|
|
樓主 |
發表於 2011-4-5 15:54:57
|
顯示全部樓層
2011年3月3、4日。在夏威夷举行了纪念张立青的汉语拼音国际研讨会。这是美国吴文超（Apollo Wu）先生在会上交流的论文。他引用了“中国语言论坛http://bbs.cnlangs.com/”中葛遂元的《小学拼音课本困境与出》等文来说明一问题。下面是他的原文：
Pinyin Assisted Chinese Text Processing
A paper to be presented at the Zhang Liqing Memorial International Conference on Pinyin
Honolulu, Hawaii April 3 – 4, 2011
Apollo Wu
Abstract: This paper provides an overview of the current use of Pinyin and Hanzi in China. Especially in today’s communication intensive environment, word separation in Hanzi should be compatible with that of Pinyin. To enhance Chinese text input, word-based Pinyin without tone marks but with specialized tonal spellings – as supported by the Chinese input method, Dianpin – is designed to enhance the accuracy of Chinese input and to promote the development of Chinese Digraphia.
Keywords: digraphia, Hanyu Pinyin Orthography, word-based Pinyin, machine illiterate, Chinese word separation (segmentation) , Dianpin, Pinzi, machine translation, human-machine communication,
Background: The Current Status of Pinyin and Hanzi Usage
A word-based approach of Pinyin orthography was generally used up to the late eighties, but Mainland Chinese textbooks and official publications have since switched to a syllabic approach. Among Chinese language teachers, some argued that it is simpler to teach syllabic Pinyin, and have attempted to downgrade the use of Pinyin in the name of reducing student’s burden. 1 Another argument is that since Pinyin is not a language, it doesn’t need writing orthography which is required only for the development of Chinese digraphia (一语双文). Since the nineties, the Chinese language textbooks in China have changed to using syllabic Pinyin for phonetic annotation. On the other hand, the Chinese government launched the ‘Basic Rules for Hanyu Pinyin Orthography’ in January 1996 which came into effect in July of the same year. 2 This paper surveys the Chinese language environment in Mainland China with emphasis on the effectiveness of word-based Hanyu Pinyin.
The Mega Trend in Chinese Language Usage
With the widespread use of electronic mass communication, ranging from television, computers, mobile phones, Internet and GPS, the reliance on spoken Chinese and the use of Pinyin input has increased rapidly. More and more Chinese admit that they can’t write Chinese from memory – 提笔忘字。 In a talk delivered at the Hong Kong Chinese University, visiting professor Lu Jianming (陆剑明) of Peking University told a vivid story of提笔忘字. He recalled a teacher who disciplined a high school student by ordering him to write a self-criticism. The student stood there motionless. When questioned by the teacher as to why he has not started writing his self-criticism, the student asked if he can use teacher’s word processor. With teacher’s consent, he sat down and rapidly typed out his self-criticism. Although the loss of written skills exists in English too, this problem is more serious with the complicated Chinese writing system.
Due to the de-emphasis of teaching word-based Pinyin, the Chinese are not naturally inclined to use word-based Pinyin text, or to use this approach to facilitate Chinese text input. Together with the reduced use of character text, there seems to be a double loss situation with the Chinese language usage – the loss of the effective use of both Hanzi and Pinyin.
On the other hand, the 4th edition of the Oxford Concise English-Chinese Chinese-English Dictionary (2009) provides Chinese explanations together with word-based Hanyu Pinyin. Popular Pinyin Chinese input methods all use word-based Pinyin database to facilitate inputting Chinese. Search engines such as Baidu and Google also contain extensive Pinyin glossaries. The inputting of word-based Pinyin, such as Jiuzhaigou, will immediately lead to information related to 九寨沟，the famous scenic destination. All these show that word-based Hanyu Pinyin is taking on more and more practical uses.
Two Types of Illiterates
A language teacher at the Hanyu Teaching Conference held at the Suzhou University 3 said that one still will be illiterate 文盲even after mastering Hanyu Pinyin because Chinese characters are everywhere. This is true, but those who only know Hanzi and not Pinyin are to a large extent ‘machine illiterates’ or 机盲. As a non-alphabetic language, there is no such thing as ‘what you type is what you get’ for Chinese – one has to learn a coding method or use a cumbersome writing pad to input Chinese. Since character coding methods depend first on the knowledge of Hanzi, these input methods should be considered as secondary symbol systems. Moreover, using such a coding method to input one’s language is just not natural nor intuitive, interrupting the thought process of writers. 4 As Pinyin is a primary symbol system directly link to spoken Chinese, it doesn’t disturb one’s thought process.
As traditional and simplified Hanzi are often written differently, each character form must have different coding systems. On the other hand, Pinyin can be used to input both forms of Chinese. A survey done in Guangdong a few years ago indicated that 97% of mobile phone users input Chinese with Pinyin. 5
The Case for Word-based Hanzi Writing
In 1987, almost a quarter century ago, Professor Chen Liwei (陈力为) of Peking University delivered his keynote speech at the Chinese Language Information Processing Conference held at Friendship Hotel 友谊宾馆in Beijing, in which he argued for the necessity of word separation in writing Chinese. Despite the obvious drawbacks of the existing use of syllabic based Pinyin, no movement towards word separation has taken place. On the contrary, an erroneous theory of the Chinese language needing to be character-based has stirred up heated debates among language professionals.
However, with the increased use of Chinese in electronic communication, the need for Chinese word separation has become more urgent. Existing practice of writing Chinese can be considered as text that is all linked together or all separated by individual characters, depending on one’s perspective. There maybe advantages for human comprehension of short, continuously linked Chinese text, for example, 生产和服务, that can be understood readily. However, machine comprehension would probably misunderstand it as 生产 和服 务, with 和服 referring to the Japanese kimono. The four Chinese characters of 环院长跑 is difficult even for humans to differentiate, with环 院 长跑 or环 院长 跑 being both valid options, but with different meanings and pronunciations.
Another example, 亚历山大王 can be read as亚历山 大王 or亚历 山大王and not the correct rendition of 亚历山大 王. In order to read accurately, Chinese readers require prior knowledge of word separation. One wonders why the Chinese accepts such ambiguity after thousand years of use. As pointed out by Prof. Chen Liwei, Chinese did not have punctuations some 75 years before 1987. Ironically, modern Chinese today seems to have more punctuation marks than English. How much longer is needed for word-based Chinese to become a universal practice?
Develop Accurate Pinyin Input Method with Specialized Spellings and Utilizing Techniques of Machine Translation
With the help of the eminent Pinyin scholar, Mr. Yin Binyong 6, the author has developed an enhanced Chinese input method call Dianpin 电拼 7. With Dianpin, tonal information is incorporated into a Pinyin syllable for purpose of disambiguation. The tonal row for syllable ‘shi’ 是thus created would be sshi师, shir石, shii使, shih势. By replacing ‘i’ with ‘y’ in the syllable ‘shi’, another tonal row would also be created as: shy诗, sheer时, shy史, shyh世. Together with additional specialized spellings of ‘shyc’ for 实， ‘shyx’ for试and ‘shx’ for 适, the Chinese sentence ‘这个 产品 需要 经过 试用，如果 适用 和 实用 的话，我们 才会 使用。’ can then be inputted as ‘zhege chanpin xuyao jingguo shyxyong, ruguo shxyong he shycyong dehua, women caihui shiyong’. Specialized spellings can readily disambiguate ‘他把它给她’ as ‘ta ba tta gei tha’.
With Dianpin’s one-to-one correspondence with more than 2,000 monosyllabic Chinese words - covering about 99% of modern Chinese - the accuracy of Dianpin is significantly higher than Pinyin. For inputting a large number of rarely used Chinese character such as颉 镛 瀛, Dianpin places a ‘v,’ as a marker, in front of the Pinzi (i.e. using Pinyin to spellout the components of a Chinese character) of the word components to form vjiye, vjinyong and vshuiying respectively. Although the Dianpin design is elegant and linguistically sound, its use will depend on promotion at the government or policy level and acceptance and adoption by the 300 million Chinese who are now accessing the Internet access and communicating online.
In order to further reduce the need for homophone selection, machines translation techniques maybe employed. For example the expression ‘yaoshi wo you yaoshi’ can be correctly translated into ‘要是我有钥匙’.
The Sad Case of Losing the Word Information in the Chinese Output
Powerful Pinyin input engines such as that of Sogou and Microsoft routinely use word information to facilitate the input of Chinese. Users may easily select correctly the right word to use. However, Chinese text output strips away any word information that is provided at the input level. However, Chinese text can just be presented with an extra space – just like in all other Roman languages and Asian languages like Korean. Alternatively, should half-space word segmentation be desired, as some Chinese prefer, only a simple programming change is needed to create this with Chinese text output. Prof. Peng Zerun of Hunan University has written a linguistic textbook with explicit Chinese word separation. 8 He felt that as Chinese word separation appears too wide with existing word-processors, he expressed the wish that software developers can provide a downward adjustment of the word-separation space by 1/2 or 1/4 spaces to conform to readers’ reading habit while retaining the word information for error checking and other IT applications such as information search, machine translation, artificial intelligence and word count, etc. This rational step in Chinese text processing would enhance the accuracy of the Chinese text, particularly with respect to human-machine communication.
Further Developments in the Direction of Chinese Digraphic
The author summarizes the issues discussed above into the Chinese catch phrase: ‘汉字分写，拼音连写’, which means that if we consider existing Chinese text are all linked together, then what is needed is to break it down into meaningful word segments. With regards to the syllabic Pinyin recently employed in Chinese language textbooks, what is needed is to linkup the Pinyin syllables into meaningful word segments. The result will be highly compatible text formats between the Chinese and Pinyin, hopefully both in word processing and in daily usage. This is illustrated by the following sentence outputted by the online GoChinese learning platform 9 – the Pinyin input text and Chinese output text are respectively: ‘jintian tianqi hen hao’ and ,’ 今天 天气 很 好。’
Footnotes:
See葛遂元《小学拼音课本困境与出》http://localhost/cnlangs/bbs/viewthread.php?tid=15283&page=e
For the 汉语拼音正词法基本规则, see http://www.pinyinok.com/GB_GBT.HTM
2010年12月3日至7日由蘇州大學、香港教育學院主办的「語文及國際漢語教育的反思」國際研討會。
See “现在的初学者最好别选五笔输入法”
http://www.lixiaolai.com/index.php/archives/8338.html
CHINESE PINYIN INPUT METHOD FOR MOBILE PHONE http://www.isca-speech.org/archive_open/archive_papers/iscslp2000/PSB2/113.pdf
See the information about Yin Binyong in Baidu http://baike.baidu.com/view/63033.htm
For more information about Dianpin and free download of Dianpin software, visit
http://www.yukexue.org/jianti/index.php, the 下载中文输入法-电拼 is on the right side of the webpage.
彭 泽润、李 葆嘉 两 位 同志 主编 的《语言 理论》，由 全国 几十 所 大学 的 语言学 概论 教师 协作 编写 Please visit: http://www.yywzw.com/jt/zerun/zerun050929.htm
For more information about GoChinese, please visit
http://hk.gochinese.net/index.jsp. GoChinese & the software, KEY Multimedia, use similar word segmentation engine.
Note on Author: Apollo Wu 吴文超 apollowu@gmail.com
Hong Kong mobile phone: (852) 9238 4174.
The author is the chairman and director of Gowell Software Ltd. in Hong Kong and is a retired former Chinese translator at the United Nations Headquarters in New York, U.S.A.
|
|
IP卡
狗仔卡
發表於 2010-4-17 01:32:31
提升卡
置頂卡
沉默卡
喧囂卡
变色卡
千斤頂
显身卡
發表於 2010-4-18 02:50:04
發表於 2010-9-14 23:24:58