Prescriptive and descriptive language, and what it means...

2010-09-21 20:39

There’s two basic schools of thought about grammar and language usage, although very few people adhere rigorously to either.

Prescriptive grammarians attempt to define rules for correct use of language, and view a rule as defining usage which is correct; failure to follow the rules is incorrect. Descriptive grammarians attempt to describe language correctly, and view a rule as an attempt to describe usage; if someone uses language in a way inconsistent with the rule, that means the rule is wrong.

The same thing applies to dictionaries. Dictionaries tend to list all the meanings a word has. However, not every usage listed is one you should expect other people to understand or accept. The purpose of the dictionary is to help you guess what someone meant when they said something, by telling you what they might have meant. It is a category error to then treat this as evidence that they were “correct”. Insofar as a dictionary is merely reporting on existing usage, it isn’t offering an opinion as to whether that usage is “correct”.

Here’s the thing. Both views are useful. They may be helpfully understood in terms of Postel’s Law:

Be conservative in what you send; be liberal in what you accept.

In general, if you are speaking or writing, you will be much more effective in communicating if you stick to primary, well-established, meanings of words. If a word has conflicting or contradictory meanings, you may be best off avoiding it entirely, or at the very least indicating clearly the sense in which you mean it. It also helps to know your audience; you may find it useful to adopt local usages when communicating with specific people, even if those usages would make your message less clear to other audiences.

By contrast, if you are reading or listening, you will usually be more effective if you allow for “nonstandard” or uncommon usages, and keep an eye out for them. It helps a lot to know the usages of the people who are communicating to you — you want to understand what they mean, not what someone else would mean.

When people get up in arms about a usage being “incorrect”, it may be because a secondary meaning or popular misunderstanding imposes the risk of losing any communicative value from a word. Consider the now moderately commonplace of “literally” as an intensifier for hyperbole. That is to say, consider the use of the word “literally” to mean “figuratively”. Once you have that usage in mind, how exactly do you communicate that something is “not figurative”? Well, you don’t. There’s no longer a word for “not figurative” in that dialect of English, because the word which used to mean that now means “either not figurative or figurative”, which tells you very little.

Ultimately, it’s up to you how you want to communicate, or whether you care. However, it’s worth remembering that a general description of how people appear to talk is not an assertion of correctness, and a set of guidelines for how to write is not an assertion that no one will write in any other way.

Peter Seebach




  1. You said, “Descriptive grammarians attempt to describe language correctly, and view a rule as an attempt to describe usage; if someone uses language in a way inconsistent with the rule, that means the rule is wrong.”

    “Accurately” would be better than “correctly,” because it does not presuppose that there is necessarily a correct description.

    More seriously, the paraphrase is misquantified; there’s always someone who uses language in a way inconsistent with whatever rule is under discussion. Rather, say “If a descriptive rule is inconsistent with the predominant usage, then it is the rule, not the usage, that is wrong.”

    Linda Seebach · 2010-09-22 11:16 · #

  2. Actually, strict adherence to current grammatical rules often gets in the way of comprehensibility. For example, the singular ‘they’ is considered incorrect (though it was acceptable before Victorian times) but ‘he’ suggests a particular gender, and ‘he or she’ gets really unwieldy.

    Another one that comes up a lot is calling a piece of a fruit a fruit. It’s just too unwieldy to always be talking of apple slices when the kids know they’re being offered a slice.

    My rule of thumb is that you should follow grammatical rules unless you have a reason not to. That is, you should know the rule you’re breaking.

    Dave Leppik · 2010-09-23 12:54 · #

  3. As a side note, I find singular-they more confusing than using either of the pronouns which may or may not have been intended to imply gender. I don’t care if I get peoples’ gender wrong, I do care if I get the number of people wrong.

    I think the “they” thing may be a good example of a case where there’s a compelling reason for language to change, and the amount of data we lose by it is relatively small — it’s fairly easy to tag something for number of people if you need to. (I sorta wonder why we didn’t end up with “they is” for gender-unknown singular and “they are” for plural, though…)

    My blog has become a no-kill shelter for abandoned pet peeves.

    Peter Seebach · 2010-09-23 15:37 · #

  4. I don’t usually complain about what prescriptivists would label “incorrect language”; I do lament the loss of precision in the vernacular that is so well illustrated by the word “literally”. Whenever I encounter such a wounded word, I find myself resorting to awkward combinations like “quite literally”.

    Caelan · 2010-10-01 14:44 · #