The world's languages are expanding while contracting

World culture in the 21st century will be dominated by 10 or 12 major languages. Most 21st-century citizens will speak at least one of them, and virtually all other languages, including some traditional European ones, will ''probably be reduced to something like local dialects.''

So says Antonio Houaiss, a Brazilian linguist who has been studying world language trends for more than 40 years. He speaks half a dozen European languages fluently and reads more than 20, including Arabic, Sanskrit, and ancient Greek. Among his accomplishments is a 1961 translation of Ulysses.

Currently head of a Brazilian Academy of Letters project whose goal is to codify the Portuguese language before the end of the century, Mr. Houaiss sees the world's linguistic future evolving more rapidly than most people think.

''A steady unification of world language and culture is taking place today,'' he says. ''Some languages will emerge in the early 21st century as major world tongues, probably 10 or 12, and the rest will probably be reduced to something like local dialects. Almost everyone will have to speak at least one of the 10 or 12 dominant languages, and many people will speak two or three.''

''Certainly Spanish will be one,'' he says, ''because it is already widely spoken and growing. Obviously, English will be another, because of its economic power; Chinese will be one because of the sheer force of its size. Russian will obviously be important. Arabic is extensive and growing. Japanese is secure, because of the economic power of Japan.''

Houaiss's native language is also a contender for inclusion on the list.

''Portuguese is in a very strong position,'' he says. ''Brazil now represents about 3 percent of the world's population. And the Portuguese-speaking world includes much more than Brazil. There is Portugal itself, plus Brazil, Angola, Mozambique, and a handful of former possessions.

''Angola is a large and growing country in which Portuguese is now asserting itself as the dominant language over traditional tribal languages. This is mainly because of a deliberate effort on the part of the government there to entrench Portuguese as the national tongue. And finally, of course, there is Brazil itself. If present demographic trends continue, Brazil alone will represent 10 percent of the world's population by the year 2030.''

It is precisely these demographic and national-development trends that will make so-called third-world languages, such as Portuguese, Spanish, and Malay, more prominent, while some traditional languages may decline, in Houaiss's view.

Although French and German will hold their own into the 21st century, ''Romanian, Italian, and Dutch are not very secure,'' he says, adding, ''Dutch, in particular, may soon become another accident of history.''

Part of the problem, of course, involves demographic trends. While the total number of French-speakers in the world is hardly growing at all, languages such as Portuguese and Hindi are expanding dramatically, because of sheer population growth.

But there is another factor. Not only are populations growing; so are languages themselves. Says Houaiss, ''It took modern languages about 1,000 years to go from 4,000- to 90,000-word vocabularies by the mid-19th century. But it has taken only 100 years to expand to 400,000 or 500,000 words.''

It is absolutely essential, in his opinion, for every important world language to fully codify its vocabulary. ''Any culture, in order to progress, has to resolve its linguistic problems,'' he argues, ''and it can only do so by means of a comprehensive dictionary. If the language isn't codified, then it runs the risk of a crisis of comprehension.''

With modern vocabularies expanding geometrically because of technological development, the need to codify becomes increasingly urgent. ''Some people say you can only develop a comprehensive dictionary after 1,000 years of accumulated literature,'' he says.

''The people who put out Webster's Dictionary say their book includes some 4 million quotations. But they say they had to read 82 million items to get those 4 million references. In other words, they only used 5 percent of the material.''

And Webster's isn't comprehensive. Houaiss says: ''The Oxford English Dictionary is, of course, definitive. But the OED is 16 volumes of 1,000 pages each.''

According to him, English and French have ''already effectively resolved their linguistic problems.'' But he says Portuguese, Spanish, and Italian have not. He adds, ''Italian is a language in serious trouble, partly because it isn't organized very well and partly because the number of Italian-speakers isn't growing very fast.''

He believes the Portuguese-speaking world could codify its language ''in 15 or 20 years.'' His basic proposal is to create ''a central computer bank with all of the necessary information. With the bank hooked into major libraries throughout the Portuguese-speaking world it would be possible for people to consult the comprehensive dictionary just as easily as if there were a written dictionary. And the computer dictionary would have the advantage of being flexible. You wouldn't have to wait 20 or 30 years for the next edition. The changes brought about by cultural flux and new research could be included immediately.''

One of Houaiss's major frustrations has been his inability so far to obtain the backing ''of large-scale institutions'' to undertake the project. He adds, ''In the meantime, I'm sorry to say, the best dictionary in the world for references in Portuguese is the OED.''