Controlled Language 2-98 (July 1998) by Wladyslaw Janowski Poland

Controlled Language (CL) is a controversial issue for linguists, editors, readers, but also for firms. Costs, marketing and sales figures are at stake.

Why did I select "risks and side effects", from the numerous problems involved, for my contribution? I am convinced that CL will be successful because positive / financial arguments prevail. Consequently, we will have to avail ourselves of CL, and identify and realize the risks involved and potential vicious side effects.

Compromise Between Efficiency and General Use of Language

• intelligibility,
• repeatability,
• low cost, balanced against potential risks
• high error rate, and
• probable rejection by target groups like private or professional product users.

• excessive simplification, and
• the deprivation of everyday colloquial.

A society using CL in everyday communication would -- horror of horrors -- be setting a standard. The danger is imminent because people tend to imitate stereotypes. Therefore CL must be applied sensibly and pragmatically.

Some Languages Are Not Fit for CL?

Let me give an example:

There have been Windows applications in the Polish language for years. Version 3.1 asks: "Exit Windows?" before closing a session. As "Exit" means end of function or process, translators chose the Polish concept "Koniec" ("end"), instead of the word by word translation of "exit" which would be "leave". Applying CL rules, one should stick to "End Windows" or "Close Windows" instead of "Exit Windows".

The adaptability of CL is different in different languages. It might be extremely difficult to find acceptable CL expressions in some languages. A probable solution might be worldwide application of International English, even though that might prove dangerous to culture and language. On the other hand, there are certain advantages; which means that, willingly or unwillingly, International English is being introduced, step by step, not only in Technical Communication but also in the fields of the media and colloquial language.

Fairly old concepts like GerEnglish, Franglais or Germish hint at definitely related phenomena which are not directly at work though, when people use International English in other languages, they are endangering the English language itself. As more and more people use International English worldwide, it’s even more necessary that the English and Americans learn to understand and use a version of their own language that is not only simplified but also nonstandard from their point of view.

CL and Translation Tools

Let me quote an example from this field: After analysis of syntax the translator is offered the separate syntactical unit "assemble", but there is no context. No wonder this expression will later turn up in different forms, such as

• title of paragraph
• part of title, e.g. "assemble automatic gear"
• assignment, like "assemble module x and check seal"

So I translate into Polish the equivalent of "montage" which proves correct and is fed into the data base.

Then this syntactic unit is used for a preliminary translation of a new text, and the program qualifies its output as either

• Correct, or
• Problematic (to be checked), or
• Untranslated.

Even if I look through the data base, I cannot change this process because the machine-translation is correct on a given syntactic level -- but not in this special context. It’s only when proofreading that I may realize that, by taking into account the context, something is wrong. By then it would be too much trouble to correct the data base. Or delete and start again? Would that be sensible?

Where does this risk derive from? In translation memory systems you must use the shortest possible syntactic units (sentences as a rule, but also individual words). The first and most important selection of segmentation criteria takes place in the original language.

During translation you may interfere (i.e. take steps) to correct wrong or dangerous segmentation. But that means additional work; work you wanted to avoid. Thus the need to interfere may render the use of translation-memory-systems senseless.

Finally, syntactic rules in different languages are different. What might function well with English-German-English translations might prove difficult with other languages, or result in comical or dangerous mistakes, all of which increase costs immensely.

Conclusions

Should we use CAT-systems? In principle yes, but........

Should we use International English in the CL-version as an alternative for translation in the technical field? Why not, but.........

Are there "risks and side effects"? Yes, but -- whom should we ask?

"transline Deutschland - Übersetzungsdienst für technische Übersetzung"