I interviewed Simon Winchester some years ago for the City Arts and Lectures series in San Francisco, just after the publication of his book The Professor and the Madman (British title The Surgeon of Crowthorne). He's a personable and engaging story-teller, and of all the interviews I've done in that series, from Robert Pinsky to A. S. Byatt, his was the easiest and most entertaining (I said afterwards that it was like pitching batting practice to Barry Bonds). A few years later he published The Meaning of Everything, a very readable book about the creation of the OED, and the one I usually recommend to people who are interested in the topic. So he was a very good choice to review Jonathon Green's Dictionary of Slang for the New York Review of Books a few weeks ago. The review took an unfortunate turn, though, when Winchester brought in Jonathan Lighter's still uncompleted Historical Dictionary of American Slang and compared it invidiously, and quite unfairly, to Green's work. It's another in a long line of ill-conceived evaluations of dictionaries by writers who mistake their literacy and passion for the language for lexicographical expertise—think of Dwight Macdonald on Webster's Third, for example. I wrote the following letter to the New York Review. They haven't run it (not surprising, considering its length and the relative marginality of the topic), but because I think the review did an injustice to Lighter, I'm posting it here.
To the editor: When it comes to the topic of slang, even writers as imaginative as Emerson, Chesterton, and Anthony Burgess have had only two or three things to say. You can celebrate the poetry and effervescence of the language of the common folk, you can revel in raffish identification with long-gone rakes and rowdies, and you can proclaim your embrace of slang in defiance of the (even longer gone) pedants and purists who disdain it. The thing can only be done badly or well. So one could do a lot worse than assign the review of Jonathon Green's Dictionary of Slang to Simon Winchester, an engaging writer who has produced two very readable popular books about dictionaries.
Everything was fine until Winchester brought in the unhappy story of Jonathan Lighter's unfinished Historical Dictionary of American Slang, which has been able to find no publisher to take it the rest of the way from P to Z. Winchester may have been led there by his ear for a good tale, but it isn't clear why he took it on himself to compare Lighter's book invidiously with Green's—or more accurately, to say that there is "no real comparison between the two books." Winchester is not a linguist, a lexicographer or a student of the history of English, and he's way out of his depth in making that call, which is gratuitous and, well, wrong.
As it happens, I’ve just finished a manuscript on the word asshole and other obscenities that were converted to social descriptions over the twentieth century, a development largely unprecedented in English or other languages. I relied heavily on Lighter’s work in preparing in the historical sections of the book, and when Green’s dictionary appeared, I quickly acquired a copy, thanks to a credit I had at Oxford for some reviewing work, and examined the relevant entries in it as well. It was very helpful, but for purposes of serious scholarship I don’t find it quite at the level of Lighter’s dictionary, which is more rationally organized and tends to have more and earlier citations for most of the important words. (Take the use of asshole to refer to the most unpleasant spot in a region, as in "the asshole of the universe." Green illustrates that sense with seven citations; Lighter has fifteen, plus a lovely anticipation using bung-hole from Oliver Wendell Holmes.)
Critics of dictionaries are inevitably obliged to cherry-pick their examples, but it's a sign of how far Winchester is in over his head that the very entries he offers in evidence of the superiority of Green's work actually testify to its limitations. Both dictionaries begin with slang uses of the letter "A," he notes, but "here, crucially, Lighter stumbles—for despite his being a compendium of American slang only, the first entry in Green for the letter A is an American usage that doesn’t appear in Lighter at all."
But not so fast. The entry that Winchester is referring to involves the use of A as shorthand for the Model A Ford. But why would you include that in a dictionary of slang, any more than the abbreviation Z for the Datsun 280Z, which Green also lists, or for that matter than Jaguar's E and XKE, which he doesn't? They're not colorful or colloquial, not restricted to a particular group, and they're devoid of attitude. Nor is there the slightest slang feel to the abbreviation a. and b. for “assault and battery,” which also appears on Green's first page, with a first citation from a 1928 legal journal. If a. and b counts as slang, why are there no entries for M and A, R and R, and PB and J?
What's more, it's not likely that A was ever actually used in that way. Green gives only one citation for it, from a Stephen King novel written in 1986. But that's almost sixty years after the first of 5 million Model A's rolled off the assembly line, and if they were commonly referred to as A's, you'd figure Green would have located a citation for the usage well before then—particularly now that lexicographers have literally millions of books and newspapers from that period available for searching online. A more wary compiler would have surmised that the usage was just another of the many specters created by King's fecund imagination, at least until earlier evidence for it turned up.
The very next item in Green is even more curious—an entry for the final -a of woulda, shoulda, and the like, which Green describes as “used to denote a colloquial or slangy pronunciation." Linguistically speaking, that entry is a mess. In the first place, shoulda isn't slang but informal (all speakers produce the form in conversation, whether they're aware of it or not). In the second place, what's informal here is the pronunciation, not the way writers transcribe it. And in the third place, the -a isn't a separable element—you might as well give an entry to n’, since that is "used to denote a colloquial or slangy pronunciation" of words like going, not to mention separate entries for the apostrophe in li’l and the -uz in wuz.
Those first few entries demonstrate two endemic (though far from fatal) difficulties with Green's work. His bower-bird instinct has him scooping up slang, colloquialisms, nonstandard and informal items with undiscriminating enthusiasm. And he too often relies on documentation that's thin or implausible. A large number of his entries are drawn from slang lists or glossaries of dubious accuracy and don't seem to have any other written attestation—a necessary expedient for the slang of earlier periods but not in the age of the web. Take the entry for claven to mean someone who is long-winded, carried over from the 2005 Cassell's slang dictionary that Green did, for which he cites a lone entry by one "Mudflap" at the anyone-can-play site urbandictionary.com. I haven't been able to find it anywhere else on the Web. Green has it as "ety. unknown," though a second urbandictionary entry for the form, dated several years after the 2005 Cassell's dictionary appeared, relates it to the surname of the loquacious mailman on Cheers—actually spelled Clavin—though that doesn't seem to be attested anywhere else on the web either. In any case, you'd want to nose around a bit more before awarding it an entry.
The effect of both of these practices is to somewhat inflate the size of Green's dictionary, which has a great many items that are absent in Lighter—though the vast majority of these are legitimate, and of course Green is documenting the slang of the entire English-speaking world, not just America. Winchester attaches great importance to the items missing in Lighter (there are only about two or three things that writers have to say about dictionaries, too). But lexicographers know that comparing word-counts isn't a very useful standard for evaluating a dictionary, even if their advertising departments insist on stressing the counts in their publicity. By that standard, after all, we'd have to judge Johnson’s 1755 dictionary as inferior to that of Nathan Bailey's of 1730, which contained 48,000 words to Johnson's 40,000. And it’s particularly pointless to count words when we’re talking about a category as elastic as slang, which doesn't offer even the illusion of the "clear and unmistakable nucleus" that James Murray discerned for the standard language.
Does any of this really matter? Certainly not to casual browsers of Green's dictionary, who will find it a delight to rummage around in, provided they take some of the unattestested items with a grain of salt. But every dictionary is an implicit theory of its compass, which is why critics have taken dictionaries from Johnson's to Webster's Third to task for including items they held to be unsuitable for use in public discourse. And it matters how we define and circumscribe slang, too, a category that has historically embodied changing assumptions about class, deviance and criminality that are particular to the English-speaking world. (That's why the word slang itself doesn't have an exact translation in French or Italian, for all that Green asserts in his introduction that it’s a universal category.) Whereas Green's expansive coverage seems to blur the category of slang even beyond its naturally fuzzy boundaries.
Winchester stresses that Green, like himself, is not an academic, but "a child of Sixties counterculture," whose popular volumes on sex, cannabis, and other topics were all derived "from personal, streetwise experience." But a Bohemian youth isn't much of a qualification for the painstaking work of lexicography, and Green to his credit relies not on his streetwise Sprachgefühl but on the same musty sources that other lexicographers use. I suspect there's a touch of projection in Winchester's suggestion that Green "seems not to take slang quite so seriously," which if you took it at face value would render Green's decades of prodigious devotion to the project a matter for professional intervention.
I'm not trying to be captious. Overall, these are venial shortcomings in an impressive and important work that should be on the shelves of any word-lover who's in a position to spring for the nearly $600 that Oxford is asking for it. But Winchester's ill-considered comparisons to Lighter do casual injury to the scholarly reputation of someone who has already endured a great deal. Winchester rightly indicts two of the culprits behind the abandonment of Lighter's dictionary: first Bertelsmann, who dropped it when they acquired Random House, and then the Oxford University Press, who negotiated a deal to publish the remainder but then stopped publication. We can add to these the National Endowment for the Humanities, which decided not to help Oxford complete the project—partly on the grounds, I'm told, that it was felt to be inappropriate for the NEH to make grants to a foreign company, even if the result would be a unique record of American cultural and linguistic history. So for the foreseeable future we are left without an authoritative historical slang dictionary of our own, or at least for the terra incognita beyond O. It takes nothing away from Green's admirable work to say that it doesn't go far enough towards filling that gap.
Given the sensitivites involved here, I'll welcome comments on this one that are specifically relevant to the issues at hand, but will keep a firm hand on the spigot.