Not everything is grammatical

This is a follow-up on my last post about grammatical rules. There I argued that we should figure out the grammar of a language by observing how native speakers use that language. One of the replies to that post argued back that this approach seems to dispense with grammar all together. If grammar can be inferred from usage, then whatever anyone says is correct. In other words, anything goes and grammar becomes a free-for-all.

I can see how someone might draw that conclusion, and in fact this is probably the most common response I get to this idea. But adopting a usage-based approach doesn’t mean that a sentence is correct just because someone said it. We’re free to say any kind of garbage we want, like

the tea up before again Mary because bites

But just uttering words doesn’t make them grammatical. I think we all agree there’s something wrong with this sentence. The mere fact that we can come to this agreement suggests that grammar does exist. If there was no grammar, if language was truly a free-for-all, then that sentence ought to be totally unremarkable.

This is the important point: even though it’s true that people could say whatever they want, they don’t. There are sentences that people don’t, or wouldn’t, say. People are really good at keeping their auxiliaries in the right order, and no one would ever say *I been may have working. Relative clause markers consistently go at the beginning of the clause, and I never hear sentences like *that’s the house we bought which. No one violates island extraction constraints to ask questions like *What does Fred like the woman who wears?.

And it’s the absence of sentences such as *I been may have working that allows us to know there’s a grammar at work in the first place. If absolutely every arrangement of English words was possible, then the descriptive enterprise would be sunk, or at least we’d have very different goals. But that’s not how English, or any language, works. There are rules for which strings of words are allowed (among other things).

Let me put forth the following definition for a rule: a rule is a claim about a language that could be verified through observation of native speakers. It’s a hypothesis about grammar, if you like. I want it to be crystal clear that my definition of rule is not the same a rule that you “have to follow”. There’s no obligation to do anything.

When I say a sentence is “grammatical”, I mean that it’s a sentence which follows all of the rules I’ve claimed about English. When I say a sentence is “ungrammatical”, I mean that it has some properties that are not predicted, or which contradict, my rules. So if I have a rule “articles come before nouns”, then the sentence *I ate soup the is ungrammatical, because it doesn’t fit with my rule. Again, this isn’t a judgement about the sentence or the speaker. It’s a descriptive statement.

(I say “my rules” because this is my blog and this is just an explanation. In real life I’m not so pompous and when I don’t understand something then I check a descriptive source, this is a good one, or a linguistics paper, or I find a syntactician.)

Let’s walk through how a rule could be figured out in the first place. Consider the words while and during. How are they different? And what kind of rule could we come up with that describes this difference? Here’s some data to think about:

(1) I read the newspaper while I ate lunch.
(2) *I read the newspaper while lunch.

(3) I read the newspaper during lunch.
(4) *I read the newspaper during I ate lunch.

Sentences (1) and (3) are grammatical. That is, I claim that speakers of English would accept and use such a sentence in conversation. They sound like “normal” English.

Sentences (2) and (4) have a * symbol, which means “ungrammatical” (linguists have lots of these kinds of symbols). I’m claiming these are not sentences that any English speaker would sponatenously utter in a normal conversation. If you are a native speaker, you probably have a “gut feeling” that something has gone wrong with them. (And if you have some reason to challenge these claims, maybe you find these sentences perfectly acceptable, please post a comment.)

The ungrammaticality of sentences like (2) and (4) shows that while and during cannot simply be interchanged. There is some pattern to their distribution; a rule if you will. If you look at just the data above, the pattern is that during is followed by nouns and while is followed by sentences.

And that’s a rudimentary “rule” of English that describes the differences between those words. As you include more data, you made need to change your rule, and in this case we would definitely have to change our rule. It happens that while can atually be followed by other things too, like prepositional phrases (“while on vacation”) and verb phrases (“while walking my dog”). But I don’t want to turn this post into a syntax homework problem.

Now obviously you can say things like (2) and (4). There’s no physical, legal, or moral restriction on it. You can say anything you like. But it’s not normal behaviour, expect a stare if you use them in conversation, and if you search a corpus of English, you won’t come up with many or any results (and you’d need to check the context to see if the speaker corrected herself afterwards). So if someone once uttered (2), it wouldn’t mean we’d have to include it in any description of English.



Filed under Linguistics

2 responses to “Not everything is grammatical

  1. Thanks for clarifying. I read the “Not all errors…” and most of the issues don’t bother me, as the sentences are still clear. I don’t believe in prescriptive quibbling, but I do believe there is wrong language that – like a wrong turning – doesn’t get you where you want to go.

    An example that comes to mind – semantic, but word meaning evolves just as grammar does, so I’d say the same principles apply – is “unsaturated” fatty acid. Probably 99% of native speakers haven’t a clue what unsaturated means. Perhaps “healthy” or “good for you”, or “still hungry” or “dissatisfied”. But it has a precise meaning – containing one or more double bonds (actually because the molecule is still able to absorb H+). If we drop the rule that it means exactly that, just because most people don’t know it, we quickly lose the ability to express anything we can’t say with an emoticon or internet abbreviation.

    Of course, you could make this a “dialect” – precise English, or whatever. I don’t object to people using it to mean “hungry” at the dinner table. But in a food store or hospital we need to be able to express ourselves precisely and rely on being correctly understood. So changes need to be slow enough for the meaning at any given time to be clear.


  2. I appreciate your thoughtful comments, Delft. And I like linguishtick’s post explaining how we linguists think about what is “grammatical.”

    I also understand the importance of precise meaning in certain situations. To me, what you describe is “communication,” which I’ve defined elsewhere like so:

    Communication means shared meaning. For example, you create shared meaning when you use words (or nonverbal cues like eye contact) your audience understands. Saying (or writing) “cat” to another speaker of English means I’ve created some shared meaning—both of us share some meaning for “cat.” However, the more my audience knows about me and my experiences with cats, the more meaning we will have in common when I use that word—i.e., the more I communicate. In other words, there is more shared meaning between my husband and me when I say “cat” than there is between you and me.

    To create precise shared meaning requires much more than an agreed upon definition of a word.


Tell the world what you think!

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s