Swiss 'strippers' remove leaves

The lexicographer's lot is not a happy one. No dictionary can be complete because language is forever changing, evolving. No dictionary seems to be internally consistent, however much love and labour has gone into it. And all manage, somehow, somewhere, to frustrate their users.

Look up " étrive " in the French-English volume of this hefty unabridged edition of the Harrap French Dictionary : it is a nautical term that means "throat-seizing". You expect to find "throat-seizing" in the English-French volume, translated as " étrive ", but it is not there. What is throat-seizing anyway? Seeing that it is absent from the single-volume Collins English Dictionary , you turn to the ultimate reference, the unabridged Oxford . It is there, of course, not as a separate entry such as "throat-band", "throat-boll", "throat-cutter", but deeply buried under "throat", between "throat-rupture" and "throat-strap". Having prised it out, you are rewarded with a cryptic " Naut .: see quot." So you scan down to the bottom of the column for the promised quotation, noticing that they are in alphabetical order, which helps a bit, and soon you find it in the next column: "1867 SMYTH Sailor's Word-bk., Throat-seizing , in blocks, confines the hook & thimble in the strop home to the scores." Turning in despair to Le Petit Larousse Illustré you learn that " étrive " means either the angle of approach in a nautical manoeuvre, or the tying together of two mooring lines crossing each other, and you are left wondering which meaning "throat-seizing" covers, if any. The lot of dictionary users is not a happy one either, but if you are inclined to burn the Harrap for such lapses you must, in all justice, burn the Oxford too.

With this warning in mind, flip now through the two volumes of this new Harrap . Its clean layout and innovative features jump at you at every page. Open the English-French volume at, say, page 835. A boxed note prominently titled "OXBRIDGE", quite impossible to miss, details in French the historical background and the cultural implications of the word. This note complements the short entry for "Oxbridge" itself, just above. There are many such notes, from "THE ABDICATION CRISIS OF 1936" on the English side to " ZONE " on the French side, via "CARNABY STREET" and "SAINT-GERMAIN-DES-PRES".

This, then, is not a mere dictionary, it is also a historical and cultural encyclopedia. Is this appropriate in a bilingual dictionary? Emphatically, yes. A language is not a mere collection of words and grammar rules any more than a human being is a fleshless, soulless skeleton. Proverbs and literary allusions are displayed in grey-tinted boxes, with ample explanations. Thus Mark Twain's "Rumours of my death have been greatly exaggerated" and the very much more recent " Les feuilles mortes se ramassent à la pelle ".

Translations of the titles of works of art, films and songs are provided, highlighted by telling icons: two masks for theatre plays, a frame for paintings, an open book for novels, a clapperboard for films. An overkill? Not in the least. How would you translate The Taming of the Shrew into French? " Le Domptage de la Harpie " at best, " de la Musaraigne " at worst, certainly not " La Mégère Apprivoisée ".

This concern with helping to bridge cultural differences seems to extend to the "normal" entries as well. To the French, English cooking is bland, nay, awful and boring, and Scottish food unheard of. The Collins-Robert Dictionary tells you that haggis is "haggis ( plat national écossais )". The Larousse by Marguerite-M. Dubois is a bit more informative: haggis is "haggis, panse de brebis farcie ." But only the Harrap shatters your Gallic prejudices while tickling your taste buds: "Haggis is made of minced offal, and very spicy."

Acronyms and even personal names are also there, to remind you that Lyons is Lyon in French, Gaddafi is Khadafi, an ATM is a GAB, RAM is MEV, but, curiously, ROM is ROM. You will find regional variations carefully recorded, between British, American and Australian English and French, Canadian and Swiss French. How else would you know that " une action " in Switzerland is " des soldes " in France, and " un char " in Montreal is what they call " une voiture " in Paris but is a game akin to nine-men morris in Geneva? Or " une effeuilleuse " is not at all a stripper (" strip-teaseuse ") in Switzerland, but "a woman employed to strip vines of unwanted leaves".

Slang and argot are not forgotten. It is a difficult choice here. How far should you record this eminently ephemeral form of speech? This dictionary lists all the meanings of French "coco", some of them already obsolete, from coconut to a sort of haricot bean, via egg, communist, cocaine and its original meaning (a drink made from liquorice and water), but it stops short of "petrol" and "booze". There are some words of verlan (French backward slang, from " l'envers " pronounced à l'envers , that is, back to front), but only those that, like English rhyming slang, cannot be guessed at, such as " meuf " ( femme , woman), " keuf " ( mec , bloke, fellow), keuf ( flic , "copper"). On the English side, there is some rhyming slang, such as "cocoa" (think so), and "china" (for china plate = mate), but without any explanation as to their formation, which is a bit of a pity, and quite a puzzle to any French speaker who has never heard of rhyming slang - there is nothing similar in French. This should see you through most modern thrillers and comic strips. If it does not, then you will find it all in the Harrap Slang Dictionary/Dictionnaire - quite an outstanding, and affordable, paperback.

This, then, seems quite a comprehensive work. But is it, as a French person would say, "convivial" (user-friendly)? Words such as "get" have long been the bane of learners of the English language, of users of bilingual dictionaries, and of lexicographers, extending over page upon page upon page, through which you have to wend your way painfully.

In the Harrap , the entry "get" is introduced by a menu of 25 possible translations, from " recevoir " to " réussir à ", each referring you to a numbered sub-entry. A helpful scheme, inspired from computer menus, but all is not well. Some translations point to several sub-entries, and this without a hint as to their contents. Thus " recevoir " points to five scattered sub-entries, and it is only after referring to them one by one that you discover that 1A(a) is " recevoir " meaning "to receive a gift, a letter, a phone call", that 1A(d) is " recevoir " meaning "to obtain in exchange", and so on. Dubois's dictionary, published by Larousse in 1981, adopted a slightly different approach, where these menus were actually tables of contents. Sub-entries were numbered sequentially, with both number and keyword in bold face, which made them very easy to spot. The menus were likewise sequential, and following each sub-entry number was a word or a category summarising the meanings contained there. Thus, for instance, you had "1.2. Se procurer, acquérir " and so on to "20. ECCLES." where you would find the American expression "to get religion". The layout was better than Harrap , where sub-entries do not stand out sharply enough.

It is not only words such as "get" that are the curse of the English language, it is what comes after them, in the form of get off, get up, get down, get... Here, Harrap, like many other bilingual dictionaries, choses to list those compounds, those verbal phrases, alphabetically, after the main entry for "get". Not only did Harrap avoid here the silly mistake of the 1979 Collins English Dictionary where "Gethsemane" occurred between "get down" and "get in", but they went as far as listing separately such phrases as "get away" and "get away with". A laudable innovation.

Can this dictionary, then, be recommended without reservations? No. There is room for improvement, and there are errors. One glaring error is the entry for " sapeur ". A sapeur is a "young, well-dressed African man". True, with reservations: a black African man: the term is not used, in my experience, of North Africans, Algerians or Moroccans. But this is recent argot ( c . 1970), from " se saper " meaning "to dress up". The other, original meanings are missing. A sapeur is first a sapper, then a military engineer. The culprit here is certainly computerisation, as the editors explain: "The computerisation of the editorial process, and the existence of texts in searchable databases, have allowed this dictionary to be produced in a fraction of the time required to write the Standard ." Yes, computers are a boon. They are also a bane. Look up "spring" in the English-French volume:

"spring ( pt sprang or sprung, pp sprung)". So far, so good. But what is this? "( season ) printemps ... ( device, coil ) ressort "? By courtesy of some programming glitch, spring, the verb, has been relegated to after spring, the noun.

There is nothing, however, worth damning in this dictionary. For, despite its occasional shortcomings it is (pardon my English) a damn fine dictionary, and it is only " des fouille-merde " (see page 532) like yours faithfully who are likely to find nits in it to pick.

Jacques B. M. Guy is a French-born computer scientist interested in natural language understanding, who holds a PhD in linguistics from the Australian National University.