Chinese Impacts on the Japanese Language
-
Chinese impacts on the Japanese language
Shibatani Masayoshi
Due to the historical accident that the Japanese decided to absorb Chinese culture and adopt Chinese characters as a means to represent their language, colloquial speech and the written language took different paths of development, though some genres of writing were much closer to the colloquial speech of the time than others. The divergence between the colloquial form and the written form originates in the development of the written style called Sinico-Japanese (or kanbun) tradition. Sinico-Japanese developed as a response to the desire to write Japanese in Chinese characters and in Chinese syntax! Such a desire was strongly felt during the Heian period (800‒1186), when the influence of Chinese culture was at its peak. Much as Latin was cherished in the medieval world of scholarship in the West, Chinese letters were both a symbol of learning and a means of recording official matters.
However, as Japanese and Chinese happened to be radically different in syntax, writing Japanese in Chinese and reading the resulting texts in the Japanese style involved a number of complications. Writing was perhaps more like translating Japanese sentences into Chinese, but reading the resulting text as a Japanese text required not only inversion in word order but also the supplying of inflectional endings and particles that were foreign to the Chinese language. It is from this cumbersome way of writing Japanese that the tradition of Sinico-Japanese has developed. Due to the imperfect learning of Chinese by the early practitioners and subsequent internal developments, Sinico-Japanese is neither pure Chinese nor Japanese, and thus neither Chinese nor Japanese are able to read Sinico-Japanese texts without special training.
Another literary tradition developed with the invention of the kana syllabary around the tenth century, and the language of this tradition, as represented by the language found in theGenji monogatari (The Tale of Genji), is believed to have been much closer to the colloquial speech of the time. However, official writing was carried out within the Sinico-Japanese tradition, and a few distinct writing styles that developed within this tradition were maintained in the written language until the late nineteenth century, when a movement for matching speech and writing (gen-bun itti) was started by literary men.
In Japan, Modern Japanese is often contrasted with Late Old Japanese and Middle Japanese in terms of gendai-go ‘modern language’ (or koo-go ‘colloquial language’) versus bun-go ‘literary language’. The English term “Classical Japanese” largely corresponds to bungo, though different scholars use the term more specifically to refer to the language of any period between the eighth and the fourteenth or the fifteenth century.
1.1 Overview
Instead of describing the characteristics of the language of each period, the following exposition traces the entire course of development of a number of salient features of Japanese. Following this, a couple of topics of general appeal are taken up in detail.
The major achievement of the Japanese of the Nara and the early Heian period is the acquisition of a writing system. The writing system evolved from both semantically and phonetically based uses of Chinese characters. The present-day writing system is the result of simplification of the original Chinese characters in two ways.
Alongside the use of the original Chinese characters, the Japanese developed two syllabaries. The system know as katakana developed from the abbreviation of Chinese characters, while the other version of kana called hiragana resulted from the grass (cursive) style of writing characters. Much more recently Chinese characters themselves were also simplified leading to a system that diverges from the original characters. Because of the recent abbreviation of the characters in mainland China, there are now three different sets of Chinese characters: the original complex forms used in Taiwan and the Republic of Korea (in formal writing), the medially complex complex forms used in Japan and the Republic of Korea (informal writing), and the most radically abbreviated forms used in the People’s Republic of China. The evolution of the writing system will be taken up separately below.
Contact with a superior culture means the borrowing of, among other things, lexical items. Japanese is no exception to this truism, and it borrowed a large number of Chinese words in early phases of its history. The systematic borrowing of Chinese words took place in three waves, though sporadic pre-historic borrowing leaves such deeply ingrained words asuma ‘horse’ and ume ‘plum’. The beginning of the first wave of borrowing occurred in the pre-Nara period when Buddhism was introduced to Japan in A.D. 552 (or 538). The sounds or readings of the Chinese characters attributed to this first wave of borrowing are known asgo’on (‘go sound/pronunciation’). It is generally believed that go refers to the Wu area of southern China, and thus the go’on reading is thought to reflect a southern Chinese dialect of the Six Dynasties period (ca. beginning of the fourth century to the mid or late sixth century).
The second wave of Chinese borrowing took place during the Nara period, when many students and court officials went to study in Luoyang and Chang’an, the two great metropolises of the Tang dynasty. The new pronunciation associated with this second wave of borrowing during the early phase of the eight century is called kan’on and is considered to reflect the standard language of the Tang period. Then, during the fourteenth century, a new vocabulary and pronunciation were introduced chiefly by the followers of the Zen sect of Buddhism. This new pronunciation, believed to represent that of the Hangzhou area of the period, is called tÅ�-sÅ�’on (‘Tang-Song pronunciation’).
While some characters have three ways of reading, reflecting the three waves of borrowing, most have only the go’on and/or kan’on pronunciations, indicating the smaller influence of thetÅ�-sÅ�’on pronunciation. Also, different pronunciations are associated with different spheres of learning: go’on with Buddhism, kan’on with Confucian and other secular learning, and tÅ�-sÅ�’on with Zen Buddhism. However, all this is a historical fact, and average native speakers of Japanese do not know about these historical distinctions in pronunciation; they simply learn that a given character has two or more different readings (see below).
The first systematic contact with the Western world was brought about in 1543, when a drifting Portuguese merchant ship reached an island off KyÅ«shÅ«. The Portuguese introduced to Japan not only guns and other Western objects but also Western words as well as Christianity. The Spaniards and Dutch followed suit. The second wave of Western borrowing had to wait until the beginning of the Meiji period, for during the Edo period the country was closed to the outside world except for limited contact through the Dutch in KyÅ«shÅ«, who represented Western interests. Since the Meiji period English, German, and French have usurped the position of Dutch as a language of foreign learning, and after World War II, the third onslaught of foreign loans came primarily from America. Thus since the Meiji period English loans have claimed the lion’s share in the vocabulary of Western origin. A thorough description of the role of foreign loan words is given separately.
As the vocabulary has been constantly enriched by borrowing throughout history, the phonological system has also been affected with the importation of new syllable types. New syllable types also evolved as as result of internal changes, but some of these internally created new syllable types are attributed to influence from the foreign syllables. The canonical syllable structure of Old Japanese was (C)V, all syllables and words ending in a vowel, and vowel sequences being avoided. However, Chinese loans brought in new syllabic types, those ending in consonants and sequences of vowels. Among the new additions were the syllable-final nasal and other syllable-final consonants, as represented by hon ‘book’, and kekka ‘result’ (and more recent forms such as beddo ‘bed’). These subsequently gave rise to independent moraic units. Also forms such as keidai ‘temple compound’ introduced vowel sequences. It is generally believed that these imported syllables were instrumental in the internal development of syllable types after the Heian period, e.g. yonde < yomite ‘read CONJ’, totte < torite ‘take CONJ’, kaite <kakite ‘write CONJ’, siroi < siroki ‘white ATTR’. Besides this kind of enrichment of the phonological system, there are also aspects of simplification of the system. Especially noteworthy is the reduction of the number of syllables distinguished. Old Japanese of the Nara period is phonologically most interesting in that it distinguished the greatest known number of native syllables. The earliest written document of significance, the Kojiki (Record of Ancient Matters) (A.D. 713), distinguished eighty-eight syllables, whereas the Nihon shoki (Chronicle of Japan) (A.D. 720) distinguished eighty-seven syllables obliterating the distinction between the syllables mo1 and mo2. (See Table for syllable types, and the o1 and o2 and similar distinctions.) The significantly larger number of syllables of Old Japanese in contrast to that of Modern Japanese, which distinguishes sixty-two native syllables, is due to the distinctions made between the so-called koo (A series) and otu (B series) vowels observed in Old Japanese. A distinction, for example, was made between ki1 and ki2, which merged into ki in the ensuing Heian period. It is due to these A and B distinctions that the hypothesis has been advanced that Old Japanese may have had more vowels than the present five. For example, the syllableki in certain words was written with the character 伎 and in some other words with the character ç´€. Thus the first syllable in kimi ‘emperor’ was written with the first character as 伎美 (ki-mi), but the second syllable in the word tuki ‘moon’ was written with the second character as 都紀 (tu-ki). Distinctions such as these were fairly consistent and the characters were not freely interchanged.
The early Heian period saw the obliteration of the A and B series distinctions noted above, and further merger of e ([ịe]?) and je, and o ([ṷo]?) and wo resulted in the radical reduction of the numbers of syllables, of which there were now sixty-six. This was followed by the merger ofi and wi and e and we during the Kamakura period. Entering the Edo period, zi merged withdi and zu merged with du, resulting in the present-day number of sixty-two native syllables.
The accentual system is also believed to have developed in such as way that newer systems are the results of mergers of earlier distinctions. While this appears to be the case with many modern dialect systems, there is considerable debate as to whether the Ky�to‒Ōsaka system, which makes more distinctions, is older or newer than the T�ky� system. This issue and other notable phonological changes are taken up later, where the geographical distributions of their remnants are discussed.
In the realm of syntax there are both stable and fluctuating aspects. Among the most consistent features is word order, which has been SOV from the time of Old Japanese, though observed throughout history is the inverted word order, which places a subject or object nominal after the verb, for the expression of afterthought, emphatic assertion, etc. Other phenomena related to the order of the major constituents are also systematically exhibited in accord with the basic word order, making Japanese one of the ideal SOV languages. Thus, modifiers, e.g. adjectives and adverbs, precede their heads, e.g. nouns and verbs. Auxiliaries in the form of suffixes consistently follow the main verb, and adpositional elements in the form of particles are consistently postpositional. Despite this overall consistency that run throughout history, the contents of these categories have undergone considerable degrees of change.
In the case of auxiliary elements, the system has moved from a richer to a more impoverished one. Especially noteworthy is the decline in the systems of tense, aspect, and mood. In Modern Japanese there is only one past tense suffix ta, which is also used in an expression of the perfective sense, but in Old Japanese there was a clear formal distinction between the perfective aspect and the past tense. Tu, nu, tari, and ri were perfective suffixes, the first two having an added meaning of confirmation. The past tense was expressed by ki andkeri. Nu, tu, ri, ki and keri disappear during the Kamakura and Muromachi periods, whiletari begins to develop into the present-day past-tense form ta at the end of the Heian period;tari itself survives until the end of the Edo period.
In the area of mood, again there were many forms expressing conjecture, but they too died out before the end of the Muromachi period. Ramu and rasi contrasted in terms of firmness of evidence, the former being dubitative and the latter implying firmer conjecture. Meri was for a conjecture based on visual evidence, while nari was for one based on auditory evidence. The present-day form rasii ‘(it) seems’ has no directly connected historical antecedent. Although there was the form rasi, which shows semblance to rasii, in the Nara and Heian period, it dies out at the end of the latter period, and rasii only emerges in the Edo period.
Compared to these, voice suffixes are fairly stable. The spontaneous/potential/passive/honorific raru and the causative/honorific sasu, which emerged in the Heian period, survive and are vigorous today in the forms rare and sase, though the latter has shed the honorific meaning (except in the subject honorific expression asoba-su (play-CAUS) ‘do’). The desiderative tai is a new development from tasi, which took place toward the end of the Kamakura period.
Particularly interesting is the rise and fall of the honorific system, since this reflects the social organization of each period rather clearly. While honorific forms are seen in the earliest history of the language, the elaboration of the system started in the Heian period, when the court-centered society came to maturity. New developments during this period are 1) two levels of honorifics, 2) combination of subject and object honorific forms, which is not possible in Modern Japanese, and 3) polite expressions directed toward the addressee. Among the forms that had two levels of honorifics are: ‘to be’ owasu (ordinary honorific): owasimasu (super-honorific), ‘to give’ tamau (ordinary): tamawasu (super).
The elaboration of the honorific system continues throughout the Kamakura and Muromachi periods, when political power was transferred to the warrior class, which is characterized by a strict hierarchical rank system within each clan as well as across different clans. Howerver, the zenith of the continued elaboration is witnessed by the Edo period, when the Tokugawa government issued a decree for the formation of social classes, which consisted in the decreasing order of 1) warrior-bureaucrats, 2) the peasants, 3) the artisans, and 4) the merchants. While there seemed to be a clear class distinction between the first category and the rest, there does not seem to be any significant social difference among the people of the latter three classes. Thus for practical purposes there were two classes, the warrior-bureaucrats (busi and kuge) and the townsmen (tyoonin). As can be surmised, it is among the townsmen that newer honorific forms developed, as the warrior-bureaucrats continued to use the earlier honorific forms. In this period, certain expressions developed three degrees of honorification. For example, for the verb ‘to be, to come, to go’, gozaimasu is the super honorific form,gozaru the ordinary honorific form, and mesaru the intimate honorific form. Also developed were those forms that reflect the social class, the sex, and the profession of the speaker. Ozyaru ‘to come’ and tukamaturu ‘(I) do’ are forms used by warriors, syansu ‘to do’ and the polite endingsgozansu and yansu are exclusively for female speakers, and the polite endings gozarinsu andyasu were those belonging to the language of the amusement quarters.
The elaborate honorific system comes to an abrupt end when the Meiji Restoration (1867) abolished the Tokugawa social class system, thus putting an end to feudalism. Though a new class system consisting of noblemen and the common people was instituted, the tide of Western democracy and compulsory education had the effect of simplifying the honorific system considerably. The present-day honorific expressions are generally descendants of the newer Edo expressions. The subject honorific form o/go … ni naru and its precursor o/go …nasaru are seen in the dialect of Edo (presently TÅ�kyÅ�), but its wide use started only in the middle of the Meiji period, when the polite ending desu also began to be widely used. The spread of the object honorific form o … suru was even later, i.e. in the late Meiji period.
In the case of particles, a general trend has been that the adverbial particles have become less active, while case particles have become more rigidly integrated. This change in the system of particles has been interpreted as a reflection of an underlying structural change in the Japanese language by Sakakura (1977).
The use of certain adverbial particles had the effect of cutting off the flow of the predication process, which allowed the insertion of various adverbial expressions and clauses within a clause. The insertion of complex adverbial clauses was made largely because certain adverbial particles participated in a phenomenon called kakari-musubi, whereby the occurrence of a particular adverbial particle correlated with a particular inflectional form of a predicative element; thus even if the subject nominal marked off by an adverbial particle was separated from its predicate by an intervening adverbial clause, the subject and the verb could be paired up because of the rules of kakari-musubi. Sakakura calls this characteristic of Late Old, or Classical Japanese of the Heian period “open structure”. The disintegration of the kakari-musubi phenomenon and the concomitant weakening of many of the adverbial particles during the Kamakura period ushered in the developement of the nominative case particles, which had the effect of tying the subject and the predicate in a closer relationship because the nominative particles developed from the attributive particles that related two nominal elements under various semantic relationships. This shift, according to Sakakura, has turned Japanese from an open-structure language to a closed-structure language, which does not allow insertion of complex adverbial clauses within a clause, which was formerly permitted.
Sakakura believes that this is a major overall structural change that has taken place in the history of Japanese. This change, which took place over the span of the Kamakura and Muromachi periods, coincides with a time of great socio-political upheaval, when the seat of government was transferred back and forth between Ky�to and Kamakura, crossing a major bundle of dialect isoglosses. One can agree with Sakakura that major changes in the Japanese language occurred during this period without necessarily subscribing to his more speculative assessment of the cause for the changes, which he attributes to the theory that the socio-political upheaval of the period brought together people of different social and regional backgrounds necessitating a more logical discourse punctuated more clearly by the use of case particles and characterized by simple, straightforward sentences.
Having presented an overall picture of the changes that have taken place in the history of Japanese, the following sections examine in detail too topics of general interest. The developments of principal particles are discussed larer.
1.2 The writing system
The earliest extant systematic written records of the Japanese language date back to the eighth century. The oldest among them, the Kojiki (Record of Ancient Matters) (A.D. 712), is written in Chinese characters. The preface to this work is written in Chinese syntax as well. Thus, the document is not readily intelligible to those who do not know how the Chinese ordering of elements corresponds to the Japanese ordering; i.e. while Chinese is basically an SVO language like English, Japanese is an SOV language. Furthermore it is not clear how these characters were read; they may have been read purely in Chinese style in imitation of the Chinese pronunciation of the characters, or they may have been read in a Japanese way, i.e. by uttering those Japanese words corresponding in meaning to the written Chinese characters and inverting the order of elements so as to follow the Japanese syntax. Perhaps both systems were practiced.
By the time the Man’yÅ�shÅ« (Collection of a Myriad Leaves), an anthology of Japanese verse, was completed (A.D. 759), the Japanese had mastered the use of Chinese characters as a phonetic means of writing Japanese. For example, the Japanese word yama ‘mountain’ could be written phonetically by using a combination of characters with a sound similar to ya (e.g. 夜 ‘evening’) and another character with a sound like ma (e.g. 麻 ‘hemp’) as 夜麻. In other words, what stands for ‘mountain’ could now be written in two ways. One used the Chinese word and its character å±±. The other way was to choose Chinese characters read as ya and ma. This practice helped establish a relation between the original character and its Japanese semantic equivalent leading to the practice of assigning a dual reading to Chinese characters: one a Chinese reading (on-yomi) and the other a Japanese reading (kun-yomi). This means that a character such as å±± ‘mountain’ was read both as san, the Chinese reading, and as yama, the Japanese reading. This practice of reading Chinese characters both in the Chinese way and the Japanese way persists even today (see below).
The phonetically used characters are known as man’yoo-gana among Japanese grammarians because their most diversified use is seen in the Man’yÅ�shÅ«, though their earlier use is seen in other writings, e.g. the songs in the Kojiki. This kind of phonetic use of Chinese characters as purely phonetic representation of words is most likely not the invention of the Japanese at the time. It was practiced in China in rendering foreign names, place names, official titles, and especially in transliterating Sanskrit Buddhist terms. The same was done by the Koreans who used Chinese characters phonetically in indicating grammatical particles and inflectional endings as an aid in reading documents written in the Chinese manner. Since the early tutors of the Chinese language in Japan were Koreans, it is reasonable to assume that their practice at home influenced the rendering of Japanese in terms of Chinese characters.
Next came the development of syllabary writings known as kana, originally pronounced askarina. Two kinds of kana have developed as a simplification of man’yoo-gana. Originally,kana were used as mnemonic symbols for reading characters and were written alongside them.Katakana ‘partial kana’ developed as abbreviations of characters, and they retain a squarish shape. Hiragana ‘plain kana’, on the other hand, have been developed by simplifying the grass (i.e. cursive) style of writing characters, and they are round in shape. Examples of these two types of kana and their sources are shown below, and the complete contemporarykana syllabaries are set out in Table 1.1.
(Voicing oppositions, where applicable, are indicated by diacritical dots on the upper right-hand corner of each kana, e.g. gi ã�Ž, ã‚® as opposed to ki ã��, ã‚. ã‚’ and ヲ, used only in writing the accusative particle o are pronounced the same ways as ã�Š and オ, which are used in all other instances of the o sound. ã�¯ and ãƒ� are pronounced as wa when they are used to write the topic particle wa. ã�¸ and ヘ are pronounced as e when used to write the directional particle e. Among the missing syllables in the table, the kana for wi and we exist but they are no longer in common use, as the sounds have merged with i and e.)
The orthodox Sinico-Japanese, or kanbun, manner of writing was a formal writing method. The phonetic kana wrting was considered only “temporay” or mnemonic in nature. Despite this secondary status of kana, they were eagerly used by women with literary aspirations, who were discouraged from learning Chinese characters, which belonged to men’s domain of learning and official writings. Indeed the
-
>>>
The orthodox Sinico-Japanese, or kanbun, manner of writing was a formal writing method. The phonetic kana wrting was considered only “temporay” or mnemonic in nature. Despite this secondary status of kana, they were eagerly used by women with literary aspirations, who were discouraged from learning Chinese characters, which belonged to men’s domain of learning and official writings. Indeed the Genji monogatari (The Tale of Genji), considered by many to be Japan’s greatest literary achievement, was written almost entirely in hiragana by Lady Murasaki at the beginning of the eleventh century (ca. 1001‒10).
Despite this early separation between Chinese characters and kana, the contemporary practice is to use Chinese characters, called kanji, for content words, and hiragana for grammatical function words such as particles and inflectional endings. Katakana are now used to write foreign loan words, telegrams, and certain onomatopoeic expressions.
In addition to these writing systems that originate in Chinese characters, there is rÅ�maji, which is another phonetic writing system using the Roman alphabet. The first systematic transcription of Japanese in the Roman alphabet dates back to the late sixteenth to early seventeenth centuries, when Portuguese and Spanish missionaries transliterated Western stories, such as Aesop’s fables, and Japanese tales. Also, grammatical treatises by these missionaries contain Romanized Japanese expressions. While the system of transcription was largely based on the Portuguese spelling pronunciation, the materials are of great philological importance, since they indicate the pronunciation that was used during the early seventeenth century but which has been lost since then. For example, in perhaps the most impressive grammatical treatise on Japanese in a Western language of the time, Arte da lingoa de Iapam by Joãa Rodriguez (1561‒1633), which was completed between 1604 and 1617, the s-beginning syllables are transcribed as sa, xi, su, xe, so, indicating that the se syllable, which is now pronounced as [se] in the standard dialect, was pronounced during the period with a palatalized sibilant as [ʃe], just like the si syllable, which is still pronounced as [ʃi]. (The [ʃe] pronounciation occurs dialectally even today.) Also di and zi, which have now merged and are pronounced as [dÊ’i], are distinguished as gi and ji; and du and zu, as zzu/dzu and zu, indicating that all these syllables were distinguished at the time when transcriptions were made. (Again, there are dialects that distinguish these syllables to varying degrees.)
Other attempts to devise r�maji systems based on Spanish and Dutch were made, but the arrival of a system with wide acceptance and endurance had to wait until the middle of the Meiji period, when the American missionary James Curtin Hepburn (1815‒1911) developed a system for his Wa-ei gorin shūsei (Japanese‒English Glossary), first published in 1867. The present Hepburn system (known in Japan as Hebon-shiki) is largely based on the system adopted in the third edition of the above dictionary published in 1886. There have been several additional systems proposed since then, but in 1937 the Japanese government attempted to unify all the different systems and promulgated what is referred to as the kunreisiki (cabinet ordinance system). The essential difference between the two is that while the Hepburn system is based on the system of English spelling pronunciation, the kunreisiki is phonemic in nature, largely based on the arrangement of the hiragana chart (see Table 1.1). The Hepburn system, which is arguably easier to read for those familiar with English spelling pronunciation practice, has retained its popularity, and in 1954 the government promulgated a revised version of thekunreisiki system, which consists of two charts, Chart 1 listing all the syllables in thekunreisiki and Chart 2 listing those syllables that are written differently in the Hepburn system. The preface to the revised version stipulates that in transcribing the Japanese language Chart 1 be used, but those in Chart 2 may be used if the change cannot be made at once due to customs and considerations of international relations. Thus, the Japanese are stuck with two overlapping official systems of Romanization, which are used mainly in signing documents written in Western languages, foreign acronyms, e.g. ILO, IMF, and names of stores and firms on signboards and in advertisements.
The essential difference between the two systems lies in the transcription of some twenty syllables. The syllables [tsɯ], [ʃi], [tʃi], [φu], [dÊ’i], etc. are written as tsu, shi, chi, fu, and ji in the Hepburn system and as tu, ti, hu, zi/di in the kunreisiki system. Thus, for example, the family name pronounced [tsɯtʃihaʃi] is written as Tsuchihashi in the Hepburn system and as Tutihasi in the kunreisiki system. Despite the government’s decree for the use of the kunreisiki system, the Hepburn system is still widely used and perhaps more popular than the kunreisiki system. One area in which the Hepburn system is consistently used is in the writing of Japanese railroad station names as an aid for foreigners. The transliteration of Japanese expressions in linguistics treatises is often done in the kunreisiki system. However the practice here is not consistent even among specialists of the Japanese language. While Roy Andrew Miller’s The Japanese Language (1967) uses the Hepburn system, Samuel E. Martin’s A Reference Grammar of Japanese (1975) uses the kunreisiki system. The style sheet of the journal Papers in Japanese Linguistics stipulates that the transliteration be done in the kunreisiki system. The method adopted in this book and some others basically follows the government decree using thekunreisiki system for cited forms and the Hepburn system for proper nouns, such as people’s names and names of books.
Thus the Japanese essentially have four ways of writing their language. The word for ‘mountain’ can be written as å±± in kanji, as ヤマ in katakana, as ã‚„ã�¾ in hiragana, and yama inrÅ�maji. Though the domains of their use are fairly distinct, as indicated in the above exposition, sometimes all these four ways of writing can be found in one sentence. For example, the sentence ‘Hanako is an OL ( < office lady, i.e. office girl) working in that building’ is most likely to be written as follows, where four writing systems are mixed.
花å� ã�¯ ã�‚ã�® ビル ã�§ åƒ�ã�„ã�¦ã�„ã‚‹ OL ã�§ã�™ã€‚
Hanako wa ano biru de hataraiteiru ooeru desu
TOP that building at is working OL COPULA
The traditional way of writing is vertical, with the lines progressing from right to left, and books accordingly open from the reverse direction from English. Today both vertical writing and horizontal writing are practiced, though the formal way is vertical writing, as reflected in newspaper articles and formal letters.
As can be surmised from the above discussion, learning how to write Japanese involves considerable effort. Japanese children must master all four ways of writing by the time they complete nine years of Japan’s compulsory education. (RÅ�maji is briefly taught in the Japanese language class in the fourth grade of primary school.) Of these, the most difficult is the kanji. Not only are kanji difficult to write correctly according to the correct order of strokes, but they are also hard to read. For most kanji at least two ways of reading must be learned: one the on-yomi, the Sino-Japanese reading, and the other kun-yomi, the Japanese reading. As mentioned above, for the character å±± ‘mountain’, san is the Sino-Japanese reading and yama the Japanese. Normally, the Sino-Japanese reading is employed in compounds consisting of two or more characters, while in isolation, the Japanese reading is adopted. Thus, again for å±±, the Sino-Japanese reading is employed for a compound like 山脈 sanmyaku ‘mountain range’ and the Japanese reading for a form like å±± ã�Œ yama ga ‘mountain NOM’, where the character occurs alone together with the nominative particle written in hiragana.
An additional complication is the multiplicity of Sino-Japanese readings. This is due to the fact that, as discussed earlier, Chinese words and their pronunciations were borrowed from different parts of China as well as at different times. Thus, dialectal differences in pronunciation also had to be learned by the Japanese. The problem is not simple because one and the same character may have two or three different readings. For example, the character è¡Œ, which has the Japanese reading ik-u as in è¡Œ ã�� ‘to go’, has the go’on reading gyoo as in 行儀gyoogi ‘manner’, the kan’on reading koo as in 行進 koosin ‘marching’, and the tÅ�-sÅ�’on reading an as in 行脚 angya ‘pilgrimage’. Unlike the Japanese versus Sino-Japanese reading, there is no systematic rule for determining whether a given expression is to be read in kan’on, go’on, or tÅ�-sÅ�’on; each expression must be learned as to which way it is read. That is, the go’on/kan’on/tÅ�-sÅ�’on distinction is purely historical, and speakers of Japanese must simply live with the fact that in addition to the Japanese way of reading, most kanji have two or more Chinese ways of reading them, and that the kanji is likely to be pronounced differently depending on the expression in which it is used.
Because of the complexity caused by retaining all these writing systems and different readings of kanji, there have been movements for abolishing Chinese characters in favor ofkana writing, or even more radical movements for completely Romanizing the Japanese language. All these, however, have so far failed, and it is safe to say that the Chinese characters are here to stay. What has been done instead of abolishing Chinese characters altogether is to simplify their shapes and limit the number of commonly used characters. In 1946, the Japanese government issued a list of 1,850 characters for this purpose. The list was revised in 1981, and the new list, called j�y� kanji hy� (list of characters for daily use), contains 1,945 characters recommended for daily use. This is now regarded as the basic list of Chinese characters to be learned during primary and secondary education. Also, most newspapers try to limit the use of characters to these 1,945 characters; when those outside the list are used, the reading inhiragana accompanies them.
One of the most inefficient aspects in dealing with Japanese-language materials has been typewriting, for it involves a large tray of characters from which one has to find and type out the desired characters. This is very time-consuming compared with using a European typewriter; and official documents must be typed. This problem, largely caused by the presence of Chinese characters, has been solved to a large extent due to progress in electronics technology and the development of Japanese word processors, some of which can handle both Japanese and European languages and are nearly as efficient as European typewriters. The process of converting kana sequences into the normal sequences of mixed kana and characters is, however, an additional step still required.
1.3 Loan words
The Japanese lexicon contains an extremely large number of loan words. Japanese has borrowed words from neighbouring languages such as Ainu and Korean, but by far the most numerous are Chinese loan words. Traditionally, the Japanese lexicon is characterized in terms of three strata. The terms wago ‘Japanese words’ or Yamato-kotoba ‘Yamato words’ refer to the stratum of the native vocabulary, and kango ‘Chinese words’ refers to loan words of Chinese origin (hereafter called Sino-Japanese words or, simply S-J words). All other loan words from European languages as well as from Korean and Southeast Asian languages are designated by the rubric gairaigo ‘foreign words’ (lit. ‘foreign coming words’). The relative proportions of these loan words in the Genkai dictionary (1859) were: Sino-Japanese words ‒ 60 percent, and foreign words ‒ 1.4 percent. Although the proportion of foreign words has steadily increased, that of Sino-Japanese words remains fairly constant.
The effect of loan words on the Japanese language is not insignificant. In particular, the effects of S-J borrowing have been felt in all aspects of the Japanese language, including syntax. Restricting our discussion to the domain of the lexicon, however, S-J and foreign loan words have resulted in a large number of synonymous expressions. This demonstrates that Japanese has borrowed even those words whose equivalents already existed in the language. This may appear at first to be unmotivated and uneconomical. However, ostensibly synonymous words are often associated with different shades of meaning and stylistic values, thereby enriching the Japanese vocabulary and allowing for a greater range of expression. For example, Sibata (1976) makes some interesting observations about the following sets of synonymous triplets:
(1) Gloss ‘inn’ ‘idea’ ‘acrobat’ ‘detour’ ‘cancellation’
Native yadoya omoituki karuwaza mawarimiti torikesi
S-J ryokan tyakusoo kyokugei ukairo kaiyaku
Foreign hoteru aidea akurobatto baipasu kyanseru
In general, the native words have broader meanings than their loan counterparts. For example, torikesi can be applied to various kinds of cancellation-type acts, even in taking back one’s words. The S-J word kaiyaku is normally used with reference to the cancellation of contracts and other formal transactions. The foreign word kyanseru, on the other hand, is used only for the cancellation of appointments or ticket reservations, etc. The S-J words, which generally convey a more formal impression, tend to be used with reference to higher quality objects than do the native equivalents. On the other hand, the foreign words have a modern and stylish flavor.
Though various factors can be pointed out to account for the ready acceptance of loan word in Japanese, the main linguistic reasons have to do with the lack of nominal inflection and the presence of a syllabary writing system. Since Japanese does not mark gender, person, or number on nouns, and since cases are indicated by separate particles, a loan word can simply be inserted into any position where a native nominal might appear, with no morphological readjustment. For the borrowing of verbal expressions, Japanese utilizes the verb suru, which has the very general meaning ‘do’. This useful verb can attach to the nominal forms of loan words to create a verbal expression, e.g. the S-J word hukusya ‘copy’ yields hukusya-suru ‘to copy’, and the English loan kopii ‘copy’ yields kopii-suru ‘to copy’.
Adjectives are borrowed into the category of adjectival nominals, which take the -da ending in the predicative function. The word soft is thus borrowed as sohuto-da ‘(it is) soft’. The ending-da becomes -na in the attributive function, as in sohuta-na ‘soft’. The adverbial ending -ni may be substituted for -da/-na to yield the adverbial form sohuto-ni.
The presence of a syllabary writing system also facilitates the borrowing of foreign words which can be adapted into Japanese at the phonetic level, though the original pronunciation is often grossly altered. The possibility of adapting foreign words without the semantic interference entailed by adapting Chinese characters, which are largely ideographic, gives Japanese an advantage over Chinese in the domain of lexical borrowing. Indeed, in Chinese those loan words that are given characters without semantic consideration tend not to stay, being replaced by new words created by the translation of the original foreign words.
1.4 Sino-Japanese words
It is generally believed that Chinese words were first introduced into Japan during the first century A.D., or possibly even before that. According to the Nihon shoki (Chronicles of Japan) (A.D. 720), a systematic introduction of the Chinese language occurred around A.D. 400, when Korean scholars brought Chinese books to Japan. While Chinese characters and words remained primarily the instruments of recording official documents and of scholarly writing for a long time, they were gradually absorbed into everyday Japanese language. By the end of the Edo period (1603‒1867), they had penetrated the colloquial language and local dialects. However, many words that are presently classified as kango (Chinese words) are actually not of Chinese origin.
Following the Meiji Restoration (1867), the new government, in its march toward modernization, renovated the system of government as well as that of various academic fields. A part of this modernization effort included the invention of new terms that were, for the most part, translations of English terms. These new terms were created by using the existing Chinese characters, and accordingly, they belong to the stratum of S-J words despite their origin. In other words, the name for this stratum kango refers to those words that utilize the Chinese meaning of characters. Some of these Japanese-made words, such as syakai ‘society’ andkagaku ‘science’, have been introduced to and adopted by the Chinese language.
J.C. Hepburn noted in the 1886 revision of his dictionary Wa-ei gorin shūsei (Japanese‒English Glossary) that he had added more than 10,000 new words since the first edition, published in 1867. This dramatic increase in the proportion of S-J words in the Japanese lexicon is thus attributable to the modernization effort of the Meiji period. The proportion and the status of S-J words in Japanese are strikingly similar to that of Latinate words in English. Ueno (1980) points out that the proportion of Latinate words in the English vocabulary is 55 percent, while that of Germanic (Anglo-Saxon) words and of other foreign loans is 35 percent and 10 percent, respectively. Statistics of these different lexical strata incorporating their frequency of occurrence, on the other hand, show Germanic words occupying 85 percent of the total word count. Though the proportion of Japanese words in actual texts is not as high as that of Germanic words in English, the proportion that S-J words occupy in the Japanese vocabulary is quite comparable to that of Latinate words in the English vocabulary.
Furthermore, the status of S-J words in Japanese is quite similar to that of Latinate words in English. S-J words tend to express abstract concepts, and academic vocabulary is mainly comprised of S-J words. These differences can be seen when native and S-J synonyms are compared, e.g.
(2) Gloss Native S-J
yesterday kinoo sakuzitu
language kotoba gengo
play asobi yuugi
receipt uketori ryoosyuusyo
difference tigai sooi
form katati keitai
forest mori sinrin
Native words such as kinoo and asobi generally belong to the colloquial language, while their S-J counterparts are found in the literary language or academic vocabulary. Compare this with the following list in which Latinate words are contrasted with their Germanic counterparts in English.
(3) Germanic Latinate
help aid
begin commence
hide conceal
happiness felicity
deep profound
wish desire
Native words and Chinese vocabulary are also differentiated with regard to the syntactic categories that they belong to. The majority of S-J words (as well as other foreign word) belong to the category of nouns, with some included in the adverb category. Most of these nouns express abstract concepts. In order to function as verbs, they must be compounded with the native word suru ‘do’, which carries tense; for example, kenkyuu ‘study(N)’ becomes kenkyuu-suru‘study(V)’. This special category of verbal nouns requires the dummy verb suru ‘do’ in the predicate function. They are typically S-J (Sino-Japanese) words, but certain native nouns as well as foreign loan nouns that are derived from verbs also belong to this category. Kenkyuu-suru ‘to study’ and soodan-suru ‘discuss’ are S-J words and typical of the predicates formed with verbal nouns. However, a native compound such as azi-tuke ‘seasoning’ and a foreign loan such as kopii ‘copy’ may also take suru and become verbs: azi-tuke-suru ‘to season’, kopii-suru ‘to copy’. Notice that regular nouns cannot be verbalized by suru: hon ‘book’ *hon-suru,yama ‘mountain’ *yama-suru, syatu ( < shirt ) *syatu-suru. Verbal nouns are like other nouns in that they can function as subject or object, taking that particles ga or o respectively: kenkyuu ga hituyoo da ‘studying is necessary’, kenkyuu o hazimeru ‘begin studying’. This noun-like property of verbal nouns yields analytic syntactic expressions alongside the compound predicate expressions: kenkyuu-suru ‘to study’: kenkyuu o suru ‘to do studying’. On the other hand, words in the categories of verbs, adjectives, and adverbs and other minor categories such as conjunctions mostly belong to the native vocabulary. The other special category of nouns is called, in traditional Japanese grammar, keiyoo-doosi ‘adjectival verb’ by Martin (1975). This category is really more like nouns than verbs, contrary to what the traditional term seems to suggest. Many Western grammarians refer to this category as “nominal adjectives”, but we will follow Martin here, and use the term adjectival nouns, since these, unlike regular adjectives, take the copula. E.g. (a) Ano hito wa kirei da. (kirei da = adjectival nominal predicate) ‘That person is pretty.’ as compared to Ano hito wa gakusei da. (gakusei da = nominal predicate) ‘That person is a student.’ as compared to Ano hito wa utukusii. (utukusii = adjectival predicate) ‘That person is beautiful.’ (b) kirei na hito (adjectival nominal attribute) ‘pretty person’ as compared to gakusei no Taroo (nominal attribute) ‘Taro, who is a student’ as compared to utukusii hito (adjectival attribute) ‘beautiful person’. Adjectival nouns also differ from regular nouns in that they cannot possess the grammatical functions of subject, objec
-
>>>
Adjectival nouns also differ from regular nouns in that they cannot possess the grammatical functions of subject, object, etc. Thus, they cannot directly take case particles such as the nominative marker ga or the accusative o; gakusei ga ‘student NOM’ but not *kirei ga ‘pretty NOM’. However, some adjectival nouns have the corresponding nouns as well, and these forms function like a noun in this respect. For example, genki ‘vigorous/vigor’ and kenkoo ‘healthy/health’ are both adjectival nouns as well as regular nouns. There is one derivation suffix that generates adjectives from nouns, namely -rasii (the last i being a tense-marker for adjectives). Thus, nouns, such as gakusei ‘student’, yield adjectives as gakusei-rasii ‘student-like’. Neither adjectival nouns nor adjectives can use the suffix: *kirei-rasii, *utukusii-rasii. Once a noun, e.g. gakusei ‘student’, is made into an adjective by the suffixation of -rasii, it can then be nominalized by -sa like any other adjective: gakusei-rasi-sa ‘student-like-ness’. The special categories of adjectival nouns and verbal nouns have important functions in the Japanese lexicon. For one thing, they are the categories to which certain newly derived forms belong. More importantly, these categories are most actively used in borrowing and compounding. Foreign adjectives are generally borrowed into the adjectival noun category, e.g. tahu da ( < tough), sikku da ( < chic). As noted above, foreign verbs are borrowed into the verbal noun category, e.g. zerokkusu-suru ‘to xerox’, taipu-suru ‘to type’. Many noun compounds involving the adverbial, or infinitive, form of a verb also function as verbal nouns, e.g. haya-gawari-suru (quick-change-do) ‘to make a quick change’, tati-yomi-suru (stand-read-do) ‘to read while standing’.
From a morphological point of view, S-J words are more analytic and therefore semantically more transparent than native words. In general, S-J nouns consist of two or more morphemes, with the last morpheme often indicating the object category of the designated referent. For example, the morpheme ki (æ©Ÿ) means ‘machine’, and words referring to various types of machines contain this morpheme, e.g.
(4) hikoo-ki ‘airplane’
soozi-ki ‘cleaner’
densan-ki ‘computer’
syok-ki ‘spinning machine’
Similarly, the terms for different branches of learning all contain the morpheme gaku (å¦) ‘study’ as in:
(5) suu-gaku ‘mathematics’
rekisi-gaku ‘history’
kaikei-gaku ‘accounting’
tookei-gaku ‘statistics’
syakai-gaku ‘sociology’
Furthermore, since the meaning of each character is fairly stable, one can often correctly guess the meaning of a word even upon encountering it for the first time. For instance, even if one did not know the meaning of the word gengo-gaku (言語å¦), one would know that gen (言) and go (語) both mean ‘language’ or ‘word’. From this, one could guess that the word means something like study of language or words; and the word indeed means ‘linguistics’. This process, of course, is analogous to an intelligent guess that an English speaker might make on encountering a Latinate word like linguistics. However, this parallelism between S-J words in Japanese and Latinate words in English, which tend to be more analytic than Germanic words, is only apparent, for the meaning of Chinese characters is far more transparent to the ordinary Japanese than Latinate formatives are to ordinary speakers of English; and, thus, the Japanese have much more advantage in guessing the meaning of unknown words written in characters than English speakers faced with unknown Latinate words.
1.5 Affixation
Japanese is an agglutinative language and its morphology involves both prefixes and suffixes. In examining the process of affixation, several issues need to be taken up. Among them, we will examine the questions of (a) productivity, (b) category change, and (c) the interactions of morphology and syntax.
If one analyzes Japanese verbs morphologically, various affixes can be identified. However, not all these affixes are productive to the same extent, in that some of them co-occur only with specific roots. The transitive/intransitive suffixes are of this type. A comparison of forms such as ak-u ‘to open’ and ak-e-ru ‘to open X’ reveals that -e is a transitivizing suffix. However, many other transitive verbs do not use this suffix and instead have others, e.g. tob-u ‘to fly’: tob-as-u ‘to fly X’, or-i-ru ‘to come down’: or-o-su ‘to bring down’, sinu ‘to die’: koros-u ‘to kill’. In other words, the suffixes seen here, -e, -as and -os, occur only with specific verbs, and there is only a limited range of subregularity. This is similar to the state of affairs with intransitive/transitive pairs in English.
Compared to these, the causative suffixes -sase, -se and the passive suffixes -rare, -re, both of which are classified as auxiliary verbs in traditional grammar, are quite regular, and can be suffixed to any verb stem as long as no semantic constraints bar them. Thus among the suffixes, we must distinguish between productive and non-productive suffixes.
This same division applies to prefixes. The prefixes su- ‘bare-’ and hi- ‘slight-’ can occur respectively with nouns such as asi ‘leg’ and te ‘hand’, and with adjectives such as yowai ‘weak’ and karai ‘spicy’, yieldig the forms: su-asi ‘bare foot’, su-de ‘bare-hand’, hi-yowai ‘weak and slight’, and hi-garai ‘tingling’. Their productivity, however, is extremely limited, and we find that they cannot be attached to many other nouns such as atama ‘head’ or to the adjectives such as takai ‘high’; *su-atama, *hi-dakai. On the other hand, there exist other prefixes such as the honorific o- and S-J prefixes han- ‘anti’ or hu- ‘un-’, which are very productive.
Generally speaking, the S-J affixes are more productive than the native Japanese affixes. From the historical point of view, productive affixes are those newly introduced, whereas the non-productive ones are remnants of old native forms. In the case of the intransitivizing and transitivizing suffixes, we see some historical connections between them and the currently productive passive and causative suffixes, respectively, indicating the possibility that the former were once productive morphological processes.
While there are both nominal and verbal suffixes, most prefixes are affixed to nouns. The prefix hi- mentioned above, the honorific o-, and ko- ‘slight’ are about the only prefixes that attach to adjectives. The honorific prefix o- also attaches to verbs in honorific verbal expressions such as o-ik-i ni naru (the subject honorific form of ‘to go’), o-tasuke suru (the object honorifc form of ‘to help’). But the verbs in these expressions occur in their adverbial, or infinitive, forms (ik-i ‘going’ and tasuke ‘helping’) which are actually to a great extent like nouns. Indeed, many adverbial verb forms have become nouns. Iki ( < ik-u ‘going’), tasuke ( < tasuke-ru ‘help’) used above and a host of others like yorokobi ‘joy’ ( < yorokob-u ‘to be glad’),nezi ‘screw’ ( < nezi-ru ‘to twist’), and asobi ‘play’ ( < asob-u ‘to play’) are listed as nouns in dictionaries. It is because of this that the inflectional category of adverbial form is often called “infinitive” in Western literature, as we occasionally do in subsequent discussion for ease of identification of the form and its partial function for the Western readers. Certain other adverbial/infinitival forms such aruk-i ( < aruku) ‘walking’, and ne ( < neru) ‘sleeping’ do not occur as nouns by themselves, but do form nouns through compounding, e.g. sozoro-aruki ‘leisurely walk’, hiru-ne (day-sleep) ‘nap’.
Nouns are also derived from adjectives with the suffixes -sa and -mi. The former is quite productive, but the latter’s occurrence is limited to a number of specific adjectives suchatataka-i ‘warm’: atataka-mi ‘warmth’, uma-i ‘tasty’: uma-mi ‘taste’. These adjectives can of course by nominalized by -sa as well, yielding atataka-sa ‘warmness’ and uma-sa ‘tastiness’. As the glosses indicate, the -sa forms are analogous to nouns derived by the productive English suffix -ness, and they, just like the -ness forms, tend to be more abstract in meaning than the -mi forms. Contrast the English forms beautifulness: beauty.
Other noun-forming suffixes have specific meanings. The suffix -te (lit. ‘hand’) refers to a person who does what is specified by the verb, just like the agentive suffix -er in English:hanasi-te (speaking-hand) ‘speaker’, okuri-te (sending-hand) ‘sender’. S-J suffixes -nin ‘person’, -in ‘member’, -sya ‘person’, -syu ‘hand’ etc. all have a similar function: uketori-nin (receiving-person) ‘recipient’m gesyuku-nin (boarding-person) ‘boarder’, soodan-in (consulting-person) ‘consultant’, hanzai-sya (crime-person) ‘criminal’, unten-syu (drive-hand) ‘driver’. Just as it is difficult to determine the basis for the distribution of the several agentive suffixes in English (painter, consultant, linguist, logician, etc.), these S-J agentive suffixes have no apparent logical distribution and must be learned separately.
Verbs are formed with the suffixes -garu, -meku, and a few others, but these each have specific meanings associated with them. -garu, as in samu-garu ( < samu-i ‘cold’) ‘appear/show signs of being cold’ and uresi-garu ( < uresi-i ‘glad’) ‘appear glad’, has the meaning of ‘appear to be X’, and -meku has the meaning of ‘become like X’ as in haru-meku ‘become spring-like’ andaki-meku ‘become autumn-like’.
The productive S-J suffix -teki produces adjectival nouns, which, as seen above, function both as predicates and as noun modifiers together with the copula da, e.g. kagaku-teki ‘scientific’, goori-teki ‘rational’. Verbal nouns are produced from nouns by the suffix -ka ‘-ize/-fy’ and some others, e.g. kindai-ka ‘modernize’, eki-ka ‘liquefy’, dorama-ka ‘dramatize’, eiga-ka ‘make a movie out of (a novel)’.
The prefixes do not generally change lexical categories. Notable exceptions to this are the S-J prefixes having negative meanings: hu- ‘un-’, mu- ‘-less’, hi- ‘non-’, mi- ‘not yet’. These change nouns into adjectival nouns, e.g. tyuui ‘attention’: hu-tyuui ‘careless’, rieki ‘profit’: mu-eki ‘profitless’, zyooken ‘condition’: mu-zyooken ‘unconditional’, zyoosiki ‘common sense’: hi-zyoosiki ‘senseless’, kaihatu ‘development’: mi-kaihatu ‘undeveloped’.
1.6 Compounds
Compounding is by far the most productive process of new word creation. In Japanese, compounding is a particularly productive process for it combines all categories of elements as the examples below indicate.
(6) Native compounds:
a. N‒N : aki-zora ‘autumn sky’
kona-yuki ‘powdery snow’
b. A‒N : maru-gao ‘round face’
tika-miti ‘short cut’
c. Vi‒N : watari-dori ‘migratory birds’
nomi-mizu ‘drinking water’
(Vi = infinitive, or adverbial, form or a verb)
(7) Sino-Japanese compounds:
a. N‒N : ki-soki ‘rule’
hu-bo ‘parents’
b. A‒N : syoo-u ‘slight rain’
koo-ri ‘high interest’
c. V‒N : si-ketu ‘stopping of bleeding’
satu-zin ‘manslaughter’
(8) Hybrid compounds:
a. S-J‒native: dai-dokoro ‘kitchen’
b. native‒SJ: to-kei ‘clock’
c. S-J‒foreign: sekiyu-sutoobu ‘oil stove’
d. foreign‒S-J: taoru-zi ‘towel cloth’
e. native‒foreign: ita-tyoko ‘chocolate bar’
f. foreign‒native: garasu-mado ‘glass window’
g. foreign‒foreign: teeburu-manaa ‘table manner’
Among the compounds in which the second element is a noun, noun‒noun compounds are the most numerous. Enlarging the class of compounds to include forms in which the second element is the infinitive form of a verb or an adjective stem, we find that the N‒Vi type is the most numerous, e.g.
(9) a. N‒Vi : yuki-doke ‘snow-melting’
tume-kiri ‘nail clipper’
b. A‒Vi : naga-tuzuki ‘long lasting’
haya-zini ‘premature death’
c. Vi‒Vi : tati-yomi ‘reading while standing’
hasiri-zukare ‘fatigue due to running’
d. N‒As : hara-ita ‘stomachache’
&n
-
>>>
d. N‒As : hara-ita ‘stomachache’
iro-ziro ‘fair skinned’
e. As‒As : taka-hiku ‘high-low’
f. Vi‒As : kire-naga ‘long-slit (as of eyes)’
(Vi = infinitive form of a vern, As = adjective stem)
Most of the resulting compounds belong to the categories of nouns or verbal nouns. The native compounds having nouns as their second element, e.g. (15), are all nouns. Among the S-J compounds, those that do not have a verbal element are all nouns, while those which do, e.g. (16c), are either verbal nouns, such as si-ketu ‘stopping of bleeding’, or nouns, as satu-zin ‘manslaughter’. All hybrid compounds are nouns. For compounds having the infinitive form of a verb or an adjective stem as their second element, it is difficult to predict which category they will belong to. Forms with adjective stems as their second element, (18d‒f), are, for the most part, nouns. Those forms having as their second element the infinitive form of a verb, such as the N‒Vi forms in (18a), are either nouns or verbal nouns. The examples in (18a) all happen to be nouns and do not take the dummy verb suru: *yuki-doke-suru. However, such forms as ki-zukare ‘mental weariness’ and mizu-asobi ‘water play’ are verbal nouns and, consequently, can take suru: ki-zukare-suru ‘to be mentally weary’, mizu-asobi-suru ‘to play in the water’. On the other hand, A‒Vi forms and Vi‒Vi forms typically fall into the category of verbal nouns.
A comparison between S-J and native Japanese compounds indicates that the order of the elements in a compound reflects the normal word order of Chinese and Japanese. Both these languages exhibit a modifier‒head linear order; accordingly, the typical adjective‒noun and adverbial‒verb order is reflected in compounds, e.g.
(10) a. aki no sora ↔ aki-zora ‘autumn sky’
autumn of sky
‘sky of autumn’
b. maru-i kao ↔ maru-gao ‘round face’
round face
‘round face’
c. wataru tori ↔ watari-dori ‘migratory birds’
migrate bird
‘birds that migrate’
d. pen de kaku ↔ pen-gaki ‘pen writing’
with write
‘write with a pen’
e. Hawai de yakeru ↔ Hawai-yake ‘Hawaiian tan’
Hawaii in get-tanned
‘get tanned in Hawaii’
f. ki ni noboru ↔ ki-nobori ‘tree climbing’
tree to climb
‘climb up a tree’
g. gaikoku kara kaeru ↔ gaikoku-gaeri ‘foreign returnee’
foreign country from return
‘return from a foreign country’
h. otoko yori masaru ↔ otoko-masari ‘mannish’
man than surpass
‘surpassing a man’
i. hito to tukiau ↔ hito-zukiai ‘socializing’
person with associate
‘associate oneself with people’
j. tyotto miru ↔ tyotto-mi ‘glimpse’
slightly look
‘look slightly’
The Japanese subject‒verb order is also reflected in the following:
(11) a. Yuki ga tokeru. ↔ yuki-doke ‘snow-melting’
snow NOM melt
‘Snow melts.’
b. Asi ga nagai. ↔ asi-naga ‘long-legged’
leg long
‘Legs are long.’
The major difference between S-J and native Japanese compounds with respect to the order of compounded elements occur in forms that reflect the verb‒object relationship. While Japanese is an OV language, Chinese exhibits VO order. This difference is accordingly reflected in compounds: the native compounds have N‒Vi order, and the S-J compounds V‒N order. Compare the following synonymous words in which the (a) examples are native and the (b) examples Sino-Japanese.
(12) a. hito-gorosi ‘manslaughter’
person-killing
b. satu-zin ‘(ditto)’
kill-person
(13) a. ti-dome ‘stoppage of bleeding’
blood-stop
b. si-ketu ‘(ditto)’
stop-blood
(14) a. tosi-kosi ‘passing of a year’
year-pass
b. etu-nen ‘(ditto)’
pass-year
(15) a. iro-zuke ‘coloring’
color-apply
b. tyaku-syoku ‘(ditto)’
apply-color
Since the S-J pattern of V‒N is deeply ingrained in the Japanese language, even recent S-J coinages made within the realm of the Japanese language follow this pattern. In the case of S-J forms that reflect subject‒verb order, the situation is not quite as clear. Many words, as can be seen below, have the V‒N order of the transitive pattern seen above, although N‒V forms, e.g. (25d), occasionally occur:
(16) a. raku-yoo ‘falling leaves’
fall-leaf
b. syuk-ka ‘the outbreak of a fire’
erupt-fire
c. gyoo-ketu ‘the coagulation of blood’
coagulate-blood
d. niti-botu ‘sun-set’
sun-set
The V‒N pattern of the forms based on the intransitive structure seen above may be understood either as being due to the influence of the transitive pattern, which predominates in the compounding phenomenon, or as a reflection of the Chinese word order of a special type of sentence. In the tradition of Chinese grammar, the intransitive sentences that typically depict the occurrence of a phenomenon, such as the falling of rain and the outbreak of fire, are called “phenomenon sentences”, and they, unlike other intransitive sentences, exhibit the verb‒subject (i.e. V‒N) order. Since all these intransitive subject‒verb compounds are in fact analogous to the phenomenon sentences, in which the subjects are patients rather than agents, the latter interpretation appears to be correct.
For two-word S-J compounds then, the V‒N pattern is predominant. However, when longer transitive compounds are made, the Japanese pattern, reflecting N‒V (of object-verb) order, prevails, e.g. [yu-syutu]-[sei-gen] (export-limit) ‘limiting of export’, [buk-ka]-[too-sei] (price-control) ‘price-control’.
There are two major issues relating to the semantic interpretation of compounds involving nouns and verbal elements. One, a major issue in recent discussions of compound formation, has to do with the grammatical function of the nominal element that can be compounded. Examination of the correlations shown in (19) between the compounds and their phrasal paraphrases reveals that in the case of Japanese, the nominal elements of compounds paraphrase into various grammatical relations. Despite this wide possibility, there is a definite pattern recognized. First, as in other languages, nominals corresponding to the subjects of transitive verbs are lacking, though there are some sporadic forms violating this general pattern, e.g. musi-kui (bug-eating) ‘eating by a moth’. Even those nominals that correspond to the subjects of intransitive verbs are of specific type; namely, they are patients rather than agents. That is, while forms like ame-huri ‘rain-falling’ and kata-kori ‘shoulder-stiffening’ abound, we do not find compounds like *kodomo-naki ‘child-crying’ and *uma-hasiri ‘horse-running’. Among the non-subject functions, the object is the relation most frequently associated with the nominal elements of the compounds under discussion.
The other issue concerned with the semantics of compounds that merits further study has to do with the denotation of the resulting compound forms and the semantic connection between the compounding elements. In the Noun‒Noun compounds, forms ending withnin ‘person’ or sya ‘person’ normally designate the agent of the action denoted by the first element of the compound. Kenbutu-nin (sightseeing-person) is a sight-seer, and annai-nin (guide-person) is a guide. However, there are forms in which the denotee is the patient, such as siyoo-nin (use-person), which means employee, not employer, which is siyoo-sya (use-person). Similarly perplexing is the contrast between words such as asobi-nin (play-person) and asobi-tomodati (play-friend); the former is a “playing-person”, i.e. a person who seeks easy money, but the latter means a playmate.
Understanding the connection between compounded elements often requires extremely elaborate explanations. Tyoo-musubi (butterfly-tie), for example, is a way of tying a ribbon such that the result looks like a butterfly (i.e. a bow knot). One of the more horrendous compounds of this type is uguisu-bari (nightingale-flooring); its meaning is something like: ‘the boarding of a floor such that when people walk on it the boards squeak, emitting sounds like the singing of a nightingale as a warning of their (enemies’) approach’. (Actually, the sounds of the uguisu-bari floors, found in castles and temples, are more like the squeaking of mice.)
Despite occasional oddities like those above, most compounds with the infinitivized form of a verb as their second element have fairly regular patterns of meaning. The most basic and perhaps the original meaning of a given compound of this type, from which all other extended meanings were derived, is the denotation of the event itself expressed by the elements of the compound. These are like the nominalization of a verb phrase by the suffixation of an infinitive, or adverbial, ending, though the results, containing no particles and modifying elements, are distinctively words rather than phrases. Thus, mizu-maki (water-spraying) is the act of spraying water, and yama-nobori (mountain-climbing) is the act of mountain climbing.
Many forms, however, take on additional, extended meanings. For example, hito-gorosi (man-killing) can be either manslaughter or a killer. That is to say, the agent of the designated action, as well as the action itself, are both denoted by the same compound word. Very often this agentive meaning asuumes the entire range of the meaning, supplanting the denotation of the action. E-kaki (picture-painting), for instance, denotes a painter, and no longer the act of painting itself. (Curiously, the basic meaning denoting the act of painting obtains when the beautification prefix o- is attached as in o-e-kaki ‘painting’.) Sumoo-tori (wrestling-doing) ‘SumÅ� wrestler’, kami-yui (hair-dressing) ‘hairdresser’, uta-utai (song-singing) ‘singer’ are some additional examples of this phenomenon.
Note that many of these agent-denoting compounds are translated into the agentive -er forms in English. In English this suffix can refer not only to a human agent, but to an instrument used to carry out the specified action as well. For example, washer is an instrument for washing, and a (floor) waxer is a machine used for waxing floors. Interestingly, this same extension of the agentive meaning to cover instruments has also occurred in Japanese. Tiri-tori (dust-collecting) is a dustpan, keito-arai (yarn-washing) is a detergent for washing woolen clothes, and ha-migaki (tooth-polishing) is toothpaste.
There are, as well, certain forms that denote the location in which the specified action takes place. Mono-hosi (thing-drying) is a place for drying laundry, te-arai (hand-washing) is a lavatory, and huna-watasi (boat-crossing) is a ferry terminal. Since these forms can themselves be compounded with ba ‘place’, as in mono-hosi-ba (thing-drying-place), we might consider the locative compounds listed above to be either abbreviations of the ba-compounds themselves or further extensions of meaning to the instrumental reading. As can be seen from the above examples, the compounds derived from transitive verbs generally denote agents or instruments rather than objects or patients. Hito-gorosi (man-killing) ‘manslaughter/killer’, for example, does not refer to the victim. Likewise umi-zuri (sea-fishing) ‘fishing at sea’ cannot mean that which is caught at sea. There is, however, one semantic field in which the compounds almost exclusively do denote objects, rather than agents or instruments. Curiously enough, it is the field of culinary terms. For example, kaki-age (stirring-frying) ‘(a kind of tempura)’, asa-zuke (light-pickling) ‘pickles lightly pickled’, and tako-yaki (octopus-grilling) ‘round dumpling-like snack with octopus bits in it’ all refer to something that is made rather than the person or an instrument that makes it.
Characteristically, the first elements of all these compounds denote modification of the verb or subsidiary materials used rather than the objects affected. Kaki-age is a tempura dish made of vegetable pieces that are stirred and mixed in the dough (the kaki part), and then fried (the age part). Tako-yaki (octopus-grilling) is not really grilled octopus. Rather, it is a round dumpling-like food that has octopus bits as its main ingredient. Otya-zuke (tea-dipping) and nori-maki (seaweed-rolling) are of the same type; the former denotes cooked rice, often with dried-fish shavings, seaweed, or other flavorings, over which Japanese tea is poured, and the latter,sushi rolled with seaweed. (Among non-food objects, tyoo-musubi (butterfly-tying), uguisu-bari (nightingale-flooring) mentioned earlier, as well as ittoo-bori ([one knife]-carving) ‘carving with one knife,’ and mae-kake (front-hanging) ‘apron’, belong to this class of compounds.)
However, compounds that denote objects of actions are not all like this, and there are in fact many forms in which the first element does denote the very object affected, the meaning of the compound then being the effect of the verbal second element upon the nominal first element. Ika-yaki (squid-broiling), unlike tako-yaki (octopus-grilling) mentioned above, is actually broiled squid, tamago-yaki (egg-frying) is a rolled-up fried egg, yasai-itame (vegetable-sauteing) refers to sauteed vegetables, and ume-bosi (plum-drying) denotes dried and pickled plums. Again, some of the compounds in this class display, at best, a rather far-fetched connection between the elements compounded, and the meaning of the compound elements. If one asks for tai-yaki (‘seabream-baking), one can expect neither baked seabream nor a pie with seabream meat in it; but rather a kind of baked snack filled with bean-jam, which is made in the shape of a seabream (the favourite fish of the Japanese).
Earlier, we noted the parallel extension of agentive meaning to the instrumental meaning that has occurred with some Japanese compounds and in English with the forms having the -er suffix. It also happens that certain English -er forms express objects rather than agents or instruments. For example, broiler denotes not a person who broils, but a chicken for broiling. (Broiler can, of course, denote the broiling instrument itself.) Likewise, a frier can be either a person or an implement that frieds, or something to be fried. A prisoner is not a janitor, but the one who is imprisoned.
As discussed more systematically in the next section, the kinds of compound being discussed here are lexical compounds such that they are listed in the lexicon. However, when similar compounds are newly created, they sometimes interact with syntactic elements, e.g. [izime-rare]-kko ([bully-PASS]-child) ‘a child tending to be bullied’ and [mirai no hazu]-sagasi ([future of husband]-search) ‘search for a future husband’. In the first example, the first element is an infinitival verb form containing the passive morpheme izime-rare ‘bully-PASS’. This form itself is not an independent noun and occurs only as a compound element, as in the above exmple or in [izime-rare]-zuki ([bully-PASS]-one who likes) ‘one who likes to be bullied’. The form [mirai no hazu]-sagasi ‘search for a future husband’ is also interesting, for it contains a phrase with the particle no ‘of’ within its first element. Another example of this type of compounding, found in a newspaper advertisement, is [zyosei no kimoti]-doroboo (woman’s heart-thief) ‘a stealer of a woman’s heart’ used in reference to a popular movie actor. Compounding in general, however, does not allow the involvement of a phrasal element, and indeed, forms such as those cited here are quite rare in Japanese as well. When they do occur, they tend to appear in commercial messages created for the nonce or in broadcasting of sports events, which contain certain lexicalized forms, e.g. [tiisana henka]-kyuu ([small change] ball) ‘a ball that deflects slightly as in a curve or shoot’, which originate in instantaneous compound creation.
There are two additional types of compounds in Japanese that are not very frequently seen in English or in other European languages. One is the so-called dvandva compound that expresses a coordinate relationship among the compound elements, while the other is the verb compound.
In English the compound in which two members have the same status as head ‒ the dvandva compound ‒ most frequently appears as a prenominal modifier: e.g. mother‒child interaction
-
>>>
In English the compound in which two members have the same status as head ‒ the dvandva compound ‒ most frequently appears as a prenominal modifier: e.g. mother‒child interaction, employer‒employee relationship. The use of dvandva compounds in Japanese is much wider and more frequent. While many have three or more elements, as in oya-ko-mago ‘parent-child-grandchild’, the most frequent ones have just two. The ordering of elements is fixed and is determined largely on the basis of the notion “priority”: positive > negative (yosi-asi ‘good and bad’), up > down (ten-ti ‘heaven and earth’), right > left (migi-hidari ‘right and left’), male > female (dan-zyo ‘man and woman’), old > young (oya-ko ‘parent‒child’) ‒ but new > old (sin-kyuu ‘new and old’) ‒ superior > subordinate (zyoosi-buka ‘superior and subordinate members of an organization’), inside > outside (uti-soto ‘inside and outisde’) ‒ but far > near (atti-kotti ‘there and here’) ‒ etc. (cf. Kageyama 1982).
Compound verbs are of several types. Some are noun‒verb combinations, while others are adverb‒verb combinations. Still others are verb‒verb combinations. Among these, the first two classes reflect the syntactic order of sentence elements, as shown below:
(17) a. Subject‒verb
Awa ga tatu ↔ Awa-datu ‘bubble up’
bubble NOM rise
Iro ga tuku ↔ Iro-zuku ‘become coloured (as of flowers)’
color NOM put on
b. Object‒verb
Na o tukeru ↔ na-zukeru ‘to name’
name ACC attach
Yuuki o tukeru ↔ yuuki-zukeru ‘to encourage’
courage attach
c. Adverb‒verb
tikaku ni yoru ↔ tika-yoru ‘come closer’
near to come
wakaku kaeru ↔ waka-gaeru ‘to become young again’
young return
Once again, the combination of a subject and transitive verb is almost entirely lacking. These noun‒verb and adverb‒verb forms, however, are not as numerous as the nominalized compounds discussed above. Indeed there are a great many gaps between the nominalized compounds and verb compounds of this type. For example, in the case of awa-datu ‘bubble up’ seen above, there is a corresponding nominalized form, awa-dati ‘foaming’. The compound iro-zuku ‘put on a color’, however, does not have a corresponding nominalized form iro-zuki. On the other hand, there is the nominal niku-zuki (flesh-put on) ‘well-fleshedness’, but no corresponding verbal form niku-zuku. In the majority of cases, an N‒Vi compound exists but the corresponding N‒V verbal form does not, e.g. hon-yomi (book-reading) but no hon-yomu (book-read).
Far more numerous than the N‒V or Adv‒V compounds are verb‒verb compounds. The verb‒verb compounds can be classified into three groups on the basis of the semantic relationship that holds between the first verb and the second. First, there are those in which the first verb modifies the second. This type is similar to English expressions such as break open, in that the first verb expresses the manner in which the action denoted by the second verb is carried out. Let us represent this type of compound as V/m‒V. The second type is the inverse of the first in that here the second verb modifies the first. Typically, the second verb expresses any of various aspectual meanings. Let us then call this the V-V/m type. The last type contains two verbs that possess an equal level of importance, and we well therefore refer to it as the V‒V type. Some examples of these three classes follow:
(18) a. V/m‒V type:
naguri-korosu (beat-kill) ‘kill by beating’
kiri-taosu (cut-fell) ‘fell by cutting’
kami-kiru (bite-cut) ‘cut by biting’
b. V‒V/m type:
kaki-ageru (write-raise) ‘write-up’
ii-tukusu (say-exhaust) ‘say exhaustively’
koware-hazimeru (break-begin) ‘begin to break’
c. V-V type:
naki-wameku (cry-shout) ‘cry and shout’
tobi-haneru (jump-spring) ‘jump and spring’
Compounds of the V/m‒V type allow paraphrasing such as nagut-te korosu (beat-CONJ kill) ‘kill by beating’ in which the second verbal element alone has the predicate function. The V‒V/m type compounds do not allow this kind of paraphrasing, as kaki-ageru ‘write up’ does not yield *kai-te ageru. The second verbal element in compounds of this type is more like a suffix in that it often does not preserve the original meaning of the verb. For example, ageru of kaki-ageru ‘write up’ means ‘to raise’ in isolation. However, in compounds such as kaki-ageru and si-ageru (‘do-raise’) ‘finish up’, ageru does not possess its isolate meaning; rather it expresses the notion of completion, just as up metaphorically takes on the meaning of completion in English glosses. Secondly, transitive aspectual verbs such as hazimeru ‘begin’ and tuzukeru ‘continue’ are used in compounds, irrespective of the transitivity of the compound. Hazimeru, for example, is transitive as opposed to its intransitive counterpart hazimaru. But in the compound forms, the transitive form hazimeru is exclusively used, even when the sentence itself is intransitive by its nature, e.g. Ame ga huri-hazimeta ‘It began to rain’ (lit. ‘The rain began to fall’), Hana ga siore-hazimeta ‘The flowers began to wilt.’ These facts indicate that in V-V/m compounds, the second verbal elements are in the process of losing their dependent status, and may before too long be reduced to verbal suffixes having aspectual meanings.
V‒V type compounds are rare compared to the other two types. They might be paraphrased as ‘do X-ing and Y-ing’, e.g.
(19) naki-sakebu ↔ nai-tari saken-dari suru
cry-shout cry-and shout-and do
‘do crying and shouting’
1.7 Post-syntactic compounds
As noted earlier, some of the noun compounds discussed above seem to show some interaction with syntax. However, such an interaction is rather sporadic, and this limited interaction of the compound elements with syntactic elements outside the compounds indicates that they are by and large formed in the lexicon. We present here a case of compound formation that is claimed to take place not in the lexicon but at the post-syntactic, and possibly phonological level. On the one hand, the post-syntactic compounds are words, for their formation obeys the constraints on word formation to which lexically formed compounds (hereafter “lexical compounds”) are subject, and on the other, they are distinct from lexical compounds in that the members making up the compounds have independent syntactic status. Our claim, then, is that in Japanese there are two types of compounds, those formed in the lexicon and those formed post-syntactically outside of the lexicon.
What are claimed to be post-syntactic compounds are of two kinds. One type derives from a clausal structure embedded under a noun phrase and headed by formal nouns denoting various notions of time relations, e.g.
(19) [ … ] no N (native)
a. [ … ] no ori ‘(on) the occasion of’
b. [ … ] no setu ‘(at) the time when …’
c. [ … ] no akatuki ‘(at) the happy occasion of’
(20) [ … ]-N (Sino-Japanese)
a. [ … ]-tyuu ‘in the middle of’
b. [ … ]-go ‘after’
c. [ … ]-sidai ‘as soon as’
The clauses headed by these nouns are regular clauses except that their predicates are Sino-Japanese verbal nouns, which do not express tense by themselves. (They must be supported by the verb suru ‘do’ in order to carry the tense.) Now, these clauses may consist of the predicate and nominal arguments with appropriate case particles, and in this shape they are no different from other clauses except for the two properties mentioned, i.e. they involve S-J verbal nouns as predicates and they do not carry tense in the relevant construction. However, they may also appear in the form in which the case particle of one of the nominals is missing. Thus, we observe the following synonymous pairs of expressions:
(21) a. [sensei ga kaigai o ryokoo] no sai
teacher NOM abroad ACC travel of occasion
‘(on) the occasion of the teacher’s traveling abroad’
b. [sensei ga kaigai:ryokoo] no sai
‘(ditto)’
(22) a. [Yamada-san ga tyuukosya o hanbai]-tyuu
Mr. NOM used car ACC sell middle
‘in the middle of Mr. Yamada’s selling used cars’
b. [Yamada-san ga tyuukosya:hanbai]-tyuu
‘(ditto)’
Notice that, although the clauses in question are translated in the nominalized form in English, they are in fact regular clauses in Japanese exhibiting the regular case-marking patterns. What is peculiar in the (b) versions above is that the object nominals do not have the expected accusative case particle o. Our claim is that these combinations of the object nominal and the predicate, indicated by the bold face, are compound words created post-syntactically by deleting the case particle.
The other post-syntactic compounds derive from the noun phrase consisting of two nominals mediated by the genitive particle no. Again, the particle-less forms are claimed to be compounds formed post-syntactically.
(23) a. biru no kansei
building GEN completion
‘the completion of a building’
b. biru:kansei
‘(ditto)’
(24) a. betonamu no sinryaku
Vietnam GEN invasion
‘the invasion of Vietnam’
b. betonamu:sinryaku
‘(ditto)’
Our argument for the claim that these noun‒verb combinations are indeed compound words takes the form of presenting evidence that these combinations both show the formal characteristics of a lexical compound and obey constraints on word formation which the lexical compounds also obey. For the sake of simplicity, we mostly concentrate on the first type of post-syntactic compounds (see Shibatani and Kageyama (1998) and Kageyama and Shibatani (1989) for further detail).
There are five major considerations that point to the conclusion that the present phenomenon constitutes a word formation process. The first consideration is formal, and has to do with the lack of particles. As discussed elsewhere, the particles such as the topic wa, the nominative ga, and the accusative o are dropped in Japanese, but this is a characteristic of informal colloquial speech. However, one domain in which these particles are legitimately absent is the compounds, in which nouns and verbs are joined together without intervening particles. This formal property of the lexical compounds is shared by the post-syntactic compounds. These post-syntactic compounds are, in fact, not part of informal colloquial speech, as they typically occur in formal letters or newspapers.
A second consideration has to do with one of the principles of word formation that prohibits tense inflection on an element of a compound. Thus, taxidrover, church-wenter, etc. are not possible. This is also true with the Japanese lexical compounds in that the compounds involving the verbal elements have these in the infinitive, or adverbial, forms. Thus, yama-nobori ‘mountain climbing’ is possible but yama-nobo-ru (mountain-climb-PRES) is not. The post-syntactic compounds also obey this constraint, and the expressions involving tense-inflected forms do not undergo compounding even though the head is one of those that trigger post-syntactic compounding.
(25) a. [Nakasone-san ga Amerika o hoomon-suru] sai
Mr. NOM America ACC visit-do occasion
‘the occasion when Mr. Nakasone visits America’
b. *[Nakasone-san ga Amerika:hoomon-suru] sai
The (b) version, which contains the tensed verb (as expressed by the dummy verbal element suru), is only possible as an informal colloquial expression. Compare this with the perfectly well-formed formal expression in (26b) below, which lacks the tense in the embedded clause.
(26) a. [Nakasone-san ga Amerika o hoomon] no sai
Mr. NOM ACC visit GEN occasion
‘the occasion when Mr. Nakasone visits America’
b. [Nakasone-san ga Amerika:hoomon] no sai
‘(ditto)’
A third morphological principle that the post-syntactic compounds obey has to do with the grammatical relations of the nominals involved. As seen above, the Japanese lexical compounds involve nominals holding various grammatical and syntactic relations. However, most predominant are the ones involving the relations of the subject of an intransitive verb and the object of a transitive verb. Extremely rare are those involving the relation of the subject of a transitive verb. This is in line with the general properties of word formation such as compounding and noun incorporation (see Mithun (1984) and the section on noun incorporation in Part 1 of this book). This systematic avoidance of the transitive subject in word formation processes is also shared by the post-syntactic compounds, as evidenced by the impossibility of (27c).
(27) a. Intransitive subject
[hikooki:tuiraku]-go
airplane crash after
‘after the airplane crashed’
b. Transitive object
[sensei ga kaigai:ryokoo]-tyuu ni
teacher NOM overseas travel middle in
‘in the middle of the teacher’s traveling overseas’
c. Transitive subject
*[Nakasone-san:Amerika o hoomon] no sai
d. Transitive subject (OSV order)
*[Amerika o Nakasone-san:hoomon] no sai
The impossibility of (27c) may be explained in terms of another principle discussed presently, but the fact that even the subject contiguous to the predicate, as in (27d), cannot be compounded shows that the grammatical relation plays an important role here. Another manifestation of the principle excluding the transitive subject is that the derived compounds involving a verb whose transitive and intransitive forms are identical yield the ambiguous readings that the involved nominal holds the subject relation of the intransitive version of the verb or that the nominal is the object of the transitive version. This is the case in (23b) above, which can be used in the sense that the building has come to the completed state or that someone has completed the building. That the transitive subject interpretation is impossible is seen in the example in (24b) above, which only means that Vietnam is the party to be invaded (the object reading), while the phrasal counterpart, (24a), is ambiguous in the two readings of Vietnam’s being invaded (the object reading) and its invading another country (the subject reading).
The next lexical property that the post-syntactic compounds also show has to do with the morphological integrity that the words in general exhibit; namely, words do not permit adverbs or parenthetical elements to intrude into them. Thus, a compound word such as theater-going does not allow an adverb like often to intrude, as shown by the ill-formedness of *theater-often-going. The (b) example below shows the applicability of this condition to post-syntactic compounds.
(28) a. Yooroppa o nonbiri ryokoo-tyuu ni
Europe ACC leisurely travel middle in
‘in the middle of traveling Europe leisurely’
b. *[Yooroppa: nonbiri ryokoo]-tyuu ni
Finally, there is the principle that prohibits a phrasal element from being involved in word formation. A phrase such as sly fox cannot be part of a compound, as sly fox-hunting is not a possible compound, where sly is interpreted to be forming a phrase with fox. The same is true with the Japanese lexical compounds, and a form like ookina yama-nobori (big-mountain climbing) is impossible. Now, the post-syntactic compounds too obey this restriction against involving a phrase in word formation, as (29b) is not well-formed.
(29) a. [muzukasii zikken o syuuryoo]-go
difficult experiment ACC complete-after
‘after the completion of a difficult experiment’
b. *[muzukasii zikken:syuuryoo]-go
All these facts show that what we are calling post-syntactic compounding is indeed a word formation process rather than a simple particle deletion phenomenon. Particle deletion in informal speech or in newspaper headlines does not obey these principles of word formation. Our claim that the present phenomenon is post-syntactic, rather than lexical, word formation is burdened by the responsibility for showing, on the one hand, that the phenomenon is a word formation phenomenon like lexical compounding, and that it is difficult from lexical compounding, on the other. That is, we now need to point out those properties of the post-syntactic compounds that are not shared by lexical compounds and that are explainable if we assume the syntactic basis for them.
The first property discussed points out a phonological property that contradicts the phonological notion of a Japanese word. But our claim is that this property is due to the post-syntactic nature of this compounding process. The property in question has to do with the pitch pattern. In Japanese a word (more strictly a minor phonological phrase, which a word typically constitutes) has at most one stretch of high pitch. The lexical compounds, and those syntactically formed words follow this pattern, but the post-syntactic compounds do not.
(30) a. Lexical compounds
-
>>>
(30) a. Lexical compoundskaigai- ryokoo ‘overseas travel’
LH HH H LL
katei- hoomon ‘home-visit’
L HH HL LL
mimi-soozi ‘ear-cleaning’
L H HL L
b. Syntactic compounds
aruka- se-ru ‘make X walk’
L H H H L
nagura- re- u ‘to be beaten’
L H H H L
naguri- ta- i ‘want to beat’
L H H H L
c. Post-syntactic compounds
kaigai:ryokoo (tyuu) ‘(while) traveling overseas’
HL LL HH
Amerika:hoomon (tyuu) ‘(while) visiting America’
L H H HLH HH
mimi : soozi (tyuu) ‘(while) cleaning an ear’
L H LH H
This disparity in the pitch patterns between the lexically and syntactically formed words on the one hand, and the post-syntactically formed words, on the other, can be easily accounted for if we assume that post-syntactic compounds are derived after the phonological rules have assigned pitches to the phrasal constituents. That is, the post-syntactic compounds, which show the same pitch patterns as the corresponding phrases (except for the presence and absence of the particle), simply inherit the phrasal pitch patterns, while the lexically and syntactically formed words obey the regular phonological rules that apply to the word unit.
Another property of the post-syntactic compounds that suggests their syntactic origin has to do with the fact that the number and the nature of the nominal arguments that co-occur with the post-syntactic compounds are exactly the same as those that co-occur in the corresponding clausal expressions, except for the one that is compounded. Thus, if an intransitive clause undergoes post-syntactic compounding, there will be no independent argument left, if a transitive clause is involved, the subject nominal remains, and so on (see (27)). This can be most straightforwardly accounted for if we assume a syntactic (or clausal) basis for the post-syntactic compounds.
The post-syntactic compounds also show the results of their interaction with other syntactic rules. Thus, their verbs may have undergone honorific marking at the time when compounding took place. Such an interaction is not possible with the members of lexically formed compounds, as indicated by the following contrast.
(31) a. [sensei ga kaigai:go-ryokoo]-tyuu ni
teacher NOM overseas HON-travel middle in
‘in the middle of the teacher’s traveling (honorific) overseas’
b. sensei no seito-sidoo
teacher GEN student guidance
‘teacher’s student guidance’
c. *sensei no seito-go-sidoo
HON
Finally, the post-syntactic compounds, as opposed to the lexical compounds, do not form an anaphoric island such that their nominals can form anaphoric relationships with pronouns occurring elsewhere. For example,
(32) Taroo wa senzitu, [tyuukosyai : hanbai]-tyuu ni sorerai
TOP the-other-day used car sell middle in them
no itidai o kowasite simatta.
GEN one-car ACC damaged ended-up
‘The other day, in the middle of selling used cars,
Taro ended up damaging on of them.’
That a lexical compound forms an anaphoric island can be seen from the following, where the lexical compound tyuukosya-hanbai ‘used car sale’ is used.
(33) *Amerika de wa, tyuukosyai-hanbai o suru toki wa, sorerai
in TOP used-car-sale ACC do when TOP them
ni hosyoo o tukenakereba naranai.
DAT guarantee ACC put-not-if won’t do
(Lit.) ‘In America, when you do used-car sales,
you must put a guarantee on them.’
Again, the account that the post-syntactic compounds are syntactically based can explain the difference easily by saying that the surface structure, which has not undergone post-syntactic compounding, is the input to the anaphoric interpretation. Thus, all in all, the phenomena presented above show that, in addition to the lexically formed compounds, Japanese possesses post-syntactically formed compounds. The fact the both types of compound formation obey the same principles of word formation has a significant theoretical implication. Namely, the principles of word formation are not localized in the lexicon; rather they form an independent module which controls word formation in various domains of grammar. Thus, whether word formation takes place in the lexicon, in the syntax, or in the post-syntactic phonological component, it is susceptible to the principles of word formation. What we have observed here are lexical word formation and post-syntactic word formation. There is also word formation that takes place in the syntax. Japanese, thus, exhibits extremely interesting properties of word formation, as it takes place in three major components of grammar.
1.8 Abbreviation
As seen above, compounding is one of the favorite word formation processes in Japanese. One of the consequences of this is the need to accommodate lengthy words. However, Japanese, unlike, for example, German, does not seem hospitable to lengthy words. Native Japanese words are not very long, generally being somewhere between two to four moras. The most common morpho-phonological shape for Sino-Japanese words is that of two morphemes and of three or four moras. Indeed, there are very few single-morpheme, two-mora S-J words, which is perhaps a reflection of the phonological preference manifested by the Chinese language. In fact, just so as to maintain the phonological balance, many S-J words are even composed of two morphemes which have the same meaning, e.g. ki-soku (rule-rule) ‘rule’, too-ka (light-fire) ‘light’, setu-dan (cut-sever) ‘sever’. The vast majority of non-derived words have two moras or less, and noun compounds typically have four moras. These two factors appear to have led Japanese to set its standard word size at two to four moras. However, compounding often produces words that are longer than this standard, and there is a tendency in the language to attempt to confine them to the limits of the ideal word size. The mechanism that Japanese uses to accomplish this is abbreviation, which is extremely common and which is sometimes applied ruthlessly, practically mutilating the original forms.
Abbreviation is applied to both native and S-J compounds e.g. hiru-mesi (day-meal) →hiru ‘lunch’ (Native), seiyoo-siki (Western-style) → yoo-siki ‘Western style’ (S-J). However, since S-J words are far more capable of combining to produce longer compounds, they are more often subject to abbreviation. Another favorite target for abbreviation is foreign loan words. As mentioned earlier, foreign loans are resyllabified so as to meet the requirements of the Japanese phonological system and thereby result in forms that are often far longer than the original words in their native language. Again, since Japanese has the tendency to avoid long words wherever possible, abbreviation is applied ruthlessly, e.g. sutoraiki → suto ‘(labor) strike’.
There are several patterns of abbreviation: (i) clipping of everything after the first few moras, e.g. suto ‘(labor) strike’, panku ‘puncture’; (ii) clipping of the entire first or second part, or the middle part, e.g. ‘platform’ > purattohoomu → hoomu, ‘supermarket’ > suupaa maaketto →suupaa, ‘ham and eggs’ → hamu eggu; (iii) the selection of a mora or two from each member of the compound, e.g. kookyoo-kigyoo-tai (‘public-enterprise-body’) → koo-ki-tai, gakusei-waribiki (‘student discount’) → gaku-wari. Among these, the favorite method is the one exemplified by the last example, in which the initial moras of each member of the compound are picked and combined to form a new word. The resulting word thus conforms to the morpho-phonological pattern seen in the basic form of S-J compounds. Recent abbreviations of this type are paso-kon ← paasonaru konpyuutaa ( < personal computer), mai-kon ← maikuro-konpyuutaa ( < micro computer) and waa-puro ← waado purosessaa ( < word-processor).
Abbreviation sometimes applies even to phrases. Among recent popular expressions are:natukasi no merodii ‘long-for melodies’ → natu-mero, mune ga kyun to naru ‘the heart aches (for you)’ → mune-kyun. Notice that the resulting forms conform to the standard pattern for compounds. Indeed, certain compounds appear to have arisen as a result of abbreviation of phrases. For example, [azi-tuke]-nori (season-put-seaweed) ‘seasoned seaweed’ can be considered as an abbreviation of azi o tuke-ta nori ‘seaweed that has been seasoned’. Other examples include sumi de yaku suteeki ‘steak that is broiled with charcoal’ → [sumi-yaki]-suteeki ‘charcoal broiled steak’ and tosi ga wakai tuma ‘a wife whose age is tender’ → [tosi-waka]-zuma ‘tender-age wife’.
Abbreviated hybrid words such as natu-mero, pointed out above, confuse not only foreigners but also the Japanese themselves, and they are sometimes used without a precise understanding of their original components. Another type of mind boggler results from abbreviation of Japanese-made compounds of foreign words, such as beesu-appu ( < base-up) ‘raising of one’s base salary’ → be-a, pantii-sutokkingu ( < panty-stocking) → pan-suto, hangaa-sutoraiku ( < hunger strike) → han-suto. For foreigners, phonological alteration and semantic changes applied to words from their native languages are by themselves enough to throw them off balance. These abbreviated foreign loan compounds often lie completely beyond their comprehension.
1.9 Development of o
Among the case particles, it is generally believed that the accusative o is related to the interjectional particle of the same sound, which in turn could have evolved from an independent interjection. While some believe that the opposite is the case, Japanese grammarians in general believe that the accusative particle evolved from the interjectional particle. Before entering this main topic, a note on the orthography and pronunciation is in order. The accusative particle is written as ã‚’ (wo), which is aligned with other syllables beginning with the labial guide in the traditional Japanese syllable table. This orthographic representation is opposed to ã�Š (o), which is aligned with the other vowel syllables. From this fact, it is most likely that the accusative particle was pronounced with the labial glide in contra-distinction to the simple o. However , the two began to be confused at the beginning of the Heian period and they merged in the middle of the same period, both becoming wo, which subsequently turned to o. The accusative particle has continued to be written as ã‚’ (wo); in fact it is the only form in which this hiragana is used, as the syllable wo no longer exists. (Wo ando apparently used to contrast, and the former occurred in initial position as well in forms such as wotome ‘maiden’, now pronounced as otome). Except when a distinction is necessary for expository purposes, here and elsewhere in this book, the accusative particle is transliterated simply as o.
The earliest records already show that the particle o was occasionally used to mark the direct object (in addition to other functions to be discussed below), but apparently it was more common not to mark the direct object at all. According to Matsuo (1944), the distribution ofo in the poems and songs of the earliest extensive documents, the Kojiki (712) and the Nihon-shoki (720), is as follows: sixty-two o-marked objects and seventy-four unmarked objects. The following two examples from poems which occur next to each other in the Man’yÅ�shÅ« illustrate this optionality of object marking:
(34) a. … aretaru kyoo mireba kanasimo
ruined capital see COND sad AP
‘my heart grows sad when I see the ruined capital’
b. … huruki kyoo o mireba kanasiki
old capital ACC see COND sad
‘how sad it is to see the old capital’
In the songs of the Kojiki, however, sixty instances of o of the total of seventy were for object marking. This fact is taken as evidence by Konoshima (1953) and others for the conclusion that by the time of Old Japanese the accusative case marker o had been firmly established, though the actual marking was optional. However, subsequent developments indicate that this may not necessarily have been the case. Some reservations notwithstanding on the claim that the particle o was already a case marker, all believe that all uses of o are historically connected, and that etymological plausibility exists for the hypothesis that they go back to an interjectional particle or interjection.
The hypothesis regarding the development of the accusative o from an interjectional particle is based on the following facts: 1) as seen above, the object need not be marked byo and perhaps it was more common not to mark it (thus at an even earlier time the object was not marked at all); 2) at earlier times, a wider range of nominals participated in o-marking; and 3) there was an interjectional particle with the sound o (or perhaps wo), whose meaning (or semantic effect) was largely shared by the object-marking o. The scenario envisioned by many traditional Japanese grammarians of the subsequent development from the time when o marked the object optionally and when it had an interjectional function goes as follows. As time progressed, the interjectional/accusative o gave rise to the conjunctive particle o as well. Whereas both interjectional and conjunctive uses subsequently diminished, and the range of nominals that took on o accordingly became narrower, the need for expressing the logical relation of the object overtly was felt more strongly as the thinking process (of the Japanese of the time) became more complex, eventually leading to the present state, in which o exclusively marks the object, and in which the presence of o after an object is obligatory at least in the formal language. This development of the accusative particle, which was at first optional and then became more and more obligatory until firmly established, looks fairly simple, but in reality the progress was not as straightforward as it might first appear. Before going into this, let us first examine the connections between the accusative use of o and other uses of (or independent particles) o.
Just as in the case of adverbial particles and the corresponding interjectional particles, the criteria for distinguishing one class (or use) of particles from another are rather unclear. In the case of o, the basic criterion that distinguishes the accusative use from the interjectional use is whether or not what is marked is a grammatical object or semantic patient. Thus, those occurring sentence finally, where there is no inverted object nominal ((35a) below), and those following a constituent that is not a direct object (35b) are clear instances of the interjectional particle o.
(35) a. Ie oramase o.
house stay-HON
‘Stay in the house.’ (Honorific)
b. Hototogisu kokoni tikaku o kinakite yo.
here near come-sing FP
‘Oh, hototogisu (name of a bird), come near and sing!’
In the case of the object-marking o, however, it is not entirely clear whether it is any different from the interjectional particle o ‒ the fact that is happens to follow an object nominal rather than other types of constituent does not by itself guarantee that it is different from the interjectional particle. Indeed it must be pointed out, though not generally acknowledged in the literature, that the object-marking o, or at least some instances of it, also had the effect that interjectional particles normally have. In fact, that the object-marking o also had an effect similar to a regular interjectional particle is a necessary assumption for the widely held view regarding the development of the conjunctive use of o from the object-marking o, which is believed to have taken place in the following manner. The interjectional o, like other interjectional particles, conveys the senses of exclamation, lamentation, or wish, and such senses are generally recognized in the object-marking o as well. It is such senses of lamentation and wish that paved the way for the rise of the conjunctive use of the particle o. In the following example, the o-marked nominal kimi ‘my beloved’ is grammatically object of the verbkin ‘will wear (a garment)’. However, though there is no conjoined clause, the presence ofo and the conveyed sense of lamentation lead one to imagine an implicit expression: “But it’s too bad that you aren’t, and that I can’t do as I wish.”
(36) Kakunomi ni arikeru kimi o koromo naraba sita nimo kin
such a way be beloved ACC garment be if bottom at wear
to aga omoerikeru.
that I think
‘If you were a piece of garment, my beloved, I would have worn you
underneath.’
If the implicitly conveyed meaning is overtly expressed, a full-fledged conjunctive use develops, ultimately leading to a situation in which two independent clauses are conjoined byo, as exemplified below:
(37) Asihiki no yama nimo tikaki o hototogisu tukitatu
-
>>>
(37) Asihiki no yama nimo tikaki o hototogisu tukitatu
(pillow word) mountain to close moon-rise
madeni nanika kinakan.
before why come-sing-NEG
‘When the mountain is so close, why don’t you,
hototogisu, come and sing before the moon rise?’
The particle o marked a wider range of nominals than the o of the later periods ‒ a fact that leads us to the assumption that the object-marking o was an instance of the interjectional particle o, which occurred not only sentence finally or after adverbials but also after a wide range of nominal constituents as well. The following illustrate those nominal constituents that were earlier marked by o, but which would not be so marked in later periods: in Modern Japanese (a) would have the nominative ga, (b) the genitive no, (c) the dative ni, (d) temporal ni, (e) and (f) the comitative to, and (g) the directional adverbial form ni mukatte ‘towards’.
(38) a. otomera o sode huru yama
girls sleeve wave mountain
‘the mountain (from which) girls wave their sleeves’
b. miwakasi o turugi no ike no
(pillow word) sword GEN pond GEN
‘of the pond of the sword (of miwakasi)’
c. Oosaka ni au ya otome o miti toeba
ÅŒsaka in meet AP girl road ask COND
‘as I asked the girl met in ÅŒsaka the way’
d. Akikaze no samuki asake o Sano no oka koyuran
autumn wind GEN cold morning GEN hill cross
kimi ni koromo kasamasi o.
loved to garment lend
‘In the cold morning of autumn wind, oh how I wish I could lend
you some clothes when you cross the hill of Sano on such a morning!’
e. kuyasiku imo o wakare kinikeri
grief loved part come
‘I have parted, to my grief, from my beloved’
f. kimi o hanarete koi ni sinubesi
loved part love in die
‘having parted from you, I shall die of lost love’
g. waseko ga ie no atari o yamazu hurinisi
my wife house GEN area incessantly waving
‘my (departing) wife was incessantly waving (her sleeve) toward
the house’
The interjectional particle o disappears during the Heian period, and the range of o-marked nominals narrows considerably, while the conjunctive particle becomes more firmly established. However, the conjunctive o too dies out during the Edo period, leaving the object-marking o as the only survivor of the earlier uses of o.
It was against the background of this generally held view on the development of the accusative particle o that Matsuo (1938, 1944) cast strong doubt on the assumption that o had already acquired the status of an accusative marker by the time of Old Japanese, when the marking is believed to be optional. Matsuo’s original interest was to determine the time when object marking by o became more frequent than leaving object nominals unmarked ‒ this is presumably the time when people became conscious of object marking ‒ the grammatical consciousness shared by the speakers of Modern Japanese.
Matsuo’s research shows that the literary materials from the early Heian period clearly indicate that in prose as well as in poems, o-marked objects outnumber unmarked objects, indicating the use of the object-marking o. However, he also notices some inconsistency in the data he examined. In some works, o-marking is more frequent, but in some others, zero-marking is more frequent. Also there is a difference between the poetic materials and prose materials, which are believed to reflect the colloquial speech of the time; the former show more frequent marking by o than the latter. Here Matsuo raises a question: poetry writing is a more conservative activity than daily conversation, and thus if o-marking is a progressive aspect of the grammar of the time, it is strange that it occurs more frequently in the conservative domain of language use than the domain which should be more susceptible to the on-going change.
The second disturbing fact for Matsuo is that, if the generally held view that o-marking began to be applied more frequently in order to indicate the logical relation more explicitly as the sentence structure became more complex as a reflection of the progressive complication of thinking of the Heian people were indeed correct, then we would expect to see o-marking occurring in a complex structure more frequently than in a simple structure where the object occurs immediately next to the verb. However, this is not the case, and the use of o in simple structures outnumbers that of o used in complex structures. Thirdly, the hypothesis of the progressive spread of the object-marking o predicts that more types of object nominal would be susceptible to marking than those of earlier times, but the data show a contrary tendency; namely, in the Heian period the types of object nominals that take o became narrower. In particular, there is a tendency for the o-marking to be limited to (1) words relating to people, i.e. people’s names and personal pronouns; (2) pronouns; (3) nominalized clauses followed by the “formal” heads koto ‘that’ and yosi ‘(they say) that’, and (4) headless nominalized clauses.
These facts have led Matsuo to conclude that o was not really functioning as the accusative case marker, i.e. as an indicator of object. Although in the case of o used after nominalized clauses some functional explanation is plausible, for complex structures are functioning as objects, the cases involving personal names and pronouns, which generally are more frequent than the other cases, defy such an explanation. Instead, Matsuo suspects the o-marking had some semantic import, such as an expression of subjective feeling of the speaker/writer.
Matsuo’s suspicion has been largely confirmed by two subsequent studies, which undertook careful examination of the literary works of the middle and later Heian periods, in contrast to those from the early period examined by Matsuo. Hiroi (1957) examined the Utsubo monogatari (The Tale of Utsubo) (970‒99) and Oyama (1958) took up the Genji monogatari (The Tale of Genji) (1001‒10). These studies reveal that o is used frequently in: 1) those portions where the influence of Sinico-Japanese (kanbun) materials is obvious; 2) portions where emotional overtones exist; 3) portions depicting formal situations, especially in regard to the court; and 4) portions referring to specific or special persons. In contrast to these uses of o, zero-marking occurs frequently in the sequential descriptions of events and objects, and in sentences describing rapidly unfolding events and excited states. Also conversational materials often show zero-marking.
The influence of the Sinico-Japanese materials had been noticed by others, including Matsuo. In the reading of the materials written in imitation of Chinese syntax, word order was inverted so that the actual reading would follow Japanese syntax, and in the process relation-indicating particles were supplied so as to make the material easier to comprehend. In this Sinico-Japanese reading tradition, o was regularly employed to mark the direct object. Indeed, it is quite reasonable to assume that this use of o in fact had by far the strongest effect in the spread of the object-marking o.
The other contexts in which o-marking occurs indicate that this marking was more than a simple accusative marker; its occurrence had the effect of expressing the speaker/writer’s emotional involvement or attitude in relation to the event described or in reference to the person marked by o. As pointed out by Oyama, this is nothing other than a direct development of the earlier interjectional use of o. Oyama’s conclusion, thus, is that, in Old Japanese and Classical Japanese of at least the early Heian period, o was still a general-purpose interjectional particle, which, because of its frequent association with the object nominal, came to be used as an object-indicating particle in the reading of the Sinico-Japanese materials. During the Heian period, the interjectional o-marking became more specific as a means of expression of the speaker/writer’s emotional involvement in the described event. As time progressed, the object-marking function o in the tradition of Sinico-Japanese materials began to assert itself in other domains of language, leading to the present-day status of o as the accusative case particle.
Matsuo, Hiroi, and Oyama thus argue that the presence or absence of the object-markingo was not a matter of optionality; rather, in Old Japanese there was no accusative case marker, the object being zero-marked, and the “object-marking o” was none other than the interjectional particle o. The establishment of o as an accusative case particle, due to the influence of the Sinico-Japanese reading tradition, took place much later than hitherto assumed.
While the arguments by Matsuo and his followers that o was not an accusative case marker in Old Japanese seem quite convincing, there is in fact another possible interpretation of the development of the accusative o, which has not been entertained in the writings of the traditional grammarians, but which accommodates the two views reviewed above. This interpretation assumes that o was an accusative case particle in Old Japanese, whose application was optional, as it is in the colloquial speech of Modern Japanese. The accusative marking took on semantic significance during the Heian period, when, influenced by the maturity of the court-centred society, the grammar developed systems, such as an elaborate honorific system, for expressing the speaker/writer’s attitude and involvement in the event being described or in reference to the persons referred to in a sentence. Subsequently, the semantic significance attached to o faded away due to the increasing pressure to make its application obligatory, at least in the written language ‒ this pressure being partly caused by the Sinico-Japanese reading tradition, and partly by the grammaticization of other particles.
The observation by Konoshima (1953) that sixty out of seventy uses of o in the songs of the Kojiki were used to mark the object nominal is significant, despite the fact that a larger number of object nominals were left unmarked. When 86 percent of the use of o was associated with the object function, it is hard to imagine that no relation was felt between the form and the grammatical function. It is when such an association is felt that the category status of a form is established ‒ in this case the status of o as marking accusative case. Indeed, if such an association was not felt, we would not expect o to be employed as an object marker in the Sinico-Japanese reading tradition in the first place; otherwise, any interjectional particle could have been adopted for the purpose.
Matsuo’s point concerning the fact that the most progressive aspect of the language, namely colloquial speech, was not affected by the presumed expansion of o-marking misses the point. As noted above, even Modern Japanese allows non-marking of objects in colloquial speech, despite the fact that o-marking is obligatory in the written language and in careful speech, i.e. the formal language. Thus, the expansion and the subsequent obligatoriness of o-marking have affected the domain of the written language more than that of colloquial speech, and this explains why the conservative activity of poetry writing is more associated with o than the colloquial speech of the time.
The narrowing noted by Matsuo of the range of the nominal objects which were marked byo in the early Heian period can be interpreted as an indication that o began to take on semantic significance. That is, marking of the object by o began to assert its etymological meaning associated with the original interjectional use. Thus, only those objects referring to special persons or those involved in expressions with emotional overtones came to be marked by o. The marking or non-marking then became not a simple matter of optional marking. That a case-marking rule can be correlated with some semantic significance is not at all strange in view of the fact that in Altaic languages (e.g. Turkish) and Indo-Iranian languages (e.g. Hindi, Nepali), objects have overt accusative markers only when they are referential, e.g. definite, or animate.
The accusative marking subsequently lost its semantic significance as its use became more and more obligatory, at least in the written language, and, as noted above, this frequent use ofo-marking is very likely to have been caused by the Sinico-Japanese materials. Also, in the early Kamakura period, both no and ga acquired the status of nominative case markers. This grammaticization of no and ga as case markers may have had some influence in making the status of o as an accusative marker a firmer one, eventually leading to obligatory accusative case marking.
The above interpretation of the development of the accusative case marker o assumes that a case-marking particle may take on semantic significance and may shed it along the course of its development. That this is not an unusual development can be seen when compared with the developments of the nominative particle ga and the genitive particle no.
1.10 Epilogue
There are at least two cultural factors that are responsible for making Japanese speech or writing less clear to the ears or the eyes of a non-Japanese than might be the case in a European language. One is the Confucian tradition, which emphasizes deeds over words. Eloquence has not been one of the virtues people have been encouraged to cultivate in Japan. The European tradition of rhetoric as an art of persuasion is not taught in Japanese schools, which emphasize more the learning of the skills that produce artistic effects. In fact, persuasion of others by means of linguistic skills is largely discouraged as direct confrontation in general is avoided. Thus, another cultural factor that affects the form of discourse is the favored pattern of indirect transmission of the intended meaning. In Japan, the art of persuasion takes the form of “beating about the bush”, whereby the listener is expected to make good guesses and to arrive on his own at the conclusion intended by the persuader. It is the person’s ability to arrive at an intended conclusion rather than the persuader’s logical presentation that is evaluated. Thus, one who does not get the point by merely hearing hints is considered a dull person.
Due to avoidance of direct confrontation, which is part of the politeness mechanism that dictates the behaviour of Japanese, straightforward replies are sometimes hard to draw. A famous anecdote among international businessmen refers to the Japanese expression Kangaete okimasu ‘I’ll think it over’, which, when uttered at the conclusion of a business discussion, actually means “No”! Hedges and other ambiguous expressions are also favored, especially among politicians, to avoid any ensuing responsibility that a clear statement may engender. In other words, in Japan there is another kind of rhetoric that is perhaps diametrically opposed to the European rhetorical tradition ‒ the former emphasizes vagueness that invites the addressee to search for answers, whereas the latter emphasizes clarity that leaves no room for the addressee to wonder. In both traditions there is a logic internal to each, but they are not entirely commensurable.
Another factor that contributes to the seeming vagueness of Japanese is the high degree of contextual dependency that Japanese expressions show. A variety of contextual factors dictate the form of a Japanese expression, but the most striking structural feature relevant here is the absence of understood elements from the surface expression. Where many other languages may have pronouns, Japanese simply uses no overt expression. The absence of overt pronominal forms is often compensated for by agreement features in European, especially Romance languages, but Japanese leaves no such trace. Thus, the lack of agreement and the absence of expected pronominal forms are often equated with the absence of the grammatical subject and hence the lack of logic and the obvious vagueness a subjectless structure entails. It is true that the presence of overt pronouns and/or agreement features is helpful in some contexts, for they tell us at least the person and/or gender of the referent. Yet, even when these clues are provided, an independent system for identifying the referent of a pronoun is needed, and such a system seems to differ little from what is needed in a zero-pronoun language like Japanese. Just as pronouns are used where correct identification is possible, a zero pronoun is used only where its antecedent can be recovered. What is not obvious is the mechanism by which the hearer identifies the pronouns, whether they be overt pronouns or zero pronouns. And this is one of the areas where professional linguists have not been able to offer a satisfactory account. Once the mechanism of pronoun identification is explicated, it is expected that Japanese will turn out to be no vaguer than those languages that involve overt pronouns and/or agreement features.
While the difficulty of learning a particular foreign language is a relative matter, depending on the learner’s linguistic background and motivation, as well as the learning environment, there do exist certain grammatical and extra-grammatical features that make Japanese a difficult language for at least some non-Japanese to learn. One feature that immediately comes to mind is the writing system. Learning Chinese characters is a task that all Japanese children must contend with, but is also presents the biggest obstacle for foreign learners other than Chinese and Koreans, whose languages share a large number of characters with Japanese. Mastering all four language skills requires more time in Japanese than other languages due primarily to the presence of Chinese characters. It is commonly said that it takes more than nine years of school education before Japanese children can read a newspaper satisfactorily. Thus, Japanese, perhaps along with Chinese and Korean, is a difficult language with respect to the learning of writing and reading.