Two Canadian publishers rushed into print with weighty general-purpose desk dictionaries in the fall, some three months ahead of their 1997 copyright dates. The Gage Canadian Dictionary (hereafter, "G") boasts 140,000 entries including 13,000 added to its previous (1983) edition. The ITP Nelson Canadian Dictionary of the English Language (hereafter, "N") counts 150,000 entries in its first-ever edition. Both books highlight the word "canadian" on cover and spine, and in the swirl of media coverage that attended their launches both publishers emphasized their Canadianness.
They are equally hefty hardcovers, almost the same size, with handsome dust-jackets, but otherwise they are about as different as dictionaries can be.
N and G are both based on American dictionaries. The American bases exert an ineradicable influence on these Canadian versions. N is a newcomer, based on The American Heritage High School Dictionary (Houghton Mifflin). G has the advantage of having been Canadianizing its base dictionary for thirty years, to the point where its American source is no longer acknowledged.
Design: G is the more showy, with a red cover brighter than the red of the flag. N has a more studious air, with a sombre blue-grey jacket. It has thumb-indexing and speckled page-ends; G has neither. Bibliophiles will recognize N's trappings as the production values of a well-made book, but they carry a cost: a price $4 higher. G declares its price in a yellow flag on the front, N discreetly on the back.
Flipping the books open randomly, the contrast continues to be striking. N's lay-out is greyer, typed more densely, the typeface of its bold headwords similar in size to the definitions. G is whiter, with headwords that stand out not only in boldface but also in larger type than the definition. It is undeniably more legible, but this legibility has something of the style of a children's book; its large headwords are printed in a sans serif font that lacks class, and then the definitions switch to a smaller serif font. N uses the same serif font for both the bold headwords and the definitions. Right-justifying of the columns adds to N's greyness; G's columns are ragged on the right. N's entries have a hanging indent, which is a bit enlivening; G's are flush at the left margin.
The dense look of N is increased further by the most noticeable design feature: the pages have a wide outside margin, in which the pronunciation key and illustrations are printed. This is a striking and original feature, but it squeezes the columns together. Because of it, N's columns are only about 63 cm; G's are about 73 cm.
N's density shows up again at the pages starting and ending each letter. The letter "J", for instance, runs on from the preceding letter (so the last words of "I" are at the top of the page on which "J" starts). G starts every letter on a new page and always starts it on a right-hand page.
In N, the illustrations are put in the outer margin near the text that they refer to. In G, they are in the columns, with the text wrapped around them. N has, by my estimate, at least three times as many illustrations as G. In the letter "H", for instance, N has 115 illustrations in 62 pages and G has 32 illustrations in 64 pages.
Both dictionaries have problems with clarity or sharpness of photographs as illustrations. N's photographs are necessarily confined in width by the margin size, so that when one of them is detailed, it is hard to see. The photograph for jinriksha, for instance, with a man in a singlet pulling a contraption with a sahib in it against the backdrop of a city skyline, is not helpful unless you already know what a ricksha looks like. Also, because the illustrations are all in the margins, the shadow of illustrations on the obverse page is visible. When a photograph on one page is backed by an illustration on the next, the photograph darkens. The one that illustrates high jump is possibly the worst in the book, because it has superfluous detail and is backed by another photograph. But G's photographs are worse. All but the lightest greys turn black. The photo that is labelled Engelman spruce shows something that is vaguely evergreen but indistinguishable from any other evergreen, and the Hellebore photograph, like many others, simply looks like a black rectangle.
Front & Back: In both dictionaries, the main part of the front matter is a detailed guide to the structure and purpose of the entries. N's guide is shorter than G's, but thoughtfully Canadianized. G's is more pedagogical in form and tone. Both are interesting and more than adequate-especially considering that almost no-one ever pays any attention to such guides.
The back matter, customarily reserved for lists of things that crossword puzzle addicts might need to know, differs completely. G offers only seven pages, with things like world alphabets, chemical elements, wind chill factor, and "traditional Canadian measures" (a real anachronism, with fluid drams, hundredweights, bushels, and pecks). N has an eye-opening assortment. First there are 48 pages of lists, many of them ostentatiously Canadian (Canadian Folk Sayings, Canada at a Glance, National Parks of Canada, Canadian Charter of Rights and Freedoms, Canadian Universities, and many more). It also includes non-Canadian lists, among them a world alphabet table that is better than G's.
Astoundingly, N provides at the back of the book a course in Indo-European linguistics, with an essay on "Indo-European and the Indo-Europeans" by Calvert Watkins, a Harvard philologist, who is probably also the author of the language-family tree, table of sound correspondences, guide to Indo-European roots, and a long list of roots (no kidding, 43 pages). If these pages were extracted and reprinted as a monograph, they would make a useful supplement for university courses in Indo-European philology. Their quality is unimpeachable. But what are they doing at the back of a general-purpose dictionary?
Pronunciations: G's experience in making Canadian dictionaries finds its greatest advantage in its pronunciations. In this, it is far better not only than N but than every other Canadian dictionary, including the Collier Macmillan school dictionaries of the early 1980s-for which I, as pronunciation editor, failed to convince the editors to eradicate Briticisms and Americanisms in favour of full-blooded Canadian pronunciations.
Like all its predecessors, N sticks with a version of the traditional dictionary pronunciation key, with diacritics over alphabet symbols. For example, tense vowels are marked by a macron over the most likely vowel symbol (e for the vowel in bee, a for the vowel in pay, and so on), lax vowels are marked by a luna over the most likely vowel symbol ( e for the vowel in pet, a for the vowel in pat, and so on).
To a linguist, these keys always seem obtuse, idiosyncratic, and mystifying. Everyone who works with language professionally uses the International Phonetic Alphabet, a list of symbols with international currency and scientifically established values. The IPA also has quite simple graphics, with few diacritics, good visual discrimination, and attractive shapes. Strangely, dictionary-makers-until now-have refused to adopt IPA symbols instead of their own keys, which are always cumbersome and jerry-built.
So G's pronunciation key is a breakthrough, setting an international precedent. It uses the IPA alphabet, and does so fairly faithfully. The pay-off is evident in simpler symbolization, shorter pronunciation entries, and instant recognition by any reader with a smattering of phonetic training. (Every year, hundreds of students learn IPA notation not only in linguistics courses but also in foreign language instruction, acting and elocution classes, speech therapy, ESL, and other places. No-one is ever trained in a dictionary pronunciation key like N's; such keys have no meaning or currency outside the covers of the books they appear in.)
Not only in its key but in its main Canadian applications, G takes a giant step beyond all its competitors. One of the most distinctive features of standard Canadian English is the merger of low back vowels, so that in Canada but no other major English-speaking region, pairs of words like the following are homophones:
A dictionary that truly represented Canadian pronunciation would use only one vowel for both sets of words. American and British pronunciations both have different vowels in the two lists. Until now, no Canadian dictionary has ever dared break precedent with American and British models.
N compromises. For the cot/stock/knotty set, N lists the Canadian pronunciation (which is, as it happens, the American pronunciation as well). For the caught/stalk/naughty set, N lists two pronunciations: first, the Canadian one (which is the same as cot/stock/knotty ) and then the American one (with a rounded vowel like the one in New York coffee).
G shows the paired words on the list (and all similar words) with one pronunciation only, the same for cot and caught, stock and stalk, knotty and naughty. No compromises here. G gives them exactly the same pronunciation. This is real Canadianization.
G also scores full points for representing the most noticeable feature of Canadian accents for outsiders. We pronounce words like advice, rice, lout, and couch with a short mid-vowel, which is different from the vowel of words like advise, rise, loud, and gouge. N ignores the difference. It represents the two vowels in exactly the same way, as they would be in standard American or British pronunciations. G again does better, showing the pronunciations as different from one another. It is thus the first dictionary to represent Canadian Raising-a term I coined in a 1973 article.
But oddly enough, G does not define "Canadian Raising" as a dictionary entry while N does, as "the difference in sound quality of the vowel sound ou in cloud or clout or of the vowel sound i in eyes and ice that occurs in the speech of most Canadians."
Encyclopaedic entries: For over thirty years, desk dictionaries have been expanding such entries in number and breadth. Years ago, lexicographers debated whether names of politicians, cities, historical events, sects, and other general knowledge entries really deserved space in what is essentially (or was traditionally) a book about words. Whatever the academic arguments on either side-and surely the rational ones are mainly against these entries-the reality is that contemporary dictionaries must be encyclopaedic. They make an easy and obvious selling-point that publishers find compelling.
In this respect, N has a decided edge over G, both in Canadian names and general ones. If you want to get a quick fix on, for example, the Avro Arrow, Mackenzie Bowell, the Dew Line, Ellen Fairclough, Sir Ernest MacMillan, Marshall McLuhan, or Meech Lake, you will find them in N but not G. Similarly, if you need to identify Istanbul, Mencken, Modigliani, Vivaldi, or Yalta, N will serve your purposes but not G.
It is not that G shuns encyclopaedic entries. It has Dewar flask and Dewey decimal system where Dew Line might have been, and it has McCarthyism and McIntosh (the apple) where McLuhan (or at least McLuhanism) might have been. There are plenty of proper nouns in G, but Canadian ones are few. G's omission of entries for Canadian prime ministers seems a glaring one in a Canadian dictionary. There have been only twenty such persons, hardly enough to make a dent in 140,000 or 150,000 entries. While it is merely odd to leave out a minor PM like Bowell, G's omission of major figures such as Borden (it has Bordeaux), Laurier (it has Laurentian), King (it has King Arthur and King Charles spaniel), and Trudeau (it has trudgen stroke) just seems perverse. By my calculation, N has about four times as many encyclopaedic entries; it also has better choices.
General-purpose entries: To compare N and G as general-purpose desk dictionaries, I made a random sample of headwords from Webster's New World Dictionary Third College Edition, a fairly good general-purpose American dictionary (published in 1988) that happened to be near at hand. It is reasonable to expect a decent desk dictionary to be as comprehensive in its coverage as Webster's New World.
The breadth of coverage in N and G appears to be fairly similar to one another, and comparable to Webster's New World. Neither dictionary includes collywobbles or stoccado (a stab or thrust with a pointed weapon). In addition, N leaves out heat of vaporization and pahoehoe (solidified lava with a smooth surface) but G has both. G leaves out pair production (the conversion of gamma rays into electron and positron pairs), but N has it. If either dictionary has an advantage in the breadth of its general entries, it is undetectable.
Depth of coverage is also similar. G generally has longer entries. For some common words, the differences are considerable. For the verb look, for instance, N has 26 lines but G has 55; N lists five phrasal verbs (look into, look out, look up, and so on) and six idioms (look the other way, look-see, and others) but G lists 24 phrasal verbs and idioms. Part of the greater length in G is added on by sample sentences (He doesn't look his age, He looked through the drawer trying to find his keys), such as N never gives. G apparently expects these usage examples to provide models for some grammatical differences; they do not distinguish between intransitive and transitive verbs, as N (like most other dictionaries) does.
In some cases, the extra length of G's entries is useful, for instance in the entry for bacteria, which supplements the definition by an explanation that clarifies it beyond what we would expect. In other cases, the extra length merely supplies more detail and finer distinctions than are practical, as I will soon note for some of the Canadian entries.
N and G as Canadian dictionaries: To check the Canadianness of the two dictionaries, I made two lists of words, one for spelling differences and the other for Canadian meanings:
Spellings: Certain words (and word classes) have particular relevance for Canadian spelling practices. Dictionaries conventionally distinguish variants by listing the one considered more common or more acceptable first. Nearly equal variants come after the headword and marginally acceptable ones are tacked on at the end of the entry. Surprisingly, only G takes advantage of these conventions. N lists all variants in the headword entry.
G distinguishes among cases. It lists variants like travelling and traveling or traveller and traveler in the headword. These variants seem to be genuinely arbitrary in Canadian spelling practice, and listing them as near-equals in the headword seems fair. Similarly, G lists the -our/-or variants (colour/color, fervour/fervor, neighbour/neighbor, and all the rest) in the headword entry. This too seems reasonable, for the second spelling (the -or variant) is standard in Alberta and elsewhere, and putting both in the headword tacitly acknowledges its status in some regions of the country.
G puts most other spelling variants at the end of the entries. In this way, preference is shown clearly for the spellings calibre, catalogue, and metre over caliber, catalog, and meter. This seems a true reflection of Canadian spelling practices. The variants for these words might be seen in American publications, but they are not used in Canadian publications. N loses this distinction by giving these variants the same prominence as the more acceptable ones.
Canadianisms: The chief selling-point of both dictionaries is their Canadianness, and both deliver, with dozens of entries for old chestnuts and some new ones.
In a couple of cases, G's experience at Canadianizing its dictionary over thirty years works against it. It gets stuck with definitions that probably reflect the usage of its original Canadianizer, Walter Avis (1919-79), and now seem slightly old-fashioned. One of these is Bluenose, defined in the expected way in N as "a native or inhabitant of Nova Scotia." G goes further, saying, "a Nova Scotian, less often a New Brunswicker." G (or someone else) should check to discover if New Brunswickers still answer to "Bluenose", because the association has become so specifically Nova Scotian in the last forty years or so.
G is also old-fashioned on serviette, defined as "a paper napkin", which is too restricted. For the oldest Canadians, a serviette could be made of either paper or cloth; for the youngest, it is neither-both things are napkins. The specific meaning for the object made of paper (as opposed to cloth) held for people in the middle, now generally only people over fifty. N gives a general definition: "a table napkin". Neither dictionary seems to be aware that "serviette" as the general term for a napkin was a Canadianism. It was never used generally as such in the U.S. ("napkin" was and is) and though it had some currency in England in the nineteenth century, it fell out of general use; the Oxford English Dictionary says it is "considered vulgar." "Serviette" has been used in Scotland for centuries; Scottish-Canadians must have made it the general term in Canada for over a century, before it was replaced by the American "napkin".
G's tendency toward verbosity shows up in its entry for the hockey coinage deke, defined with admirable directness by N as "A feint or change of direction to mislead an opponent. -...To deceive (an opponent) by means of a deke." G labels the word Cdn. Slang, which seems outdated for a word that nowadays is not considered déclassé in writing. G defines the noun as "a fake shot or movement intended to draw a defending player out of position," thus making an unnecessary distinction between a fake movement and a fake shot. It then goes on to give two verb definitions, the first almost indistinguishable from the noun and the second almost indistinguishable from the other verb definition. Both dictionaries label the word as restricted to hockey, but that seems too narrow; the word can be used metaphorically for any quick, deceptive movement in any context.
Occasionally, G's extra verbiage results in outstanding clarity. I have already mentioned this for bacteria among the general entries. A similar clarity occurs in the entry for the Canadianism humidex. N tells us it is "a measure of hot weather in which air of a given temperature and moisture content is compared to air with a higher temperature and a negligible moisture content." G makes it clearer, as "an index of discomfort resulting from a combination of humidity and heat," and then goes on to explain the method of calculation and to illustrate how it works. The result is outstanding clarity for a concept that is commonly regarded as deeply baffling.
On more recent Canadianisms, the two dictionaries reflect the lack of consensus in Canadian usage. The primary spelling of the name of the one-dollar coin is given as loonie in N and as loony in G. Both list the other spelling as an alternative but N's order is correct: since 1989, "loonie" has prevailed. The two-dollar coin came into being only in 1996, when these dictionaries were already in production. As a result, there is a speculative air to their entries, though both place toonie first. Both list alternatives but the only one they agree on is the exotic twoonie. G adds both toony and twonie, thereby covering (I suppose) all imaginable eventualities. Those alternatives now look a bit fanciful, only a few months after the introduction of the toonie. They will undoubtedly look weird in a year, and zany in two.
One other Canadianism demands scrutiny. Both N and G rightly label hoser as Canadian slang. N defines it in the expected way, as "a gullible, uncouth, beer-drinking man." G does something astounding; it defines it simply as "a Canadian." Period. It is probably a good thing they didn't try putting it in a sample sentence: e.g., Vincent Massey was a debonair hoser. Or in their cover blurb: The Gage dictionary for hosers. G's definition is either an embarrassing typo or a cynical insult.
Either way, it is one measly mistake amidst thousands of details. In that respect, both books are triumphs. They cover their fields competently, and sometimes brilliantly. Which one is better? It depends where you look. N is way better at encyclopaedic entries, and G is way better at pronunciations. N is more handsome, but G is airier, more legible. As general-purpose dictionaries, N is more concise but G goes beyond definitions with sample sentences and explanations. For Canadian content, G is clearer about spelling preferences but N probably has clearer definitions.
I was pleased to find how different the two dictionaries are from each other. They present a real choice, and by no means an easy one. The choice is all the harder because these two are not alone. For years I have used (and endorsed) Funk & Wagnalls Canadian College Dictionary in the Canadianized edition published by Fitzhenry & Whiteside (reprinted 1989). I wonder if either of these will supplant that one on my shelves. A little more than a year from now, Oxford Canada will weigh in with its inaugural Oxford Canadian Dictionary (for which I serve as an editorial adviser). I wonder what it can possibly offer that these two have overlooked.
Jack Chambers is a professor of linguistics at the University of Toronto. He has written dozens of articles about Canadian English, on which he has also taught a course since the mid-seventies. He is the editor of the first book that was published on this subject, Canadian English: Origins & Structures (Methuen, 1975), and of The Languages of Canada (Didier, 1979).