High-tech tool for literati: using computers to analyze poems, prose

Robert Dilligan is a thoroughly modern scholar. Working alone in a bland room full of screens, keyboards, and printers, wearing an open-collar plaid shirt, he looks like just another computer scientist.

That's another sign of these electronic times. Dr. Dilligan is an English professor at the University of Southern California. And he is one of the distinct but growing minority of humanities scholars who are fascinated with the new tool of the scholarly trade.

Like the cloistered monks of the Middle Ages, copying texts by hand, his vocation is to ''preserve and transmit the culture.'' Gutenberg made the job easier in the 15th century, and now computers are making it still easier. Now scholars can do projects as a matter of course that once took the devotion of a career, or were beyond a scholar's reach altogether.

For example, beginning in the late 19th century, English poet and scholar Robert Bridges spent his life studying the poetic structure of John Milton's work, line by line, counting several stylistic features in each.

One of Professor Dilligan's graduate students recently put the same body of work onto a computer and charted some 40 features per line in less than two years.

Tabulations like this allow scholars to see how much a poet's style changed and - by counting the same stylistic features in samples by later writers - to gauge the influence of a poet's style on those that followed.

In Milton's case, Dilligan's student picked out the key features of the author's prosody and counted them in 1,000-line samples of other poets' work. The count showed a steady wane in Milton's influence on English verse right up to World War I.

Still, a minority of humanities scholars have taken the electronic route. Dilligan puts the number of specialists in this quantitative approach at 5 percent or 10 percent of his colleagues. The annual of the American Philological Association has one or two papers each year now using computerized studies.

But some expect the number to grow sharply in the next few years, simply because of the number of humanities scholars getting hold of personal computers.

''If you go into a faculty club these days,'' notes David Wilson, a classicist at the University of California, Irvine, ''it's almost boring to listen to the English professors yammering, not about Milton, but about what they did with their personal computers the night before.''

The biggest electronic project currently is UC Irvine's Thesaurus Linguae Graecae, a computerized thesaurus of classical Greek begun 10 years ago.

By comparison, a Latin thesaurus was begun in the 1890s by a team in Munich. Nine million or 10 million words of text were put on catalog cards in about 10 years. The interpreting is two-thirds finished, and at the current rate it will be complete in 2025. The Greek thesaurus at Irvine is eight times the size of the Latin.

''It was obviously going to be hopeless armed with pens, pencils, and students,'' says Dr. Wilson, assistant director of the project. The data were put into the computer in about two years. Interpreting must still be done the traditional way, but the thesaurus is already heavily used.

''If someone wants to study truth, beauty, and good in Plato, he can call the data in a few minutes, rather than reading the whole corpus,'' Wilson says.

Works of classical literature are being computerized at universities all over the world for electronic indexes, concordances, and bibliographies. The first such project was the keypunching and indexing of the complete works of St. Thomas Aquinas, sponsored by IBM and the Vatican in 1949.

''In the past, we had scholars with enormous memories who could make tremendous associations because they remembered,'' explains Vinton Dearing, a professor of English and computer applications in literature at the University of California at Los Angeles. ''You no longer have to have a great memory.''