Word Imperfect

Roget's Thesaurus has long been considered one of the great lexicographical achievements in the history of the English language, a reference work of astonishing ubiquity and far-reaching influence. But now the author of The Professor and the Madman—the best-selling tale of the making of the Oxford English Dictionary—questions the legacy of the definitive list of synonyms that the brilliant Peter Mark Roget compiled 150 years ago. Is the name Roget becoming a synonym for intellectually second-rate?

By Simon Winchester

The writing world may at last be having second thoughts about Peter Mark Roget, Esquire—polymath, physician, cinema inventor, slide-rule maker, chess master, lexical scholar, and the man who gave us one of the best-known reference works in the English language. One hint as to his possibly altered standing comes from the latest version of the Encyclopaedia Britannica, which, although Roget was an editor of the seventh edition and a contributor of more than 300,000 words to it, gives him somewhat short shrift today, with an entry of a mere twenty lines. On the same pair of pages onto which his life is squeezed are far-more-substantial articles about figures one must suppose are now more deserving of note: the Korean leader Roh Tae Woo; the Belgian statesman Charles Rogier; the Huguenot Duke of Rohan; the author of a book called Australian Totemism, one Géza Róheim; and a murdered Nazi storm trooper named Ernst Röhm.

But a more potent clue is to be found whenever one tries, using modern equipment, to write anything about Roget. The spell-checker that is provided on most computers these days has no listing whatsoever for Roget. The old-fashioned Oxford English Dictionary has a listing, naturally, and makes it clear that the word—as in, for example, to look it up in Roget—is now so well known as to have the status of an eponym. (It has to be said, however, that the name is not in Oxford's Dictionary of Eponyms.) But no readily available computer-program-integrated dictionary seems to agree with the OED: none I used lists the word, even though Microsoft Word (which not unreasonably sports its company name in its own dictionary) has a fair enough share of other eponymous and similar words and phrases that might be thought of as equally significant—boycott, thermos, Kodak, and bowler hat among them.

No, if you attempt to write the word Roget using any of Bill Gates's spectacular software, you get a squiggly red line underneath, indicating that you have written a word the software doesn't recognize. Even worse, if you happen to enter the same word in the thesaurus that comes with Microsoft Word (but which is made under contract by a firm with the name—somewhat less than encouraging for lexicographers—Soft-Art Inc.), you will be obligingly informed that what you must have been thinking of, when you were so slack as to write by mistake the name of literature's most celebrated helpmeet, was in fact the word rogue. Don't know a good synonym for philistine? Why not look it up in rogue?

Yet what some will see as just another melancholy comment on the degraded state of contemporary language has for me turned out to be rather helpful. I bless the operatives at Soft-Art and their fine economy of word association. For what I am trying to grapple with, and what might otherwise have resulted in little more than a hymn of praise to a man I see (for his myriad other achievements) as one of the great unsung heroes of all time, has been distilled into an elegant and nicely challenging proposition: Perhaps Peter Mark Roget actually deserves to be regarded as the very rogue that this sparkling new twenty-first-century thesaurus would make him. If so, it is because the evident decline in language is something for which at least partial blame must be laid at his door; it has its origins in the extraordinary eponymous volume that he single-handedly created almost 150 years ago.

To put it more forcefully: Roget's Thesaurus no longer merits the unvarnished adoration it has over the years almost invariably received. It should be roundly condemned as a crucial part of the engine work that has transported us to our current state of linguistic and intellectual mediocrity.

More than 30 million copies of Roget's Thesaurus of English Words and Phrases have been sold since the book was first published, in London, by the firm of Longman, Brown, Green and Longmans, in May of 1852. It is, by any standard, one of the most popular reference books ever written—a "treasure-house" indeed, as thesaurus translates from the Greek. Rare is the household without a dog-eared copy somewhere—perhaps a holdover from school days; perhaps bought years ago with good intent, along with Merriam-Webster and Bartlett's Familiar Quotations; perhaps twinned with a book of crossword puzzles or acrostics. The motives for owning Roget—improving that essay, finding that eight-letter word beginning with t, getting the mot juste for that Rotary Club or senate-campaign speech—are manifold.

There have been countless editions. Roget himself presided over twenty-five of the twenty-eight—each one subtly different—that were published during the two decades in which he continued to work on his magnum opus. The book has been published in America continuously since 1854. A considerable industry has arisen alongside Roget, devoted to books with a similar function and with similar titles. Many of these works once used the word Roget in their titles, as the name of the original author or, quite often, as a purely descriptive term. C.O. & Sylvester Mawson's A Dictionary of Synonyms and Antonyms; Being a Presentation of Roget's Thesaurus ... in Alphabetical Form (1931), is one example. Amazon lists 935 works for sale that include thesaurus in the title. (An impressive number, one might think—though it is perhaps worth noting that the same catalogue lists 21,782 products that incorporate the word dictionary in their titles, and 10,748 that call themselves encyclopedias. Roget may sell phenomenally well, but it has much less competition than do many other great works of reference.)

However defining and useful Roget's Thesaurus may have proved to be over the past 150 years, it was not the first book of its kind. Actually it would be more accurate to say that earlier books performed the function that Roget's Thesaurus is believed to perform—a distinction that makes it necessary to focus at the outset on three questions: What books of similar intent existed in published form before 1852? What exactly was Roget trying to do when he first sat down to assemble his famous work? And what did he in fact achieve?

The answer to the third question is, superficially at least, self-evident: Roget's Thesaurus is a stylish and comprehensive list of synonyms. The other questions are less straightforward, however, and to answer them we must first consider sixteenth- and seventeenth-century English society—particularly the eagerness among its most prominent members to make sure they looked and sounded elegant and learned.

This period of English history was witness to an explosion of intellectual energy. It was the time of Newton and Dryden, Bunyan and Purcell, Halley and Wren, Aphra Behn and Beau Nash. These great thinkers and creators radiated energy; the nation basked in their glory; and those who were rich and grand enough but less talented did their corporate best to sparkle in all the reflected luminance.

The style of the times was all about glitter and reflection: the houses, the parties, the gardens, the games, the flowery rituals of mannered society—all pointed to a fascination with the baroque complications of an ever accelerating civilization. The dress of the day suggested much the same. Fashionable foppishness also exceeded all reason in what may strike us today as the most ludicrous of English affectations at that time: the laboriously convoluted language. The way the upper classes and those who aspired to join them tortured the tongue positively beggars belief.

Nor was there any irony in the widespread use of what were called (dismissively, by truly learned folk) "inkhorn terms." The language spoken in the overheated "withdrawing rooms" of Belgrave Square and Pimlico was larded with them: phrases such as ingent affabilitie and dominicall superiorite; verbs such as revolute and deruncinate; and adjectives—rather fewer, mercifully—such as magnifical and splendidious. The relative lack of adjectives hints that perhaps some vocabulary god thought this nonsense was dreadful enough without further qualification.

In due course, in 1604, a Coventry schoolteacher named Robert Cawdrey came to the aid of those who were floundering. He gathered together some 2,500 words, arranged them in alphabetical order, and offered his volume,120 pages and bound octavo.

The book was called A Table Alphabeticall ... of Hard Usuall English Wordes, and it became—this first-ever true dictionary of the English language—a best seller. Cawdrey noted without a hint of condescension that he intended it "for the benefit & helpe of Ladies, Gentlewomen, or any other unskilfull persons."

For a while Cawdrey's book did the trick. But before long one of the axiomatically inherent flaws of alphabetical dictionaries in general became apparent: it is not possible to look up a word if you don't know what that word is.

"Of course," one would say today. "That stands to reason." But in the sixteenth and seventeenth centuries, when the cataloguing of words was an altogether brand-new skill, it took time to understand. And once it had sunk in, the need arose for quite another kind of book—one that would lead the inquirer to a particular word if he knew roughly what it was that he wanted to say but had no firm idea of the assemblage of letters and syllables that would enable him to say it.

The first notable book to do so was by one John Trusler, and its portmanteau title was The Difference Between Words Esteemed Synonymous in the English Language; and the Proper Choice of Them Determined. Trusler's volume (one of the first to incorporate a word derived from synonym in the title), more than any other, was the true precursor if not specifically of Roget's Thesaurus then of what the Thesaurus has since erroneously (as we shall see) been perceived to be. It was, in essence, the first real synonym finder.

Trusler's work answered the prayers of the drawing-room dandy, because it meant that at last he could look up a word without knowing what it was. He simply had to look up a word of similar meaning, and lo!—if his synonym finder was a good one (as John Trusler's was), then a suitable word would be listed; he could select one and promptly delight the world.

Trusler's book, though, was essentially a catalogue of synonyms: it merely listed them, noting their existence but making no attempt to prescribe which words should be used in which situations. It was left to a friend of Samuel Johnson's, Hester Lynch Piozzi, to create the first prescriptive work on British synonymy. Although Piozzi's book had many detractors (Roget was probably among them), it also persuaded literate Britain to start thinking hard about a question that lies at the core of this examination of Roget: what exactly is a synonym?

Piozzi's book, British Synonymy; or, an Attempt at Regulating the Choice of Words in Familiar Conversation, came out in 1794, nearly sixty years before Roget's first thesaurus was published—but, according to his writings, only a decade before Roget started planning it. The links between Piozzi's book and his are vague, but her volume's fame was such that he no doubt knew of it. British Synonymy was published in Dublin and ran to 427 pages, and it was directed, as D. J. Emblen wrote in his biography Peter Mark Roget (1970), at "those coming up in society and ... eligible foreigners" who might not be familiar with the shades and nuances of the language, and who might want to avoid embarrassing themselves.

The language of the day was becoming ever more complex: thanks to new discoveries in science and geography, and to the ever multiplying energies of writers, vocabulary choices were increasing almost by the day. To use English properly was, in a word, daunting. Opportunities for making oneself look and sound foolish were everywhere. Hence the need for the balm of Piozzi's enormous book. She would show one how to sound one's best.

The way she tried to accomplish this, mind you, ruffled some feathers. It was unimportant to her if what one said was lexically imperfect—that was just a matter of pedantry. Style and the veneer of sophistication were all. "Synonymy," she reassured her readers, with a remark that would have infuriated Roget, "has more to do with elegance than truth."

The organization of Piozzi's book is peculiar compared with that of modern synonymies. She grouped what she regarded as synonymous words together and then explained in a longish paragraph the nuances that distinguished them. Take, for example (as quoted in Emblen's book),

AFFABILITY, CONDESCENSION, COURTESY, GRACIOUSNESS
Are nearly synonymous, though common discourse certainly admits that an equal may be affable, which I should still think wrong in a printed book, and unpleasing everywhere, because the word itself seems to imply superiority. We will allow however that the lofty courtesy of a princess loses little of its graciousness, although some condescension be left visible through the exterior affability; but that, among people where talents or fortune only make the difference, a strain of polished familiarity, or familiar politeness (call it as you will) is the behaviour most likely to attract affectionate esteem.

In Piozzi's view, a slight but important difference of class association existed in the employment of these four words. The manner in which each was used—except among those whose "talents or fortune" rather than rank distinguished them—depended on who was being affable or courteous to whom: a duke might be affable to a commoner; it behooved a commoner to be courteous in return, and most decidedly not affable.

Other synonymies were published between Piozzi's and Roget's—the biggest and most durable being George Crabb's work of 1816, English Synonymes Explained, in Alphabetical Order; With Copious Illustrations and Explanations Drawn From the Best Writers. It was a cumbersome book, widely criticized for being prolix in style and maddeningly circular in argument. But it, too, answered a seemingly urgent need: sixteen editions were printed over the thirty-six years before Roget swept the board.

Crabb's success spawned other synonymies: in the decade before Roget fourteen synonym finders were published, known today to the small fraternity of thesaurus collectors by the names of their compilers—among them William Carpenter (1842), George Graham (1846), James Jermyn (1848), and James Rawson (1850). Many such books could have helped users to jog their memories. It is rather charming to suppose that such a book, slipped furtively from a partygoer's pocket, might act as a sort of hip flask for the conversationally parched. The image may strike us as more than a little awkward, but perhaps the etiquette of the time made the use of such a little book no trickier than the use of a phrase book when abroad.

But although all these volumes listed hundreds upon hundreds of almost interchangeable words and phrases, and gave helpful hints as to how each might be suitably employed, none of them—with the possible exception of Piozzi's—took care to examine the subtle notion of the synonym itself. None of the editors wondered—at least not in print—why a language as complex and finely tuned as English would include any two words that meant exactly the same thing. Was there such a thing as a real synonym? Or had every word been created for a unique purpose? Roget, who began as early as 1805 to consider the need for some formal classification of the chaotic entity that was then the English language, was fascinated by these questions. The answers he constructed led him, fifty years later, to the creation of this organizational masterpiece that bears his name.

Consider some of the words that are listed in the OED (under synonym) as examples of as-near-as-it-comes synonymy. The first are serpent and snake. Are these words true synonyms? In terms of pure definition, yes—sort of. Serpent is defined in the OED as "any of the scaly limbless reptiles regarded as having the properties of hissing and 'stinging'; Zool. a reptile of the group Ophidia; a snake." Snake is defined as "one or other of the limbless vertebrates constituting the reptilian order Ophidia (characterized by a greatly elongated body, tapering tail, and smooth scaly integument), some species of which are noted for their venomous properties; an ophidian; a serpent." The difference at this level is minuscule: one definition is more comprehensive, mentioning the vertebrated skeleton and the scaly integument; the other is somewhat more colorful, accentuating the hissing noise that the beast is able to emit.

Were this all, one might agree that the two words are perfect synonyms. But the OED, complete as always, continues its definition of serpent, observing that nowadays, in ordinary use, the word is "applied chiefly to the larger and more venomous species; otherwise only rhetorical ... or with reference to serpent-worship."

Therein is the suggestion of reptilian synonymy blown suddenly asunder. For serpent is indeed the word we choose when we want to denote a snake that is bigger and nastier than most, and snake is the word we choose to describe any smooth and elongated creature that skitters from beneath the lawnmower blades. We say, on the one hand, "There is a snake in the basement" and, on the other, that missionaries were once thrown into "pits filled with serpents." But we don't proclaim there to be a serpent in the garden shed, and if we told a listener that a missionary was in a pit full of snakes, we would realize from the resulting questions—Were they big? Were they venomous?—that we had used the word wrongly, poorly, or incautiously. (This leads to another supposed synonymy: venom and poison. The words, however, are not exactly synonymous, because one can speak with venom yet perhaps not quite with poison. Venom is both a substance and a tone; poison is more a matter of chemistry.)

Examination of any words thought to be synonymous reveals a congruency of range but not an identical meaning. Take some other illustrative related examples from the OED: ship, vessel; compassion, fellow-feeling, sympathy; enormous, excessive, immense; glad, happy, joyful, joyous; kill, slay, slaughter; grieve, mourn, lament, sorrow. Some are very close indeed; there is little to distinguish a ship from a vessel, except that one doesn't say fishing ship or war vessel—suggesting that a vessel is likely to be engaged in peaceful activities, whereas a ship can have a more menacing role. One cannot quite imagine Nelson's having spoken of vessels on the horizon off Cape Trafalgar, or any dockside idler's speaking of the handsome lines of the ship that has just brought lobsters back from the Outer Banks. (In truth, he would probably say boat.) Others on the list are more obviously distinguishable. Sometimes the distinction is a matter of degree: one kills a man; one slays his child; one slaughters the villagers who sheltered the family. On other occasions the context suggests one choice rather than another: one feels compassion for the villagers in such circumstances, but fellow-feeling for the brother of the first who had to die.

The practices surrounding synonymy, which seem to have come into being at around the time Roget was beginning to codify the language (after Samuel Johnson had created his dictionary but before the members of the Philological Society of London had commenced work on the all-encompassing OED), are really quite simple. Currently, the high priest of this field is Ladislav Zgusta, a scholar of Czech extraction (reinforcing the notion that English is often more scrupulously regarded by those who come from less than pure English stock; James Murray, of the OED, was proudly a Scot, and Roget came from a Swiss Huguenot family). To be absolute synonyms, Zgusta says, words must be the same in three distinct ways: They must have the same designatum—that is, the same essential qualities. They must have the same connotation—the same associated features of meaning. And they must occupy the same range of usage and application—the contexts in which they are generally used must be identical.

Very few words satisfy this formidable set of criteria. Most of those that do are technical. In his standard work on lexicography, Dictionaries (1984), the near-legendary authority Sidney Landau offered as an example ten absolute synonyms, agreeing in designatum, connotation, and range, for the mad-cow-disease-related human ailment that is currently terrifying England and threatening France—Creutzfeldt-Jakob disease, widely known as CJD. The variants that Landau listed include Jakob-Creutzfeldt disease, spongiform encephalopathy, and Jones-Nevin syndrome. (The disease afflicting England, with ninety-one known cases and perhaps thousands more expected, is in fact now known, confusingly for my purposes here, as New-Variant CJD—and is different from Landau's plain-vanilla CJD. The new version varies more in terms of pathology than of lexical standing.)

The synonyms listed by the OED fall short of Zgusta's standards in a variety of ways. Serpent and snake have the same designatum—each is a long, slithery, cold-blooded, and scaly reptile. And they have more or less the same connotation. But they have very different ranges of use, as I have shown. So with only two of the three criteria satisfied, serpent and snake are not absolute synonyms: they are simply near synonyms, not to be substituted for each other without care and attention to the context.

Much the same can be said for beast and brute, or mirth and cheerfulness. The words in each pair share a designatum but not a connotation or range of possible use. A brute is invariably untamed; a beast often is, but he need not be. We can say "a beast of burden" or "a beast of the field"; we would never say "a brute of burden" and probably would not say "a brute of the field"—though we might allow "brute of the forest," since the word here reinforces an image in which all is mystery and secret danger.

An awareness of the nuances of synonymy, then, is fundamental to the speaking and writing of good English. Merely cataloguing synonyms, uncritically offering up lists of alternative words from which a speaker or writer may choose, makes for something less happy. Roget realized this back at the beginning of the nineteenth century, when the literate world was awash in dictionaries, thesauri, lexicons, and other guides for the betterment of verbal display. His goal was to make sure that what was written and spoken and read was impeccable. To this end he began a study of the language with the primary aim of classifying it and then distilling from that classification a guide to how it might best be made to work.

Peter Mark Roget, born in London on January 18, 1779, was in myriad ways a most extraordinary man. I said earlier that he was a polymath, and in the literal sense of the word—"a person of much or varied learning"—he truly was. Although the term is still widely used today, few men or women in Western societies are genuinely polymathic. In Britain the physician, philosopher, playwright, librettist, and writer Jonathan Miller springs to mind as a possibility; in America, sad to say, despite some astonishingly bright people, no one fits the bill.

Roget seems to have been interested in and learned about almost everything. I first came across him when I was investigating the circumstances of the 1831 awarding of a medal by the Geological Society of London. A geologist named William Smith, who had for many years been ill-used by his peers, was, after a scandalous delay, being formally recognized. I wanted to know who those peers were, and who had so wisely engineered Smith's reappraisal. I needed a list of those who attended the award ceremony.

The Geological Society had the relevant papers in its archives, and the dusty volume holding them was duly brought to me. It was opened carefully by a librarian, who found the proceedings of the meeting and scanned the handwritten list of those present. It was much as I had expected: the meeting had been chaired by the renowned Adam Sedgwick, and the attendees included William Broderip, the Reverend William Whewell, Leonard Horner, Captain James Vetch, the Henry De la Beche who would go on to be the first head of the British Geological Survey, Professor Edward Turner, the great Silurian expert Roderick Murchison, and one Dr. P. M. Roget. The librarian gulped. "Bless my soul," she said. "Do you think it could be him?" We looked in the index of members and found that Peter Mark Roget had indeed been a fellow of the Geological Society—and of many, many other august bodies besides.

The title page of the first edition of his Thesaurus offers tantalizing clues to his brilliance: his geological inclinations are shown only in the string of initials after his name ("M.D., F.R.S., F.R.A.S., F.G.S."), but he is also identified as "Fellow of the Royal College of Physicians; Member of the Senate of the University of London; of the Literary and Philosophical Societies etc. of Manchester, Liverpool, Bristol, Quebec, New York, Haarlem, Turin and Stockholm. Author of the 'Bridgewater Treatise on Animal and Vegetable Physiology,' etc."

The "etc." shrouds an array of other achievements. Even the sparse Britannica entry mentions that Roget invented the so-called log-log slide rule, that he was the secretary of London's fabled Royal Society, and that he was a philologist. A cursory look at his scores of published papers attests to his many other interests (and the fact that he wrote some of these papers in French, German, and Latin as well as English shows his formidable linguistic abilities, stemming from his Swiss Huguenot roots).

Roget's first published works were purely medical, though wide-ranging by any standards: a study of the effects of inhaling laughing gas, a test for detecting the presence of arsenic, how the skin changes color if a patient ingests silver nitrate, on sweating, on tetanus, on voluntary action of the iris, on perception and feeling in animals, on epilepsy, on the medical care of prisoners. By midlife he was starting to branch out to other topics. In 1815, when the rest of Europe was preoccupied with Waterloo, he published a paper titled "A New Instrument for Performing Mechanically the Involution and Evolution of Numbers"—that is, the slide rule. Three years later he produced papers on the kaleidoscope and Dante. He was fascinated by chess problems, too, and one of his more amusing papers shows that it is possible (if not necessarily desirable) to move a knight across every square of the board.

Then, in 1825, came his paper "Explanation of an Optical Deception in the Appearance of the Spokes of a Wheel Seen Through Vertical Apertures," which is regarded as seminal by modern historians of the cinema. The impetus for the paper came about by chance. Early one morning soon after his marriage, at home in Bloomsbury, Roget, according to Emblen, had been gazing up idly from his basement kitchen, from behind a vertical venetian blind, at the passing traffic. The slats of the blind, he suddenly noticed, broke the movements of passing carriage wheels into a jerky series of still pictures. Depending on the speed of the carriage and the position of his eyes, the spokes of the wheel appeared to bend, to curve sometimes backward, sometimes forward. When he moved his head up and down, he noticed that the image changed. He dashed outside and paid a driver a shilling to drive his carriage back and forth along the street while Roget jotted down notes and made sketches. All the while his new wife, Mary, was upstairs tapping her feet, waiting with the breakfast kedgeree cooling in the dining room. Life as Mrs. Peter Mark Roget was, she soon came to understand, invariably odd.

The paper her husband then published led in time to what Marshall McLuhan and others recognized, in Emblen's words, as "another dimension for human existence"—the motion-picture industry. For, as Roget concluded, "an impression made by a pencil of rays on the retina, if sufficiently vivid, will remain for a certain time after the cause has ceased." Persistence of vision had been discovered; the Zoetrope, the Praxinoscope, and the CinemaScope were just waiting to be invented, and 6,000 miles westward Hollywood awaited its founding.

But there was another, more soberly philosophical side to Roget's raft of achievements and ambitions. He seems to have been—sentimental though this may sound in today's cynical climate—a thoroughly good person, suffused with ideals about the society of which he was a privileged member. He held a profound belief in the right of ordinary men and women to know things—to be able fully to appreciate the wonders and complexities of the world. He was influenced by Jeremy Bentham's ideas of utilitarianism, which sought to promote the spread of happiness to the greatest possible number of people. He offered his medical services free to those who couldn't pay. He was a keen supporter of preventive medicine—urging reforms, for example, in London's water supply (and proposing a method of water filtration through sand that is still in use today). He was a founding member of the short-lived Society for the Diffusion of Useful Knowledge, and he wrote a series of sixpenny treatises—on electricity, galvanism, magnetism, and electromagnetism—that were intended to help poorer and less educated people learn what he and his kind already had the privilege of knowing.

He also (as already mentioned) contributed about 300,000 words to the well-regarded seventh edition of the Encyclopaedia Britannica. Of these, 150,000 words went for one major treatise, Physiology; the remainder were divided appropriately among the articles Ant, Apiary, Baldinger, Sir Joseph Banks, Barthez, Beddoes, Bee, Bichat, Brocklesby, Broussonet, Camper, Crawford, Currie, Deaf & Dumb, Kaleidoscope, and Phrenology (of which he was deeply scornful). Although some of these articles have survived, much cut, into the modern era (an article on Marie-François-Xavier Bichat, the French founder of the science of histology, is still there, and twice as long as the article on the man who first authored it), Roget's name does not survive in the index of the Britannica's contributors; today "PMR" is a blameless French museum curator named Pierre Rosenberg.

It remains one of the more curious aspects of Roget's life that in this fury of intellectual and socially reforming energy his deep interest in the English language did not become fully apparent until he was in late middle age. He was nearly seventy when, still living in his London townhouse, he began work on his Thesaurus, having just been forced from his post at the Royal Society to make way for younger, cleverer, more energetic scientists.

Roget was by no means bitter—he knew well his own scientific strengths and weaknesses, and he would have been the first to acknowledge that his great achievements to date, largely in the fields of description and classification, had not been marked by brio and inspiration. He settled down to his retirement, and yet equipped himself with a grand plan: if the God in whom he believed so implicitly allowed him sufficient time and energy (he was in fact granted a further twenty-one years, all of them healthy), he, by writing and thinking and organizing, would bestow on the kingdom of language the same order that Linnaeus had given to the kingdoms of animals and plants.

Such an ordering, Roget came to believe, would not just answer an intellectual need—it might well have benefits for society as a whole. He mentioned in the Thesaurus's introduction that it might even help to fulfill his long-cherished dream of fostering the birth of a universal language—"that splendid aspiration of philanthropists."

Ever since the turn of the century, when he first began delivering medical lectures in Manchester and London, Roget had carried a succession of notebooks. Few, if any, of the originals survived the fires of World War II. The notebooks contained list after list of words that appeared to Roget to be near synonyms or—as the lists lengthened and the notebooks multiplied—to belong to the same philosophical groupings. Roget's realization as a young man that words could be placed in classes would later underpin his making of the Thesaurus. This is the essential difference between Roget's accomplishment and that of the other thesaurus makers before and since: his was a conceptual thesaurus, whereas the others were merely arranged alphabetically or otherwise organized to be useful.

There is no doubt that Roget aimed to produce a volume that was likely to be helpful to some users. In his preface he used the word desideratum—he was creating something he felt was required or desired. The book was designed "to supply, with respect to the English language, a desideratum hitherto unsupplied in any language; namely, a collection of the words it contains and of the idiomatic combinations peculiar to it, arranged, not in alphabetical order as they are in a Dictionary, but according to the ideas which they express."

Tom McArthur, the editor of the monthly magazine English Today, wrote in the Oxford Companion to the English Language that Roget's clear intent was "not ... to define or discriminate [words], but to arrange them in synonymous and antonymous groups; it serves as both a word-finder and a prompter of the memory regarding words one knows but could not recall to mind."

Yet—and this is why I feel his book proved an ultimate and unwitting disservice to the language—Roget aimed, in selecting his supposed readership, very high indeed. His sixpenny tracts for diffusing knowledge may well have been intended for the artless and the educationally impoverished. His Thesaurus, on the other hand, was meant for users equipped with more finely honed intellects and with a very real lexical intuition. He surely never imagined that businessmen, students, and politicians would one day help to make his book such a commercial success. Had he done so, he might have organized his work in a signally different way.

His nobly Platonic vision was that the language could come to be seen as an ordered part of the cosmos, amply reflective of divine will and inspiration. He took an Aristotelian approach to his task as well, marshaling his subject according to the strictest logic. His organization was clearly of the moment: he believed in all sincerity that from out of the miasma of Victorian intellectual confusion could rise a gleaming pillar of lexical glory, a totem to the God who had made it all.

Once he had established his conceptual framework, he concluded that all words could be placed in one of six classes. "The purpose of this work," he wrote in the introduction to the first edition, "is not to explain the signification of words, but simply to classify and arrange them according to the sense in which they are now used, and which I presume to be already known to the reader." Those six classes indicate his staggering polymathy: Abstract Relations, Space, Matter, Intellect, Volition, Sentient and Moral Powers.

The first three cover the external world: Abstract Relations encompasses concepts such as order, number, and time; Space encompasses words relating to size and movement; Matter encompasses the physical world and the way in which people experience it with the five senses. The second three classes relate to the interior world of human beings: Intellect encompasses matters of the mind; Volition acts of will; and Sentient and Moral Powers (or, as modern Roget editors now call it, Emotion, Religion and Morality) the more profound matters of the heart and soul. As we can perhaps already see, it would take an extraordinary mind to discern such order and to impose it on the language we use.

Some might suggest that Roget was rather out of touch in assuming that such things were already known to or understood by his potential readers. Further, such an assumption, they say, casts doubt on Roget's utilitarian ideals. For how on earth could an average user understand something that few trained lexicographers—and, indeed, few philosophers of language—can properly comprehend (let alone agree on) today?

But that seems not to matter: a glance at a modern Longman edition of Roget's Thesaurus, published now by Penguin Press (1998) and edited by Betty Kirkpatrick, shows that the classification structure survives. However incomprehensible it may be, it is still the basis for the real Thesaurus, a century and a half later. Only in matters of the heart and soul has it been slightly changed, and only, in truth, for cosmetic reasons. The words within have changed, of course, with the times and the fashions—which is true of any thesaurus, dictionary, encyclopedia, or other work of reference that passes through new editions over the years.

To illustrate the cleverness of Roget's extraordinary classification system, let us track in the Penguin edition all the way through a single class of thought to its logical conclusion—to what, if we compared this with the Linnaean system of classifying life, we would call the thought's linguistic species or subspecies. Let us take the conceptual class Volition.

"Volition: the exercise of the will" is first divided into Individual Volition (the will of one) and Social Volition (the will of many). Under Individual Volition are five subclasses: Volition in General, Prospective Volition, Voluntary Action, Antagonism, and Results of Action.

Let us look at Voluntary Action, which may be Simple or Complex, and let us select Simple. The Heads—the paragraphs full of words, organized by part of speech and according to whether they represent the idea or its precise opposite—that fall within this classification are self-explanatory: Action, Activity, Haste, Exertion, Fatigue, Agent, Workshop, and their opposites, which are Inaction, Inactivity, Leisure, Repose, and Refreshment. Agent and Workshop, being neutral aspects, do not have obvious opposites, of course.

Under one of those Heads—let us choose Haste—are about 200 words and phrases that can properly be used to express portions of the range of this single idea. Some of them are nouns: dispatch, urgency, impetuosity. Rather more of them are adjectives: hot-headed, breakneck, slapdash, immediate. About seventy-five verbs connote individual voluntary simple haste, including scurry, bustle, fret, cut and run, make oneself scarce, whip, lash, bundle off. There are a few adverbs, such as feverishly, on the spur of the moment, with not a moment to lose, and a very few interjections—Buck up! Quick march! At the double!

Within any one paragraph the words are grouped according to the manner and context in which they are used—whether, for example, they are usually employed in colloquial or in formal circumstances. Roget persuaded Longman to allow him to organize the book's Heads into two columns, with words of opposite meaning across from each other. So set against Haste is (no, not Less Speed; there is precious little wit in Roget) Leisure. Among those arrayed against over-hasty is deliberate (though with no phonetic explanation to guard against the verb). Against expedite is, among others, while away.

In formulating this dual-column plan, however, Roget very promptly came up against a problem that has long plagued semanticists interested in synonymy and antonymy: in countless groups there are in fact not two but three shades of sense—the meaning, its opposite, and a middle ground. Beginning, middle, and end are one obvious example; past, present, and future are another. And the more the nature of the middle-ground word is subject to scrutiny, the more it changes. In some trinities noted by Roget in his introduction, the central word is the opposite of both the others—concavity, flatness, convexity, for example, or desire, indifference, aversion. In some groups the middle word is the standard against which the others are measured: sufficiency exists between insufficiency and redundance.

To rub in the difficulty of trying to make semantic order out of English linguistic mayhem, what about groups in which the middle word represents an imperfect state of the other two? Dimsightedness is less perfect than both vision and blindness. Semitransparency is a less perfect state than either transparency or its opposite, opacity. To take this a little further: damp is neither wet nor dry, tepid is neither hot nor cold—and neither of these last two middle words has a precise opposite.

So should there have been three columns in Roget's book? Roget originally thought so, but he eventually recognized the practical difficulties and the cost of such an arrangement, to which his publisher was hostile. The plan was dropped. But in and of themselves, such difficulties with the internal classifications of synonymy need not have made Roget the force for literary ill that I believe it to be. If anything, those difficulties served to remind people interested in the vocabulary what a superbly complex entity the English language is. They served further to enhance public respect for the language, and to emphasize still more widely the great care that needed to be employed to ensure its best use.

No, the central shortcoming in Roget's Thesaurus, as I see it, stems not from the book's troublesome structure but from something quite different—from Peter Mark Roget's Panglossian regard for the intellectual merit of his likely readership. Roget never imagined, for instance, that an Ohio sophomore majoring in political science might one day use his book to find a word with which to pad out a paragraph in a midterm paper. Roget never envisioned that paperback editions of his work would be stuffed into millions of school backpacks and satchels from Huddersfield to Hobart, or that a barely literate board chairman bound for Liverpool would have his secretary's volume by his side as he was writing his report to shareholders on the morning express from Euston.

Roget assumed, as he organized his work, that anyone who might chance upon it would be just as clever as he, just as accustomed to precise syntax, to scrupulous grammar, and to confident and impeccable word selection. Armed with this naive set of assumptions, he produced a book that was predicated on the misguided belief that, as he wrote in his introduction, users would guide themselves through the thicket of words by relying on what he grandly termed their "instinctive tact." Thus there was no need to explain what any word meant—because his users, with all their "tact," would know that full well already.

And so an icily precise classification was all that was necessary. Definitions would be so cumbersome to include, so time-consuming to assemble, so costly to publish, so unnecessary, so—so insulting to his highly accomplished readers. Thus, crucially, Roget, who already possessed a deep knowledge of the English vocabulary, decided not to include them. And yet, of course, neither the Ohio sophomore nor the Liverpudlian businessman had such knowledge. Neither had the ability to make proper use of the volume Roget created. For such as these, the Roget that they were persuaded to buy and to use—wrongheadedly and irresponsibly, in my view—provided no more than an unexplained and inexplicable list of quick fixes. Each user had a sudden want. Each needed a word. Each reached for the Roget—and presto! The way the book is arranged makes it all appear easy, a quick solution in an efficient microsecond. And yet, precisely because the users are ill-versed, and because the book makes utterly invalid assumptions about their knowledge, and offers no help at all in discovering what anything means, the word chosen with each presto! is often wrong. Sometimes very wrong. Often slightly wrong. And at the very least, frequently, curiously, and discordantly off. For example, a freshman student of mine, who admitted to using Roget, attempted to improve the phrase "his earthly fingers" by changing it to "his chthonic digits."

Each time such a wrong is perpetrated by way of Peter Mark Roget, the language, as spoken, written, or read, becomes a little worse, a little more mediocre, and a measure more decayed, disarranged, and unlovely. And that, I suggest, is why all Rogets should be shunned.

Back in the 1850s only a few recognized the potential shortcomings of Roget's Thesaurus. One of them was Edwin P. Whipple, who wrote a perceptive essay in the 1854 issue of the North American Review. His bitingly amusing remarks (which do not dwell on the extraordinary assumptions Roget made about his customers) deserve to be quoted at length, which Emblen does.

We congratulate that large, respectable, inexpressive and unexpressed class of thinkers, who are continually complaining of the barrenness of their vocabulary as compared with the affluence of their ideas, on the appearance of Dr. Roget's volume. If it does nothing else, it will bring a popular theory of verbal expression to the test; and if that theory be correct, we count upon witnessing a mob of mute Miltons and Bacons, and speechless Chathams and Burkes, crowding and tramping into print. Dr. Roget, for a moderate fee, prescribes the verbal medicine which will relieve the congestion of their thoughts. All the tools and implements employed by all the poets and philosophers of England can be obtained at his shop. The idea being given, he guarantees in every case to supply the word ... Indeed, if the apt use of words be a mechanical exercise, we cannot doubt that this immense mass of the raw material of expression will be rapidly manufactured into history, philosophy, poetry, and eloquence.
Seriously, we consider this book as one of the best of a numerous class, whose aim is to secure the results without imposing the tasks of labor, to arrive at ends by a dexterous dodging of means, to accelerate the tongue without accelerating the faculties. It is an outside remedy for an inward defect. In our opinion, the work mistakes the whole process by which living thought makes its way into living words, and it might be thoroughly mastered without conveying any real power or facility of expression.

A reference book of this nature and with this express purpose, Whipple continued, was certain to spread the contagion of literary mediocrity. What was needed was not more information (oh! how today's users of the Internet might consider that anew) but more inspiration. Not more words to make easy the expression of ideas but more energy to make more probable the conception of ideas. Not so much message, not so much medium—but more, many more, marvels of true creation.

And Roget, Whipple implied, never had an original thought in his life. He was a mere classifier of the existing order, a pedant, a noble dullard who should no more have been let loose on the language than a civil engineer should be let loose on the west door of Chartres, or an industrial chemist on the manufacture of Haut-Brion. Whipple, amused and appalled, saw Roget as a man devoid of poetry, a sufferer from a "fluent debility,"

which never stumbles into ideas nor stutters into passion, which calls its commonplace comprehensiveness, and styles its sedate languor repose, would, if put upon a short allowance of words, and compelled to purchase language at the expense of conquering obstacles, be likely to evince some spasms of genuine expression; but it is hardly reasonable to expect this verbal abstemiousness at a period when the whole wealth of the English tongue is placed at the disposal of the puniest whisperers of rhetoric,—when the art of writing is avowedly taught on the principle of imitating the "best models,"—when words are worked into the ears of the young in the hope that something will be found answering to them in their brains,—and when Dr. Peter Mark Roget, who never happened on a verbal felicity or uttered a "thought-executing" word in the course of his long and useful life, rushes about, book in hand, to tempt unthinking and unimpassioned mediocrity into the delusion, that its disconnected glimpses of truths never fairly grasped, and its faint movements of embryo aspirations which never broke their shell, can be worded by his specifics into creative thought and passion.

Whatever the carping and the cavils, Roget and his publishers swiftly realized that their creation was a gold mine. The first edition of the Thesaurus, bound octavo, sold out its initial printing of a thousand copies by the end of the year. A second edition came out the following March, and a third—"cheaper ... enlarged ... improved"—went on the streets bound duodecimo in February of 1855. This edition—for which Roget rewrote part of the text and added "many thousand" new expressions and subsidiary headings that in his view filled gaps in the original structure—was used as the basis for so many subsequent printings that the steel-and-antimony plates were eventually worn smooth and useless.

The first American edition (1854) was edited by a man named Barnas Sears and published by Gould and Lincoln, of Boston. Roget conceived an immediate dislike for it, writing that "an imperfect edition of this work was published at Boston ... in which the editor, among other mutilations, has altogether omitted the Phrases ... and has removed from the body of the work all the words and expressions borrowed from a foreign language, throwing them into an Appendix, where ... they are completely lost to the inquirer." He would perhaps have been more pleased when the distinctly different International Edition was published in New York, in 1922, by Thomas Crowell (which had bought the rights to the book in 1886 and had published for the next three and a half decades what was essentially a facsimile of the original, despite its being manifestly designed for the English-speakers of England). The rationale behind the International Edition was abundantly clear. As C. O. Sylvester Mawson wrote in his preface to Crowell's Roget's International Thesaurus, "The English language marches with no frontiers; it is a world possession."So we find words and expressions that were much better known on the Continent than in either America or Britain. Under the heading for Haste and Leisure, for instance, we find brusquerie and its Latin converse, otium cum dignitate. (In the newest, fifth edition of the International, published by HarperCollins in 1992, both these obscure forms have vanished, though the Latin term that was under the Leisure column has been replaced by the Italian dolce far niente, which is amply supplemented by the phrases ride the gravy train and lead the life of Riley.)

Roget's son, John, took over the Thesaurus when Roget died, at ninety, in 1869. John was more than modest about his own achievements (his only other publication was "A History of the Society of Water-Colourists," in 1891) and insisted that any changes he made to the great book during what was to be thirty-nine years of work (up until his death, in 1908) were "almost entirely of a practical nature, demanding industry and attention, rather than philosophic culture or the learning of a philologist."

John Roget did, however, engineer one important organizational compromise during his tenure: he extended his father's embryonic system of cross-referencing and fine-tuned it over the years, greatly simplifying a work that was in danger of collapsing under its own weight because of the rapid proliferation of words. (We like to think that our time is producing an uncountable welter of new words; but in comparing the mere 45,000 English words recognized by the last editions of Samuel Johnson's dictionary, published in the 1860s, with the 414,825 listed in the first edition of the OED, published in 1928, one can perhaps understand the pressure under which thesaurus compilers of the time were compelled to work.)

As the editions thundered out, the new words included in them—electrolier, lorry, motor-car, veldt, and outspan were all added under John Roget's suzerainty—reflected new technologies and the war in Natal and Cape Province. When television was added, it was put into the class of concepts that is termed Intellect—the exercise of the mind. (Sour minds might wonder at the propriety of classifying one of the senses of television under Intellect, particularly if they see on the screen such puzzling phenomena as Jerry Springer, Dan Rather, and Ron Popeil's Pocket Fisherman.)

Peter Mark Roget's primary intent in creating his book was a noble one—avowedly Platonic, Aristotelian, a monument to the Almighty and His purpose. Or at least it was until the very last minute, when Roget decided to include a feature he had earlier intended to forget—a feature that changed for all time the role that this remarkable book would play: an index.

When the index was finished, it was a miserly thing—hastily done, fully advertising the reluctance of its compiler. But because it made the book much easier to use—though in a way very different from what Roget had intended—it was to swell mightily over the coming years. John Roget worked hard to increase the size and scope of the index. Today the index to the British edition is twenty pages longer than the thesaurus itself. The index to Roget's International Thesaurus, in America, though set in a typeface two points smaller than that of the main body of the book, still occupies half the number of pages the thesaurus does; it would be far longer were it printed in the same size.

The index is and always has been what everybody uses. The classification system is something of which almost no user of Roget is even vaguely aware. I defy all but the specialists among readers of this article to claim that they knew, for example, that deodorant, henpecked, box-office, and consuetude can be found in a class Roget called Volition, or that dog collar, privet, fulcrum, and clotheshorse are in the class he called Space. However noble Roget's design, no one uses it and few care about it; if there was once a Platonic ideal for his book, it is subordinate to the relentless usefulness that was brought about, at a stroke, by the inclusion of the index. As Roget might well have grumbled, that index represents the chicane that separates the original intent of the book from its present vulgar function.

This brings us again to the last of the three introductory questions: What have Roget and others in fact achieved? All the thesauri, word finders, and lexicons that compete for our attention and for hundreds of millions of dollars each year on the reference shelves and the back-to-school tables at Borders and Barnes & Noble have much to answer for. Some are crude devices for advancing lexical laziness; some offer amusing and intelligent ways of examining the marvels of the language. None has the biblical authority of Roget, however—which is why, though they merit brief consideration here, none will receive the withering blast that I think should be directed at the primary culprit. Roget may be a rogue; the others are just the naughty boys of the school yard, the hangers-on, the keepers of bad company.

There are many types, and there have been many experiments. An initial distinction should be drawn between conceptually arranged thesauri, of which Roget is the obvious example, and those—like the American Century Thesaurus, from Time Warner, and the Oxford Paperback Thesaurus, "in Clear A-Z Form with Most Useful Words First"—that have no truck with concept but are arranged wholly alphabetically. The relative benefits of alphabetical books can be fairly easily demonstrated—by seeing which words are included and which are not. Conceptual arrangements permit looking-up only in the index, and the indexes are not necessarily updated as frequently as are the word lists themselves; a new word may well be slipped hurriedly into the body of the book and left there as a serendipitous surprise.

Laurence Urdang, who founded the journal Verbatim and who has written or edited more than 120 dictionaries of various kinds, including the 1966 Random House Dictionary of the English Language—Unabridged Edition (he is currently working on a dictionary of historical nautical terms), has been responsible for various word finders, some high-minded, some pleasingly silly. His company, Verbatim Books, in 1982 published Word for Word, by Edward C. Pinkerton, one of the more serious (but far from solemn). It offers extraordinary chapters (with numbered lines, so complex a book is it) that show the links between seemingly unrelated words. The flap-copy tease is the relationship of armadillo, arthritis, harmony, reason, and rituals, all of which have some connection to the Indo-European root ar-, which means "to fit together." (On a more frivolous level, Urdang's alter ego, the suspiciously anagrammatic Claurène duGran, who is reputed to have "studied linguistics at Oxford and drama at Cambridge," put out into the world twenty years ago a book called Wordmanship, which offered the linguistically insecure a raft of suggested synonyms to make their daily lives richer and more impressive. She recommended using accipitrine for hawklike, discommode for inconvenience, and depilate for shave. The noun form of the third has in the past two decades indeed made its way into the language of Madison Avenue.) Urdang also gave us the Dictionary of Differences (1988), which offered an alphabetically arranged list of useful but easily confused words, concepts, and fancies: incredible and incredulous, baking powder and baking soda, and the pentarchy iamb, trochee, spondee, dactyl, and anapaest.

Urdang is thus in a good position to have favorites among the better-known thesauri, and one of them is Francis March's huge, out-of-print Thesaurus Dictionary: A Treasure House of Words and Knowledge (1903). White-bearded and benign, photographed for the book's frontispiece seemingly floating in an ocean of scholarly paperwork, March presides over the instructions to the reader. His aim was that of all the other thesaurus makers—though not so noble, of course, as Roget's ideal. He hoped that his method would help users "definitely express a given idea."

March's thesaurus contains a vast alphabetical list ranging from A-1 to zymotic. For each word—and, crucially, after the word's definition—it offers one or more pairs of words representing the range within which the particular word is used: for example, A-1 is given the range goodness-badness. The reader may turn to this recommended pair of words in the alphabetical list and come upon an immense secondary list of several hundred nouns, verbs, and adjectives, all with definitions, that offer synonyms and antonyms to A-1. Beneficence, bijou, value, and virtue on one hand; depravity, malevolence, peccancy, and pestilence on the other. And each of these words, too, has a definition.

But March's thesaurus requires work. One must make an informed selection from among the vitally important series of definitions offered in the range within which a word might be used. Regrettably, the demanding nature of the task seems to have ended publication of this unusual and rather wonderful book after little more than twenty years of life. Yet in terms of fostering the proper use of English, March's was in truth a very good thesaurus, for precisely the reason that Roget, which fosters showy mediocrity by offering no information, is a bad one. March suggested a range and then forced the reader to make a choice. Roget presented a list and said, essentially, pick a word, any word. March simply asked too much of the reader: in the matter of words, making choices calls for both a modicum of intelligence and a time-consuming process of thought.

Let us say you want to use Roget to find a word that means "habit" (relating not to the convent but to the ashtray or the essence of juniper). After looking up habit in the index, you find the above-mentioned consuetude; but as to what precisely consuetude means, the book is silent. Unless you take care to consult a proper dictionary, you have no idea whether it is suitable for your purpose. You are dimly aware only of its range of use, and of other words that share that range—cacoëthes, for example, surely one of the ugliest words ever made. No thesaurus except March's will tell you what on earth this word means either, of course. March's will tell you very fully: it will tell you that the word means "a bad habit" and that it lies in the usage range of habit-desuetude, and it will also inform you of the phrase cacoëthes scribendi, meaning an unbearable and indescribable itch to write—the itch, perhaps, that has compelled me to get this far along without summing up why I think Roget is, in the literary universe, such a serious force for bad.

Thesauri came about in the first place in response to the desire of men and women of society to speak and write more fluently; the immediate popularity of these books, like that of Robert Cawdrey's Table Alphabeticall, arose from that ambition. Some of them proved helpful; Piozzi's, unfashionably prescriptive though it was, suggested ways in which words might be used, and in doing so gave the language a nudge in the direction of improvement. The flowery barbarisms of inkhorn terms vanished within years of the appearance of her book—no cause and effect established, but the possibility observed.

Roget, however, was not intended to address the want of drawing-room conversation; it met a very different set of needs. The book's real popularity was achieved—and sales figures more than amply confirm this—in the immediate aftermath of one fateful Sunday in December of 1913. That was the day that the popular newspaper the New York World published in its supplement Fun a small matrix of black and white squares into which readers were directed to write words that, in some clever way, correctly responded to a list of clues printed alongside the matrix. The crossword puzzle had been born.

By the 1920s the craze had spread across America and across the Atlantic; by 1930 the venerable Times of London had a puzzle, the speed of completing which was used as a test of ability and intelligence. The provost of Eton, it was said, finished this exercise each morning in the time it took to boil his egg (from cold).

But there were, especially in Britain, strict and unspoken rules. Self-respecting people figuring out the answers to clever crossword puzzles never, ever used dictionaries—or any other reference books. To do so was an admission of defeat. It was simply not done.

From the archives:

The Atlantic Monthly Puzzler
A collection of word games by Emily Cox and Henry Rathvon from July, 1997, to the present. Each Atlantic Puzzler has its own theme, accompanied by special instructions.

"Masters of the Tiles" (June 1987)
Even to initiates, Scrabble has yet to yield up all its secrets. By Barry Chamish

Because many of these early crossword puzzles offered monetary rewards, a goodly number of stupid, callow, competitive, and greedy people did not observe these proprieties. They had a pressing need—to gain standing, to win dollars or pounds, to best the fellow next door. They yearned for a tool for looking up unknown words: a book that listed words that were similar to other words, so that if the clue was "habit," they might find the answer in a list that stretched from cacoëthes and consuetude to disposition, leaning, and mannerism. Thesauri answered the call. Scores of them were on sale—the 1920s equivalent of the 935 thesauri listed for sale by Amazon eight decades later. And the king of the hill even then was Roget.

Roget's Thesaurus, which had come into being as a linguistic example of the Platonic ideal, became instead a vade mecum for the crossword cheat.

It already had other, more insidious shortcomings. By eschewing definitions altogether, and thus suggesting no choices, it fostered poor writing. It offered facile answers to complex linguistic questions. It appealed to a growing desire for snap solutions to tricky verbal situations. It enabled students to appear learned without ever helping to make them so. It encouraged a malaprop society. It made for literary window dressing. It was meretricious.

But a tool for cutting corners? How Peter Mark Roget might have turned in his grave: the book by which he set so much store, his most lasting memorial, was being used for petty and degrading purposes—assisting no one with the language, but boosting the circulation of tabloid dross. And its usage has not widened significantly since (not, that is, in the sense of becoming more than a quick and easy remedy for the lexically distressed—for the literary poseur, if you will). A student in want of a word? No need to expend mental energy, no need to wait until blood prickles from the forehead: Roget will supply the answer, will find the syllables to plug the hole, will offer the solution with no delay or fuss. No need to bend, spindle, or mutilate.

I thought when I began this that I might telephone or write a representative few writers, to ask if they used Roget to any measurable degree. I wondered if I might drop a line to John Updike, or Saul Bellow; I wondered about those writers who obviously have exceptional vocabulary control: Richard Ford, Anthony Lane, Tom Wolfe, Calvin Trillin. Or William F. Buckley, surely the writer with the greatest store of words. I thought about Robert Lowell, Anthony Burgess, Julian Barnes. I imagined lyricists—Cole Porter, Tom Lehrer. I thought first about the translator of the Norwegian Nobel Prize-winner Knut Hamsun, and second about the most recent translator of Dostoevski. Did any of these refer to Roget? Would any of them admit it? And what would be their overall view of this century-and-a-half-old book? Did they find it inspiring? Irritating? Do they use a thesaurus as much as they do a dictionary? A phone book?

In the end I neither called nor wrote—because I knew the answer to be obvious. Everyone has the book. Occasionally one makes use of it. But one never, never relies on it to help with the making of good writing. It may be used once in a while, to jog the memory, to unstall a synaptic moment. But it should never be trawled through or mined; its offerings should never be taken and transfused into a paragraph as relief for emptiness of thought.

Whatever merit all these paragraphs may have as writing, it seems appropriate to mention that Roget was not once employed in the selection of les mots justes. Such words as these paragraphs contain came, as they should do for all writers, from within—from memory, experience, conversation, reading, imperfectly recalled strands of knowledge.

And anyway—it is not simply by richness of word supply that a great or even a good writer makes a mark. It is by the acuity with which he or she uses the words that mind and passion have placed at the ready. It is the brave employment of the words that one already knows that makes the writing fine or no. And no increase in their numbers will do very much to improve the temper or the temperature or the pace if mind and heart and argument and passion are wanting. An engineer faced with the challenge of working on the west door of Chartres would do no better if he had more equipment: perhaps, in fact, he would ruin the greatness there more quickly, would more efficiently reduce it to something utterly without inspiration or art.

Consider, in conclusion, two simple passages, both drawn from the first chapter of Sir Ernest Gower's The Complete Plain Words (1954), a British-government-issued book that is still the Bible of the best English writing. The first is from a Shakespearean sonnet that begins, "Full many a glorious morning have I seen / Flatter the mountain-tops with sovereign eye," and continues, as Gower cites, "Kissing with golden face the meadows green / Gilding pale streams with heavenly alchymy."

There was no Roget when Shakespeare wrote those lines—or, for that matter, any dictionary. The lines were written perhaps a decade before Robert Cawdrey and his Table Alphabeticall, 250 years before Peter Mark Roget and his high lexical ideals. And yet the writing is perfect, in choice and arrangement of words, in thought, timbre, address, and note.

As, on altogether another level, is this—a notice that was once placed inside all British post offices: "Postmasters are neither bound to give change nor authorised to demand it."

No Roget was employed here, either. The civil servant who penned those words knew just what he wanted to say, and had a mind lithe and educated enough to come up with a sentence in which every word counts, not one is superfluous, and the whole has a harmony that in its modest way achieves the greatness of poetry.

So, indeed, Peter Mark Roget, physician, chess genius, expert on bees, phrenology, and the kaleidoscope: for all your noble ideals and Aristotelian logic, your book offers comfort only to the few—some clues for crossword cheats, some natterings for speechwriters, and some quick and easy solutions for the making of the middlebrow, the mindless, and the mundane. Roget has become no more than a calculator for the lexically lazy: used too often, relied on at all, it will cause the most valuable part of the brain to atrophy, the core of human expression to wither.

To make us think a little more, to make us wonder a little longer, might it not now be for the best for your book simply to vanish, and for the name that is at the root of the eponym to be banished from the lexicon for all time? Perhaps the Encyclopaedia Britannica is right to give you only twenty lines. Perhaps Microsoft has the meet and proper solution: to help render Roget into a nonword once again, and let the memory of the man and his creation fade, so that we can clamber back to a plainer, simpler world, where dictionaries and encyclopedias alone are set to guide us, and where our literary powers are born not out of banal and mediocre suggestion, not out of lexical shopping lists, but out of passion, thought, and intensity of feeling.

This article available online at:

http://www.theatlantic.com/magazine/archive/2001/05/word-imperfect/302226/