The Word Detective

by John Simpson

  Back in the dictionary offices, we were getting used to our new open-plan environment, which seemed fitting for the collaborative expansion and scholarly openness we were keen to promote. Several huge pillars holding up the ceiling stood in the centre of the office. On these—in keeping with our low-tech coloured progress strip of the 1980s—we plastered paper progress charts which the editors marked in pencil as they reached the end of a task. Somehow it meant we were still in control of things and hadn’t ceded responsibility. Once we had become adept at researching our words in an expanding online environment, we started to turn our attention towards online editing of dictionary entries. Up to this point in the early 1990s, the extensive changes to the text of the Second Edition of the dictionary had been marked by editors on double-spaced printouts, which had then been passed around the various editorial groups (etymology, bibliography, science, new words, etc.) who needed to review and revise different aspects of the text. We had text-editing software from the old days in the 1980s for merging OED and Supplement text, but this wasn’t available to all the editors, and our marked-up printouts were keyed to our dictionary database by specialist keyboarders.

  Later in the 1990s our text-editing software became resilient enough to allow online editing by all editors. For the publisher, online editing meant cutting one step in the publication process, with the possibility of saving money; for the editors it provided a higher level of control and satisfaction. There were academic arguments at dictionary conferences over whether lexicography was an art or a craft. It is both, but the online text editor gave us scope—depending on the particular editor and the particular word edited—of foregrounding either the artistic impulse to present a beautiful and coordinated entry, or to weave—craft-like—the complex threads and skeins of the entry into a pleasing whole. The computer screens containing the text of a dictionary entry were themselves a mass of vivid colours. Each slice of information (the definition, the etymology, all of the parts of a quotation, etc.) was allocated its own personal colour—maybe to help us remember which section of the entry we were working in. Most editors quickly became adept at online editing.

  We have had the word line in English since the earliest days, when it mainly meant a length of string or rope. But it’s been around long enough by now to have developed many sub-meanings and new branches. One of those new branches is the word online, and because we have so many different “lines” (telephone, shipping, air travel, consumer products, etc.), it’s hardly surprising that online can live in various contexts. Obviously the digital use of online is a fairly recent innovation. Online started life as a railway term. In fact, the OED says we have to start our investigation with offline, back in 1919. I suppose that’s not really surprising, as more things are offline than are online. In early twentieth-century railway days, a warehouse was offline if it was not immediately connected to a railway, and you had to transport goods for a distance by lorry to get them to the freight train. If you were online (recorded from 1926), then things were a lot easier, as your manufacturing facility was right next to the railway track and all ready to ship freight by train.

  If we move on to the 1940s, we find a similar distinction happening on the airlines (most early air-flight terminology comes from shipping, sometimes via railways). We had shipping lines well before we had airlines; pilots guided boats before they guided planes. In 1940s’ flight jargon, a stop was “online” if it was one of the authorised stops used by an airline, and “offline” if it wasn’t.

  The big jump to computing happened soon after this. The OED records both offline and online from 1950 in relation to operations or processes that take place while connected to (or not connected to) a computer system. As with our editing, you could be “online” just by being connected to a computer system. Our research and publication was online in the slightly newer sense of being connected to the Internet.

  For the newer editors, who hardly knew the old ways, online editing was the only way forward. They were used to manipulating their mobile phones and SMS communications in a way that the old guard, such as Ed and myself, were not. One or two editors were unhappy with the way things were going. You could see that they wouldn’t last. But I was excited by the possibilities, and with working out how to make those possibilities into realities. In all areas of our work (editing, research, publication), I wanted to find ways of retaining the standards of the past, and maintaining the intellectual rigour and discipline of what we were achieving, while making the process of capturing those standards easier and extending exponentially the profusion of ways we could display, present, analyse, and reuse our data by storing it online.

  Researching and editing online created new issues, of course. New editorial and logistical problems arose daily that we had to address and solve. Different research databases had different kinds of search software and held their text in different ways. Working with them wasn’t simply a question of keying your request into an anonymous search engine. Take a simple example: when researching words, we continually wanted to search for two-word compounds in modern and historical text. But the problem was—or at least one of the problems was—that they could be found in print in any of three different forms: e.g., table cloth, table-cloth, and tablecloth. The quick-eyed amongst you will notice that sometimes the term is written as one word, and sometimes as two, and sometimes gracefully split by a hyphen. And surprisingly enough, this doesn’t only happen with the word tablecloth. It happens all the time. So when dictionary editors were searching computer text in those old days, they typically had to perform three separate searches to find all the relevant examples they needed. This was before we were all familiar with repetitive-strain injury.

  Some of the high-use research databases were controlled by us (our reading-programme database, for example). For these databases, it seemed a small thing to write a tiny computer program to allow us to search for all three formats at once, and indeed it was. But as with any simple but crucial development (I’m giving this one more credit than it’s due, as I devised it), it’s not the program that is important, but the very first spark of the idea. I remember the jaw of the editor of Toronto’s Dictionary of Old English bouncing on the ground when I explained to her that we could do this. She put her programmer on to replicating it that very afternoon.

  Alongside the innovations in research and editing, we were still promoting the transfer of the full OED to the Internet, both as a means of searching and displaying its contents, and as a way of publishing our new and revised entries incrementally. These ideas would involve the University Press in a major change of plan, and one where failure might prove embarrassing, if not catastrophic. But over the previous decade and more, the OED had attracted risk, and had survived. We argued that online publication was a new and justifiable risk, and in fact a risk that the dictionary had to take to survive.

  The prototype online OED devised in the mid-1990s was our starting point. We drew up plans to rewrite the software to industry standards (about which I knew nothing, but which was clearly the way to go). The plans involved selling the online version, as the University Press had to recoup some of its massive investment. Would universities, businesses, the public, take up a subscription scheme? And now at last we had our own marketing director, Susanna Lob, dedicated single-mindedly to convincing the universities of the world (and anyone else who wanted access) that they needed to subscribe to the OED Online. This wasn’t the sort of thing that university libraries at the time wanted to do with their annual budgets, but even they could see the wind was changing. From our point of view, we weren’t used to marketeers who understood and appreciated what the OED was up to in the way that Susanna did. But she used to say that selling the as-yet-non-existent OED Online to librarians was the easiest job she’d ever done. It more or less sold itself. People immediately sensed the possibilities—librarians could see that this wasn’t just a dictionary, but an opportunity to introduce their users to a journey through language and through the differ
ent chambers of knowledge. It was considered—by those in the know—that, rather than try to attract all types of users in the first instance, we should concentrate on whole-university and whole-college subscriptions, which would be purchased by librarians. We were experimenting as much as they were, and starting there would give us valuable experience on how to market the dictionary online in the future.

  We went on a focus-group tour across the United States in 1999 to demo what we had and to seek comments for last-minute changes. It will be a long time before I forget the audible gasp that erupted from a focus-group audience in America when I clicked a key and linked from one of the OED’s millions of quotations to a page showing the relevant passage of the book from which the quotation was taken. That sort of thing is easy now, but back in 1999 it was cutting edge.

  But the biggest change for us, editorially, was that we would no longer be running a project where nothing was published until the whole work was completed. It was hard to motivate editors who might not still be on the staff (through old age or infirmity) to see the fruits of their labour. By publishing online, we could now be publishing incremental updates to the dictionary at quarterly intervals for the foreseeable future—even after the main revision was done. No other dictionary had plans like this. It was what we should have been doing, and at last the University Press was behind it.

  Our plan was to go online in March 2000 with the full Second Edition (1989) and the very first instalment of newly revised entries. With a growing sense of concern, we looked at those early entries from M which we had edited in draft over five years earlier. Sure enough, it was clear we needed to rework them significantly before publishing, as much more information about the M words was now available on the Internet. We needed to find that data before our reviewers and users did. We had about a year left to bring this about, and so we had to get our skates on.

  Fortunately we had a cupboard full of skates, and we set about brushing up those old M entries. We decided that the first instalment would only include a thousand new entries, which didn’t take us far into M, but would serve as a clear indication of things to come. The Internet was transforming what we knew about our words, and I enjoyed pushing mad scientist back from 1940 to 1893, and finding the Mafia in Sicily in 1866 (rather than 1875, as The Times had formerly led us to believe). In the changed world, we could never have published those entries as we had first left them—although they had seemed fine at the time. There were new standards now.

  We were also drawn down some curious sidetracks. Since the publication of the Second Edition of the OED, the whole project was becoming more visible to the public, and some companies had done their best to clamber on to the bandwagon of the dictionary’s success. A couple of these ambushes were quite extreme. At one point, the University Press’s offices in Oxford were the object of a small demonstration by a potato company on behalf of their wards (the potatoes themselves). The argument—such as it was—ran along the lines that the OED was disrespecting the potato by including an entry for couch potato. Clearly, the demonstrators were people, not potatoes, as the potatoes were far too idle to get off their sofas in their own support. But things died down once the company had achieved whatever publicity it was after. We fearlessly refused to budge from our position that people had the right to read an entry for couch potato in the OED, and we returned to our lairs wondering whether to seek revenge by downsizing our entries for chip and spud.

  We had a similar altercation with some purveyors of lettucy salads. The company’s representatives complained that the OED’s definition of salad regarded it as a cold collation, whereas they (marketeers of salads of all kinds) wanted us to recognise the existence of the warm salad. Our entry for salad had not been updated since 1909, so it was not surprising that it was a bit out of date. Updating it was, after all, the purpose of the new edition we were all working on. This time we deflated our demonstrators by agreeing with them.

  When I had first taken up my desk at the OED in 1976, life was very much as it had been in Victorian Oxford. Editors researched the history of their words by pacing around the reference shelves and the card indexes of the dictionary department, and wrote hopeful little research-request notes to colleagues working in libraries in Oxford, London, New York, Washington, and elsewhere. Definitions were written out on index cards, and batches of cards were sent off at fairly reasonable intervals to the printer to be turned into OED pages.

  March 2000 arrived, and the University Press’s publicity machine geared up for the OED going online. The University Press had little or no prior experience of running a campaign around an online product, but its marketing and publicity specialists were never short of new ideas. Our publicity campaign was led by me discussing the dictionary with a somewhat precocious five-year-old descendant of the original editor of the OED. The small Murray had nothing new to say about the dictionary, but he smiled cheerfully for the photographers, and demonstrated continuity, and friendliness and charm, and whatever else the publicity machine had hoped for.

  We had wanted to open up the dictionary—to make it easier to use. We had wanted to open up access to a wider readership. We had wanted to incorporate the newly accessible information about language from the Internet. We had wanted to publish the Third Edition of the dictionary. It’s not as if everything we had ever wished for had come true, but we were on the way—and we had to let the dictionary find its place in this new world. I had spoken to the tiny descendant of the original founding editor. It was March 2000, and the dictionary was at last launched on its new life online.

  TWELVE

  Flavour of the Month

  Once we had gone online, we suddenly became flavour of the month within the University Press. The dictionary was certainly not making a profit, but we were at last returning a ripple of revenue to the organisational coffers rather than sustaining the generous losses that the dictionary had traditionally delivered ever since it was first published in 1884. The fact that almost every university in the known world—or at least the English-speaking part of it—was subscribing to the OED Online meant that the future of the dictionary suddenly felt more secure. I should really have realised that we were starting to be regarded internally as another “brand,” rather than as a research project, and that this might in future make us susceptible to “brand management.” But as usual I missed this.

  It is not unusual for expressions to arise quietly in particular contexts, but then to attract public notice to the extent that they drift into the mainstream. According to the dictionary, we have the Illinois Association of Ice Cream Manufacturers to thank for the expression flavour of the month. The Ice Cream Review for September 1946 (yes, we read everything) reported that the Illinois ice-men were giving “serious study” to a proposed “flavour-of-the-month” programme for 1947. Apparently it wasn’t until the 1970s that we even thought of using “flavour of the month” in any context outside ice cream.

  Although the OED Online had solidified our position within the University, it also created new problems. By daring to publish and update the dictionary online—and especially by keeping entries continually up to date once they had been revised and published—we were taking something of a gamble that people would be brave enough to accept that there is no absolute truth about language; that as soon as we published an updated entry, it was out of date and liable to change. Scholars—or just ordinary members of the public—could mail us with even newer and better evidence that they had managed to squeeze out of their memories or their databases. And they did. No sooner had we announced to the world that the earliest record of the word numismatics dated from 1803 than we were informed that it also had occurred in 1790—as we hadn’t thought to check in Adolf Ristell’s Characters and Anecdotes of the Court of Sweden of that year. But this didn’t annoy us—it was what was supposed to happen. We had initiated what was effectively an international challenge to beat us—and at the same time improve the record of the language. It was the sort of democratic user e
ngagement I thrived on. We didn’t seem to attract time-wasters, either. So we made these changes to the database as soon as we could, to keep the dictionary dynamic. The alternative was to stick our heads firmly in the sand, draw a firm line of demarcation (once again in the sand), or do whatever other things people did with sand to show they didn’t want to move with the times.

  This race to keep up to date created an enormous amount of editorial spade-work for us. While the Internet had made certain kinds of research much easier and quicker to accomplish, there were still a great many issues that couldn’t be solved by the Internet. Sometimes we just had to pull on our old detective leggings and venture out into the real world in search of information. One day we were working quietly away in the letter P when we came to revise the entry for pal (in the sense of “a friend” or “mate”). The OED had published its original entry for the word back in 1904. It had found that, in the vocabulary of the day, pal was a “Gipsy” word that had reverberated across Europe as the travelling community had spread it generously around, taking it into other languages as well as English. It had, the OED told us, entered English in the late seventeenth century. First of all, it had been used for the most part in quite disreputable company: a pal was a highwayman’s accomplice, for example, or a thief’s associate. It was only later, in the late eighteenth or early nineteenth century, that it settled down to have the “chummy” sort of meaning we know today.

  The very first quotation example for the word pal in the dictionary proved to be a problem. It dated from 1682 and was supposed to mean “a criminal accomplice.” This example didn’t come from a printed book, but from a set of legal depositions or statements kept in a manuscript in the Hereford Diocesan Records. At some point in the dim past, a Friend of the OED had sat for hours with his slippers and pipe in a comfy armchair in the vestry at Hereford Cathedral, or wherever these records were kept in those days, and had read through legal deposition after legal deposition looking for word usages that might be of interest to the OED editors back on dry land.