Dissertation: Machine Translation, and the Future

Pages: 24 (7864 words)  ·  Bibliography Sources: 1+  ·  Level: College Senior  ·  Topic: Communication - Language  ·  Buy This Paper


[. . .] Machine Translation: History and Current Issues

Champollion (2001) addresses the issue that is the main subject of this research, the ability of machines to replace human translators. Champollion agrees that the field of machine translation is still very young. He also feels that advances in software will eventually make the job of a translator, more of a proofreader of machine output (Champollion, 2001). Champollion makes a distinction between a computer playing chess by calculation and a human "understanding" chess. He compares machine translation to this example. The future of machine translation technology lies in the advancement of neural networks and artificial intelligence. Champollion's main argument is that in time advances in technology will make machine translation more accurate, but the human proofreader will still be needed. Although Champollion's arguments were interesting, they were for the most part, based on opinion and he gave no solid evidence that technology was moving in the direction that he claims. He references no known research and development advances that support his argument, therefore his opinions can be considered to be just that, a hopeful opinion.

Celia Perez (2001) addressed the same issue regarding the capabilities of machine translation. "Translation is both an industrial product and process, and its methods have to adapt to the new industry requirements" (Perez, 2001). The roots of machine translation are in cryptography; translating messages and then decoding them character by character. The first language translation was word for word and as we all know, language cannot always be translated word for word, without some very embarrassing mistakes.

In 1966 the ALPAC (Automatic Language Processing Advisory Committee) concluded that machine translation could not replace human translation and that it was slower, less reliable and more expensive (Perez, 2001). Perez feels that several large companies do indeed have very good translators and that they speed the process of translation. However, she also points out that they still cannot be relied upon without human assistance. Perez states that the primary differences in machine translation software currently on the market in speed and data management. Other than they are primarily the same (Perez, 2001).

The most widely used machine translation system today is Systran. It I used on a daily basis by the Periodico'd e Catalunya. It automatically translates the entire daily newspaper from Spanish to Catalan. It is considered to be perfectly satisfactory, producing a nearly exact copy every time (Perez, 2001). Perez feels that this example is a strong example of the potential for future machine technology advances to replace human translators.

This may be a good example of the capabilities of a machine to perform a routine daily task. However, the argument is not as strong as Perez would suggest. It is unlikely that we will jump for Systran to Universal translators in a short time. The translation from Spanish to Catalan is relatively simple. Catalan is different from Spanish in many ways, but many consider Catalan to be more of a dialect of Spanish proper, but don't say that in Catalana. There grammatical changes, mainly syntax changes. A search and replace command could accomplish the same thing as the current program. It would make a stronger argument for the point fi the translation were being performed on two languages that were nor so similar. A translation from Spanish to Catalan is rather simplistic in execution.

John Hutchins (1996) feels that the idea of producing perfect machine translated text has been abandoned for more realistic ideas about its proper usage and capabilities. There are approximately 1000 machine translation packages currently available worldwide (Hutchins, 1996). Many of them only concentrate on one language pair, such English-Spanish, or Chinese-Russian for example. Most of the major computer companies have a Japanese-English translator program including Fujitsu, Toshiba, NTT, Brother, Catena, Matsu*****a, Mitsubishi, Sharp, Sanyo, Hitachi, NEC, Panasonic, Kodensha, Nova, and Oki (Hutchins, 1996). The major developers and makers of these systyems are AppTek, CITAC, EJ Bilingual, LEC, Neocor, PC-Translator, and Globalink (Hutchins, 1996)..

Many of the older systems, such as Systran, Fujitsu, Metal and Logos were originally intended to run on a mainframe and were only available to large commercial firms. Due to competition, these developers are now having to produce versions that will run on a PC (Hutchins, 1996). The difficulty in this will be that these systems depend on very large databases. The challenge will be to create a system that can handle the large database, without making it exceedingly slow

One of the most important innovations, recently was when Compuserve introduced a machine translator that could translate conversational e-mail (Hutchins, 1996). This package gained wide public acceptance. According to Hutchins, PC-based systems are being purchased on a large scale in Japan and the United States, but in Europe, most systems are still large mainframe-based systems used by large companies.

Many professionals prefer translation workstations where they can choose the tool most suited to the task. The translator can choose a level from spell-checker to full automated translation. Some of the most popular workstations are Trados' Translation Workbench, IBM's TranslationManager, STAR's Transit, and Eurolang's Optimizer. This type of software is more polular in Europe than elsewhere. Translation is important in Europe, especially as the European Union attempts to become more integrated.

Hutchins reviews the new research projects currently under way. Carnegie-Mellon University ahs begun work on the JANUS system. The JANUS system is a first generation speech translator. The first step to this however, is to develop adequate speech simulator software. Speech simulator software does exist and has found some limited applications such as the JAWS system that translates the written word into speech for blind persons. Voice recognition technology is also behind automated phone systems when you hear the system say, "Please press or say one." This technology must improve greatly before we are ready to move into actual speech translation in a different language.

In the beginning, software developers became caught up in just how far they could take their ideas. However, then a practical question must be asked, if the software is available will the people use it? At a 1995 summit on machine translation Bruce and Associates (1995) asked this same question. They found that people were using the new software, but that it had was not being used as much as would be expected.

An Internet search revealed an interesting position published by a supposed group of linguists and working translators, of which none of them published their names or took credit for the article. The purpose of the article was to dispel myths regarding the true capabilities of machine translation. According to this article Mr. Amar Almasude (coordinator of instructional technology, Francis Marion University, Florence, South Carolina) made the following statement at a Translators and Computers Session, ATA November 4-8, 1998. He said "Machine translations are still considered inadequate, even useless" (Mad Translation, 2003). The original quote could not be located to determine in what context it should be taken. This article appears to be well supported. However, when it is more closely examined, the arguments lacked some key elements. Quotes were made for which no source could be found. The article gives many examples of flawed translations. They also claim to have found the following in a magazine, although they never did say which magazine,

Maybe in twenty or thirty years from now there will be super-powerful computers with much smarter Operating Systems (neuronal, bionic or whatever) that can produce high-quality automatic translations. The trouble is, you don't want to have to wait until then. What can you do in the meantime? Globalization is knocking on your door... Can you afford to wait for these (hypothetical) tools to become available? We suggest you accept help from less ambitious but more effective tools." (Mad Translation, 2003).

In addition, we will also find this bad example of translation produced by a translation program. Although again, the authors never reveal which program was used.

Our program will allow everyone to understand one another better. It will also give you a little push to help you start or increase your interactions with other countries.

Translated into French by a commercial automatic translator:

Notre programme permettra a chacun de comprendre un autre meilleur. Il vous donnera egalement poussee pour vous aider commencent ou augmentent vos interactions avec d'autres pays" (Mad Translation, 2003).

According to the authors, the French translation is non-sense and says nothing. They use the above example to make the point that translating programs are expensive and do not produce useable results.

The authors of the above mentioned article felt strongly that machine translation was a waste of time and that it was useless in its current state. They largely used opinions of others similar to theirs as support for their position, yet they never five a complete source for this information. This was one extreme view on the disposition of machine translation and was included in this literature review to make several points. The first is that the… [END OF PREVIEW]

Four Different Ordering Options:

Which Option Should I Choose?

1.  Buy the full, 24-page paper:  $28.88


2.  Buy + remove from all search engines
(Google, Yahoo, Bing) for 30 days:  $38.88


3.  Access all 175,000+ papers:  $41.97/mo

(Already a member?  Click to download the paper!)


4.  Let us write a NEW paper for you!

Ask Us to Write a New Paper
Most popular!

Translation Plagiarism and Detection Software Chapter Writing

Music Education or Cross Platform Development Term Paper

Adaptation of Advertisements in Muslim Countries Research Proposal

Translating the Iliad Into More Contemporary Language Term Paper

Virtual Memory Management Term Paper

View 81 other related papers  >>

Cite This Dissertation:

APA Format

Machine Translation, and the Future.  (2003, February 20).  Retrieved July 20, 2019, from https://www.essaytown.com/subjects/paper/machine-translation-future/3334790

MLA Format

"Machine Translation, and the Future."  20 February 2003.  Web.  20 July 2019. <https://www.essaytown.com/subjects/paper/machine-translation-future/3334790>.

Chicago Format

"Machine Translation, and the Future."  Essaytown.com.  February 20, 2003.  Accessed July 20, 2019.