One of the most common claims levelled against descriptivists, and against linguists of every stripe, is that our linguistic philosophy amounts to “anything goes”. Whenever anyone says something, the thought is, we will take it as a valid sentence in their language.

Of course, prescriptivists and other anti-descriptivists denounce this position as folly. But so do (almost) all descriptivists. The position is intellectually bankrupt. There are many reasons for an utterance not to be assumed to be grammatical. For instance, young speakers of the language speak pretty terribly (“I goed to the store”), so they clearly need to be exempted from the set of speakers establishing the grammar of the language. You will not find a linguist listening to a three-year-old and dutifully transcribing their speech as grammatical forms of the language.

But that one’s pretty obvious. In a more problematic case, we also know that people make grammatical errors that they subsequently recognize as errors. I know this especially well because every third post or so I get a comment or email asking if I didn’t make a grammatical error in a sentence, and often it’s because I did. I’m not talking about sentences that merely deviate from stylebook norms (YOUR PERIODS FOLLOW THE QUOTES YOU IDIOT!), but undeniably ungrammatical utterances like These is a big problem or worse. If it were really anything goes, you’d see linguists rushing to the defense of these ill-formed sentences even as I said “no, no, they’re not right!”

So let me try to state the maximally descriptivist position that I think a reasonable person could take. It’s that the set of grammatical utterances of a language is the set of utterances that can be made by speakers who have sufficient linguistic ability (i.e., adults who are fluent in the language) such that the speaker making that utterance does not find a problem with it after careful examination. More briefly, it’s the set of sentences that a qualified speaker would accept. But this is hardly “anything goes” — it’s more like “anything meeting certain standards goes”, and that’s a major philosophical shift. In fact, the difference between this theoretical “certain standards” descriptivist and a moderate prescriptivist is little more than a difference of what the standards are.*

And if we’re treating this as the descriptivist baseline, I have to confess that I am a bit less accepting than that. For me, the set of grammatical utterances is community-based; a sentence is grammatical in a linguistic community if and only if it is considered acceptable by a substantial portion of the linguistic community. Note that this is equivalent to the position I sketched above when the “community” is the individual; the difference is that my position does not extend the individual’s grammaticality judgments any further unless the rest of the community agrees.**

Now, the descriptivist philosophy I’ve outlined doesn’t rule out an additional prescriptive preference in stylistic matters, nor does it say that one can’t have a preference between two grammatical sentences. It is only defining the set of grammatical sentences. Most every descriptivist I know has these sorts of stylistic preferences. I, for instance, don’t like hyperbolic usages like figurative literally. Do I think they’re ungrammatical? No, not usually. But would I advise people to avoid them? Yes. And would people be right to ignore my advice? Sure, if they didn’t care what I think (and why should they?).

Lastly — and this is a point that Jonathon at Arrant Pedantry has made better in two of his posts, but it’s important enough to repeat — this means that descriptivism and prescriptivism aren’t necessarily at odds. You can be a descriptivist who acknowledges that something is an acceptable usage even as you avoid it yourself. And in fact, I know many self-described prescriptivist editors who hold this (I think eminently reasonable) position.

What’re your thoughts on the matter? If you’re a descriptivist, do you hold one of the philosophies I’ve sketched above, or something else? If you’re a prescriptivist, do you feel that your philosophy meshes with this sort of descriptivism, or do descriptivists still seem like whateverist hippies dancing in the ruins of English?

*: That little more, though, contains the philosophical difference I tweeted the other day: if usage and rules conflict, the descriptivist will base grammaticality on usage, the prescriptivist on rules.

**: To clarify, I think the maximal-descriptivist position is valid for describing idiolects, one’s personal form of the language. But for the purposes of delineating a dialect or language, it doesn’t matter if one person thinks a certain usage is good if all the rest of the world disagrees.