The Word Detective

by John Simpson

  All of a sudden, the Shark was best friends with Ed and me. I knew this for certain when he started asking me to play for his village cricket team in Oxfordshire. It was rubber-stamped in gold when he asked me to turn out a few times in the annual Oxford Publishers v. London Publishers cricket match held on the John Paul Getty estate at Wormsley in Oxfordshire. It was good to be asked, but technically speaking I was at the edge of my competence—much as I loved the game. Each team was bolstered by one or two international players of the sort you normally only see on television. The games were generally a ritual humiliation for the Oxford crew, as you can imagine that there are rather more publishers in London than there are in Oxford, so they had a far larger pool to draw from. The Shark kept wicket and nonetheless tried to encourage us to put on a good show.

  There was one event that happened at that time that symbolised two things for me rather violently: firstly, the way we were shuffling off the old traditions of printing and dissemination in readiness for the new; and secondly, a worryingly dismissive attitude held in parts of Oxford towards the icons of the past. The Shark and Ed and I had joined one of the educational visits made from time to time by Oxford editors into the print-works at the back of the University Press. Although some of OUP’s printing was done by third parties around the world, many texts were still printed in Oxford. In those days the enormous printing machines stood two storeys tall, like the massive machines you see in old clips of newspaper offices, rattling out sheet after sheet of text for binding and then publication. It was a very impressive sight, and one that was swept away by new technology and working practices at the very end of the 1980s.

  We were walking with our group along a narrow corridor on a gallery halfway up the print room when we came across an old man throwing squares of metal into a roaring furnace. On further investigation, it appeared that this was not some job-creation scheme, but part of the Oxford printer’s recycling initiative. In the past, huge storerooms had been stocked with these old copper printing plates. Over time the plates grew worn and were no longer of any use. Eventually they were recycled by being melted down and reused.

  To our alarm, we discovered that the heavy copper plates which the lonely old man was committing to the inferno were in fact printing plates once used to print the OED. Needless to say, Ed and I were horrified by what seemed to be the wilful destruction of the dictionary’s history. The Shark was also shocked: though doubtless he shared our concerns, he clearly in addition lamented the loss of the windfall revenue that this Götterdämmerung of the OED’s heritage would precipitate. We halted the burning and salvaged what we could. To recoup some revenue, the Shark sold some of the plates through the University Press’s bookshop in Oxford. Lucky lexicographers, I well remember, carted others home on their bicycles.

  Every word has a shape, or profile, and the historical lexicographer has to be attuned to this. When we looked at AIDS, we saw that it had arrived very recently in the language, had one basic and very strong meaning, and that it had developed a wide-ranging collection of compounds, such as AIDS virus, etc. If we visualise it as a tree, it has a short trunk and a large canopy of leaves. Words which entered English in the Middle Ages or earlier have had a long time to develop, and their profiles often resemble old oak trees, with extensive ramifications (literally, branch-like structures).

  Inferno is an important, powerful word today, but curiously, it does not have an extensive profile. According to the OED it has a single meaning, and no compounds worthy of note. It is more like a nursery sapling placed in the ground awaiting the chance to develop than a bona fide tree. And the reason for this odd profile is that inferno entered English late (1834), and as a result of conscious literary borrowing from Italian. We knew the word inferno before the 1830s, but only as the title of the first part of Dante’s Divine Comedy—the section where Virgil takes Dante down through the nine circles of Hell (the Inferno), to find Satan bound in its very heart. But until the early nineteenth century we do not seem to have thought to use the word for Hell itself (or anything resembling it, such as a fiery furnace); we reserved its use for the title of Dante’s text (which alone wouldn’t merit inclusion in a dictionary—which is not intended to be an encyclopaedia of knowledge). Inferno derives directly from Italian inferno, itself developed—after the Fall of the Roman Empire, from late Latin infernus (= Hell).

  Why didn’t we use inferno earlier in English? Well, being a culture with a long Judaeo-Christian tradition, we already had a long stream of words for “Hell” (the OED lists ninety-seven, starting chronologically with the good Old English Hell, then perdition, welling woe, Tophet, Avern, Hades, Sheol, etc.). There wasn’t necessarily room for a new one until Gothic literary types from the early nineteenth century saw fit to introduce another.

  On the other hand, inferno slid into English relatively easily, as we were already perfectly familiar with its relations. Inferno is a cousin of the adjective infernal, which we absorbed in the Middle Ages from French infernal, where it had developed by the twelfth century from Latin infernalis (“of the realms below”—i.e., Hell). Italian inferno came by a separate route from Latin to that taken by infernal. Infernal is a word with a strong profile in English, to the extent that we sometimes forget its link to Hell when we use it in exasperation to mean “blasted” or “confounded.”

  With the new OED computerisation project approved, and with the Shark on our side, the message reached Ed (as chief editor) and me (as New Words editor) back in our offices that if we could manage to pull off the coup of computerising the dictionary on schedule and on budget—which became the mantra of the day—then we could more or less choose what we wanted to do in the future: the resurgence of the dictionary would be assured, as would our place with it. Promises like that have a short half-life, but we believed them enough to want to get on with the job.

  The OED had reached a turning point. Over the past year our new managing director had galvanised support for the ageing and rather infirm dictionary both within the Press and (more importantly) more widely in the University. We had worked on plans to transfer the dictionary on to computer, and had found that its contents would fit reasonably easily into a database structure. We—and here I need to give much more credit to Ed and the Admiral than to myself—had conjured up project plans which the venerable scholars of the University were prepared to back. Before the Delegates of the University Press, orchestrated by Richard Charkin, had voted funds for what became known as the New OED project, we had been given to understand that there was a serious possibility that the grand old OED would simply be allowed to sink into obsolescence and decrepitude. The Press had no stomach for other long-drawn-out supplements. But now, at last, there was a chance that the dictionary might just pull through.

  SEVEN

  OED Redux

  The new OED computerisation project received the official blessing of the University Press, and that approval came with an understanding—agreed by all parties—that the Second Edition of the OED (as it came to be called) should be published in 1989. So we had about five years—working with IBM, International Computaprint, and the University of Waterloo—to put the text of the dictionary and its supplement on to computer, to merge these two texts together into one seamless whole, and to do whatever else was needed to publish the dictionary in print and on time. It would not be a new dictionary editorially. The key point for the future was that the merged text was to be published from the computer database, not from copper plates, or by phototypesetting from paper-based camera-ready copy. This had never been done before for such a large reference work, so we were venturing into the unknown. Once the Second Edition had been published in print, we could then think about disseminating the text as a searchable CD. That would come after the print publication, but as part of a separate project.

  We set about announcing the project to the scholarly community, at first through journals, newspapers, and budget-free word of mouth. (You will see that the syntax of word of mouth is extr
aordinarily un-English; not surprisingly, it is a straight translation of the medieval Latin verbum oris, first recorded in English in the 1450s.) I think someone had already booked the dining room in Claridge’s Hotel in London for the launch party in 1989, just so we appreciated that the date was serious and inflexible. Sometimes it seems that publishing projects creep on beyond their planned completion time. The Shark and the Admiral constantly reminded Ed Weiner and myself that that wasn’t an option for us.

  In order to help us concentrate even better, we were moved into new offices—new offices being generally thought to be motivational. And so the New OED project was now proudly housed in what was then known in the University Press as “The Old Post Room.” This was not a title of antique distinction, like “The Old Rectory” or “The Old Bakery,” accorded the room to remind us of its picturesque aspect and quaint and charming history. It was called “The Old Post Room” because it was a big, square, utilitarian room which used to be the Post Room, in which old-fashioned incoming post or mail was sorted and delivered to the Press’s departments, until the Post Room staff found more suitable premises in the basement of the Press. At this stage there were never more than a handful of us, and I was officially moonlighting between the OED’s new offices and our old offices back in St Giles’, where the New Words group was still located. Our new premises had little character and we had no budget to create one.

  For me, the OED was at last heading in the right direction. It wasn’t that I thought the Second Edition was the be-all and end-all of lexicography. In truth, for me it was a mechanistic project—though tough and imaginative for all that. I wanted to see the text of the dictionary safely housed on computer so that we could start updating the OED comprehensively, rather than making piecemeal additions through supplementary volumes. It was all, of course, still very uncertain. Ed and I didn’t really know whether the University Press would countenance a big update project in the 1990s. Even if it did, we didn’t really know how to carry out such an extensive historical overhaul of a dictionary that had arguably been neglected for decades. We hadn’t sat down and thought out in detail how we might adjust the editorial policy for a modern national dictionary.

  But first of all, we had to get the dictionary on to computer by 1989. If we couldn’t do that, there was little point in trying to write a happy ending beyond. I thought of the project until 1989 as Phase One; Phase Two was our secret dream of what we wanted to do with the dictionary in the longer term. We thought that perhaps we could engineer an OED redux—an OED reborn. If you wait long enough, most things make a comeback.

  I had worked on the curious word redux in the late 1970s, for the Supplement to the OED. The original edition of the OED contained just a single meaning for the term, magisterially and rather obliquely defined as “Of crepitation or other physical signs: indicating the return of an organ to a healthy state,” and squeezed in between the words reduviid (“belonging to a family of predaceous bugs”) and redvore (“a variant of radevore; the precise sense is not clear”). The editors found that redux was a Latin word that had been borrowed directly into English by learned medical men during the Renaissance. By the twentieth century the OED editors knew that this old medical term was falling out of use, but that a new use—reborn from the original Latin and easier for the ordinary person to understand—was now known, from the titles of novels such as Anthony Trollope’s Phineas Redux (1873) and John Updike’s Rabbit Redux (1971): “brought back; restored, reborn”—as when a character in an earlier novel by the same author reappears and is “reborn” in a novel later in the series. It was naturally a usage associated mostly with literary types. I liked to think that we were going to be taking the dictionary through a process of rebirth.

  Quietly, under cover of the Old Post Room, we planned a revolution in lexicography. This was the office where the Second Edition of the OED (OED2) would be prepared, spun from the text of the old dictionary, and nursed through to publication by Ed, myself, and Yvonne Warburton. We were joined by several other hand-picked colleagues, including Julia Swannell (silently present earlier in this narrative as the person who was actually appointed to the post on the Pocket Oxford Dictionary for which I long ago applied) and Veronica Hurst (later to become the OED’s bibliographer). If we could pull it off, then dictionaries could take a quantum leap to the benefit of both readers and editors. Up until now, dictionary readers had been restricted to “single look-up” (finding out information about the specific word they wanted to look up) and serendipitous browsing; now they would be entering into what we thought of as a new language universe, based around the OED. For editors, we would be creating a massive, dynamic, and updatable language resource: we wouldn’t be restricted anymore to alphabetical, paper-based updating, but could edit in whatever mode we wanted. That was the hope, anyway.

  You need to bear in mind that this emphatically wasn’t a project to put the dictionary online in the way we understand that term today. Going online wasn’t something that anyone knew about back then; the Internet didn’t exist if you were not part of the US military. The Shark’s big project involved putting the dictionary on computer and seeing how the land lay computationally when that had been achieved. The real possibility of online publication came later. We talked then about computerising the dictionary. It sounds old-fashioned now: later on we understood that we were digitising the dictionary. It’s much the same thing.

  Little words like same cause enormous problems for lexicographers. This is because, however it is used, same always means the same thing. It is a word with Germanic antecedents which has had a long time to develop in English, as first records of same date from the Middle English period (though Continental parallels indicate that it is likely to have been known but unrecorded in Old English). The OED defines it with a broad, catch-all sweep as “not numerically different from an object indicated or implied; identical.” You can see it’s having difficulties from the very start, because it begins its definition with a negative (“not numerically different”). That is generally regarded as bad practice.

  But lexicographers like to be helpful. They know that the contexts that same can be used in differ widely. Here are two sentences: “The same observations are true of all other contracts similarly circumstanced”; and “All the planets travel round the Sun in the same direction.” In the first sentence, “the same observations” carries with it a reference backwards in time (probably earlier in the paragraph) to other, comparable “observations.” In the second sentence, the planets are orbiting in the “same direction,” not in reference to anything in the past, but in reference to each other. Some people like this sort of distinction, and they become lexicographers.

  Here is a slightly different poser, but an easier one. Can you see how you might go about describing the difference between these two usages: “the sailors in the fleet all received the same pay as the soldiers”; and “the Greeks and Macedonians . . . looked on the Egyptian Ammon as the same god with their own Zeus”? Same has a very similar meaning in both sentences, but in one it is used in a syntactic construction with as, and in the other with with. There is enough of a chink of light between the two for the kindly lexicographer to wish to explain the history of each separately.

  The end product is that if you examine the documentary records for same, you can separate the data into tens of different contexts and syntactic constructions, and define them all by describing the context and not the meaning. This happens with lots of little words, such as prepositions and conjunctions, and also with many adjectives and adverbs. Here is how one of those examples of same we just looked at is defined: “Followed by as. Now the commonest construction.” If you look at entries such as same, you’ll find this sort of technique used regularly. It reminds you that lexicography and language aren’t only about meaning, but also about context.

  Ed had relocated his office to the Old Post Room a few months before me. With the assistance of various computer types from left (OUP), right (IBM), and ce
ntre (Waterloo), he had devised a routine for transferring the OED’s text safely and elegantly on to computer in preparation for OED2. Under his oversight, the twelve volumes of the very old dictionary (1884–1928) and four volumes of the Supplement were being keyboarded in America to be ready for the publication of OED2 in 1989.

  The very basic decision to have the full text of the dictionary keyed manually (rather than automatically scanned) on to computer was one that the project team had made early on. But before we made it, we listened to advice from industry insiders. Several told us that we would be daft if we didn’t scan the text on to computer (using optical character recognition), rather than waste time and risk introducing error by having it painstakingly keyed by hand. We always gave these obliging consultants a few pages of the OED, and said we would be delighted to check over the results when they had run it through their state-of-the-art text-scanning equipment. Curiously, we never saw any of them again. Perhaps some went mad, and maybe others were sighted selling ice creams on the sea-front in Brighton. But they certainly didn’t come back to us with scanned text that we could make use of.

  In the end, we had the whole text keyed on to computer by hand. But even then it transpired that there was no company in Britain big enough and courageous enough to take on the job of keying the entire text of the OED. Fortunately this was an activity that one of our potential partners offered to do: International Computaprint Corporation, in the home of the brave and the land of the big (Pennsylvania). They had our text keyed down in Florida by around 150 of their keyboarders, working over eighteen months. We required a high accuracy rate: no more than 7 errors in every 10,000 keystrokes. That in itself was suitably challenging. And at the same time the typists had to introduce tags before each element of the dictionary entry (the definition, the etymology, etc.), so that a computer could recognise it for what it was.