Dictionaries

English lexicography​ knocks Johnnie Walker into a tricuspidal fedora. Over four hundred years, and going stronger than ever.

Of course, in the 16th century the market was for ‘bilingual’ dictionaries (especially Latin-English). We had to wait upon Robert Cawdrey in 1604 for a ‘monolingual’ model – aimed at ‘Ladies … or other unskilfull persons’.

But the principles and goals are essentially the same. We don’t look up door to find that it means the chunk of material that seals off rooms and fridges. We know that. We look up meretricious for its meaning, fuchsia for its spelling, controversy for its pronunciation (to correct somebody else’s). The words from Latin (and so forth) today correspond to the words in Latin at the time of Hooker and Shakespeare: the language most familiar to the most educated, least familiar to the least. This is something of an oversimplification, but not all that much. The tradition is rich and unbroken: in Thomas Elyot’s Latin dictionary of 1538 we find aedificium ‘building’, in Bullokar’s English dictionary of 1616 the minimally anglicised edifice ‘a building’, in the 1979 Collins edifice ‘a building’.

For all their bright newness, dictionaries of the 1970s (such as the Webster 8th Collegiate, the Longman Modern English Dictionary, the revised Chambers, the 6th Concise Oxford, the Longman Dictionary of Contemporary English, the Oxford Paperback) are still basically concerned with translating a relatively foreign language into a relatively familiar one. And – despite the claims of radical differences – there is a striking resemblance between them. As between members of a family. Nor is this specially surprising. However hermetically distinct the financial structure of the publishing firms concerned, the people who actually make the dictionaries are a fairly small group of people who know each other and who are as mobile as musicians. If Solti is with the Chicago Symphony one week and with the LPO the next, so also do the professional lexicographers like Clarence Barnhart, Sidney Landau and Lawrence Urdang move from one dictionary house to another. Oxford (with R.W. Burchfield and John Sykes) is comparatively stable.

When work began on the new Collins, Paul Procter and Della Summers were young conductors under impresario Urdang, and they later moved on to make dictionaries for Longman. Patrick Hanks was recruited to complete the Collins when he had finished a somewhat similar job for Hamlyn. Both Urdang and T.H. Long were earlier on the Random House Dictionary. All very cosy. But while it desirably makes for shared knowledge and a solid tradition (a euphemism, some would say, for massive reciprocal plagiarism), it is not exactly a prescription for exciting new departures.

Nonetheless, the marketplace demands that for each new dictionary, claims must be made about its uniqueness. For Collins, these rest upon coverage (words from wherever ‘English is spoken as a native language’) and size in relation to provenance (the biggest English dictionary ‘to be originated in Britain’ since 1933). This is technically true no doubt, though given the American leadership of Urdang (and Long), the American sources that constituted the lexical materials, and in any case the essentially amphi-Atlantic nature of all such enterprises today, the claim is perhaps more one of investment than lexicography. And ‘biggest’ is by no means as simple a measure as it sounds.

There is certainly some additional coverage, with a fair number of words that I haven’t come across in earlier dictionaries – or anywhere else if it comes to that: for example, grovet (wrestling). But the size (nearly twice as ‘big’ as the Concise Oxford) is not primarily on this account. It is achieved in part by including people, places and the like (about one sixth of the whole book, I estimate), and in part by avoiding space-economisers like swung dashes and abbreviations. The latter principle has the laudable aim of making the entries more quickly comprehensible but the cost is considerable – especially in making etymologies (surely used chiefly by readers who can cope with ‘OE’ and ‘OHG’) far lengthier than usual: ‘from Old French grouchier … compare Old High German … ’ The former is also for the convenience of Everyman, for whom the dictionary may be his sole reference book.

Combining the dictionary and the encyclopedia is no new idea (Cockeram did it in 1623), and the distinction between the two types of information is by no means as clear as conventional wisdom would have it. Indeed, it is least satisfactorily upheld by the lexicographers who uphold it most primly. Thus the Oxford and Webster-Merriam tradition is to include Kafkaesque but not Kafka, though if ‘linguistic’ principles were strictly observed, only -esque (‘in the manner of’) should be listed. Nor is the distinction any easier to maintain with words having no initial capital. Collins defines motorcycle as ‘a two-wheeled vehicle, having a stronger frame than a bicycle, that is driven by a petrol engine’. Much of this is obviously ‘encyclopedic’ (indeed incidental), threatening the definition of other words with hair-raising implications which fortunately are not often realised. (We are not told, for instance, that a bus has a larger engine than a car.) The semanticists have long grappled with such problems. What is the linguistic meaning of carrot or radio as distinct from the encyclopedic meaning? We competently choose, use and refer to these things without necessarily knowing that the one is an ‘umbelliferous plant’ (Concise) or that the other ‘demodulates electromagnetic waves’ (Collins). Nor do these ‘meanings’ help to explain why we can talk of carroty hair or a radio personality.

So Collins is sensible enough in effectively dismissing the distinction and putting Guillaume de Lorris between guileless and guillemot. The trouble is knowing where to stop. It is hard enough to establish which words and which meanings of words to list, but if a careful search (especially of other recent dictionaries) gives assurance that every word in yesterday morning’s paper will be securely in and adequately defined, the lexicographer has some kind of rough-and-ready guide. But will the names of every person and place in yesterday’s paper provide a similar check? Obviously not: yet on what principle shall we include Margaret Thatcher and Bessie Smith (both in Collins) and exclude some John Smith who had to be rescued after a fall in Snowdonia?

The answer is, of course, common sense – and on the whole there seems to have been a good supply in Aylesbury (largely, one gathers, that of Ms Lucy Liddell). Leaders in politics, writing, theatre, music, painting and sport get in; quite a number of relatively obscure figures from the past too (for example, Guiscard, the 11th-century Norman who became a sort of Sicilian Godfather); and a wide selection of places, from Cottbus on the Spree to Whitehall, both as a street and – I am happy to see – in its metonymic use as ‘(British) Government’.

Goodness knows what is meant by the intriguing claim that ‘these items play an increasingly important part in communication today.’ The growth of name-dropping? But inevitably in so subjective an area the coverage is uneven, and the chap or the spot you’re looking for may not be there. Comics come off worse than straight actors (Harry Lauder makes it but not George Robey, Tommy Handley, or Morecambe and Wise). So in other fields. Bobby Charlton is in; Geoff Boycott isn’t. Piggott but not Carson. John Williams (the guitarist) but not Shirley. Healey but not Howe. John Le Carré but not Naipaul or Storey. Bud Powell but not Sandy. We even have Bovril but not Oxo. Pink Floyd but not the Hallé Orchestra. The Rolling Stones but not the Amadeus Quartet.

These last examples are more than incidental in their suggestion of trendiness. Bob Dylan gets eight lines; Dylan Thomas four. George Solti, Colin Davis, Janet Baker and Joan Sutherland are all treated with tight-lipped and austere brevity as compared with Ringo Starr, Paul McCartney and Mick Jagger.

Even where relatively lavish space is given to unquestionably significant people, it is not so unquestionable whether the space has been well used. Writers and composers tend to have their works listed, directory fashion, rather than their work assessed. For example, Shakespeare gets 20 lines, but 14 of these consist of titles.

Yet the content of entries – whether encyclopedic or lexical – is clearly something to which the editorial team have given great attention. Two claims are made for the special quality of definitions: that they are in ‘lucid prose’ (p. xv), reflecting progress made in the study of semantics (p. vii); and that they are ordered with priority for the sense that is ‘most common in current usage’.

The first of these certainly represents a desirable goal. How well it is achieved is another matter. My impression is that the editors have been more successful with concretes than with abstracts and attributes: the definition of engine is shorter, easier and more effective than the one in the Concise Oxford. On the other hand, the converse seems true for the adjective enervate: ‘lacking vigour’ (Concise); ‘deprived of strength’ (Collins) with an unfortunate and unwanted suggestion of some external agency. But for the most part, I doubt whether the defining skills are better or worse than in the general run of recent dictionaries, and despite the claims about freshening things up, one is struck by the survival of the most traditional type of defining language. Often ignotum per ignotius. Take anus. The first dictionary I ever used (and still go back to occasionally: Ogilvie-Annandale in a ‘new’ edition of 1895) at least justifies its ‘inferior opening of the alimentary canal’ by first putting something a bit more straightforward. The Concise Oxford has only ‘terminal excretory opening of alimentary canal’ – which isn’t going to help the scared patient trying to make sense of something he heard the doctor say about his piles. Collins offers no breakthrough: ‘the excretory opening at the end of the alimentary canal’. The reader who understands the words in definitions like these is unlikely to be ignorant of the words they define. This is a trap laid not merely by the centuries-old fustian tradition but also by the lexicographer’s neglect of recent work in semantics and lexicology. By contrast, in the Longman Dictionary of Contemporary English, Paul Procter insisted on all words being defined within a restricted vocabulary, and this had the predictable effect of extracting elemental meaning in the simplest language. In consequence, anus is defined as ‘the hole through which solid food waste leaves the bowels’.

The second of the defining characteristics is more controversial. There are three fairly obvious choices in handling polysemous words. First, lay out the meanings in the order which is most explanatory semantically. This is the least popular option, doubtless because it is most demanding. But the second is in effect closely similar: lay out the meanings in historical order (thus frequently explaining how one meaning has developed from an earlier more ‘basic’ one). This is the principle largely adopted by Merriam-Webster, and the most serious objection to it is that it can obviously entail giving priority to a meaning that is now relatively unimportant. My favourite example is in Chambers (‘entirely new edition’ 1972) which for the first two meanings of sad gives ‘sated’ and ‘steadfast’, with ‘sorrowful’ not making it till the fourth line.

The third way is the one claimed for both the Concise Oxford and the new Collins: ‘commonest meaning first’. Superficially attractive for obvious reasons, this is probably the least satisfactory of the three. There is no technique currently available to establish frequency of meanings – even if we could agree on the principles and the type of discourse to be used for the inquiry. (For example, there must be plenty of people for whom ticket has ‘parking summons’ as the commonest meaning in ‘I got a ticket’, but ‘theatrical admission’ in ‘I’ve got some tickets’.) This is perhaps fortunate, since the principle would produce chaotically mystifying dictionary entries if it were seriously applied. But of course it isn’t. The Collins definition of paper begins with the ‘substance made from cellulose fibres’ (and not, say, with the sense of ‘today’s paper’). The definition of crane begins with ‘long-legged wading bird’, and I would bet that there aren’t many people for whom the ‘lifting machine’ isn’t commoner. The treatment of crash begins with the acoustic sense, though the compounds that follow (like crash barrier, crash helmet) might sufficiently suggest that the ‘destructive impact’ sense is linguistically dominant.

But I don’t want these remarks to sound too negative. For all that we may react against the brash stridency with which publishers plug ‘big is beautiful’ and ‘new is great’, the brisk market for dictionaries is healthy for lexicography. Signs of originality and progress are far more modest than blurbs and forewords would suggest, but they are real. And this is more true of the new Collins than of some other recent dictionaries. For a long time, ‘dictionary’ has meant ‘Oxford’ in this country almost as automatically as it has meant (the far more polysemous) ‘Webster’ in the United States. Oxford deserves its fine reputation – and, if for no other reason, it equally deserves more insistent competition than it has been used to. The various alternative approaches now being pressed by such houses as Longman and Collins will have a beneficial effect upon lexicography as a whole.