The Definitive Word on English. Second edition of the 100-year-old Oxford English Dictionary took five years to complete. UPDATED DICTIONARY
LONDON
WRITING a comprehensive dictionary of the English language is not what it used to be. Yes, word lovers still send thousands of hand-written quotation cards to editors at the Oxford University Press, adding fresh raw material to the files just as free-lance contributors have done since the 19th century.
And editors still agonize over words and their pronunciations as Sir James Murray did 100 years ago, when he began editing the first edition of the Oxford English Dictionary (OED).
Yet the scholars who worked for five years to bring out the second edition of the world's most authoritative dictionary of English are masters of more crafts than one.
They must be computer jockeys as well as word buffs, translating their meticulous scholarship into the language and logic of the electronic age.
They must be as comfortable with sophisticated software programs like OEDIPUS - the Oxford English Dictionary Integrating, Proofing, and Updating System - as they are with tracing data on the latest additions to the language like ``perestroika'' and ``yuppie.''
They must also yield to the organizational talents of a project director who coordinated Oxford's largest-ever dictionary publishing venture, at a cost of some $18 million.
``It was pretty clear that the only way ahead if we were to do anything more with the dictionary was to computerize it,'' says Timothy Benbow, project director on the second edition. ``But,'' he adds, ``that wasn't so simple.''
The original 13-volume dictionary (published between 1884 and 1933) and the four-volume supplement (published between 1972 and 1984) existed only on paper. These texts of some 350 million printed characters had to be keyboarded into a computer, integrated into one continuous version, and then updated with new entries, revisions, and corrections.
For guidance in this massive task, Mr. Benbow consulted management theory for complex civil engineering projects.
``No one had ever done a scissors-and-paste job using a computer to integrate two large dictionaries,'' says co-editor Edmund Weiner. ``The project was basically a question of dialogue between lexicographers on the one hand and computer people on the other. Both sides had to be reeducated.''
The beauty of the second edition is that, through its 2 million quotations from scholarly and popular literature and thoughtful cross-references, it has preserved and extended the history of the English language by using the same principles and format pioneered by its original editor.
``The remarkable thing about the OED is that Sir James Murray devised a structure for each entry that would work for all his entries, whether they were a few words long or thousands of words long,'' Dr. Weiner says with respect for the Victorian scholar most responsible for the first edition. ``The form was so logically thought out it was convertible to a computer system.''
Keyboarding the first edition and supplement onto electronic tape was the first step, which took a United States company, International Computerprint Corporation, a mere 18 months. Using dozens of keypunch operators in Tampa, Fla., the company kept within Benbow's stringent guidelines for accuracy, and the uncorrected-error rate was only 4.5 keystrokes per 10,000. ``They did an absolutely wonderful job,'' he says.
The keyboarders also added basic computer ``tags'' that identified the main elements of some 300,000 entries. The University of Waterloo in Canada and Oxford's computer department refined the tagging system to identify the sub-elements in more detail, including 137,000 pronunciations, 249,000 etymologies, 577,000 cross-references, and those 2.4 million quotations.
A computer could then integrate the supplement with the original dictionary. This was followed by many weeks of tidying up by on-line editors. ``Inevitably, there were areas where the computer didn't know what to do and did something we had to unscramble,'' says Weiner, who was responsible for the integration process.
For the crucial on-line editing and revisions, International Business Machines UK Ltd. donated equipment and expertise to help Oxford develop the OEDIPUS software. Despite its name, says Weiner, the software has proved so reliable that it is now used with other dictionary projects at Oxford.
Co-editors Weiner and John Simpson then personally proofread every page of the final text, some 59 million words in all, the equivalent of reading the entire Bible through each week for a year. The integrated, revised, and updated dictionary is a reference work unrivaled in its completeness and scholarly importance, even if it was completed in less time than the dictionary's first editor sometimes took for research into a single letter of the alphabet.
``By putting all this material together you actually create links throughout the language from about AD 1150 [the earliest reference in the dictionary] to the present day,'' says Dr. Simpson. ``It will force people to realize that the language of Chaucer and Shakespeare is not unrelated to the language of the 20th century.''