When it happens I feel as if I have stepped into a Far Side cartoon. I am a magazine editor, and the galley of an article will come back from a proofreader with a low-frequency word circled and this comment in the margin: “Does this word even exist?” or “Is this a real word?”

Usually the word’s meaning is perfectly self-evident, and the word itself is relatively simple like “unbuyable,” if not deliberately goofy like “semi-idiotic-like.” And I think to myself, of course it exists. Look, there it is, right in front of us.

Sometimes the reader puts his or her suspicion differently and asks, “Is this word in the dictionary?” Having recently spent a large amount of time researching how a particularly well-known American dictionary was made, I have a very different notion of what a word’s presence, or even its absence, in a dictionary implies.

Don’t get me wrong: I like dictionaries, including several that I consult online and most of the 11 that are sitting within arm’s reach as I write this. But my recent affair with lexicography has left me certain of a couple of things.

One is that no dictionary contains every word in the language. Even an unabridged dictionary is, well, abridged. The sciences, medicine and technology generate gobs of words that never make it into a dictionary; numerous foreign words that appear in English-language contexts are left out. A great many words are invented all the time, whether for commercial reasons or to amuse one’s friends or to insult one’s enemies, and then they simply vanish from the record.

Another is that dictionary users and dictionary makers sometimes have very different notions of what a dictionary is for. One may think of it as a legal code for language; the other considers it a very partial report. One wants unambiguous answers about spelling and meaning and grammar and usage; the other aims for neutrality, and the more serious he or she is, the more wary the person is of imposing his or her own notions of good English on the language itself.

From online dictionaries we have learned that the most frequently looked-up words sit along the bookish fringe of everyday language. Among the top ten lookups on Merriam-Webster Online at this moment are holistic, pragmatic, caveat, esoteric and bourgeois. Teaching users about words they don’t already know has been, historically, a primary aim of lexicography, and modern dictionaries do this well.

But, say, you’re already a person of wide reading, and it’s rare that you require such help. And spell-check (despite its own problems with low-frequency words and proper names) is getting you past the usual pitfalls of silent letters, double consonants, indiscernible vowels and other orthographic peculiarities. Perhaps you write for a living. But occasionally, before committing to a word, you like to stop and commune with it, give it a look-over and see what the dictionary has to say about it.

The lexicographer, without a lot of space to work with, has reduced the word to what he takes to be its essential meanings. You ask yourself if the relevant sense matches your proposed use.

Of course, consulting dictionaries in this way is part of our intellectual and cultural training. It goes back to Language Arts homework (“use the word in a sentence”) and vocabulary tests. But the committed writer should be loath to substitute the lexicographer’s (no doubt well-informed but hardly infallible) sense of a word for his own. Not that you get to choose, according to your own whims, what words actually mean, but there is always much more to know about a word than what a dictionary can tell you. For example, to read in certain genres and areas is to see, on a regular basis, words used in ways that lexicographers must ignore or struggle to keep up with.

Dictionary publishers love to send out a press release when they’ve caught on to an important new term from social science or youth culture or technology or politics, which is all well and good, but in following Webster’s you’re following the followers. Language is profoundly conventional, so few of us can claim to be innovators, but the ambitious writer tries to avoid saying what has already been said. This is true for ad copy, political speeches, quality nonfiction and most other types of writing. Journalism, obviously, rests entirely on the claim to be delivering something new. And what is new should sound new, seem new and maybe require quotation marks, your copy editor thinks.

Lately I have been reading “Reporting: The Rolling Stone Style,” published in 1977, which collects feature essays from the magazine’s first 10 years. What strikes me about it is that the reporters do not sound like trade journalists circulating information within a community. Rather they sound like explorers returning from far-off lands, breathless with discovery. Their writing is for people in the know, yes, but to a much a greater degree it is for people who are not in the know.

“Don’t misunderstand, they aren’t your traditional Hilton rubes, this Pasadena burgher and the little woman, they have viewed with compassion the Louds and wouldn’t be caught dead in New York in madras shorts or cameras on straps like talismans, but this, this, it does give them pause and they freeze at the curb like Lot’s wives, hit full-face by the nightmare custard pie of it . . . .”

This was an opening sentence in an article by Tom Burke describing a gay pride parade from the point of view of two well-meaning out-of-towners. Few dictionaries would tell you what the writer means by Hilton, Pasadena, little woman, Louds, “this,” Lot’s wives ornightmare custard pie. Accounting for what words can add up to when writers explore and combine their obvious and non-obvious aspects is not what dictionaries do well or, in most cases, at all.

Early in my career as an editor I grew frustrated with what dictionaries could tell me about words and usage. An ideal dictionary would present an array of real-world examples, weighted somewhat to favor the professional over the amateur but also showing everyday usage alongside more literary examples drawn from books and movies and television and so on.

With the rise of serious online dictionaries, there are some that approach this ideal, including Wordnik and Merriam-Webster’s new unabridged version. They provide more raw information about words, enabling users to draw their own conclusions.

What a dictionary is for is rarely what a writer needs: basic help in using individual words. Of course, when a writer needs such help it is critical that he or she gets it. Only it should be kept in mind that good writing may exceed the boundaries suggested, if not intended, by dictionary definitions. As one lexicographer put it, “Nothing worth writing is written from a dictionary.”

David Skinner is the author of “The Story of Ain’t: America, Its Language, and the Most Controversial Dictionary Ever Published.” He is the editor of Humanities magazine, a member of the usage panel for the American Heritage Dictionary and a contributor to The Weekly Standard.