The OED and innovation

Back in the day

The New English Dictionary (as the OED was rightly called back then) grew out of a Victorian desire to classify. Its scope is astonishing: to provide a brief, scientific account of the history and usage of all the words of the English language, wherever and whenever they were spoken.

Common sense caused the original editors to draw back slightly from this absolute all-inclusiveness, but when the first instalment of the dictionary was published in 1884 readers were amazed at the elegance, detail, and accuracy of the text – and looked forward to a long ride through the alphabet (which in fact took another forty or so years to complete).

The OED wasn’t the first national historical dictionary. The German and Dutch dictionaries, for example, started publishing earlier. But almost as soon as the OED appeared it became the model for serious scholarly work in historical lexicography. And the OED was completed in only 44 years of publishing: remember that the German and Dutch equivalents both took over one hundred years to complete – and you could claim that English is the more complex language.

Another remarkable fact: the OED was compiled largely from materials sent to the editors by enthusiastic members of the public, in their thousands. The first Appeal for Readers (in fact, specifically American readers) dates from 1859. James Murray issued a new appeal when he took over the editorship of the dictionary in 1879, and this was followed by further appeals and lists of words for which evidence was sought. The first edition of the dictionary represented a fifty-year survey of the language by the people who spoke it.

And once the dictionary was published, it became an international obsession to try to ‘beat’ the OED by finding earlier evidence for the words and meanings it included.

Spreading the net

By the time the first edition of the OED was finished in 1928, it had spawned a new network: the Oxford Dictionaries of current English that we know today. The first edition of the Concise Oxford was edited by H. W. and F. G. Fowler and published by OUP in 1911 (updated in 1929, shortly after the OED was finished); in 1924 we saw the first Pocket Oxford; by 1933 the historical Shorter OED was ready to set sail. The range of work developing around the OED lodged Oxford well and truly at the centre of the dictionary-making world.

Relighting the fire

When the first edition of the OED was completed in 1928 the job was considered done, at least for the present. But soon it became clear that the OED would need to find a way of keeping up to date with the language if it wasn’t going to exist simply as a monument to the past.

In 1957 the University Press fanned the OED’s embers by appointing Bob Burchfield to produce a supplement of modern words. Bob had to start more or less from scratch, planning a remarkable survey of the English of the late 19th and 20th centuries. Again it was a grand vision, and typical of the radical ideas Oxford was prepared to back.

As the vision took hold, the supplement increased in size from one to four volumes (5,730 pages, 69,000 entries, and just over half a million quotations). By the time it was completed in 1986 Oxford had its second major survey of English.

Handheld technology

Back then, the curious innovation that caught the imagination of the dictionary-buying public was the magnifying glass. In 1971 OUP had boarded the crest of a wave by publishing the first edition of the OED in ‘micrographically’ reduced format – nine pages of the original to one page – with an elegantly boxed magnifying glass to help the reader unlock the tiny mysteries of the dictionary. What might seem like publishing madness struck a chord with the dictionary-buying public, many of whom wished to own the full OED, but much preferred to house two volumes than thirteen. The image of the dictionary and the magnifying glass does not seem to go away.

Real technology

By the time the modern supplement was finished in 1986, the Press had already decided to embrace the emerging technologies in order to keep the dictionary up to date and to ensure its future. Instead of publishing supplement upon supplement (the old model), it took the challenging decision to leap on board the new technology and see how the dictionary’s future might look as a machine-readable text.

There was no ‘online’ then. The Second Edition of the OED was published in print in 1989, but – most significantly for what was then a massive database – it was published from tagged, machine-readable text. The entire contents of the OED (67 million keystrokes) had been keyed to computer over 18 months by 120 typists. Not only that, the keying had incorporated a basic tagging structure, with an eye to future information retrieval.

In fact, the OED was – during the 1980s – one of the principal partners of the new Standard Generalized Mark-up Language (SGML) project, and experience with the OED’s database helped to clarify the SGML rules, which became the industry standard.

State-of-the-art computing work was being carried out on OED data at the University of Waterloo (UW) in Canada in the 1980s. The research objective was to establish the most efficient data structure for holding a large-text database like the OED. The software developed in the process was later put to successful commercial use by OpenText, a spinoff company at UW. Furthermore Tim Bray, manager of the UW Centre for the New OED, was soon to become a leading proponent of SGML’s successor, XML. Looking back, Tim notes “It would be entirely accurate to say that the work with the OED was a significant input to the development of XML” (October 2010).

With the OED data on computer, it became possible to develop the first CD-ROM of the dictionary, again a pioneer in the field of academic research when it was first released in 1992 (a prototype had even been released in 1989).

Editing and processing online

By this time, the world was moving (although it didn’t know it) towards an online future.

Editorially, the OED immediately recognized the value of searching historical and modern texts by computer, as an integral part of its mission to provide the most complete information available about each word. Editors worked in collaboration with the Oxford Text Archive (another innovator in its field) to collect and tag texts in SGML. Suddenly the OED’s new entries started to benefit from this enormous stream of new data. It was all part of the editorial objective of keeping well ahead of the pack.

On to the Internet

It was early days on the Internet. Alongside the CD-ROM developments, we saw the possibilities of the OED’s presence on the web, and wrote a prototype online site – fully searchable – for the project’s editors to experiment with as early as 1994. The Director of the OED’s North American Reading Programme at the time, Jeffery Triggs, writes (October 2010): ‘We were easily among the first 400 sites in the web… Even among those, most at the time were sites with sets of static pages. We were one of the few back then to have a dynamic site accessing backend resources to produce HTML pages on the fly, in our case the entries of the OED and other dictionaries.’

That led to March 2000, when the OED became the first major national, historical dictionary to present itself on the Internet. More than that, it was the first updatable dictionary of its kind up there. What you saw one quarter wasn’t the same as what you saw the next. The dynamic platform was unsettling for some, but it properly reflected how language changes, and how editors respond to this. And things were happening behind the scenes, too. In 2005 the OED switched to a new state-of-the-art editorial computer system (Pasadena), which gives the editorial staff a very high level of support – updating cross-references automatically, flagging possible errors, etc. – leaving the editors free to research their words and work on their definitions. Meanwhile a fresh XML information design enables the OED’s data experts to build more searchable features in to the core dictionary text.

With the relaunched online site, the OED goes a step further: the dictionary is no longer only a resource you approach for information about a word. The interweaving of the Historical Thesaurus of the OED and other linking features ensure that readers can be taken on a journey through the language, from their first point of contact on through textual, visual, or graphical links which all help to illuminate our understanding of the language, culture, and history of English speakers around the world.

John Simpson
October 2010