/******************** Article ID: YZYREP-language-future-n-world-government.en Chinese Article ID: YZYREP-language-future-n-world-government.zh Title: The Future of Human Languages and the Possibility of a World Government (2008 - 2050) Chinese Title: 人类语言的未来及世界政府之可能 (2008 - 2050) This copy's version: v0000 (The author believes that it can be a good methodology to "organically grow" a creation such as an article, a software program, a company or even a Roman Empire, i.e. the first version can be a minimalist expression of the idea which is so simple that it can immediately be made (even if it's just a technology demo), and each subsequent version enhances and revises that creation, and upon the completion of every version we should immediately go into a "Profit!" process (a profit can even be the presentation of a demo to the academia).
To see all versions, Google for the Article ID or visit (can subscribe via email/RSS): http://groups.google.com/group/yzy-organic-writings-en http://groups.google.com/group/yzy-organic-writings-zh (Chinese) Send your comments to Yao Ziyuan: [EMAIL PROTECTED] About the author: Yao Ziyuan, born 1984 in China, was once a good kid and a "good" student. In Grade 5 he won a First Prize in a regional student programming contest of Jiangsu Province (NOIP 1996); in the third year of middle school he won the 33rd place in China's National Olympiad in Informatics (NOI 2000). Since high school, he rejected and escaped all classes, homework, exams or other school activities except afternoon free-play-on-the-playground classes; instead, he developed and started to sell a couple of shareware programs on the Internet and gained economic independence and strength. In the second year of high school he was again allowed to compete in programming contests and finally won the 21st place in NOI 2002. Therefore he was admitted to Fudan University (China's third best and southern China's best university by government sponsorship and public opinion) for a computer science major. He escaped everything in college again after a week, finding classes are no! t useful and even harmful for his lifestyle. And he instead conducted his own research in a rented apartment near campus, mostly problems related to natural language, logging his research notes onto the newsgroup list.linguist which is archived by Google Groups, under the accounts [EMAIL PROTECTED], [EMAIL PROTECTED] and [EMAIL PROTECTED] He also went to the student residential park for a play, on a daily basis. After two years and a half, he was formally expelled due to too many class absences, and continued his research at home till now (March 2008). Today, he publishes some of his major research achievements to the world in a series of "Yao Ziyuan Reports" and starts to develop them into software. Yao Ziyuan thinks these areas have room for further research or development: computer-aided development (such as writing and software engineering), computer-aided problem solving (such as research in natural and social sciences), natural science itself, and democracy! in China and the whole world. ********************/ The Future of Human Languages and the Possibility of a World Government (2008 - 2050) Executive Summary ==================== Researchers and practitioners in natural language processing and machine translation should take note of several important ideas which will impact and reshape the field significantly. These ideas include: * "Automatic code-switching" (ACS) can enable people to learn a foreign language very effortlessly and very efficiently. Imagine this scenario: A Chinese person is reading a Chinese web page in his Web browser. The Web browser has a plugin (or "extension", in Firefox terminology) that automatically selects a small number of Chinese words from the Chinese Web page and annotates, or even replaces, them with their English equivalents, as soon as the Web page is loaded into the browser. This way, the Chinese person can naturally acquire English vocabulary (and in fact this method can teach grammar and other linguistic knowledge too, and machine translation experts should be able to see the underpinning know-hows here). The general idea of rewriting portions of a native language text in a foreign language for teaching this foreign language was first proposed by American anthropologist and linguist Robbins Burling in the 1960s, who dubbed it "diglot reader/diglot weave" and was in! spired by a "Learning Chinese" book series published by Yale University Press, where new chinese characters are gradually replacing Romanized Chinese in a text. There are academic literature discussing and commercial products (e.g. www.power-glide.com) commercializing this method but I am probably the first to both independently rediscover this method (twice, as recorded in an Oct 2004 and a Apr 2007 post at the newsgroup list.linguist) and instinctly assume that it is possible to be adapted to an automatic code-switching software program (like the browser plugin scenario above). I have solved problems associated with the automati adaption attempt and recorded them on the newsgroup list.linguist since the two times of discovery. [More details will be available in future versions of this article.] * A "user-input-driven syntax aid" and a "user-input-driven ontology viewer" can enable people to write in a foreign language with correct word-specific syntax and with the right terms for the current context. These two tools for foreign language writing can also be applied to "intermediary language writing" where the user uses his native vocabulary but a formal grammar similar to that of programming languages (See: Universal Networking Language). [More details will be available in future versions of this article.] * Full-automatic high-quality machine translation might be a wrong question asked at a wrong time and given a wrong priority. Maybe the right question is "With the uncertainties caused by syntactic and semantic ambiguities unresolved, how best can we *present* information of a foreign language text to a reader who doesn't know that foreign language". Let's analogize the information in a foreign language text where there are unresolved ambiguities to an image where there are some unrecoverable small holes. By default we want the user able to perceive the "gist" of the image in a comfortable manner. One way is to "zoom out" the image, to an extent that the small holes shrink to individual pixels or even sub-pixels or even invisible, while the thumbview of the image is still a well-perceivable image. This actually corresponds to the "abstraction" approach of getting an abstract semantic representation of a text in the field of automatic text summarization. Another way is not to! try to zoom out, but to "fuzzify" the holes, so that the user can still perceive a rough-less image. How do we "fuzzify" an ambiguity in machine translation? We can express the ambiguity in a vague manner in the target language, EVEN IF THE LOSS OF INFORMATION SEEMS UNACCEPTABLE IN STANDARD MACHINE TRANSLATION DOCTRINE. For example, suppose there is an English sentence "I am going to that bank." and suppose the computer can't confidently disambiguate "bank" anyway, and suppose "bank" could be "riverside" or "financial institution", our fuzzification would be to find as much information as can be ascertained, i.e. at least we can assure it is a "place" whether it be a riverside or a financial institution. Some people may argue that this "find a most specific common hypernym" procedure already exists in the treatment of ambiguity between close enough polysemous senses, but my point here is that we can use it even if the information loss seems extraordinary. I propose "full-a! utomatic layered-quality machine translation", where the user by defau l t is presented with an overview of the information ("zoomed-out thumbview"), and if he wants to explore a detail, ambiguities in the detail can by default be presented in a "fuzzified" manner, and if he still wants to further explore the possibilities of a low-level ambiguity, the computer can show him all possible exact translations for that ambiguity. Also note that if an ambiguity exists in a "significant portion" of the text, such as in the title, the first paragraph, or the first sentence of any paragraph, that ambiguity may be by default displayed as an expanded list of all possible exact translations, instead of as a fuzzification. [More details will be available in future versions of this article.] * As ACS makes it so very cheap for people to acquire a foreign language, there will be profound implications for world politics, economy, culture and the whole human civilization. For the first time will all humans be able to communicate in English (however I propose to rename "English" as "A Random Code System", or Arcs/Arx, to reflect the truth that all natural languages are based on root words whose sounds were randomly designiated and to promote globalism and demote nationalism here), and this can lead to a free and democratic world government (if the other barrier, dictatorship countries, are gone by then). And English will not be the ultimate "world language" really, because once English is used by all the world's people, it will also mean English is taken over from native English speakers to all the world, which in turn means if a significant majority of the world's population coordinates a movement that gradually transforms the English vocabulary to something "more ! fair for the rest of the world", they will succeed. [More details will be available in future versions of this article.] _________________________________________________________________ Helping your favorite cause is as easy as instant messaging. You IM, we give. http://im.live.com/Messenger/IM/Home/?source=text_hotmail_join _______________________________________________ Mt-list mailing list
