ASK Western scholars about Japanese classical poetry - the precursor to haiku - and most would probably say that the works are alliterative (sounds repeat).
That's certainly the impression. Scholars have confined their research to a few famous alliterative poems. When Jon LaCure of the University of Tennessee in Knoxville put all the poems on his computer, he expected his machine to spit out more patterns of alliteration. To his amazement, he found just the opposite. The vast majority of the poems avoided repeating sounds.
Professor LaCure's computer overturned conventional wisdom.
Slowly, a small coterie of scholars is using computers to develop new insights into literature. Some professors are putting poems on spreadsheets (each syllable gets its own position). Others are using CD-ROMs (optical disks) of dictionaries to study word origins. Still others are tackling the basic task of getting texts into computers.
That's harder than it sounds.
Researchers don't want just words on a screen. They want the computer to distinguish between chapter headings and regular text, between adjectives and verbs. Sometimes it's hard to tell what's what. In the sentence "Suzie visited Essex," is Essex a person or a place? Sometimes the author is deliberately ambiguous. So the computer has to show that too.
Several projects are putting printed works into machine-readable form. The Oxford English Dictionary is now on CD-ROM. So are the works of Greek writers from Homer to the 6th century. Chadwyck-Healey Ltd., based in Cambridge, England, aims to stuff everything written by Western authors between AD 200 and 1400 onto five CD-ROM disks.
Soon the computer "will be a regular tool within the reach of researchers," says Joel Goldfield, a professor of French at Plymouth State College in Plymouth, N.H. With the technology to search vast amounts of material, humanists will ask new questions.
"Computing is just getting powerful enough for humanists actually to use," says Edward L. Ayers, a University of Virginia history professor.