Tiun, Hak-khiam. 1998. Writing in Two Scripts：A Case Study of Digraphia in Taiwanese. Written Language and Literacy.1(2)：223-231. The Netherlands : John Benjamins Publishing Company.
Writing in Two Scripts:
A Case Study of Digraphia
Written Language and Literacy Vol. 1(2), 1998, 225-247
John Benjamins Publishing Co.
Three writing systems are currently available for writing Taiwanese, the variety
of Southern Min Chinese which is spoken in Taiwan. Traditionally, it is written either in choan-han ‘all character writing’ or in choan-lo ‘all Roman script’; however, a mixture of these two scripts, called han-lo, has been developed in recent decades. This article evaluates these three writing systems from linguistic and sociolinguistic perspectives. It is argues that han-lo can efficiently achieve the goals of corpus planning: graphization, standardization, and modernization. The educational implications of a mixed writing system are also discussed.
As Cooper 1989 has pointed out, there is a close relation between language planning and social change. Language planning must be understood within a socio-historical context. Social change in Taiwan in the past decade can be characterized in terms of democratization, modernization, and localization (Li 1994). Recent social changes have been translated into language policy in
Written Language and Literacy Vol. 1(2), 1998, 225-247
John Benjamins Publishing Co.
Taiwan. As a response to the people’s call for bilingual education, Minister of Education Guo Weifan has officially announced the inclusion of mother tongue education in the curriculum. While the Taiwanese language gians new status in education, the problems of written Taiwanese need to be addressed by corpus planning.1
There are two serious problems in using 漢 字 han-ji ‘Chinese Characters’ to write Taiwanese. The first problem is the difficulty of graphic representation. Many Taiwanese morphemes simply do not have any characters to represent them — a phenomenon generally called 有 音 無 字 u-im-bo-ji ‘sounds without characters’. Another problem encountered in written Taiwanese is the lack of standardization, especially the chaotic use of han-ji as compared with 羅 馬 字 lo-ma-ji ‘Roman script’.2 This paper will show how han-ji are employed to put Taiwanese into written form, and will discuss problems and limitations of 全 漢 choan-han ‘exclusive use of han-ji’ for written Taiwanese. Furthermore, we will describe and evaluate the other two kinds of writing systems used to write Taiwanese: 全 羅 choan-lo ‘exclusive use of lo-ma-ji’, and 漢 羅 han-lo ‘mixed han-ji and lo-ma-ji’. It is suggested that the han-lo writing system, which incorporates both Chinese characters and Roman script, can efficiently achieve the goal of graphic representation and standardization, and can strike a balance between the need of national tradition and modernization.
Examples of Digraphia
Digraphia can be defined as the use of two or more different writing systems for a single language or varieties of a language (Dale 1980, DeFrancis 1984b). If we look at the development of societies that use han-ji — such as China,
Tai-gi ‘Taiwanese’ is a variety of Southern Min Chinese, spoken in Taiwan. Speakers number about 16 million, i.e. 73.3% of the Taiwan population. Other terms for Taiwanese are Tai-oan-oe, Ban-lam-gi ‘Southern Min’, and Ho-lo language’.
The best-known and established lo-ma-ji system for Taiwanese is called 白 話 字 peh-oe-ji ‘vernacular writing system’ or 教 會 羅 馬 字 kau-hoe lo-ma-ji ‘Church Romanization’; it is used in this paper and is referred to simply as lo-ma-ji (for equivalences with IPA, see Appendix). The Romanization system used here for Mandarin is based on the pinyin system, without tonal indication. Abbreviations used in this paper include: A: adjective, N: noun, V: verb, ASP: aspect marker, NOM: nominalizer, PREP: preposition, POSS: possessive marker.
Hong kong, Japan, Korea, and Vietnam — digraphia seems to be the rule rather than the exception. Generally speaking, two types of scripts have been developed in these areas: han-ji and non-han-ji phonetic symbols (han’gul, kana, pinyin, Roman script), yielding three kinds of writing systems: exclusive use of han-ji (as in Chinese), exclusive use of non-han-ji phonetic symbols (as in the case of Vietnamese Quoc Ngu’ and Han gul in North Korea), and the mixture of two scripts (in Japan and Southe Korea). The general trend is toward the addition of a non-han-ji phonetic writing system, either used separately or mixed with han-ji.
Problems in Writing Taiwanese
A salient characteristic of written Taiwanese is inconsistency in the use of han-ji. Often a single morpheme is written with several different characters, even by a single writer in a single text. The linguistic reasons for these inconsistencies have much to do with the nature of Taiwanese and the characteristics of han-ji.
Taiwanese and Mandarin are different but related languages, differing in phonological, morphological, and syntactic features. Phonological and morphological differences are the two main causes of the difficulty of writing Taiwanese in Chinese characters. Compared with both classical Chinese and modern Mandarin, Taiwanese has many of the following phonological and morphological characteristics (R. L. Cheng 1990:222-24):
Preservation of Ancient Chinese morphemes.
Characters with distinct colloquial vs. literary readings.
Taiwanese morphemes without standardized characters.
Japanese and English loans, with most of the English loans being borrowed via Japanese.
Loans which are written with Japanese characters, but have Taiwanese pronunciations.
It is reported by R. L. Cheng 1987 that content words of Taiwanese and Mandarin tend to share a higher ratio of the same etymons (84%) than do the function words (49.5%). In addition, S. Chen 1989 has found that function words are the least standardized category. She compared two dictionaries and found that, of the 104 most frequently used function words,
62.5% were written differently. Function words, therefore, should receive special attention.
According to R. L. Cheng 1989, only 5% of Taiwanese morphemes in the total lexicon lack suitable or standardized characters to represent them. However, since most of them are words of high frequency, they account for 15% of the total number of characters in running texts.3
Orthographically speaking, the multiplicity of principles uses in character formation and adoption is the main reason for the chaotic situation in using characters to represent Taiwanese special morphemes (TSM). According to R. L. Cheng 1989, the methods used to create new characters to represent TSM include 形 聲 xingsheng ‘phonetic compunding’ and 會 意 huiyi ‘semantic aggregation’. Besides these, there are three principles for adapting old characters to represent TSM: 假 借 jiajie ‘phonetic borrowing’, 轉 注 zhuanzhu ‘similarity of sound and meaning’ (in Taiwanese), and 訓 用xunyong ‘similarity of meaning’ (in Mandarin or classical Chinese). All these principles cause uncertainty in the representation of TSM with characters.
The choice of characters of TSM is also conditioned by socio-educational factors (R. L. Cheng 1989). In general, language specialists tend to use xingsheng or zhuanzhu characters; but highly educated non-language specialists tend touse xunyong; and the general population tends to use jiajie characters. Since different principles are used in writing TSM by writers of different socio-educational background, the chaotic situation appears inevitable.
4. History of Written Taiwanese Taiwanese provides an excellent example of digraphia in actual use. There are three kinds of written Taiwanese: exclusive han-ji, exclusive lo-ma-ji, and mixed han-lo.
See R. L. Cheng (1989:325) for an estimate of the stability of han-ji by part of
speech. C. Zhou 1993 lists 453 Southern Min morphemes that have no corresponding han-ji. The amount of these so-called “no-character morphemes” would be greater if we included the many cased of onomatopoeia, loans, and contractions existing in Taiwanese.
All lo-ma-ji text:
Lan beh u chun-giam, lan e gi-gian ai siu chun-tiong; lan khoaN-tang lan e gi-gian, lan choe chio ai ti lan a hak-hau thoan hou lan e te ji tai, ai iong I lai sia chhut lan e sim-siaN.
‘To have dignity our language must be respected. To show our respect to our language, we must teach our mother tongue to our children at school and learn to express ourselves through written Taiwanese.’ (R. L. Cheng 1993:11).
Character writing of Taiwanese has the longest history — about 400 years —
As well as the greatest amount of publication. According to van der loon (19924-8), the extant Min-nan literature (after the Ming period) included stage plays, string-puppet plays, glove-puppet plays, shadow-figure plays, and ballads. During Japanese colonization, many Taiwanese textbooks were written in han-ji with kana annotation and Japanese translation. In the 1930s, a cultural movement called the Tai-oan-oe-bun un-tong ‘Written Taiwanese Movement’ proposed the use of Taiwanese as a literary language. Aside from discussions about the problems of writing the language, efforts were made to collect and transcribe Taiwanese folksongs, nursery rhymes, folktales, jokes, riddles, and proverbs. Although the main contribution of this period was the transcription of oral literature, new genres of Taiwanese literary creation also began at this time, including novels, poems, and prose. Unfortunately, the Written Taiwanese Movement ended at the outbreak of the Sino-Japanese war. Despite the theoretical discussions and experiments of the pre-war period, the problems of writing Taiwanese in characters remained unsolved.
Writing in lo-ma-ji has about 150 years of history. Taiwanese lo-ma-ji originated in the church, and was mostly used for Christian writing. However, it also included many non-religious publications, including textbooks of nursing, translations of the Chinese classics, Taiwanese newspapers, textbooks for learning Taiwanese, and dictionaries. In fact, the first newspaper published in Taiwan, Tai-oan hu-siaN Kau-hoe-po ‘Taiwan Prefectural City Church News’ (1885-1942), was written totally in lo-ma-ji.4 In 1922, Chhoa-phoe-hoe
4. The name of this newspaper has been changed several times. In 1905, it was changed to Tai-lam Kau-hoe-po, and in 1913 to Tai-oan Kau-hoe-po. Finally, in 1932 it was renamed Tai-oan Kau-hoe-kong-po.
Initiated a lo-ma-ji movement, the purpose of which was to promote Taiwanese culture and education, and his proposal was adopted by the Tai-oan bun-hoa hiap-hoe ‘Taiwanese Culture Society’. However, the lo-ma-ji movement was banned by the Japanese government. From 1989 to 1992, a lo-ma-ji magazine, Hong-hiong ‘Direction’, was published. A Taiwanese discussion group now also uses lo-ma-ji in the e-mail environment. However, exclusive use of lo-ma-ji only appears sporadically in current written Taiwanese. In most cases, lo-ma-ji are used as phonetic annotations for characters, functioning something like Japanese furigana ‘lateral kana’; or they are used to replace some characters, forming the mixed system called han-lo.
Han-lo appeared as a new writing system about three decades ago. The first proposal for it was published by Ong Iok-tek in a series of lectures entitled Tai-oan-oe kang-cho ‘Lectures on Taiwanese’, which appeared in Japan in 1964, in the magazine Tai-oan chheng-lian ‘Young Taiwanese Magazine’ (cf. Ong 1993). However, Ong did not write in this system. The first experiment in han-lo writing did not appear until 1967 in the same magazine. The second in han-lo writing movement was advanced by the Tai-oan gi-bun goeh-khan ‘Taiwanese lanugage monthly’, published in America in 1977. The idea of han-lo writing was first introduced to Taiwan by TeN Liong-ui, through the publications of this system and the exposition of its theory in the late 1980s. Through TeN’s advocacy, han-lo writing enjoys great currency in contemporary written Taiwanese; it is used in the writing of poems, novels, and prose, as well as in academic writings, Taiwanese language textbooks, and religious works. It appears in newspapers, bulletins, and books. Tai-bun thong-sin ‘Taiwanese writing forum’ is one of the leading publications advocating the use of han-lo. It is the writing system preferred by most of the advocates of written Taiwanese (e.g. Xu 1992, Ong 1993, R. L. Chneg 1989, 1990), and it has also gianed support from experts in information processing. Judging from its vitality, it is likely the han-lo writing will become the mainstream system for writing Taiwanese.
5. Three kinds of Written Taiwanese
The following sections will describe and evaluate the above three kinds of written Taiwanese: exclusive use of han-ji; exclusive use of lo-ma-ji, and mixed use of the two scripts.
5.1. Exclusive Use of Han-ji Characters used to represent TSM fall into four categories: 訓讀字 hun-thok-ji ‘semantic borrowing characters’, 本字 pun-ji ‘etymological characters’, 假 借字 ka-chioh-ji ‘phonetic borrowing characters’, and 本土字 pun-thou-ji ‘dometic characters’. The overall trend of han-ji used, in descending order, is hun-thok-ji ＞pun-ji ＞ ka-chioh-ji ＞ pun-thou-ji (Huang 1993:383).
5.1.1. Hun-thok-ji ‘Semantic Borrowing Characters’ Hun-thok-ji refers to han-ji that are borrowed into written Taiwanese for their semantic values, but not for their phonetic forms. For example, beh ‘want’ can be written as 要, borrowing its Mandarin meaning; the Taiwanese reading iau is irrelevant here. Moreover, beh can also be written with the classical character 欲, pronounced iok in Taiwanese. Hun-thok-ji have been most widely used in written Taiwanese in both the Japanese and contemporary periods.
Although hun-thok-ji have uses in the graphic representation of Taiwanese, there are some undesirable results. First, there is a chaotic use of han-ji: Because of synonyms in the source language, different writers may employ different characters to write a single Taiwanese morpheme. For example, hou ‘passive marker’ is written with three different characters in Tai-jit Toa-su-tian (Taiwanese-Japanese dictionary, 1931): 讓,使,and令. Hou can also be written with 給,被, or 俾. Another case of hun-thok-ji which causes confusion is the use of the same character to write more than one Taiwanese morpheme. For example, 給 is borrowed to write both ka ‘disposal marker’ and hou ‘passive’. Furthermore, different generations of writers draw on different source languages — Classical Chinese, Japanese, and modern Mandarin — causing inconsistency in use (R. L. Cheng 1989). So, even using the same hun-thok principle, the use of han-ji still shows inconsistency, and is sometimes unintelligible because of generational gaps.
Second, the hun-thok principle causes a great divergence between the written and spoken languages. Since only semantic value is borrowed, hun-thok-ji cannot show the exact pronunciation of Taiwanese morphemes. For example, if we say ou-to-bai ‘automobile’ and write 摩托車, the reader has no way of knowing the intended pronunciation.
Third, hun-thok-ji sometimes cannot express the exact meaning of Taiwanese morphemes. For example, chhu ‘house’ and tau ‘home’ become indistinguishable when we use the hun-thok-ji 家.
5.1.2. Pun-ji ‘Ethmological Characters’
The frequent use of pun-ji reflects the fact that Mandarin and Taiwanese share many cognate morphemes. However, it is misleading to say that every Taiwanese morpheme has a pun-ji.
First, even though certain characters have been proven to be pun-ji, they are not necessarily preferred by the public in general practice. For example, instead of using the etymological character 農 for lang ‘people’, the semantic borrowing 人 is generally used in written Taiwanese. Second, many so-called pun-ji have low frequency, e.g. 骹 kha ‘foot’ and蠓 bang ‘mosquito’. Third, because of historical development, the semantic and phonetic values of some characters have changed significantly, so that the use of some pun-ji in modern times has become impractical and unintelligible. Finally, the fact that different scholars may find different cognates for a single Taiwanese morpheme makes us question the capacity of pun-ji to solve the problem of standardization. The etymological study of characters is an interesting research topic; however, we should be careful not to confuse scholastic pursuits with practical use.
If writing is to represent visually the speech sounds of Taiwanese, then ka-chioh-ji could be ideal for written Taiwanese. The question that arises is: Can ka-chioh-ji be effectively used to represent Taiwanese morphemes without causing misunderstanding and mispronunciation? Unfortunately, the use of ka-chioh-ji seems to have been ineffective, ambiguous, and confusing. One interesting phenomenon is that, besides borrowing the han-ji reading in Taiwanese, modern ka-chioh-ji also employ the phonetic value of Mandarin; e.g., the Mandarin morpheme 贏 ying ‘win’ is borrowed to write Taiwanese eng ‘spare time’.
As Defrancis (1984a:147) correctly points out, han-ji are “morphosyllabic” rather than “ideographic”. That is, Chinese characters use syllable graphs to represent sounds. However, Chinese syllabic writing does not provide reliable information in representing the pronunciation of characters. According to Huang (1993:387), the effectiveness of han-ji in represnting sound is only about 35%. Indeed, to enhance the ability to represent sound has been the main reason for using lo-ma-ji, which has an almost perfect correspondence between graph and sound.
Since there is no consensus in choosing among phonetic alternatives, any choice can only be arbitrary and unsystematic. Thus the passive marker hou can be written with the following homophones: 戶 ‘family’, 互 ‘each
Other’, 雨 ‘rain’. Hence, even using the same ka-chioh principle, different writers may employ different character. Another problem is the interference of the original meaning in phonetic borrowings. In using ka-chioh-ji, only the phonetic value of the character is borrowed; but unfortunately, the interference of the original meaning of the character seems to be inevitable, thus causing confusion and ambiguity. For instance, chhit-tho ‘to play’ was written 七 桃 in a traditional play; however, this writing can also be interpreted as ‘seven peaches’ in Taiwanese. This interference may be one reason why semantic borrowings are preferred to phonetic borrowings.
5.1.4. Pun-thou-ji ‘Domestic Characters’
This class of characters has been created especially for written Taiwanese. The creation of such new characters follows the principles of xingsheng ‘phonetic compounding’ and huiyi ‘semantic aggregation’. An example of xingsheng is the character 睭, as in 目 睭 bak-chiu ‘eye’, with 目 bak ‘eye’ as its semantic element and 周 chiu as the phonetic element. An example of huiyi is 身長 lo ‘tall’, which combines 身 ‘body’ and 長 ‘long’.
Pun-thou-ji are “homemade” characters, and so give us some sense of authenticity. However, they are not widely used at present in written Taiwanese. Their limited use has much to do with the difficulties they cause in word processing: Many pun-thou-ji simply cannot be found in the computer. Furthermore, the use of pun-thou-ji creates a learning burden for students. Finally, many Taiwanese morphemes simply lack any character to represent them — whether pun-thou-ji, or other characters used according to the principles mentioned above.
In sum, Chinese are useful in writing shared morphemes between Mandarin and Taiwanese; however, they are ill-suited to writing Taiwanese special morphemes. The multiplicity of principles governing character formation and adoption had caused standardization based on han-ji to be a formidable, if not impossible task.
5.2 Use of Lo-ma-ji P. Chen (1994, 1996) identifies four possible functions for a new script to perform in relation to Chinese characters: auxiliary, supplementary, alternative, and superseding. What follows is a description of these as they may be performed by lo-ma-ji in writing Taiwanese.
5.2.1 Auxiliary Fuction
Lo-ma-ji can serve to annotate the sound of characters; i.e., they can be used as auxiliary phonetic symbols, rather than as an independent writing system. Many all-character texts make sporadic use of lo-ma-ji for sound annotation; others use zhuyin fuhao, a set of simplified characters used for sound annotation, and taught in primary school as an aid to learning Mandarin. However, because of the phonological differences between Mandarin and Taiwanese, modifications and new symbols have to be added when using zhuyin fuhao for Taiwanese.
5.2.2. Supplementary Funcation
Lo-ma-ji can be used as a supplementary script in two ways. One way is to place it above or on the side of every character in the text. Textbooks used in mother-tongue education or readers intended for children generally use lo-ma-ji in this way.
(4) Hian-chai goa hoa-gi e-hiau kong, tai-gi ma e thong.
現 在 我 華語 會 曉 講 台 語 嘛會 通
‘Now I can speak Mandarin as well as Taiwanese.’(Gou 1993:343)
Another type of supplementary function can be found in the mixed use of lo-ma-ji with han-ji; this is the han-lo system described in §5-3, below.
5.2.3. Alternative Function
Some Taiwanese texts consist of two parts: han-ji and lo-ma-ji. Here the latter is treated as a writing system parallel with han-ji. These texts can be regarded as biliterate written Taiwanese; readers who are literate in either writing system can have access to the text. There are different arrangements of the han-ji and lo-ma-jiparts. Generally, han-ji are placed on the left side and lo-ma-ji on the right side, or vice versa. The following excerpt from a Taiwanese poem demonstrates this type of written Taiwanese:
‘If you remember / please tell me / all the world’s happiness /
beautiful as the colorful sunset’ (Tan Lui 1997).
5.2.4. Superseding Function
A clearer example of the autonomous function of lo-ma-ji can be found in
Texts written solely in that system. Ex. 6 is an excerpt from a paper discussing the advantage of phe-oe-ji (Roman script):
Peh-oe-ji ia si bun-ji-tiong e chit lui, I e li –ek lang, pi pat-ho e ji iau u khah-kin khah-khoai …
‘Peh-oe-ji is also one of the writing systems. People can benefit from the use of Peh-oe-ji. It can be learned and used faster than other kinds of writing systems …’
(from Lau Bou-chheng, Peh-oe-ji e li ek ‘The advantage of Roman script’, Tai-oan-hu-siaN Kau-hoe-po ‘Taiwan Prefectural City Church News’, 1886).
5.2.5. Lo-ma-ji: Autonomous or Superseding Script?
So-called u-im-boji ‘sounds without characters’ constitute a serious problem in the use of characters. Texts written in lo-ma-ji have no such problems. As Ong points out (1993:31), “it is only through lo-ma-ji that Taiwanese can be fully expressed.” The advantages of lo-ma-ji over han-ji can be shown by the following comparisons. First, lo-ma-ji, unlike han-ji, have an almost perfect correspondence between symbol and sound, and thus can achieve graphization ver easily. Furthermore, though there are various lo-ma-ji systems, standardization based on lo-ma-ji can be much easier achieved than by han-ji. Third, the small number of graphs (18 letters) makes it easier to learn and process into computers.
However, despite such advantages, it is not feasible at present to adopt lo-ma-ji as a superseding system. If we consider the sociolinguistic situation of Taiwan, it is apparent that the society as a whole is not prepared for such a drastic change. The lifelong habit of using han-ji simply cannot be changed overnight. Since characters are still the main medium of writing and are widely known through compulsory education, complete replacement of han-ji by lo-ma-ji would not be likely to succeed.
Instead, complementary relationships between han-ji and lo-ma-ji offer a more favorable solution. As shown in the first issue of Tai-oan hu-siaN kau-hoe-po ‘Taiwan prefectural City Church News’ (1885), the introduction of lo-ma-ji was intended to add an easier and more efficient writing system, rather than to replace characters. That is , the idea of digraphia was encouraged, with lo-ma-ji being given the highest priority in teaching first literacy; those who were literate in characters were also encouraged to learn lo-ma-ji. Since lo-ma-ji was mainly related to church practices and religious publications, its reception was quite successful within the church. In the past, however, lo-ma-ji was not
Well received by the general public on the ground that it was a foreign writing system, learned and used only by Christians and the illiterate (IuN 1993:61). Another hindrance to the development of lo-ma-ji cam from the governmental prohibition: Since it was feared that the use of a Romanized written form might inspire separatism, lo-ma-ji as made illegal in 1969. These socio-political factors have been largely responsible for the invisibility of lo-ma-ji as an autonomous writing system.
However, the following sociolinguistic changes might be conducive to the reception of lo-ma-ji as an autonomous writing system in Taiwan. First, because of cultural and languistic contact with Western societies, Roman script is now frequently used in everyday literacy practice. Second, since Taiwan is becoming a pluralistic society, governmental and popular attitudes toward this imported script are likely to change from xenophobic sentiment to a more favorable attitude. Third, the importance of lo-ma-ji in meeting the needs of modernization, especially in the information age, is gaining more recognition. The reception of lo-ma-ji
Depends very much on the extent to which it is seen as an addition to linguistic capital and a technical improvement, rather than as a detriment to social and cultural integrity.
What functions can lo-ma-ji perform? In some domains, it may be preferable to adopt its exclusive use. First, it can provide children and adult illiterates with easy access to initial literacy, and will also help them to learn characters as second literacy. Second, for foreigners who do not know characters and whose main purpose in learning Taiwanese is to communicate orally, lo-ma-ji-only textbooks are preferable for reasons of efficiency. Third, with the coming of the information age, lo-ma-ji also takes on a new function, namely its use in Taiwanese information processors: TW301, Hotsys, and Dai-im. All of them use lo-ma-ji as their input method. Compared to shape-based input systems, lo-ma-ji are very convenient for this purpose.
Although writing in lo-ma-ji is easy and fast, reading texts written exclusively in it is generally felt to be difficult. This may be because lo-ma-ji reading materials are not widely available; thus most people do not have enough practice in reading such texts. However, using lo-ma-ji in international communication, like electronic mail, is convenient; for example, it is the most frequently used writing system in the Tai-gi-ban ‘Taiwanese discussion group’ (email@example.com).
In summary, political and cultural conditions are still not favorable for further development of lo-ma-ji as an autonomous writing system. At present, such an independent system can only have limited use. Perhaps a moderate
Apporach, nmaely the supplementary schme, may serve as a starting point toward a more radical reform of written Taiwanese.
5.3. Mixture of Han-ji and Lo-ma-ji The difficulties of using han-ji in representing Taiwanese special morphemes have pushed people to introduce the lo-ma-ji writing system in s supplementary function. In this system, morphemes shared between Mandarin and Taiwanese continue to be written in han-ji, while lo-ma-ji are used to replace unstandardized han-ji. This system is similar to the combination of kana with kanji in Japan, or the mixture of han gul and hanja in South Korea. Such scripts can be termed hybrid or mixed writing systems.
The combination of two traditions, han-ji and lo-ma-ji, has yielded a writing system which not only recognizes the tradition of digraphia in written Taiwanese, but also overcomes the shortcomings of its two components. The result is a writing system that is easy to learn, easy to write, easy to read, easy to standardize, and easy to process in the computer age.
Generally speaking, lo-ma-ji are used whenever the use of han-ji is inaccurate (in terms of phonetic or semantic value), unsuitable (in terms of typographic considerations, such as low frequency, or complex shape and multiple stokes), or unstandardized (when more than one han-ji exists for the same morpheme). The use of lo-ma-ji falls into four categories: grammatical function words, loan words and English phrases and sentences, onomatopoeia and contractions, and native contexnt words.
5.3.1. Grammatical Function Words
Function words play an important role in reading comprehension; they are also words of high frequency. Unfortunately, their use in han-ji is also the least standardized in written Taiwanese, so function words are generally written in lo-ma-ji. Exx. 7-8 demonstrate the use of lo-ma-ji to represent Taiwanese function words (with glosses) in a han-lo text.5
5. Chaotic use of han-ji, especially in function words, is the rule rather than exception (for a list of function words used by different authors and dictionaries, see R. L. Cheng 1989: 383-406).
(7) An 頂高傳落來e kan-taN是hou 人聽著會起雞母皮e
from NOM ADV PASS POSS
Kama-kahi teh 奸臣笑
‘From above cam the eerie sound of Kama-kahi’s mocking
(8) 伊 ti 恁學校teh教台語
‘He is teaching Taiwanese at your school.’(Cheng et al. 1990:28)
Two lo-ma-ji are used in ex. 8: ti ‘at’ can be written as 滯, 佇, or 在, while teh ‘progressive aspect’ can be written as 值, 在, or塊. It is obvious that the use of lo-ma-ji is already more standardized than the use of han-ji in Taiwanese; it thus facilitates reading and writing. Other frequently used function words that are written in lo-ma-ji included in ‘third person plural’, kap ‘and’, e ‘nominalizer or possessive’, chia ‘here’, hia ‘there’, ka ‘disposal marker’, hou ‘passive’, kah ‘extent marker’, kam ‘interrrogative particle’, koh ‘again’, nia-nia ‘only’, and sentence-final particles.
5.3.2. Loan Words, English Phrases and Sentences
One characteristic of Taiwanese morphemes is the abudance of English and Japanese loans which have been assimilated. They play an important role in everyday life, especially in working language. Solving the problem of writing loans is of paramount importance, both from educational and economic/ technical viewpoints.
Generaaly speaking, loans can be adapted to Taiwanese either by translation or by transliteration with characters. However, both translation and transliteration present problems of standardization, because of the use of characters as a writing system.6 Again, the most efficient way to solve the problem is the adoption of lo-ma-ji to facilitate rapid standardization.
Lo-ma-ji are used to write loan words in at least two ways. One is to use the original alphabet, as in the following example (loans are underlined);
6. For the discussion of the problems in using characters as a medium of borrowing, see Novotna 1967, R. L. Cheng 1985.
‘This type of tale is found in the folkore of many lands, but certain features of Aukele are uniquely Plynesian.’ (Taibunnun 1997:iii)
This kind of borrowing is good for international coomunication, because it uses a famliliar writing system. Another way in to transliterate into lo-ma-ji, conforming to Taiwanese pronunciation. For instance, ‘handle’ can be written as han-to-luh in Taiwanese (an English loan via Japanese). Other examples are ma-la-song ‘marathon’, siat-ta ‘shutter’, Khiu-ba ‘Cuba’. Such transliteration has the advantage of representing the sounds actually used, and it is widely used in writing Japanese loans.
Some further examples of loans written in lo-ma-ji, collected from the Taiwanese Writing Forum 1993 and recent publications of the 5% Translation Project in Taiwanese Languages, are ba-su ‘bus’, gu-lin ma-khi-tin ‘green marketing’, bu-lo-ka ‘broker’, Bian-lu ‘Benz’, ou-ji-sang ‘old gentleman’ (from Japanese), ne-ku-tai ‘necktie’, la-ji-oh ‘radio’. Some loans are directly borrowed with Taiwanese pronunciation attached, such as pie (phai), Canada (Kha-na-ta), Columbian (Kho-lam-bi-an). Comapared with the translation approach employed by Mandarin, the Taiwanese way of borrowing by transliteration has the advantages of facilitating borrowing and enhancing international cultural exchange.
Lo-ma-ji are also used to write English phrases and sentences. This is generally done for humorous effect, as in the following example.
是新的台灣神學理論。因為台灣牧師及長執若用英語司會，十個九個攏是ka會眾講: Let’s sin to praise God!
‘Is “Let’s sin to praise God” a new Taiwanese theological theory? Because every time the Taiwanese pastor or elder is reponsible for the English church service, nine times out of ten, he will say to the church goers: “Let’s sin to praise God.”’ (S. Cheng 1993:138)
5.3.3. Onomatopoeia and Contractions
Chinese characters are a poor tool for indicating sounds. Thus it is extremely difficult, if not impossible, to write down onomatopoeia or contractions, which are abundant in everyday speech. Some onomatopoeia used in written Taiwanese are ang-kong-kong ‘deeper red’, chheN-leng-leng ‘green’, sio-thng-thng
Contactions are another category which han-ji fail to express; examples are siang (contraction of siaN-lang ‘who’), loai (contraction of loh-lai ‘come down’), and chang (contraction of cha-hng ‘yesterday’). Such words are frequent in colloquial Taiwanese, but are seldom written down. The use of lo-ma-ji can easily solve this problem.
5.3.4 Native Content Words
Native content word are special Taiwanese morphemes which do not belong to the above three categories, but are written in lo-ma-ji because their han-ji counterparts are inappropriate, inaccurate, or unstandardized. These content words can be classified by parts of speech: noun (N), adjective (A), and verb (V). Some examples of their use in written Taiwanese are the following (native content words are glossed).
(11) toe ti 伊 kha-chhng 後面 peh 起來。有千千萬萬人
V:follow N:buttocks V:climb up
‘There are millions of people climbing after him.’ (Taiwanese Writing
Forum 29:1, 1994)
(12) 石頭仔 khng ti lak-te-a 內。歌利亞 chong 來e時…
V:place N:pocket V:rush about
‘He placed some stones in his pocket. When Ko-li-a rushed about…’
(Taiwanese Writing Forum 29:2,1994)
(13) 路枯x掠hou-liu…囝仔做伙拍kan-lok…溪溝hou 魚
N:loach N:spinning top V:scoop
‘Catching loaches in the mud … Children played spinning tops
together … scooped fish in the river …’ (Taiwanese Writing Forum
sng 數字 e GAME … ngaiio ngaiio …
‘Play the game of number … However, everyone seems to feel
uncomfortalbe about this game …’ (Loonng 1993:7)
The advantages of using lo-ma-ji in these cases are threefold. First, standardization can be easily achieved. Second, the burden of learning han-ji can be lessened. Third, romanization can solve the problems of word processing
caused by the use of rare words or “homemade” characters which are not available in the character coding systems.
5.4. Principles of Han-lo Writing To provide more specific principles for han-lo writing, R. L. Cheng (1990:228-29) suggests appropriate situations for using han-ji and lo-ma-ji, respectively. Following are the situations in which he suggests using han-ji:
Characters that have difinite etymological rigins.
Characters that are highly standardized in written Taiwanese.
Characters that are shared with Mandarin and Hakka.
Characters that have simple shape and few strokes.
Characters that have already been learned by the students.
For the following situations, lo-ma-ji are recommended:
(a) Characters that have unclear etymological origins.
Characters that are not yet standardized in written Taiwanese.
Characters that are not shared with Mandarin and Hakka.
Characters that complex and rarely used.
Characters that phonetic borrowings.
Characters that are likely to be mispronounced and misunderstood.
Transliterations of foreign place names or personal names.
Characters that have not been learned by the students.
Of course, if we follow these principles completely, the number of lo-ma-ji would be great indeed. However, the choice between han-ji and lo-ma-ji is quite flexible: Writers can make their own decisions, according to their familiarity with han-ji and lo-ma-ji, or based on considerations of the intended readership and the topics concerned. This flexibility of using han-ji vs. lo-ma-ji is similar to the use of kanji vs. kana in writing Japanese. At present, about 10-15% of lo-ma-ji are use in han-lo running text. In Japanese mixed writing, kana and kanji occur in the ratio of 5 to 3; i.e., kana comprises 62.5% of the text (Liu 1969:50). What is important is that the writer be consistent in using han-ji or lo-ma-ji within a single text to facilitate understanding by the reader.
6. Educational Implications of Digraphia
The best medium for teaching children is their mother tongue (UNESCO 1953). One advantage of mother tongue education is that it allows children to learn to read and write the way they speak. However, the promise of reading and writing efficiently in one’s mother tongue is seriously damaged by using only Chinese characters, which are often characterized as being difficult to learn and inconvenient to use. As shown above, han-lo, which combines Chinese characters and Roman script, can help to reduce the difficulties posed by the exclusive use of characters. Digraphic writing, han-lo, thus has an important educational implication, namely empowering students to read and write the way they speak.
To be sure, Taiwan has been touted as a success story in achieving literacy with characters. However, this has not been done without cost; one need only recall how much time and energy students have to spend in learning characters. It is estimated that about one-third of all class hours in China are spent in learning the Chinese language, where much of the time is used for character learning (Le Page 1992:123, P. Chen 1996:10). Table 1 compares the number of characters that Taiwanese and Japanese students have to learn in primary school.
These data show that Taiwanese primary school students have to learn 2.6 times as many characters as Japanese students. The huge inventory of characters has posed a formidable obstacle to the learner, whether in reading or writing. The burden of learning characters and the insistence on a
Table 1. Characters that Primary School Students Have to Learn in Taiwan and Japan
(Taiwan data from R. L. Cheng 1994, Japan data from Y. Zhou 1992:134).
characters-only approach influence the linguistic and intellectual development of students. Introducing a phonetic writing system to supplement characters can significantly reduce the burden of character learning and leave more time for other subjects.7
The inefficiency of using character as the only writing system can be shown by the following statistical data concerning the frequency of characters. According to Lin (1980:136, 138), of the most common 6359 characters, 1000 characters account for 90% of character occurrences, 2500 characters for 99% of character occurrence, and 3800 characters for 99.9% of occurrences. Most significant is the fact that 2555 characters account for the remaining 0.1%. these data show that the use of characters as the only medium of literacy can place a tremendous learning burden on students of han-ji. A han-ji-only approach, if adopted for Taiwanese, can only increase the burden of han-ji learning, since many Taiwanese morphemes are not shared with Mandarin. In order to learn to read Taiwanese in han-ji, according to R. L. Cheng 1994, a learner who already knows 3000 characters needs to learn 500-1500 characters ore, while one who already knows 5000 characters needs to learn 300-1000 more; but to read han-lo, both learners need to learn only 0-20 new graphic symbols.
From the above comparison, it is clear that the use of han-lo has a distinct advantage over han-ji-only text. In a han-lo text, learners of Taiwanese with knowledge of 3000 or 5000 characters need only add some 20 Roman letters. By contrast, a characters-only text may require learning hundreds or thousands of additional characters. If we compare the learning of a small number of roman letters with that of learning thousands of characters, it is clear that han-lo helps to reduce the learning burden, and makes the reading and writing Taiwanese more efficient.
As Y. Zhou has pointed out (1992:222), digraphic writing, especially the use of kana, has helped Japanese children acquire more knowledge and develop their language skills faster than Chinese children. An education experiment in Heilongjiang (in the People’s Republic of China) has tested precisely the hypothesis that pinyin can function like kana, and can increase the quality of education without placing too much burden on students. The
According to Y. Zhou (1980:92), the number of words used in primary school textbooks in the PRC (up to fourth grade) is 160,000; while in the USSR, up to the same grade, a total of 920,000 were used. That is, PRC textbooks use only 17% of the number of words used in the USSR.
experiment — called zhuyin shizi tiqian duxie (ZSTD for short, ‘phonetic annotation for recognition of characters in order to speed up reading and writing’) — proved successful (DeFrancis 1995:28). The main idea of ZSTD is what can be called an educational policy of “walking on two legs”: a policy of digraphia, with the goal of consolidating a command of pinyin. That is, at the beginning stage, emphasis is placed on the learning of pinyin; then characters are gradually introduced.
This educational reform has changed the previous emphasis from the learning of characters to the development of language skills. The delay caused by the characters is removed by the use of pinyin; thus children can learn to write whatever they can say. Furthermore, through extensive writing and reading, they also learn to read and write more characters than the comparison group. It is clear that the difficulties of using characters to write Taiwanese can be solved in a similar fashion. The adoption of digraphic writing system has educational benefits.
7. Conclusion: Toward a Digraphic Taiwanese
Since written Taiwanese has not been standardized, efforts should be mad to create norms for the language. As Cooper points out (1989:122), the language planner’s job is to create corpora that “reflect an indigenous or a classical tradition, or the values of modernity, efficiency, transparency, and the like.” How to choose among these complex and sometimes conflicting values is one of the main problems a language planner has to face.
As we have seen in the previous discussions, a combination of characters and Roman script help to create a writing system that retains the value of tradition while meeting the needs of language adaptation, in which writing systems evolve to meet the needs of the people using them. Both Japan and South Korea have succeeded in adapting characters to write their languages with the additional aid of non-han-ji phonetic symbols. Taiwanese can benefit from the experiences of corpus planning in these two countries. The adoption of digraphic writing can not only solve the problems of written Taiwanese, but can also contribute greatly to educational, cultural, technical, and economic development in Taiwan.
Vernacular languages in Taiwan have suffered from the Mandarin-only policy, and bilingual education has been suggested as a solution to the problems caused by the monolingual policy. Similarly, the problems of
written Taiwanese can be solved by digraphia. Institutionalized bilingualism and digraphia are necessary to achieve the goals of democratization, modernization, and localization of the Taiwan languages.
This is a revised version of a paper read at the Fourth Annual East-West Centerwide Conference January 23-27, 1995 East-West Center, Honolulu, Hawaii; the earlier form was entitled “Digraphia in Taiwan”. I am grateful to Carol Eastman, Robert L. Cheng, Ying-che Li, and Michael Forman for their comments on earlier versions of this paper, and to Jeffrey Hayden and Debbie Hua for editing help.
Dept. of Language and Literature Education
National Taitung Teachers Collage
Taitung 950, Taiwan
Chen, Ping. 1994. Four projected functions of new writing systems for Chinese. Anthropological Linguistics 36:366-81.
——. 1996. Toward a phonographic writing system of Chinese: A case study in writing reform. International Journal of the Sociology of Language 122:1-46.
Chen, Suyue. 1989. A study of written Taiwanese. M.A. thesis, Fu-Jen Catholic University.
Cheng, Robert L. 1985. Group interest in treating words borrowed into Mandarin and Taiwanese. Anthropological Linguistics 27:177-89.
——. 1987. Borrowing and internal development in lexical change: A comparison of Taiwanese words and their Mandarin equivalents. Journal of Chinese Linguistics 15:1.105-32.
——. 1989. Chau-hiong piau-chun-hoa e Tai-oan-oe-bun [Essays on written Taiwanese]. Tai-pak: Chu-lip chuut-pan-sia.
——. 1990. Ian-pian-tiong e Tai-oan sia-hoe gi-bun [Essays on sociolinguistic problems of Taiwan]. Tai-pak: Chu-lip chuut-pan-sia.
——. 1993. Tai-gi tian-nau bun-su-chhu-li su-jip-hoat he-thong [Manual for TW 301: Taiwanese word processing]. Tai-pak: Coconut City.
——. 1994. Characters and Taiwanese special morphemes. MS.