by Kory Stamper
We trot this rubric out all the time as lexicographers—widespread, sustained, and meaningful use—and then follow up with “If a word meets the criteria for entry, then it’s time to draft a definition.” But there is an intermediary step we don’t talk about at all: you have to define before you can define.
All words are defined contextually, so before you can know that there’s a new sense of a word to be entered, or a new word, you have to read the collected evidence for it and determine if that marked use is covered by the existing entry. It’s not always an easy task: just as words can slide around between parts of speech, so can they slide around between meanings. For instance, what do we do with the “cynical” in this sentence: “It was concluded that students experiencing loneliness report a greater level of unhappiness, display signs of detachment during social interactions, and are more cynical and dissatisfied with their social network.” This use of “cynical” doesn’t seem to quite fit the existing “contemptuously distrustful of human nature” meaning, as in “Voters have grown cynical about politicians and their motives,” nor does it fit the meaning that refers to human conduct being primarily motivated by self-interest, as in “It was just a cynical ploy to win votes.” It seems rather to mean “pessimistic”—not that the students are contemptuously distrustful of human nature, but that they are certain that the terrible status quo will continue to be terrible. But if one is distrustful of human nature, and contemptuously so at that, then wouldn’t it follow that the cynical person doesn’t expect things to turn up sunshine and flowers? Back to the citation. Do you shrug and toss it onto the “contemptuously distrustful” pile? Is this a new, possibly emerging meaning? If it’s not, do you need to revise the entry to cover this use?
The longer you define, the better you become at determining how close to a definition a citation falls and how far away it needs to be for it to be its own sense. Meaning is a spectrum; you are only describing the biggest data clusters on that spectrum. Madeline Novak puts it this way: “There’s a meaning there, and it could be sliced up any of a variety of ways, none of which really capture the whole thing. You’re going to be dissatisfied with it no matter what, so you’re kidding yourself if you think you’ve pinpointed it. There’s still stuff oozing around the edges.”
—
All the editors at Merriam-Webster go through a few months of training to learn how to parse the grammar of a word, how to read and sort the citations, and then how to write a definition that fits any new uses of that word.
There are a handful of tricks we use to get us in line, the first being the formulaic definition. You know what these are: they’re dictionaryese. “Of or relating to” blah-blah-blah; “the quality or state of” yadda yadda; “the act of” et cetera. They exist not because we’re lazy but because they can be helpful ways of tagging a definition as a particular part of speech, and definers can use formulaic definitions to help orient themselves. They can also be helpful tools to link together linguistic cousins. If I define “devotion” as “the quality or state of being devoted to (someone or something),” I’ve subtly communicated to the reader that “devotion” is closely related to the word “devote.”
While formulaic definitions are helpful at times, they can also be too restrictive. This is when lexicographers lean on their second tool: substitutability.
The idea behind substitutability is that a well-written definition should be able to slot into a sentence in place of the word being defined, and while the result may be verbose, it won’t sound wrong (as the definitions given for that “hella” T-shirt above do). Substitutability can actually help you scale down a long definition. If I were defining “hella” without using substitutability, I would probably come up with something long-winded and full of dictionaryese, like “to an excessive degree.” And it works: “That T-shirt is [to an excessive degree] awesome.” But using substitutability, I’m more likely to define “hella” as “very” or “extremely”: “That T-shirt is [very/extremely] awesome.”
While you are getting the hang of writing nounal and verbal definitions, you must also master dictionary house style. Every publisher has a house style, or a house stylebook: that’s where the publisher lists how a particular word, or compound, or bulleted list will appear in all of its books. This is for consistency, one of the Holy Grails of practical lexicography. None is pure enough for such a quest; at Merriam-Webster, this is where the Black Books come in.
Every editor at Merriam-Webster deals with the Black Books at many points during their tenure. The Black Books are the in-house set of rules for writing a dictionary (commonly called a style guide) as conceived and written in punctilious detail by the former editor in chief Philip Babcock Gove, for the creation of Webster’s Third. They are named for the black clothbound binders in which the single-spaced memos find their eternal repose.
The Black Books are formidable. They are, for the most part, the brainchild of Gove, and so reflect his borderline pathological attention to detail. Gove wrote a memo for our editors called “PUNCTUATION and TYPOGRAPHY of VOCABULARY ENTRIES” that begins, “This memo is concerned primarily with how and does not, unless expressly stated, cover when or why,” and he means it. In that memo, Gove goes on to lay out, in excruciating detail, the basic pattern for all Merriam-Webster definitions. The memo is thirty-three pages long. It even lays out for you where in definitions you are allowed to put spaces.*3
Don’t think that Gove was a windbag: he was a New Englander and valued sparse efficiency in all things (including lexicography). So it says something that the memos are so long. To be thorough as a lexicographer is to pay attention to the smallest of details. This is not a haphazard, minor-leagues approach to defining. Go granular or go home.
The Black Books also reflect Gove’s notoriously brusque manner, no doubt gained from an early career in the Navy. Memos begin, “Editorializing has no place in definitions,” or “Godlove’s psychophysical defs of color names and their references had better be regarded as sacrosanct.” Sir, yes, sir!
There is an otherworldly quality to the books themselves. For such important documents, they are housed in odd nooks and crannies in the office—in a short bookshelf near the former editorial secretary’s desk, on top of a deserted maze of cabinets that contain the paper galleys for earlier copies of the Collegiate. One set resides with the president of the company; another wanders the office and shows up in odd places without warning. One morning I left my cubicle and paused to see what grotesquerie the life sciences editors had been looking up in the medical encyclopedias near my desk*4 when I saw two of the Black Books lurking dustily on the corner of the encyclopedia table. I flipped the cover of the topmost volume open; it was a collection of defining technique memos. Just opening the cover, I could smell 1952 waft out: typewriter ribbon and mimeograph paper tinged with stale tobacco, laid over a base note of Gove’s perpetual disappointment and irk.*5
Individual books (and each of their subsequent editions) have their own style guides, because what worked for Gove in the 1950s doesn’t always work for us today. That said, each of those style guides, no matter how many hundreds of pages it runs to, relies ultimately on the Black Books. If you have a question about anything having to do with the mechanics of defining, you’ll find the answer there. Obviously, no one wants you to stutter up to their cubicle covered in flop sweat and asking how to write an adjective definition again; it’s much easier to direct you (wordlessly, scowling) to the Black Books for your answer.
—
The only real way to master house style is to practice it, which is what all editors at Merriam-Webster get to do when they are plucked, fresh-faced and wet behind the ears, from the vast sea of editorial candidates and plopped, still wriggling, down at desks with editing homework. The best way, our senior editors felt, to learn how to define was to spend some time in quiet, reflective imitation.
First we begin by going through the fifty-odd pages of defining theory that Gil and Steve have come up with
over the years. This gets into the minutiae of the definitions and defining process itself, and it is based—you guessed it—on the Black Books. You are then handed a sheaf of paper and asked to edit the definitions found on it, all of them having been taken from earlier Merriam-Webster dictionaries. You get another such worksheet for where to enter phrases, another one that asks you to fix the capitalization of older entries, and another that asks you to fix the inflected forms of older entries. This is existentially unnerving, because these definitions are taken from published dictionaries, which means they were written by people who had more training and practice than you do. It is your memento moron: no matter how smart and excellent, remember that you, too, will fuck up.
Soon, you learn how to tell the difference between a boldfaced colon in four-point type and a lightfaced colon in four-point type. You learn to tell when something wasn’t written using the right formula or when a definition accidentally editorializes. You learn when words get labeled “usu. cap” and when they get labeled “sometimes cap.” You even go over the standard proofreader’s marks so you can read galleys and old defining slips and know the difference between your and italic. And you feel as though you are scrambling up an avalanche, desperately looking for any foothold you can.
Once the worksheets are done, every new hire is given a batch of practice defining to go through; mine was a chunk of B and some sundries in other parts of the alphabet. You will define these words as if they were being entered: you must write your definitions on buffs, you must make sure you’ve got everything right, and you will date stamp them just so the practice becomes muscle memory. Then you will hand them over to Gil or Steve to be thoroughly smashed to bits.
Because I am a sentimental pack rat, I kept my practice defining slips and every once in a while pull them out and marvel at how bad a definer I was. I routinely forgot important style issues: half of my definitions don’t start with the boldfaced colon, and I used sense numbers for definitions that only have one sense (a major no-no: “What’s sense 2? You can’t have a sense 1 if you don’t have a sense 2,” Gil would harrumph).
More frustrating were the things that I did wrong that weren’t easily corrected:
I have committed a venial sin against the style guide (that damned sense number), but two mortal sins against the evidence. First, the subsense after “especially” is too broad as I’ve written it: the most common use of the word “jugate” doesn’t refer to just any collectible, like a spoon or ceramic salt and pepper shakers, but to a campaign button. I’ve covered the broader use in the first sense, but the point of the “especially” is to zero in on the most common and specific use. The subsense after the “especially” should read “a campaign button showing the heads of a presidential candidate and his running mate.”
But that’s not the worst of it. I have also, as Gil pointed out, gendered the definition when I didn’t need to. He struck the “his” from “his running mate.” “It is conceivable,” he said afterward, while going over my definitions with me, “that a woman will run for president at some point, and if she does, this definition will need to be revised. So why not write it in a way that the gender of the candidate doesn’t matter.” I was gobsmacked: here I was, a recent graduate of a women’s college, getting schooled on gendered language by an old guy. And rightly so: I brought to the definition my own assumptions about the gender of people who have in the past run for president. But a good lexicographer weighs the untold future as well as the told past and present: Is it so inconceivable that a woman could run for president? By assuming that the gender-neutral “he” was going to be fine here, because presidents are all men, I had committed the sin of editorialization.
I also had problems—as most new definers do—figuring out just what information to include in a definition. My scratch definition of “naja” is “a crescent-shaped pendant made by the Navajo people.” Is it just made by the Navajo, or do the Navajo wear them as well? Is it worn by other people? If a white tourist stops at one of the ubiquitous roadside gift shops you find throughout the Southwest and buys a necklace with a naja, then wears it, does it cease to be Navajo? Does it cease to be a naja? Would it be better to say it’s characteristic of the Navajo? And what does that mean? Joan Narmontas, our senior life sciences editor, says of defining, “It’s complex systems made simple,” but my definition is its inverse: a simple system made needlessly complex. There’s not quite enough information in this scratch definition; it ends up leaving the dictionary user with more questions than answers.
And I brought plenty of my own assumptions and biases about what makes a good dictionary definition with me, as we all do. I was tasked with defining “outershell” and gave the definition as “a protective covering.” Gil revised the definition to read “an outer protective covering.” At our next meeting, I protested—you’re not supposed to use the word you’re defining in the definition! It is a truth universally acknowledged (and enforced) by American language arts teachers everywhere!
“Well,” he said, grinning, “ ‘outer’ is not the word being defined here.” I pressed anyway—does “outer” need to be there? It looked a little, well, lazy. It was important, he said: you need to convey where the covering is, that an outershell is, in this case, something you wear over your clothes or put over an item.
I was perturbed. “Covering,” to me, already conveys outsideness, not insideness. It is covering something; there is something inside it; it is outside the thing it’s covering. Q.E. Motherfucking D. Gil ignored my botheration and we moved onward, but I secretly thought he was needlessly nitpicking.
Pride goeth before a fall: that afternoon, while proofreading, I discovered that there is a covering around the heart called the pericardium, but it is not outside; it is inside the body. The pericardium is inside the body, and the heart is inside the pericardium. I threw my head back and indulged in a quiet bleat of frustration. I was never, ever going to get the hang of this.
Two days later, I was sitting in class with Gil while we began going through our B batches. I trained with two other editors, and it was common for Gil to have us read our definitions aloud so that we could benefit from hearing how dumb we all were. We went through a few entries, then got to “birdstrike.” The other two editors went first, with Gil giving them some constructive critique about scope, usage, whether you need to split the word into two senses, and how we can probably just make do with one. Then he looked at me:
Gil rolled my definition around in his mouth for a bit. “All right,” he said. “That’s a pretty good definition.”
I’m sure he gave some constructive critique, but I didn’t hear any of it. I was on cloud nine. After months of banging my way through training and feeling, at almost every turn, like I knew little about most things and definitely nothing about English, I had some glimmer of hope: I drafted a pretty good definition. It wasn’t all needless nitpicking; it was needful nitpicking. We finished the session, and I went back to my desk, pulled my standard-issue monthly planner toward me, and wrote in the square for September 1, 1998: “Gil/birdstrike: PRETTY GOOD DEFINITION.”
I still pull the calendar out at times, to remind myself that it can be done.
—
There are more oddities in lexicography. The first is the weirdest and appears to be consistent across all traditional dictionary publishers: no one starts writing a dictionary in A. Ever.
When I asked Steve about this—dumbfounded, because where else do you start but at the beginning?—he gave two answers. First, every dictionary—not dictionary publisher, but every book that publisher puts out—has its own style guide, and it takes definers a while to settle into that guide. Hell, sometimes it takes a few letters for the style guide to even be finished. Starting about a third of the way through the alphabet, in short letters like H or K, gives editors time to settle into the new style before getting to some of those bigger letters. Many dictionaries are revised from H (or thereabouts) onward to Z, and then A through H (or there
abouts), with the first middle bunch of letters (H, maybe I) getting another revision.
History bears out this process. Thomas Elyot, a sixteenth-century lexicographer, talks about this in the (incredibly long) dedication and preface to his dictionary. He started in A, farted around a bit until he got the hang of defining, and then, he says, “Wherfore incontinent I caused the printes to cesse, and beginninge at the letter M, where I lefte, I passed forth to the last letter with a more diligent study. And that done, I eftesones returned to the fyrst letter, and with a semblable diligence performed the temenant.” If eftesones returning to the fyrst letter was good enough for Thomas Elyot, it’s good enough for us.
The second reason we don’t start in A is a little more mercenary: back in the days of yore when dictionaries were actually reviewed, reviewers would inevitably start looking at definitions in the first chunk of the book. It takes so long to write a dictionary, and the style file will inevitably change as you go along. You don’t want reviewers catching a style change midway through A, do you? A through D would be, as the last letters worked on, as close to stylistic perfection as possible. No reviewer is going to look too closely at K.
You also discover that not every letter in the alphabet has the same number of entries, and therefore lexicographers look forward to (or dread) certain letters. Pick up any desk dictionary, disregard the front and back matter, and pinch letters A through D between your fingers: you’ll find that A, B, C, and D make up about a quarter of your dictionary. E, F, and G are middling. H is long; we can chalk that up to the surprisingly large number of words that begin with “hand-” and “hyper-.” I, J, and K are relatively tiny. Then you begin the long middle section of the alphabet: L, M, N, O, and P, which always seem longer than they should be, probably because they go by so quickly in the ABC song. Q is a barely registered dip in the road, and you’re back into R, velocity maintained, corners rounded. T is a decent size, and U surprises you (all those “un-” words). V is a comparative breeze; W is about as long as it sounds—double U. X, Y, and Z are nothing, nada, the long stretch after the marathon.