in his 1959 edition.)
Another case in point is that of the iterative stem formant suffix OJap. +F- (Bentley, pp. 200-203; Vovin, p. 323). The analysis of both authors obscures an original thematic vowel -a- by treating it as part of the suffix (so also with the -i- theme in the case of +t-, +n-); they also conveniently (but incorrectly) overlook forms with an -o- thematic vowel. Bentley prints one, uturoF-, p. 201, but does not explain it. Vovin claims none exist, which is not true. Ikeda Toshio, Hamamatsu chunagon monogatari sosakuin (Tokyo, 1964), p. 27, registers five passages with this same uturoF- from this, one of Vovin's three major text-sources.
Throughout both these books stress is placed upon presenting analyses that do more than merely reproduce those of "traditional Japanese grammar." But it is to "traditional Japanese grammar" that we must look for the failure to recognize these thematic vowels for what they are, not only in the case of this iterative +F- but in many other forms as well; so also for the reason why we learn nothing here of the equally important employment of this formant as a denominal verb suffix (utaF-'sing' < uta 'song', etc.). Lewin, pp. 129-30 is informative in this respect; these two grammars are not, just as "traditional Japanese grammar" also is not. In this sense, they merely perpetuate the failures, if not the analytic excesses, of the inherited lore.
Both Bentley and Vovin want to help students of Japanese language and culture. The former writes, "[i]t is hoped that a study of norito [i.e., his book] will re-ignite excitement in the linguistic, literary, and cultural aspects of this often neglected text" (p. 262). Simultaneously with the publication of his book, Vovin wrote elsewhere of his disappointment with the "many errors" and "low level of reliability" of the information about earlier forms of Japanese now available to those wishing to study the genetic relationship of the language (Nihongo keitoron no genzai [Tokyo, 2003], p. 29 n. 3), as well as of his hope one day to remedy this unfortunate situation. As he works toward that end, one can only urge upon him more study of, and if possible eventual mastery over, the hiragana syllabary.
ROY ANDREW MILLER
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||A Reference Grammar of Classical Japanese Prose|
|Author:||Miller, Roy Andrew|
|Publication:||The Journal of the American Oriental Society|
|Article Type:||Book Review|
|Date:||Oct 1, 2003|
|Previous Article:||Yuan Hong: a case of premature death by historians?|
|Next Article:||Ancient China: Chinese Civilization from Its Origins to the Tang Dynasty.|