The Price of Tradition
Clearly, some computer tasks that would be very easy and cheap to do using a phonemic representation of Japanese data are difficult and expensive to do using kanji. But difficult does not mean impossible, and judging whether or not a particular price is fair is a notoriously subjective matter. Current trends in computer usage in Japan suggest that the Japanese are willing to pay anything for the dubious privilege of being able to use kanji on computers. In countless everyday applications, output involving kanji is not needed, yet Japan seems prepared to foot the bill for lavish output devices and all the extra costs associated with maintaining kanji databases. When it comes to input, things are even worse. Why not use standard alphanumerics? This simple bit of decidedly low-tech engineering would cost next to nothing in terms of hardware, resolve all the problems touched upon in Part I, and be extremely easy to implement, since any native speaker of Japanese can learn how to transcribe the language in roman letters quickly and accurately in a few hours. Why cling to kanji under these circumstances?
At one level, the answer is obvious: kanji are part of Japanese culture; Japanese don’t want to give them up just because of computers. But this is inadequate: there are many areas of Japanese life in which people seem quite happy to let kanji usage fall into decline. The use of kana and rōmaji where kanji were once de rigueur is now accepted practice in Japanese typography. The mass media are using ever greater numbers of gairaigo (Passin 1980: 48). Chinese-style word formation is fast losing its position as the principal means of making up new words for new things and ideas (Passin 1980: 49–61, Kabashima 1981). In most situations, Japanese have abandoned Chinese methods of writing numbers and calculations. No one seems to feel that these things are detrimental to Japanese culture. But the suggestion that it might be wise to use standard alphanumerics in those computer applications that do not specifically require kanji output leaves most Japanese bewildered. Almost everyone seems willing and able to recite a whole litany of reasons for retaining kanji: The language contains too many homonyms to be written phonetically. The native vocabulary, after centuries of erosion by Chinese, is too impoverished to meet the needs of a modern society. Japanese ways of thinking depend on schooling in kanji. All such statements (these are but a sampling) are simply false. Why do the Japanese keep repeating them?
One line of thought, put forward forcefully by linguist Roy Andrew Miller, holds that “[t]o the Japanese today, the Japanese language is not simply the way they talk and write. For them, it has assumed the dimensions of a national myth of vast proportions” (Miller 1982: 5).
The essence of this myth is that the Japanese language is different from all other languages to a superlative degree: all languages differ from one another, but Japanese is unique “to a higher order” than all the rest. From this idea flow claims to the effect that the Japanese language is exceptionally difficult in comparison with all other languages; or that the Japanese language possesses a kind of spirit or soul that sets it apart from all other languages, which do not possess such a spiritual entity; or that the Japanese language is somehow purer, and has been less involved in the course of its history with the normal process of language change and language mixture that has been the common fate of all other known human languages; or that the Japanese language is endowed with a distinctive character or special inner nature that makes it possible for Japanese society to use it for a variety of supralinguistic or nonverbal communications not enjoyed by any other society—a variety of communication not possible in societies that can only employ other, ordinary languages. (Miller 1982: 10–11)
There is a sizeable grain of truth in this. Like residents of any smallish country with a long history, the people of Japan tend to run very hot or very cold about their native tongue. Either it is the most refined, most subtle, and most esthetically perfect language in all the world, or else it is the most difficult, most illogical, and most mysterious. Either way, the point is that one’s own language is “unique” in a sense very close to “superior,” though the hubris of the sentiment is masked by the choice of words. All this is not particularly Japanese and Miller is not the first to observe it.
Miller’s account of the causes of these ethnocentric language attitudes, however, is new. We must digress to examine his interpretation in some detail, for two reasons. The first is that Miller, who enjoys considerable academic stature in the U.S. (he is currently president of the American Oriental Society), has chosen to make his case in bitterly sarcastic terms, often embroidering the facts for the sake of heightened rhetorical effect.1 Indeed, his tone in Japan’s Modern Myth is so caustic that, as of 1985, it was rumored that no less than three Japanese publishers interested in bringing out a translation backed down once they had had a chance to read the original carefully. As one sympathetic but critical reviewer notes, the “bellicose and exaggerated style” of the book “makes Miller look like a grumpy old man” (Chew 1984: 479).2 Because we will be making many critical observations about contemporary Japanese attitudes toward kanji, we must take care to distinguish our interpretation of these observations from Miller’s.
Miller’s explanation of Japanese views on language is one calculated to appeal to Americans, especially if they lived through World War II. Unfortunately, it is also the explanation least in consonance with the historical facts. This is the second reason it deserves special attention. According to Miller, “[t]he nationalist-fascist dictatorship that led Japan into its ill-advised military adventures” made use of “an elaborate sustaining myth” that “may be broken down into three equal parts.”
One comprised selected elements drawn out of indigenous Japanese religious beliefs; the second embraced a selection of nineteenth-century nationalistic fantasies borrowed from Western Europe, particularly from Prussia; and the third was rooted in a number of Japanese elaborations and perversions of the already quite sufficiently perverted “master race” or “superman” ideas that had by then begun to become popular in the Third Reich. (Miller 1982: 35)
This ideology, according to Miller, did not perish in the cataclysm of 1945. Rather, it underwent a metamorphosis in which its focus shifted to the Japanese language.
The desolation of the bombed-out cities of postwar Japan was at once both the most eloquent and the most painful metaphor for the spiritual and intellectual desolation of the Japanese themselves. In city after city, nothing of all that had seemed so very substantial was left standing. What little had survived the nightly fire raids had immediately to be pulled down before it collapsed of its own damaged weight. The day the war ended, nothing of Japanese life seemed to have endured—except the language. (Miller 1982: 36)
Daily encounters with Occupation troops heightened Japanese awareness of their linguistic isolation and encouraged “the emergence of the Japanese language as the fetish focus for a new national sustaining myth,” but “[t]he old myth had been rich in elements relating to the language,” which provided the seeds for the new (Miller 1982: 38–39). Thus, Miller’s theory is essentially one of transference and continuity—and this is why it ultimately fails as an account of contemporary Japanese attitudes towards language.
In the first place, it oversimplifies and distorts the history of the pre-1945 myth, which Carol Gluck has painstakingly traced through its formative years in the Late Meiji Period (1867–1912). Although she devotes a chapter to “the language of ideology” in the metaphorical sense of “language,” nowhere does she uncover the kind of ideas about language, in the strict sense, that Miller alleges were there all along. Concepts such as “national essence” (kokusui), “distinctive national character” (kokuminsei), and “Japanism” (Nihonshugi) were not tied to language, and were in fact first articulated by intellectuals opposed to either Shintō or Confucian national doctrine (Gluck 1985: 112–113). These concepts changed in meaning over time and became part of the vocabulary of Miller’s “old myth,” but prejudices specifically about language are simply not “among the dependent clauses of the Meiji ideological utterance” (Gluck 1985: 286) that managed to survive 1945.
The second problem with Miller’s surrender-trauma theory is that it doesn’t jibe with chronology. Some of the strongest believers in linguistic nativism today are Japanese who have no adult experience of the war or Occupation, while many otherwise conventional Japanese of the prewar period held views about language more liberal than those of their postwar counterparts. Certainly, the Occupation was a turning point, but as far as attitudes toward language are concerned, it was a transition from a period of intellectual diversity and boldness to one of uniformity and complacency. Miller’s theory gets the order backwards. Today’s linguistic nativism is not a remnant of ultranationalism clung to by defeated Japanese who lost everything else, but a distinctively postwar conservatism that appeals to affluent Japanese who never learned (or have chosen to forget) the lessons of the past.
Finally, Miller’s argument fails to discriminate among the different kinds of nativist thinking he identifies. Japanese attitudes toward the writing system, foreigners, literature, education, and so on are for him just different aspects of a single phenomenon. In fact, belief in the indispensability of kanji cuts across the spectrum of Japanese public opinion in a way that most if not all the other ideas Miller attacks do not. There are plenty of liberal-minded Japanese who are dissatisfied with the way English is presently taught in Japanese schools, who laugh at the suggestion that foreigners can never really learn Japanese, who are opposed to the government’s discriminatory policies towards resident Koreans, and who scoff at the theories of Tsunoda Tadanobu.3 In short, plenty of Japanese reject one or more of the attitudes that Miller imputes to the nation as a whole. Yet these same people, when asked about kanji, will repeat, as if rehearsed, the whole host of stock explanations for the use of kanji mentioned earlier.
What, then, lies behind the Japanese attachment to kanji? The answer, as we shall see, is to be found in the combined effects of the educational and script reforms introduced under the Occupation. These reforms have greatly increased the ability of the school system to inculcate the population with what might be called the folklore of kanji, not only by increasing the number of students in school and the years they spend there, but also by demanding more of the individual student while cheapening the social value of being literate.
Kanji and Literacy
Today’s Japanese writing system is a remarkably complex and unwieldy thing, but it must be kept in mind that it is a simplified version of the system that existed prior to 1945. Following World War II, the Japanese government took several major steps towards script reform, culminating a long period of both public and official discussion of the many forms and styles of Japanese writing that stood in the way of full economic and social development (Twine 1983, Seeley 1984). Without those decades of preparation, it is unlikely that the government would have been able to respond to Occupation recommendations for reform as promptly and effectively as it did. Between 1946 and 1959 it moved to limit the number of kanji in daily use, eliminate almost all archaic kana spellings, simplify hard-to-write kanji shapes, place restrictions on kanji readings, and standardize okurigana usage. Seeley calls these the TK Reforms because the 1,850 approved characters were known as the tōyō kanji.
Many scholars who have written about Japanese literacy are of the opinion that the TK Reforms were rather superfluous, something that was done to appease Occupation personnel who were advocating much more radical kinds of reform, such as romanization by edict. In this view, Japan was a relatively literate nation even at the end of the Tokugawa Period (1600–1867); if there were any problems with kanji before 1945, the TK Reforms eliminated them. There is, however, an important distinction to be made between literacy as usually defined in technical studies (minimum ability to read and write) and literacy as a vehicle for full and free participation in society. The use of kanji makes this distinction crucial: one might be able to write and read kana and know the principles underlying the use of kanji yet be unable to read texts without furigana or write an acceptable petition to the authorities. Before 1945, this discrepancy was greatly intensified by the heavy influence of written Chinese on Japanese style. Although writing in colloquial Japanese certainly existed, formal writing demanded familiarity with kanbun, and although the kanbun tradition gradually waned in the twentieth century, traces of it can still be found in ordinary kanamajiribun writing. Chinese literary studies were a major part of the prewar higher school curriculum (Roden 1980), and ability to use kanbun expressions is still regarded as a sign of erudition.
In Western countries, there is a similar gap between mere literacy and the level of education at which the doors to social equality and remunerated creative accomplishment are opened; but the Western gap is negligible compared with the one that existed and continues to exist in Japan, where all literacy, and hence all education, is grounded in kanji. Claims of literacy rates in excess of 99 percent (e.g., Sakamoto & Makita 1973: 444) are grossly misleading if not downright false. As Miller notes (1982: 186), this figure must be wrong simply because the percentage of handicapped persons physically unable to read must exceed one percent in any normal population. More to the point, however, the reason Sakamoto, Makita, and other Japanese claim super-high rates of literacy is that they accept a definition of literacy that totally ignores its social significance.
Some historians of the “modernization” school, which has dominated postwar scholarship in the U.S., have also missed this point. InBlack et al. 1975, a comparative study of Japanese and Russian modernization, Russia fares poorly whenever literacy is mentioned, yet virtually no hard evidence is offered for the sanguine view of early Japanese literacy. For example,
The difficulty of the Japanese writing system seems not to have been an overwhelming obstacle in the spread of literacy. Japan in the first place shared with the rest of the East Asian cultural zone a boundless respect for the written word and for scholarship. One may further surmise that the very difficulty of the Japanese writing system made it a great challenge, and hence desirable, while at the same time it provided a level of discipline and patience in the acquisition of literacy that doubtless had positive effects in the realm of social stability. Writing in particular was a highly refined artistic tradition, so much so that all artistic production in premodern times was intimately associated with the written word. In Japan literacy received an aura of prestige far greater than it did in Russia. (Black et al. 1975: 108)
The difficulty of the writing system certainly was, and remains, a force for “social stability,” but what does that, or “artistic production,” have to do with literacy? The key question is how many Japanese were able to read and write, and at what level. The evidence presented is too impressionistic to allow for any meaningful conclusions. One cannot, for example, infer from the use of “signboards to communicate the will of the shogunal and domain authorities” (Black et al. 1975: 109) that there was “an increasing reliance on the written language in the political system of Tokugawa Japan”; more likely than not, the ordinary citizen learned what a signboard said because some educated person read it out loud to the anxious crowd that had gathered to see it posted. Again, the fact that “Fukuzawa Yukichi’s Conditions in the West … is reported to have sold 150,000 copies in its first edition in 1867” (109) is meaningless in isolation and must be interpreted in relation to the extraordinary novelty and timeliness of Fukuzawa’s subject and the number of potential readers.4
A more revealing use of publication statistics is found in Kinmonth’s study of the self-help literature of the Meiji Period, where the question of literacy is mentioned in connection with the influential pamphlet Gakumon no susume (The Advancement of Learning) by Fukuzawa Yukichi. This was actually the title of a series of seventeen pamphlets published between December 1871 and November 1876. But the first outsold the rest: by 1880, Fukuzawa estimated that 200,000 legitimate and 20,000 unauthorized copies were in circulation. On this basis he concluded that one out of every 160 Japanese had read it. Kinmonth goes further, claiming that “perhaps as many as 1 out of 10 or even 1 out of 5 who could read it had done so” (Kinmonth 1981: 45; emphasis added). Particular attention should be paid to the way in which Kinmonth arrived at these figures.
For a rough calculation of potential audience, I assumed, as did Fukuzawa, a total population of 35 million. From this I deducted one third for infants and nonliterate juveniles, one-half of the remainder for females, and assumed that 20 percent of the remaining population (2.3 million) could read Gakumon no susume in its original form. Ronald Dore has placed late Tokugawa literacy at 40–50 percent for males, a figure that has been criticized by Japanese scholars as being somewhat overly optimistic. See Fukaya  p. 55, n. 127. Considering that in the 1890s conscription tests were turning up total illiteracy rates of 25 percent (seeFukaya  pp. 285–286) and that Gakumon no susume was fairly difficult in its original form (so much so that some illicit versions were simplified), 20 or even 10 percent seems a more reasonable rate. (Kinmonth 1981: 45–46.)
Even with “total illiteracy rates of 25 percent,” nineteenth-century Japan may very well have been more literate than Russia. But the significance of this purely quantitative observation pales in the light of the enormous qualitative difference between the Japanese and alphabetic writing systems.
This is hardly a new or radical insight, although views such as those of Black et al. are all too common. Other scholars have been more sensitive to the problems inherent in the definition of literacy. For example, in his 1965 Society and Education in Japan, Herbert Passin stresses the growth of education during the Tokugawa period, the diversity of its intellectual life, and the role of its educational institutions as precursors of later educational reforms (Passin 1982: 13–61). He concludes that “[w]hen feudal Japan first blinked her eyes open on the modern world,” roughly four of every ten Japanese adults were literate (Passin 1982: 56–57; seeTable 7); however, he is quick to note that “[m]uch of it was certainly of a very low order—perhaps the bare ability to write one’s name or to read simple materials with effort” (Passin 1982: 58). Taira (1971: 375–376), working with Ronald Dore’s (1965: 321) optimistic though qualified estimate of school attendance by the end of the Tokugawa Period and later school-attendance statistics, puts the male and female literacy rates in 1868 at about 35 and 8 percent respectively. Although he notes a rise to about 75 and 68 percent respectively over the next forty-five years, he emphasizes the need to analyze “Meiji Japan’s progress in education and literacy without the glow of ‘rapid’ economic development in the background.”
|Social Group||Estimated Literacy (%)||Proportion of Total Population (%)||Rate of Literacy (%)|
|Peasants in the more isolated areas||“20”||8.7d||1.7||1.7|
|Note: Figures in quotation marks are from Passin 1982: 57; other figures are estimated or calculated.
Quantitatively, the spread of literacy in Meiji Japan was credible enough. Qualitatively, however, the “compulsory” education imposed on the unwilling populace without a full commitment of public resources (“compulsory” but not “free”) was painful as well as wasteful. (Taira 1971: 372)
Taira specifically identifies the use of kanji as a factor contributing to the questionable qualitative results of the prewar educational system (390–391), and takes pains to explain the attitudes that made kanji and kanbun sacrosanct:
Spoken Japanese was only a tool of oral communication, and even lowly beings like beggars and thieves were capable of speaking it. Writing, therefore, had to show the mark of accomplishment with rules of its own sharply distinguished from the ordinary manner of speech. Often, written Japanese placed style, grammar and aesthetic value above the requirements of clear communication. Written Japanese was, therefore, grossly inadequate as a means of conveying precise technical information, though enormously effective as a tool of moral exhortation or of emotional appeal. But to be influenced by written Japanese, the ordinary souls had to have the text read to them by a person with the appropriate level of literacy. For the same reason, those who were literate considered themselves many flights above the ordinary people. The standards of literacy for written communication were so high that many, despite good elementary education, gave up their efforts to rise to that level of literacy and consequently allowed their elementary knowledge of written Japanese to atrophy. (Taira 1971: 392; emphasis added.)
Although Gluck paints a somewhat brighter picture of the same period, emphasizing the growth of the press as a factor in social change, she notes that “[t]he establishment of the written colloquial language (genbun itchi) and the widespread use offurigana glosses increased the accessibility of newspaper and popular works” (Gluck 1985: 172–173). This is something of an understatement: accessibility could not have increased otherwise. As Twine remarks in her study of pre-twentieth-century script reform movements,
Years of arduous study were required to master the literary forms and script of officialdom, and only the upper classes had the leisure to devote to it. The degree of literacy attained by the commoners was usually just sufficient for the small concerns of everyday life and the perusal of popular fiction. (Twine 1983: 116)
As always, we must not forget that the use of both kana and kanji makes the definition of literacy in the case of Japanese unusually broad.
Lower-class education extended to little more than the kana scripts; even upper-class children, ostensibly receiving a thorough Confucian education, often merely learned to recite passages by heart rather than actually read and understand them. Hours of concentrated study were required to memorize characters before the contents of books could be absorbed. (Twine 1983: 117–118)
Despite Japan’s rapid economic development during and after the turn of the century, reliable figures on literacy are hard to find. Government reports were—and still are (Booth et al. 1984: 7)—little more than estimates based on school attendance. Writing in 1947, the only statistics that John DeFrancis found worthy of quoting were from the German study cited in Chapter 1 (Scharschmidt 1924).
Before the war the requirements for graduation from lower primary school, which is all the education received by most Japanese, included the ability to read and write 1360 Sino-Japanese ideographs and to recognize another 1020, or a total of 2380 in all. Tests conducted when male youths were called up for military service years later showed that twenty-year old youths with public school education remembered how to write on an average only 500 or 600 and to recognize only 1000 of the 2380 ideographs which they had once learned. (DeFrancis 1947: 220)
Figures on the rate of “practically illiterate” young men turned up by conscription tests, such as those cited in 1931 by Nitobe Inazō, also suggest that the educational system was not meeting its stated goals (see Table 8). In a 1929 work, Nitobe refers to “an illuminating study made of the language instruction as given in Japanese and European, especially Bavarian, schools,” which showed that “Japanese children spend 44% of their school days in learning their mother tongue as against 31% by Europeans” (Nitobe 1972: 4.433). He continues,
The result of investigation seems to show that the vocabulary and the reading capacity of an ordinary Japanese youth at the age of fifteen is about on a level with the average German child of eight.
A curious corroboration of this statement is furnished by observations of the blind, who learn to read by the Braille system only the 47 characters of the Kana and who are not taught Chinese ideographs. It has been repeatedly proved that the blind acquire, in the same length of time, more solid knowledge than ordinary children—be it of history, geography or literature. (Nitobe 1972: 4.433–434)
|Year||Number Tested||Number Illiterate|
|Note: Figures for 1929 do not include Tokyo and Osaka Prefectures; all figures are based on Nitobe 1972: 3.245.|
Unfortunately, neither Scharschmidt nor Nitobe cite sources for their information.
Other evidence of the period just before and during World War II is frankly anecdotal. The following example comes from the writings of the late Tōdō Akiyasu, one of Japan’s foremost Chinese linguists:
Postwar Japan was able to rebuild rapidly. Could it have done so had it not cut military expenditures to the bone under the protection of the [new] Constitution? Across the country, young and old alike came to be able to read newspapers and magazines and to write letters and diaries. Would this have happened without the revolution in Japanese orthography? When those of us who know the past think back to before the war, surely these are the facts which make the deepest impression. I was drafted into the 36th Regiment of the Mie Prefectural Infantry; among the 150 men in my company, some from the Shima seacoast and the mountainous areas of Iga couldn’t read a whole sentence. I was ordered to teach ten or so, who could not even write kana properly, every evening. The company commander made them learn the Imperial Rescript to Soldiers and Sailors (Gunjin chokuyu) by rote, and they would bark it out with their eyes fixed on the ceiling: “Military men should set their mind on nitrogen [chisso] (which should have been ‘frugality’ [shisso]).” Whenever I went back to the country for the weekend, old couples in the neighborhood would bring me letters from their sons and ask, “Please read this for us.” (Tōdō 1982: 173–174; trans. JMU)
Personal reminiscences—even accurate ones—are certainly not the best sort of evidence, but when a scholar who has devoted his entire life to the study of the Chinese language and writing system comes down squarely on the side of script reform, one must pause to reflect. Moreover, illiteracy and the poor technical education of the average conscript figure in other wartime reports. In one case, a soldier sent to the rear to fetch a replacement part for a damaged artillery piece returned with a completely different part; he had forgotten the official kango name he had been told—to him, it was just a meaningless string of syllables (Hoshina 1949: 209). Incidents like this often led to disaster (Hirai 1948: 330), and by 1940, the army had limited the number of kanji for weapon parts to 1,235 and was studying the possibility of cutting that number in half (Hoshina 1949: 210, 215). Ironically, the military continued to pepper its reports in civilian newspapers and magazines with obscure, hard-to-read kanji in the belief that this would to impress and cow the general public (Hirai 1948: 327–328).
The first full-fledged attempt to measure literacy was carried out in 1948. This survey, conducted by the Civil Information and Education Section of the Occupation, involved the testing of about 17,000 Japanese men and women between the ages of fifteen and sixty-four throughout the country. According to Ishiguro Yoshimi, who chaired the survey’s Central Planning and Analysis Committee, it was a first not only by Japanese but also by world standards (Ishiguro 1951: 181). It is of particular interest not only because of its statistical thoroughness but also because some Japanese (e.g., Ishii 1985: 22) cite it as proof that the level of literacy of the majority of prewar Japanese was actually quite high. Certainly that is the interpretation that Joseph C. Trainor, who worked in the Education Division of CI&E from 1945 to 1952, chose to give it in his memoirs:
The results showed conclusively that with the exception of one group, those who had had no formal education, Japanese of all ages, degrees of schooling, occupation, or whatever other characteristic, possessed the ability to read, write and comprehend materials of the type which regularly appeared in their newspapers. The performances on the tests varied in relation to the amount of schooling and the age level; but to the extent that ability to read newspaper material was a measure, the Japanese could without hesitation be described as literate. Results of measurements in the other skills were similar and it seemed possible to put forward as a tentative conclusion the contention that the Japanese people possess the ability to use their language effectively in their social lives. (Trainor 1983: 323)
Trainor’s views obviously support the usual appraisal of prewar Japanese literacy. It must be remembered, however, that he was a civilian employee of the Occupation and strongly believed in the official policy of not interfering in Japanese affairs unless absolutely necessary for the fulfillment of specific Occupation aims such as fostering democratic institutions, eradicating the effects of ultranationalist propaganda, and so on. Trainor himself was not surprised by the results of the survey; they “were startling and disturbing to many who had preconceptions regarding the reading ability of the Japanese people” (323, emphasis added). By this he clearly meant reform-minded officers of the Language Simplification Branch, for whom he had little respect (300–308), and the various Japanese proponents of romanization, whom he considered “extremists” (57, 315, 325). Even if his opinions did not bias his judgment of the survey results, there are three other reasons for discounting his assessment. He was not involved with the survey itself (Yomikaki Nōryoku Chōsa Iinkai 1951: 32); his understanding of the Japanese writing system and general linguistics was deficient (e.g., Trainor 1983: 214, 313, 315, 318, etc.); and some of his statements (written in 1952–1953, immediately upon his return from Japan, though published later) are contradicted by documents of the time that have recently come to light (cf. Nishi 1982: 199–205).
Most important, the conclusions of the literacy survey itself (Yomikaki Nōryoku Chōsa Iinkai 1951: 425–430) do not agree with Trainor’s. The survey found that the rate of illiteracy (monmōritsu, complete inability to read or write) was indeed very low;but it also concluded that only 6.2 percent of the population were literate in terms of the survey definition, which was liberal. Full literacy was defined as answering all questions correctly; illiteracy was defined as scoring zero. By today’s standards, all the questions were very simple. The ability to write kanji from dictation (kanji no kakitori), which was identified as the single most important skill tested, was found to be “remarkably low” in all groups surveyed. Performance was closely correlated with levels of formal education; subjects whose education had been disrupted by the war did significantly poorer in all areas. This showed that mastery of the basics, without years of supplementary instruction in kanji, was inadequate for full literacy. Finally, the claim that the average Japanese experienced trouble dealing with the media of mass communication, which had long been made by script reform advocates, was deemed proven.
Clearly, although the rate (quantity) of complete illiteracy before World War II was low, so was the level (quality) of literacy achieved by the average Japanese. It might be thought that, thanks to the postwar reforms of the schools and the writing system and the general improvement in the Japanese standard of living, disparities in levels of literacy have been eliminated. The truth, however, seems to be quite the opposite: even today, despite the thoroughness of Japanese secondary education, the quality of literacy it imparts is not uniformly high.
Sixty years after Scharschmidt, and thirty-seven years after his own research on the subject, DeFrancis could report no change in the level of individual student accomplishment.
Sato Hideo, head of the Research Section for Historical Documents, National Institute for Education Research in the Japanese Ministry of Education, has estimated that public school graduates, who now receive nine years of compulsory schooling, retain a recognition knowledge of the 1,945 kanji but soon forget how to write all but 500 or so (1980: personal communication). (DeFrancis 1984: 217)
What has changed is not average student performance but the number of students involved, or, to put it another way, what is expected of the average Japanese youngster. Compulsory education was extended from six to nine years soon after the war. The Report of the U.S. Education Mission, submitted to General MacArthur on 30 March 1946, noted that “approximately 85% of Japanese children terminated their formal education after leaving elementary school” (U.S. Department of State 1946: 21); by 1974, more than 90 percent of all students were graduating from high school (twelfth grade), a higher percentage than in the U.S. (Rohlen 1983: 3). As a result of this change, the burden of learning thousands of kanji, once shouldered by only a small, virtually all-male fraction of the school-age population, must now be borne by nearly all Japanese children. Students who cannot make the grade, once a minority within a small minority, now constitute, in absolute numbers, a substantial group. What is more, the level of literacy that these students are having a hard time reaching no longer commands the kind of respect it once did.
Postwar Japanese education has been very effective in many ways. It is a mistake, however, to suppose that cultural homogeneity carries over into uniform academic achievement.5 While cram schools (juku) flourish and competition among diligent students remains keen, the willingness of young Japanese to read material written at a high level, apart from schoolwork, has clearly declined over the past twenty years. Comic books, sports newspapers, and sensational weeklies are crowding out serious books and magazines, and many of the larger, older publishing houses are said to be ailing financially. In one of the Kōbe high schools Rohlen studied between 1974 and 1975, an extreme but not necessarily extraordinary case, the level of student literacy was appalling.
The “discussion” comes to an end. Everyone opens his book to the appropriate page, and the teacher appoints a student to read aloud. His progress is slow as he stumbles over three or four characters per sentence. I am shocked to find that I know some characters the students are regularly missing. And so it goes for the remaining class time: seven lines of text covered in thirty minutes. The almost constant corrections of pronunciation make the lesson unbearably tiresome and boring. There is no time to discuss the meaning of the essay, its style, or its charm. The fact is that the students simply cannot read it. (Rohlen 1983: 29)
Sakura is a night-school, the least prestigious of the five studied by Rohlen. “Rarely are there more applications than openings to Sakura, and to fill its official quota of eighty freshmen the school has a late application period and has been willing to accept students clearly incapable of high school work” (Rohlen 1983: 32). But these very qualifications only go to show that the stock claims of 99 percent literacy cannot be taken seriously.
One indicator of extreme stress is suicide. Student suicides in Japan are often ascribed to the intense pressures of “examination hell,”6 but as Rohlen (330–331) points out, police officials classified only 26.6 percent of the suicides committed during the first half of 1977 by persons under twenty-one as “school-related.” And according to an earlier national study, “although 90 percent of the fifteen- to nineteen-year-old group were in school in 1974, 24 percent of the suicides in this group were committed by people not in school” (Rohlen 1983: 333). The greater concentration of cases among those few who had already dropped out of high school or left school after ninth grade (the last year of compulsory education) suggests that the rate of Japanese juvenile suicide is not as directly tied to the mad rush to pass entrance exams to elite high schools and universities as is usually thought. If the exams are not the immediate problem, what is? Rohlen (1983: 334) mentions “the custom of not keeping slow students back and the relatively undeveloped use of screening techniques to discover (and perhaps treat separately) students with learning disabilities.” Another, simpler explanation is possible, however: If students don’t keep pace in the learning of kanji, they gradually and inevitably fall behind in all areas of study. Unless they are obviously handicapped, society makes no allowance for them because the writing system has become a sacred cow.
An examination of postwar juvenile suicide rates implicates the writing system in another, more subtle way. The peak years for juvenile suicides were 1955 through 1958 (Rohlen 1983: 329). There was a marked drop after 1958, which as Rohlen points out contradicts the hypothesis that exam pressures are the prime cause of youth suicides: “[a]s the percentage of young Japanese involved in the pressures of the exam system increased, the suicide rate for their age group declined” (328). But what about the dramatic rise up to 1955–1958? It was during this period that the first wholly postwar educated students were leaving school. For them and their successors, the TK list of 1,850 kanji, though intended as a maximum limit on the number of kanji, was actually a minimum standard of literacy. They were, in this sense, the first generation in Japanese history to carry the full weight of kanji in a society that all too often equates knowledge of many kanji with real intelligence. There were undoubtedly other factors that contributed to the cresting of the juvenile suicide rate around this time, but it is hard to believe that the stress and anxiety of this unprecedented situation did not have something to do with it. A mere twenty years before, it was “common sense” that only the handful of young men who went on to higher schools and universities could read and write thousands of kanji or needed to; now, in the middle of hard times and scarce jobs, that was the yardstick by which everyone was to be measured.
Significantly, the first postwar surge of popular interest in the Japanese language occurred at just this time. The press called it a Japanese language boom (Nihongo būmu), and it was marked by bestsellers such as Nihongo (The Japanese Language) byKindaichi Haruhiko (1957, 1978), Nihongo no kigen (The Origin of the Japanese Language) by Ōno Susumu (1957, 1970) and a host of similar books. The Ministry of Education carried out a second literacy survey (1955–1956), but the findings turned out to be rather similar to those of 1948. “The results of [both] the surveys well illustrate the phenomenon of ‘restricted literacy.’ Between 20 % and 50 % of Japanese language users were described as experiencing intense or noticeable problems in the use of the written language” (Neustupný 1984: 118). But there was much more interest in the uniqueness of the Japanese language than in these mundane and rather unpleasant statistics. As uncertainties about the new educational system were rendered moot by more and larger graduating classes, Japan settled down to the happy myth of 99 percent literacy. “No new survey of literacy has been conducted since 1956” (Neustupný 1984: 120).
This period also marked the beginning of a political backlash against the limited script reforms that had begun with the tōyō kanji list of 1946 and culminated with the okurigana rules of 1959. Some of the Ministry of Education officials, such as Shiraishi Daiji (interviewed 8 January 1986), and others who were responsible for the TK reforms had mistakenly believed that they must accept a restriction in the number of kanji in order to forestall an Occupation romanization edict (never a serious threat). In fact, however, there had been years of preparation for limited script reform; only Japanese decided exactly what steps would be taken. Recommendations were made by a fairly representative and independent Japanese Language Council (Kokugo Shingikai) in a relatively apolitical atmosphere.
All this changed toward the end of the fifth term of the JLC, which lasted from 30 March 1959 to 22 March 1961 (Ōkubo 1978:105–133).
In early 1961, four conservative JLC members, with the acquiescence and probably the cooperation of responsible government officials, used filibuster tactics to prevent the JLC from electing a nominating committee. This gave the Minister of Education a free hand in appointing new JLC members for the sixth term, which began in October. In March 1962, the new council chairman announced that he, the vice-chairman, and six other members whom he had selected, would serve on a newly created Executive Committee. This new committee would initiate all matters to be considered by the council as a whole. At the third Executive Committee meeting, it was announced that the ministry was going to change the regulations governing the JLC. The Executive Committee consented. The new order gave the Minister the exclusive right to make JLC appointments. It took effect on 27 April 1962, without approval of the full Council. Chigusa Tatsuo, a justice of the Tōkyō Supreme Court and a member of the Executive Committee, pointed out the consequence of this action at the third Executive Committee meeting: “If the Minister [of Education] is to make the appointments, there is reason to fear that only those whose opinions are identical with those of the current Minister will be appointed” (Ōkubo 1978: 117; trans. JMU). He elaborated:
How shall we formulate a national language policy that is consistent not just with the language of the past but with the present and future [needs] of the people? This is the weighty charge of the Japanese Language Council. If national language policy swings to the left or right with every change of Minister, it is obvious that consistency will be lost, and that major educational problems will result as well…. I would like to point out that if Council members are selected by the Minister alone, then, whenever there is any complaint in the Diet about this Council, the Minister of Education will have to bear the entire responsibility. (Ōkubo 1978: 118–119; trans. JMU)
Chigusa’s words were prophetic. In less than ten years, the tōyō kanji (TK) would be superseded by the jōyō kanji (JK). Superficially, this made little difference; the JK list is the TK list plus an additional 95 kanji. The preamble to the new list, however, effectively repudiated the spirit if not the substance of the previous reforms. Describing the JK list merely as “a guide” (meyasu), not a definite recommendation to eschew unlisted kanji, the preamble is laced with vague language that undermines the ostensible purpose of selecting kanji for general use in the first place. Thus, “it is expected that kanji use will conform to this List as far as possible,” “there is suitable scope for reflection in its application, according to the circumstances at the time,” “there is no objection to deviation in certain areas from the way kanji are treated in this List,” and so on (Seeley’s translations [1984: 284]). The TK reformers, implementing an idea proposed in the 1930s by the author Yamamoto Yūzō, had recommended that furigana be avoided lest writers rely on them to justify the use of obscure kanji. The JK preamble, by contrast, first damns with faint praise the concept of limiting the number of kanji, and then superciliously remarks, “In cases where kanji seem difficult to read, one method might be to consider using furigana where necessary.”
As already observed, the TK list had been intended to set a cap on the number of kanji for general use. Instead, it ended up establishing a floor under the number of kanji, a plateau which one had to reach in order to be in the running for admission to top universities and top-level jobs. The JK preamble carries this subversion of the TK principles a step further. It addresses itself to “usage by persons who have to some extent experienced life in actual society or educational institutions after finishing study in the period of compulsory education,” implying that, if anything, formal schooling should aim beyond the limits set by the JK list.
Small wonder that, when an Ad Hoc Education Council (Rinji Kyōiku Shingikai) was convened in 1985, the fundamental issues of script and literacy were not even on the agenda. This Council was commissioned mainly because the government faced declining school enrollments and the prospect of an oversupply of teachers in the years ahead, and wanted expert sanction for a major overhaul of the educational system. That was, however, not the only reason for the creation of the Council. There is also the increasing incidence of school-related violence (including assaults on teachers, unwarranted corporal punishment, so-called bullying [ijime] among students, etc.); outright refusal to attend school (tōkō kyohi—different from simple truancy); steadily growing dependence on cram schools (juku); and the frequent re-entry problems of Japanese children who have attended schools overseas. The possibility of a connection between kanji-based literacy and at least some of these problem areas is not hard to see, and it is remarkable that the council, although it has recommended greater emphasis on creativity and less on rote learning, has chosen not to examine it.
Perhaps even more telling is the most recent JLC pronouncement on kana spellings. On 6 March 1986, the JLC issued relaxed guidelines on kana usage. Just as the JK list seems to be only a slightly expanded TK list, the new recommendations on kana usage appear to entail only trivial changes. Indeed, the only major change is to permit the use of hiragana zi and zu for etymologically correct di and du in certain words; however, this minor adjustment is accompanied by a general statement that effectively opens the door to the reintroduction of the prewar kana spellings done away with during the TK reforms. Interestingly, the constituency most pleased by the JLC’s action is the word-processor manufacturers. They do not want their machines to disappoint customers; the relaxation of the official rules means that they can now, in good conscience, program input routines that accept grammatically incorrect strings. For anyone who takes the problems that motivated the TK reforms seriously, however, the new JLC announcement is just one more regrettable step backwards.
Thus, the anti-TK backlash, which began just after the first “Nihongo boom,” has won the day. Although the writing system remains simplified in one sense, it has become a greater burden than even before. The attitudes of Japanese toward language and writing reflect this change. Before the war, even ultranationalists thought nothing of forcing Chinese and Koreans to learn Japanese and abandon their native languages; today, the man in the street is genuinely amazed at foreigners who speak and read Japanese. Before the war, some of Japan’s best minds were among the supporters of script reform; today, only a few show any interest in the subject at all. Before the war, a thorough knowledge of kanji was a ticket to first-class citizenship, but no one ever suggested that illiterates and semiliterates formed a culturally distinct group; today, it seems that internalization of kanji is a necessary (though by no means sufficient) condition for being Japanese.
Confusing Language with Writing
An important ingredient of this postwar shift in attitude is a systematic, often deliberate confounding of language with writing. Although versions of the Ideographic Myth can be found scattered throughout the prewar literature, Japanese scholars have built theories upon it in earnest only since the first postwar “Nihongo boom.” Their articles and papers all follow a pattern: At the outset, it is assumed that kanji possess innate prelinguistic meaning. Various data are then “explained” on the basis on this false assumption, showing (circularly) that using kanji is quite different from (indeed, superior to) all other methods of writing the Japanese language.
Of the linguists who give this position its veneer of authority, Suzuki Takao of Keiō University is perhaps the best known. Certainly, he is one of the most prolific (see Suzuki 1963, 1969, 1975a, 1975b, 1977, etc.). He also played an important role in the JK affair (Ōkubo 1978: 181–182), and has recently claimed that word-processors have “clinched it for the survival of Chinese characters” (Chin & Martin 1986). According to Suzuki, the use of kanji confers a special advantage on the Japanese writing system by lending Sino-Japanese compounds a “semantic transparency” that comparable words of English allegedly lack.
[T]he on and kun readings of Chinese logograms have become so internalized in the mind of the Japanese that they believe that almost all characters have two variant readings without thinking much about the historical development…. [E]ach logogram evokes in our mind a meaning, an idea, and a concept. And to say that the Chinese characters in Japanese (though not in Chinese) have a twofold phonetic realization is tantamount to saying that the Japanese conceptual system which exists in potentia is put into the state of actualization (in actu) through the instrumentality of two different phonetic media. (Suzuki 1975b: 180–182)
What Suzuki is trying to say is that, because most kanji have both an on and a kun reading, they are endowed with special cognitive properties when used to write Japanese. Even someone familiar with Japanese linguistics, however, is likely to get lost somewhere around “in potentia.” The following illustration, makes clearer what Suzuki has in mind:
In Japanese, the Aegean Sea is written 多島海, and pronounced as ta-tō-kai. Even those who have never set an eye on this term before, to say nothing of knowing where it is, can easily explain the meaning of it, for they can automatically paraphrase the term with the kun reading of the characters as shima-ōii-umi [sic], meaning “sea of many island(s).” The ease with which the Japanese associate a phrase or a word with its literal meaning is made possible by the dual phonetic renderings of the Chinese characters in the Japanese writing. (Suzuki 1975b: 189)
Because the kun readings of kanji are more or less just Japanese morphemes, and since learning how to “spell” Japanese entails memorizing kun readings, Suzuki thinks that Sino-Japanese compounds are more “transparent” than English compounds that make use of Latin and Greek roots. This, however, confuses meaning with etymology. If etymology were meaning, we would have no need of textbooks, only dictionaries! As Miller explains in his comments on Suzuki 1975a, “Etymology tells where a word came from, not what it means. To know what it means, you have to know the language, and moreover, you have to know the word” (Miller 1982: 190). Indeed, he could have gone farther: etymology can be misleading. To take an English example, consider “anti-Semitic.” Arabic is a Semitic language, but this word definitely has no connotation of “anti-Arab.” It goes far beyond “anti-Jewish”—the Spanish Inquisition was anti-Jewish, but (in theory at least) spared converts to Catholicism. To know what “anti-Semitic” really means, one must know something about Nazi ideology, its historical background, the Holocaust, and so on. Knowing the roots involved is not enough. In the same way, the fixed readings of kanji often give only a hint—sometimes a wrong hint—of the meanings of the words they are combined to represent.7
Developers of Japanese and Chinese word processors are another source of misinformation. Typically engineers whose knowledge of linguistics is confined to what they have picked up from writers like Suzuki, they have little awareness of the great care that must be taken in the design, execution, and evaluation of psycholinguistic experiments. For example, in one study, Doi & Yoneda 1982, twenty members of the authors’ own research organization were timed as they read Japanese selections written in different styles (customary, kana-only, rōmaji, and so on). There was no attempt to assess comprehension, no pretest training in the unfamiliar modes of reading, and no experimental control. The authors seemed to believe that the use of sophisticated statistical techniques would somehow compensate for these obvious procedural errors. All they really succeeded in showing was that their subjects could read texts written in a familiar manner much faster than texts written in an unfamiliar manner. Since it takes hundreds of hours for a native speaker of any language to learn how to read it fluently in even one script, this is just what one would expect. Yet there are Japanese computer scientists who unabashedly cite studies like this as proof that Japanese simply cannot read their own language in rōmaji.8
Reading theorists are perhaps best represented by Sakamoto and Makita, whose 1973 survey article is a classic in the genre of supposedly scientific defenses of kanji. They simply assert, without delay, that “each Kanji has its own meaning” (Sakamoto & Makita 1973: 441). They go on to make a great deal of photographs taken by Sakamoto of the eye movements of college students reading identical sentences written alternatively all in hiragana and in a normal mixture of hiragana and kanji. From this “experiment”—presumably the use of different transcriptions of identical sentences was supposed to serve as a control—they conclude that the use of kanji makes reading quicker and more accurate. In explanation of this result, they note that the use of kanji reduces the total number of symbols that must be read and provides visual contrast that compensates for the nonuse of spaces. At the top of their list (443), however, is the unfounded assumption that kanji possess innate meaning; missing entirely is any discussion of the obvious fact that no educated Japanese reads all-hiragana texts. In short, Sakamoto and Makita fail to see that it is the Ideographic Myth itself that must be subjected to empirical testing, not the “superiority” of using kanji.
This is further illustrated by their discussion of dyslexia in Japan, which they claim is virtually nonexistent because of the nature of the Japanese writing system. Their own account of how they reached this conclusion, however, puts it in doubt. For example, Makita specifically instructed the schoolteachers he surveyed in 1966 to disregard “those with intellectual retardation or visual impairments” (Sakamoto & Makita 1973: 459) instead of collecting as many cases as possible and analyzing the data himself. This was a methodologically questionable step to take in a society where schoolchildren are never kept back a grade for academic failure and teachers are held fully responsible for their charges’ progress. Negative survey results from “child guidance clinics, educational counseling services, and other child study institutions” (460) are probably also misleading because their clientele is self-selecting. In the U.S., dyslexia is often diagnosed in children referred to a specialist because of behavioral problems; it is entirely possible that dyslexic children in Japan exist but are better at compensating for their disability socially—because Japanese writing involves both kana and kanji, it may be possible to have trouble with one kind of character but not the
Even if the clinical incidence of dyslexia in Japan is actually lower than elsewhere, it is not evident what features of the Japanese writing system, if any, might be responsible. There is little unanimity among experts on the relationship between social factors and Japanese reading ability (e.g., cf. Duke 1977 and Overly 1977). Experiments comparing kana and kanji have produced some evidence that it is easier to learn to recognize symbols when they are identified with whole words rather than with meaningless subword syllables, and also that graphic complexity does not in itself make symbols harder to learn (Steinberg et al. 1977, Steinberg & Oka 1978, Steinberg & Yamada 1979a)9. But as DeFrancis (1984: 171) notes in his criticism of similar Chinese/English tests, results based on a sample of only a few characters may have little or no bearing on the real-life task of learning hundreds of characters. In any case, such results certainly do not demonstrate that Chinese characters facilitate the reading of Japanese or Chinese, nor do they imply that kanji express meaning “directly.” Indeed, what reliable experimental evidence exists suggests that, at some level of cognition, speech recoding is essential for the interpretation even of allegedly ideographic characters. Although numerous experimental studies purport to show that kana and kanji are processed differently in the brain, each and every one is flawed with unjustified assumptions and methodological oversights (Paradis et al. 1985: 57–58); none comes close to demonstrating hemispheric lateralization (dominance of one hemisphere of the brain over the other) or any other consistent difference between kanji and other kinds of writing. Nor do any of the sixty-nine clinical cases of acquired dyslexia in Japanese patients dating as far back as 1901 provide the slightest evidence that kanji reading bypasses speech recoding (Paradis et al. 1985: 196–199); on the contrary, taken together, they show that reading of kanji and kana, like that of other scripts, is a predominantly left-hemisphere activity.
The shortcomings of the work to date analyzed by Paradis and his colleagues do not deter Kaiho Hiroyuki, the psychologist at Tsukuba University who recently coined the phrase “sciencing kanji.” In his contribution to a collection of essays, which he edited and issued under this unusual title, Kaiho begins by citing an experiment that he claims proves what Suzuki would call the semantic transparency of kanji. In this experiment, a variation on a classic technique first reported by J. R. Stroop in 1935, subjects are timed as they attempt to identify colors as quickly as possible. The colors are presented in three forms: as kanji, as hiragana, and as meaningless X’s. The kanji and hiragana are purposely selected to spell the name of a color different from that of the ink used to write the word itself; the subjects subconsciously read these distracting words, and this interferes with their attempts to name the colors. According to Kaiho, a hundred trials took 80.6 seconds with kanji, 72.1 seconds with hiragana, and 56.9 seconds with the X’s (Kaiho 1983: 36). He then gives these data a most bizarre interpretation: the greater interference caused by kanji versus hiragana shows that the brain processes kanji more quickly. This follows from Kaiho’s unexamined assumption that kanji communicate linguistic meaning without the aid of language, and that a semantic conflict must be the cause of the slowdown in the color-naming task. Actually, however, the delay is better explained by the observation that shape and color recognition compete for the same cognitive resources in the brain’s right hemisphere (Yamada 1985: 257–273): kanji are processed more slowly than kana and X’s simply because they are the more complicated patterns. In addition, as noted by Paradis et al. (1985: 22), “the color terms used in these studies are those that are usually written in kanji, and hence the kana Stroop stimuli are less familiar than their kanji counterparts.” This alone might well account for the greater delay with kanji.
One contributor to Kaiho’s book deserves special attention because he unwittingly shines such a bright light on the sociolinguistic motives of the “kanji scientizers.” This is Sasaki Masato, a lecturer at Tsukuba; his chapter (Sasaki 1983) is entitled “Kanji for the Blind.” There are, in fact, two competing kanji braille systems: Hasegawa’s, based on readings, and Kawakami’s, based on graphic components (radicals). The differences between them, however, pale into insignificance when one considers the common aim shared by proponents of both systems. Evidently, they agree with the twelfth-century Chinese encyclopedist Zheng Qiao, who wrote, “The world is of the opinion that those who know Chinese characters are wise and worthy, whereas those who do not know characters are simple and stupid” (DeFrancis 1984: 1)10. Why else would anyone expect the sightless to give up a perfectly adequate kana-based braille system? (For a brief account of Japanese braille, see Unger 1984.) One wonders what prewar Japanese intellectuals would have thought of such a condescending attitude. Nitobe Inazō, for one, would have been shocked. As if to underscore his comments of 1929 cited earlier, he later wrote:
To a foreign observer, the number of years devoted to secondary education must seem strangely out of proportion to the results obtained. To the Japanese, the explanation is easy…. [T]he blind man can be better educated than his more fortunate brethren who are endowed with good sight; for the former, by acquiring the forty-seven letters of the I-ro-ha syllabary, through the Braille system, can read history, geography or anything written in that system; whereas he who has eyesight cannot read the daily papers unless he has mastered at least 2000 characters. (Nitobe 1972: 4:274–275; cf. 1972: 3.248–249.)11
The very idea of using patterns of a few embossed dots to represent kanji belies many of the stock arguments given for maintaining their use: their esthetic mystique; the ways in which they save space on paper; the alleged ease with which they are learned; their “semantic transparency”; the wholesome effects of practicing calligraphy; and so on. It shows that the Japanese attachment to kanji is intimately tied to the shared experience of mastering a complex body of knowledge that defines group membership. Whether kanji facilitate communication or not is a secondary consideration.12 What really matters, as kanji braille so clearly shows, is the sociolinguistic function of the Ideographic Myth: to sanctify the status quo.
Of course, the defense of kanji culture through allegedly scientific methods is a relatively esoteric business. More common is the time-honored practice of twisting historical and linguistic facts to meet the demands of the Ideographic Myth. Such slanting of the linguistic history of East Asia is certainly not a postwar phenomenon; however, postwar education in Japan has elevated what was once a collection of harmless fables, of interest to only a few, into a set of clichés that every Japanese schoolchild is taught as the justification for hours of copying and rote memorization of Chinese characters.
At its most moderate, the historical rationalization for kanji holds that Japanese esthetic sensibilities, intellectual perspectives, and national characteristics are so wedded to the age-old use of Chinese characters that, without them, Japanese culture as we know it would be threatened. At its most extreme, it is the impossible prediction that the Japanese language itself would perish without kanji. The comments of Dr. Uenohara Michiyuki, senior vice-president of Nippon Electric Company (NEC), one of Japan’s leading computer manufacturers, are typical.
People using European languages can engage in data-exchange and dialogue with machinery through typing with almost the speed of conversation. But this is impossible in the Japanese language. If oral input becomes possible, th[is] handicap will be totally eliminated. Because of the phonetic simplicity of the Japanese language relative to European languages, oral input will give the Japanese an advantage, reversing the present situation. The world is made in such a way that advantages and disadvantages are always offsetting relationships.
The ultimate mission of technology is to make up for disadvantages of human beings and society and bring them progress. If it ruins advantages that human beings have, it does not deserve to be called technology.
A local culture is something that has been developed through the long history of the region and human life itself. It should not be altered because of technology. Rather technology must be altered to fit the local culture. If the Japanese language is abolished for the sake of convenience of usage of computers, the Japanese will be deprived of their identity. (Gregory & Etori 1981: J40)
Even those who see the merits of script reform are not immune from the temptation to read deep historical significance into the continued use of kanji:
A reform of the written language would not only do away with a major hindrance which clogs the wheels of the nation’s affairs; it would also save years of heavy intellectual grind in the life of every successive generation. It would have even more far-reaching consequences, because the confusions and ambiguities of the Japanese script are reflected in the structure of Japanese thinking. And that, perhaps, is the cause of their instinctive resistance against reform of the script. They refuse to part with the comforts of ambiguity, and prefer the printed page, like the streets of the capital, to remain a labyrinth, where only the initiate, guided by his intuition, can find his way—or lose it in agreeable detours. (Koestler 1961: 182–183)
Uenohara, Koestler, and others who have written about the psychosocial aspects of Japanese script exaggerate the connection between kanji and culture. There are some psychological factors (notably use/reference ambiguity) that make the Ideographic Myth attractive, but they are not linked specifically to Japanese culture. When we hunt for the alleged connections between the use of kanji and the Japanese mind, there turn out to be few if indeed any at all.
We discover, for example, that although the Japanese writing system has undergone a gradual process of evolution over the centuries, it has been relatively stable compared with changes in Japanese thought and life. Japanese of the sixteenth century, for example, behaved in ways that today’s proponents of “Japanese uniqueness” would have us believe were never part of Japanese culture.
They were ardent party-goers, sophisticated hosts, eloquent discussants, and even internationally open-minded. Hideyoshi, the famous military ruler of the sixteenth century, hosted an enormous open-air tea party in 1587, to which he invited all kinds of people from the emperor down to commoners, and even foreigners. A Jesuit missionary, Francis Xavier, testified that the Japanese of his time were far from being shy and that they visited him so often to ask questions that he could hardly get any sleep. Another missionary, Organtino, reported that the two vices Japanese hated most were an impatient attitude toward other people and irrationality in discussion. (Yamazaki 1981: 65)
If the use of kanji is even remotely responsible for the alleged inability of modern Japanese to cope with plain-talking foreigners and straightforward logic, it is hard to understand how these “uniquely Japanese” deficiencies could have escaped the attention of sixteenth-century observers. Historical continuity does not preclude change, and for this reason alone, we ought to be suspicious of any claims of a direct, psychological connection between script and culture.13
We also discover that there is only a small grain of truth in the well-known tale that Chinese characters have made it possible for Japanese, Chinese, and Koreans to communicate through writing despite their mutually unintelligible languages, a state of affairs that allegedly could not have arisen had Chinese script been alphabetical. The historical and linguistic evidence flatly contradict these claims (DeFrancis 1984: 149–160). Although a common inventory of characters is a necessary condition for interlinguistic written communication, whether in East Asia (classical Chinese), the Islamic world (classical Arabic), or some other multilingual “culture zone,” it is by no means a sufficient condition. The Japanese did not merely receive kanji from China: they imported the entire apparatus of Chinese literature, historiography, science, philosophy, religion, government, and art. Classical Chinese was able to serve as a channel of communication for the Japanese elite, in the final analysis, not because of their mastery of the writing system per se, but because of their assimilation of the ideas Chinese letters brought to them. Indeed, once planted on Japanese soil, the seeds of these ideas often grew into living traditions quite different from their continental counterparts.
Lexicography is a pertinent example. The compilation of dictionaries has a history of more than a thousand years in Japan and although inspired by Chinese models, shows a remarkably original development. Of course, the Japanese couldn’t help being innovative in the treatment of their own language, but even in the case of classical Chinese, they were not mere copyists. Indeed, there is today no Chinese dictionary that can rival the thirteen-volume character dictionary of Morohashi Tetsuji (1883–1982), the pinnacle of Japanese sinology. Where the Japanese did slavishly follow the Chinese was in their uncritical acceptance of the idea that the Chinese language consists entirely of monosyllabic words. Modern linguistics shows that this idea is categorically false (Kennedy 1964: 274–322)14—yet another reason for rejecting the Ideographic Myth—but to this day, despite the internationalization of Japanese scholarship, it continues to color Japanese thinking about the nature of language and meaning.
The same pattern can be seen more generally in the Sino-Japanese tradition of teaching and study, which eschews reasoned disputation (Mizutani 1979, 1981) and ranks historical precedent ahead of empirical observation as a source of authority (Nakayama 1974, 1984). It sanctioned the withholding of information as a tool of power; thus, for example, it made perfect sense for an advocate of education for the samurai elite, Ogyū Sorai (1666–1728), to write:
[i]t is not necessary that the common people should be taught anything apart from the virtues of filial piety, brotherly submission, loyalty, and trustworthiness. Their reading should not extend beyond the Classic of Filial Piety (Hsiao Ching), the Lives of Famous Women (Lieh Nü Chuan), and other improving biographical collections which deal with the relations between sovereign and subject, father and child, man and wife. The study of other works will merely increase their cunning and will lead to disruption. (McEwan 1962: 132; also in Passin 1982: 190–191)
Today, the tradition lingers on in the way Japanese newspapers and television dote on the views of academic and literary celebrities, regardless of their specialties, while giving scant coverage to better-informed experts who lack the traditional scholarly credentials. And we see it in the obsessive cramming and rote memorization of purely factual information that takes up so much of Japanese primary and secondary education—an obsession, incidentally, also characteristic of the current American infatuation with data-processing (Roszak 1986: 156–160).
If these and other traditions were ever, in any sense, caused by the Japanese adoption of Chinese characters, they are now so thoroughly entrenched in Japanese life that a radical change in the writing system would probably not affect them in the slightest. More than a hundred years ago, this was already clear to the statesman Mori Arinori. In the early 1870s, he concluded that mere romanization of the script would not be enough to cut Japan loose from the entangling web of Chinese words and ideas that he believed had to be abandoned if Japan were to become a competitive, modern trading nation (Sakakura 1985: 66–70). Mori suggested adopting a simplified form of English as a new national language! Although he wisely dropped this idea as unworkable, his reasoning was, and still is, impeccable: to believe that the grip of Chinese tradition on modern Japanese life would suddenly loosen if the script were altered is to trivialize centuries of Japanese history.
Ironically, although the continued use of kanji serves to keep Japan linguistically isolated by holding down the number of foreigners who can read Japanese, this artificial “language barrier” has done nothing to prevent the steady erosion of both imported and genuinely native traditions. In his 1981 book How Will the Japanese Language Change?, Kabashima Tadao predicted that the influx of loanwords, particularly from English, into modern Japanese would eventually lead to the demise of kanji. Whether kanji are a boon or a nuisance makes no difference: the fact of the matter is that Japanese are adapting English to their word-making needs today with the same alacrity with which they adapted Chinese more than a thousand years ago; productive use of Sino-Japanese morphology is declining as traditional Chinese scholarship becomes less and less relevant to success at work and in society. Dr. Uenohara’s concern that technology might “abolish” the Japanese language (i.e., traditional Japanese script) is therefore misplaced. If Japan is under some kind of sociolinguistic threat, the aggressor is not technology but English and the other languages that serve as international modes of communication in the computer age.
- For example, DeFrancis (1984: 278, 296) has caught Miller putting words in the mouth of none other than Deng Xiaoping.
- Although Saint-Jacques and Suzuki (1984) are even more severe, most reviewers, for example Allen (1983), Impoco (1984), and DeWolf (1985) concur with Chew. Those unfamiliar with Miller’s other works may find it comforting to learn that he regularly savages both Westerners (e.g., Miller 1975, 1985) and Japanese (e.g., Miller 1977, 1981) with equal alacrity.
- Tsunoda claims to have proven experimentally, through the “Tsunoda Key Tapping” technique, that there is an innate switching mechanism in the brain that governs the processing of sounds, and that the language one speaks between the ages of six and eight determines how this switching mechanism operates (Tsunoda 1985a: 157, 1985b: 138). According to Tsunoda, Japanese and Polynesian languages such as Maori and Tongan lead to one type of switching, and all other languages lead to another. No one has been able to replicate Tsunoda’s results except Tsunoda, and even he has been having trouble lately; this he attributes to “the earth’s revolution, the lunar motion, and possibly other cosmic activities” (1985b: 139). One wonders what Martin Gardner (see Gardner 1957), whose mathematics column graced Scientific American for many years, would say about the recent exposition of Tsunoda’s “discoveries” in the Japanese edition of Scientific American (Tsunoda 1985c).
- A further point to consider is that people buy books for various reasons. After 1945, there were people who had received little or no education because of the war but nonetheless subscribed to newspapers lest their neighbors discover that they could not read.
- But many Japanese believe just that, as shown by a recent international incident. On 22 September 1986, Prime Minister Nakasone was quoted as saying that Japan has an advantage over the U.S. because it does not have ethnic minorities like blacks, Puerto Ricans, and Mexicans. The major Japanese papers ignored this remark (but did report other parts of Nakasone’s speech that day) until it elicited a scathing American reaction, which caught the government by surprise. Although they have since carried some criticisms of the substance of Nakasone’s statement (e.g. Kometani 1986), more space has been devoted to clarifying it. Nakasone was speaking at a Liberal Democratic Party study group, attempting to explain why the Japanese educational system was more successful than the American.
- “Examination hell” is not a postwar phenomenon. Nitobe used the term in 1931 (Nitobe 1972: 3.249), and it is doubtful that he coined it.
- Perhaps pricked by this realization, Suzuki takes a somewhat different tack in more recent writings, such as Suzuki 1977, which contains the startling claim that the Japanese writing system is a synchronic part of the Japanese language, on a par with syntax and phonology. This claim is no more compatible with the findings of linguistic science than creationism is with modern biology. The evidence that Suzuki adduces, mostly historical, simply does not require the farfetched interpretation upon which he insists.
- History as well as common sense shows that Japanese can indeed read and write their language in rōmaji. In the early part of this century, Japanese immigrants in Mexico used rōmaji in their schools and found that first-year students could read the equivalent of the first three years’ textbooks then in use in Japan (Ishida 1973: 126–134). A former member of the Kokugo Shingikai (Japanese Language Council) and a professor at the University of Tōkyō, both interviewed in 1985, have confirmed the existence of a report of similar results obtained during experimental rōmaji education in the early 1950s (referred to contemptuously by Trainor [1983: 318–319]). The Ministry of Education seems to have lost all copies of it.
- Some would not even go this far. Steinberg & Yamada 1979a elicited a blistering rejoinder from Tzeng and Singer (1979), who protested, among other things, what they felt was an attempt to slip the Ideographic Myth in through the back door. Steinberg and Yamada vociferously denied this in their surrejoinder (1979b), but in a Japanese reference to Steinberg & Yamada 1979a,Yamada (1983: 129–130) casually tacks on a statement of the Ideographic Myth.
- Interestingly, Zheng Qiao himself appears to have appreciated the power of alphabetic writing. “Now, the writings of Confucius travelled as far east as Korea, as far west as Kansu, as far south as Annam, and as far north as Inner Mongolia. All of these were our ancient feudatories. His writings did not reach beyond these ancient feudatories. How is it that Gautama Buddha’s books can penetrate the various parts of China but Confucian writings cannot reach the Airavati River? It is because there are obstacles to the passage of tones and sounds. This is the fault of scholars who came after Confucius! Wherever boats and carts can arrive, there too words and their meanings can extend. Why are there places where boats and carts arrive today but Confucius’ words and their meanings do not extend?” (Mair 1986: 134).
- Holtom (1947: 221) says he heard the same opinion (that the blind were more fortunate than the sighted in Japan) from a famous Japanese educator; perhaps he attended one of Nitobe’s lectures. In any case, since he was writing against script reform, it is noteworthy that he repeated this view without critical comment.
- Yale University professor Samuel Martin recently suggested that Japan’s international reputation for being a “secret club” was largely due to the Japanese language, specifically Japanese writing, being a “secret code” (Martin 1985: 14).
- For example, McLuhan (1962) says that China and other manuscript cultures are “audile-tactile” whereas print culture is intensely visual. Why, then, were Chinese literati so impressed by Matteo Ricci’s mastery of the art of visualizing “memory palaces” (Spence 1984)? For that matter, why did the visualization of “memory palaces” flourish in the manuscript culture of medieval Europe and fall into disuse after printed books became common? Evidently, not all manuscript cultures share the same “sense ratios.” Nakayama (1984: 78–79) criticizes McLuhan in the same vein. Incidentally, Nakayama and McLuhan (probably under the influence of Pound and Fenollosa) both accept versions of the Ideographic Myth; therefore, their differing views on Chinese culture cannot be due to differing views on the function of Chinese script.
- Chinese does have some monosyllabic words, and many disyllabic Chinese words are true compounds (like English “baseball”); however, there are plenty of Chinese disyllables (like “orange” in English) that cannot be analyzed into independent words.DeFrancis (1984: 185) quotes the eminent linguist Zhou Youguang (personal communication) as estimating that “only 2,000 or so, or about 30 percent, of the 6,800 ‘modern standard characters’ needed to write contemporary Chinese are free words.”
Full Hepburn romanization is used for Japanese unless a different spelling is given in the source. Japanese names are kept in the traditional order, surname first. All places of publication of works in Japanese are Tōkyō.
Black, Cyril E., Marius B. Jansen, Herbert S. Levine, Marion J. Levy, Jr., Henry Rosovsky, Gilbert Rozman, Henry D. Smith, II, and S. Frederick Starr. 1975. The Modernization of Japan and Russia: A Comparative Study. New York: The Free Press; London: Collier Macmillan.
DeFrancis, John. 1947. “Politics and Phonetics.” Far Eastern Survey 16:19 (November 5) 217–220.
Doi Miwako, and Yoneda Kiyoshi. 1982. “Shukanteki oyobi kyakkanteki yominikusa no hytka [Evaluation of Subjective and Objective Reading Difficulty]”. In Dai 8-kai shisutemu shinpojiumu [8th System Symposium], 19-21 August, Hokkaidō, 281-286.
——— . 1978. The Japanese Language. Translated by Hirano Umeyo. Rutland, Vermont and Tōkyō: Charles E. Tuttle & Co.
Martin, Samuel E. 1985. “Nihongo no shōrai [The Future of the Japanese Language].” In Soto kara mita Nihongo, uchi kara mita Nihongo [The Japanese Language Observed by Japanese and Non-Japanese], edited by the Kokugo Gakkai. Musashino shoin.
——— . 1977. The Japanese Language in Contemporary Japan: Some Sociolinguistic Observations. AEI-Hoover Policy Study 22, Hoover Institute Study 58. Washington, DC: American Enterprise Institute for Public Policy Research.
Mizutani Osamu. 1981. Japanese: The Spoken Language in Japanese Life. Translated by Janet Ashby. Tōkyō: The Japan Times.
Neustupný, J. V. 1984. “Literacy and Minorities: Divergent Perceptions.” In Linguistic Minorities and Literacy: Language Policy Issues in Developing Countries, ed. Florian Coulmas, 115–129. Berlin, New York, Amsterdam: Mouton.
——— . 1970. The Origin of the Japanese Language. Tōkyō: Kokusai Bunka Shinkokai.
Sakakura Atsuyoshi. 1985. “Nihonjin no Nihongo [Japanese for Japanese].” In Soto kara mita Nihongo, 47–80. See Martin 1985.
Sasaki Masato. 1983. “Mōjin ni totte kanji to wa [What Are Kanji for the Blind?].” In Kanji o kagaku suru, 225–263. See Kaiho 1983.
——— . 1975b. “On the Twofold Phonetic Realization of Basic Concepts: In Defence of Chinese Characters in Japanese.” In Language in Japanese Society, edited by Fred C. C. Peng, 175–192. Tōkyō: University of Tōkyō Press.
Tzeng, Ovid J. L., and Harry Singer. 1979. “Failure of Steinberg and Yamada to Demonstrate Superiority of Kanji over Kana for Initial Reading Instruction in Japan.” Reading Research Quarterly 14:4.661–667.
Yamada Jun. 1983. “Kodomo ni totte kanji to wa [About Kanji for Children].” In Kanji o kagaku suru, 122–152. See Kaiho 1983.