So yesterday Eugene asked me why the word inartful (meaning 'unskillfully') wasn't in any dictionary that he'd consulted, including the OED and all the dictionaries you can search through onelook.com.

He pointed out that this word has been used 600 times in 2007 alone, and that he'd found a cite going back to 1751, which seems like plenty of evidence of use to guarantee a seat at the grown-up table for inartful. So why has inartful gotten the go-by?

When thinking about how words enter a dictionary, the most important thing to understand is that there are many, many more words than there are places in any current dictionary. Because of this scarcity, lexicographers are driven to a kind of triage. Often, the question isn't 'how can I justify including this word?' but 'how can I justify EXCLUDING this word?'

It wasn't that lexicographers just weren't lucky enough to run into the word: we should assume that a lexicographer at one point saw and considered inartful, however briefly. How can I assume that? Well, Ben Zimmer at OUP let me know that the word shows up three times in the citations database the OED uses (Incomings) and shows up thirteen times in the Oxford English Corpus. In the natural course of things, then, a report should have been run on all the words in either of those databases that isn't already listed as a defined word, and the output of that report looked over.

So once inartful was on that list, why didn't it shoot straight through the process to definition and publication?

Well, inartful is fairly easily and superficially defined as "not artful". Often in- or un- words are not separately defined in dictionaries (again, the lack-of-space problem) because lexicographers assume that someone who knows in- and audible can put the pieces together on their own and figure out inaudible without our help. (These words, in lexicogger jargon, are often called derivatives or run-ons, because they are derived from and often run on to an entry, where they appear at the end, after all the definitions, in bold type.)

Of course, there are difficulties with this 'solution' to the space problem, and inartful is a good example of how this assumption of transparent meaning can go wrong. It seems that at the time inartful started to be used, artful meant, plainly, "Displaying or characterized by technical skill; performed or executed in accordance with the rules of art; artistic" [OED]. Artful had not yet taken on its later meaning of "Cunning, crafty, deceitful." [OED] So the in- + artful reading of inartful stopped working, along about the time that artful took on an unsavory character. It's not that the lexicographers were slacking off here (although it may seem that way). It may take several revision cycles for all of the in- and un- (and for that matter the re- and non-) words to get reviewed to make sure that their base word hasn't skewed off in a different direction than the prefixed version. Of course, inartful didn't even get the run-on treatment, as far as I can tell. If a word is rare enough, and it has no changes in its spelling or stress pattern or pronunciation when the prefix or suffix is added, even the run-on space may be judged as too valuable to waste on so marginal a word.

Another reason that inartful could have been left off the guest list of the A-Z party is that the word seems to be used mainly by lawyers. It's not anti-lawyer prejudice on the part of lexicographers (we tend to love lawyers, because lawyers tend to love dictionaries, and, more importantly, buy them ... ). But lexicographers know that legal terminology tends to be both contained in the law world and that the law world has good dictionaries of its own, which will provide adequate coverage. (Black's, especially, although I haven't been able to look up inartful in Black's. I don't know where my copy went!) Any kind of very specific jargon or restricted terminology won't show up in a general dictionary (I'm not really talking about the OED here, which does include a lot of specialist terminology) unless the lexicographer can show that the term does show up often enough in broader contexts.

Which should have been the case, actually, for inartful, since it was the subject of the Safire contretemps Eugene mentioned yesterday (he claimed it "wasn't a word"), AND then he retracted that claim ... based in part on testimony from a lawyer, Fred Shapiro. When a word is the subject of a public debate over its wordiness, that to me says it's a good candidate for inclusion in a general dictionary ... but I'm excusing myself from responsibility for following that particular debate, since it happened in 1985, when I was fourteen and not yet reading the Sunday New York Times on a regular basis (I don't think you could get home delivery of the NYT in Winston-Salem, N.C. in 1985).

(Although now that the NYT archive is open, does anyone want to help me do a survey of all the On Language columns to see what percentage of words discussed are, in fact, included in major dictionaries? I'll take volunteers in the comments, on a first-come, first-served basis ...)

Those are just two of the reasons you can't find inartful between inarch and inarticulate in your dictionary. They're not especially good reasons, but then again the reasons for most failures aren't especially good ones.

And, really, I don't think inartful is an especially rare case. I encounter a word that's not in a dictionary but probably could be nearly every single day. Grant Barrett has built an entire web site that lists words that aren't in the major dictionaries. Dictionaries are probably only the tip of the English iceberg — there might be as much as 90% of the language hiding below the waterline.

So tomorrow I think I should discuss how lexicographers could keep this kind of inartful failure from happening, both ideally and practically. If you want a sneak preview, you might want to check out the video of my talk at the TED conference, where I discuss this same topic. (Warning: sound starts immediately as the page loads.)