Dictionary makers try anew to define their terms

All is hush-hush at the American Heritage Dictonary offices overlooking the Boston Common. The editors are on deadline. If all goes well, and there are a minimum of interruptions, they can rush their work to the printer just under the wire - early 1991.

If nearly 10 years sounds like a long time to figure out how the 200,000 most popular words in the English language are being used, consider the Oxford English Dictionary. That most scholarly of lexicons took 1,300 free-lance readers and hundreds of editors 50 years to complete; it fills 13 volumes. This coterie of nine linguists and a secretary must choose the most ''meaningful'' words and squeeze their definitions between two covers in one-fifth the time.

Of course, that's just the major 10-year revision of their collegiate dictionary. The 1982 version came out last fall. For this, each entry in the original 1969 version was looked at, scaled down, added to, changed, or deleted. And 25,000 new words found their way in. Each year, succeeding editions have minor revisions - mostly deletions and additions. Besides the collegiate edition , there are pocket-size editions, children's editions, and younger student editions.

The nine linguists have outside help, of course. To get its last revision out , the core staff swelled to about 40. Close to the 1991 deadline, it will grow again. And in the interim, about 10 or 12 free-lance readers daily send in magazines and journals from a selected sampling of 20 to 30 publications from all over the country.

These readers have circled newly coined words - and old words with new usages - in red ink. The citations are trimmed by secretary Fran Figelman, placed on 3 -by-5 cards, and kept on file. In an attempt to get the broadest cross section of English, publications run the gamut from Science Journal to Detective Today.

''We even collect citations from menus, from underground newspapers, from current fiction,'' says editor Dolores Harris. Besides their citations, American Heritage researchers rely on a computer corpus which is a scientifically chosen sample of articles from everything from books to newspapers. Similar word citations can be called to the screen electronically.

Among the words AH wants to know about are:

* New terms or coinages, slang, technical terms, etc. New to the 1982 American Heritage Dictionary (AHD) were such words as bionic (enhanced by electronic components), ballistophobia (fear of projectiles), acid rain, gasohol , robotics, lunarnaut.

* Established terms with new senses, such as ''plastic'' meaning artificial and ''hardware'' meaning physical components of a computer. The 1982 AHD included a sixth meaning for ''sting'': undercover operation.

When it comes time to define a new word, the five or so American Heritage definers check their citation cards, choose a number of senses, and write their definitions - concisely. Each definition is checked by a peer, then must pass review by a senior editor. If there is significant conflict, the definition is discussed in a full conference of editors.

A sneak preview from the incoming citation file gives a sampling of words that may win access in the future: Ma Bell, megacorporation, megadose, megahit, meganutrition, muscle car, multi-candidate, networking, no-nukes movement, nuclear freeze, supply side, spaciness, LIFO (last in, first out accounting).

Here at American Heritage, each of the editors on the staff has a specialty and a number of secondary talents. Dolores Harris writes definitions.

''When a definer looks at a new word, he looks for two kinds of evidence: information showing that the word is widespread geographically - not just exclusively in Time magazine, for example. And you look for chronological distribution. A word which has been around only three months won't go into the dictionary. The old rule of thumb used to be that the word had to be around for seven years until it was fully accepted. With more and faster communication, things move much faster, and a certain class of words appears in the dictionary almost instantaneously.''

One example of this, she says, is ''atomic bomb.''

''With technological developments the requirements for chronological distribution aren't so stringent. When there's a significant breakthrough in science, - such as the discovery of DNA, you know the word is here to stay.''

There are two kinds of definers, Mrs. Harris says: lumpers and splitters. ''The lumpers like to have big global meanings. . . . They like to generalize, to factor out the common denominators, if you will. The splitters don't like that approach; they like very precise distinctions. They'll make distinctions sometimes which don't even exist in a language. Sometimes there are justifications for splitting, and so splitters are good to have on your staff, because they keep you thinking.''

Another primary tool in defining words is stylistic labels, such as nonstandard, slang, vulgar, obscene, offensive, and obsolete. These tell the reader that certain words are restrictive in their use. ''Vulgar'' warns of social taboos attached to a word. ''Offensive'' might be a racial slur, ''insulting and derogatory to its object and a discredit to the user as well,'' the editors say.

Many words begin as slang and then rise over the years to become part of standard usage. Lexicographers track this growth but some never make it: Duds (clothes) and dukes (fists) are still slang, but they go back to the 16th and 18 th centuries, respectively.

Critiquing Webster's Third International Dictionary when it was first published in 1961, linguist Dwight MacDonald lamented: ''The most important difference . . . is that the Third has accepted as standard English a great many words and expression to which the Second attached warning labels: slang, colloquial, erroneous, incorrect, illiteratem.'' he wrote. This ''tells us a good deal about the changes in our cultural climate since the second edition appeared in 1934.''

Marion Severynse is AHD's permanent staff etymologist (four others work on a free-lance basis). Her PhD dissertation in English at Harvard was 300 pages on three verbs of Old English, all meaning ''to turn'': cierran, hweorfan, and wendanm.

''Etymolgy starts with the most recent ancestors of each word, just like families have ancestors - mothers, sisters, and daughters,'' she says. ''Then the word is followed back in time.''To find out the derivations of words, Dr. Severynse relies not on original research but a host of library sources - most notably the Oxford English Dictionary (OED).

''One stands on the shoulders of giants in this business,'' she says. ''Fortunately, I don't have to do the original research. I rely 999 9/10ths percent on other dictionaries and other scholars' works.''

She explains that the original research going into these publications is very involved. ''The OED researchers read tons and tons of printed and manuscripted works, underlining words as they went, and making chronological lists of the usages of the words. They collected 3 1/2 million citations in a continuous history of the words from earliest instances.''

One new word she is following at the moment is ''canban.'' Used and described in the New York Times of Oct. 6, 1982, it refers to a factory management system in Japan. It is named after the small cars that are used to deliver parts to the assembly line. Should the word ever become popular enough that Dr. Severynse is asked for an etymology, she will take the number of citations collected and rely on her knowledge of the Anglicization of Japanese words - what sounds are typically transferred to which spellings. And she will back up her research by consulting Japanese scholars living in the United States.

In addition to giving the final word on spelling, dictionaries also provide correct pronunciations. They tell us the real difference between the pronunciation of Mary, merry, and marry, especially if we didn't know there were any.

Across the hall from Marion Severynse's office is that of pronunciation editor Pamela DeVinne. As with etymology, AHD does very little original research on how words are being pronounced. Instead, Mrs. DeVinne uses reference material: new word lists (compiled periodically in various areas of the country by dialecticians), scholarly works in journals such as American Speech magazine, and linguistic atlases.

''I was doing a phonology [pronunciation study] in the South,'' she says, ''and I asked one lady what's the metal that cans are made of, and she said, 'tin.' And I asked what's the number after nine? And she said 'tin.' And I said, 'So, you pronounce them exactly the same?' And she said, 'I do NOT.' ''

''This is the problem when you are doing phonetics,'' Mrs. DeVinne says. ''Is there really a distinction that, because of your own-mind set or dialect, you are missing? This happens very often in fieldwork.''

But to help readers sound out unknown words, phoneticists use words the dictionary user already knows how to pronounce. So if the reader is looking up ''lark,'' the dictionary will tell him to pronounce the ''r'' in this configuration like that in ''park.'' Regardless of the reader's accent (pock, pawk, pack), he will transfer the sound to the new word. As in words like tomato , potato, and economics, AHD includes more than one pronunciation, and they emphasize both are correct. ''We always make somebody mad, however, Mrs. DeVinne says. ''If we say 'Argenteen' [for Argentine] first, then 'Argentyne' second, the Argentyne people bury us with letters wanting to know why.''