Sunday, July 31, 2011

Language: jump in!

One of my favorite ways to think about language is to compare it to jumping. This may seem strange, but I think it makes a lot of sense. After all, jumping is something that nearly all mature humans can do, provided we have the requisite body parts. And probably, no one taught you how to jump. They may have offered advice or demonstrated their own forms, but you basically learned how to push yourself into the air by observing other people do so and trying it out for yourself. From what we know about child language acquisition, this is roughly analogous to how babies learn to speak.

It's also true of language, as of jumping, that you don't need to mentally review every step of the process before you perform it. We don't typically run through a mental catalog of the muscles we'll need to flex, or the order in which we'll flex them, before launching ourselves off the ground. And we don't need to think about the rules of grammar before opening our mouths to speak. They come naturally, like jumping, and what results is a grammatical utterance just as surely as what happens when we jump is a jump.

You can, of course, think about jumping before you begin to move, and plan out exactly how you're going to do it. You can choose to move in a precise way, more planned out and self-aware than a normal jumping maneuver would require. Language is like that too. If you think about your words carefully before speaking, you can produce a more polished utterance just as surely as you can produce a more precise jump by thinking about the matter first. But an unreflective jump is just as authentic as one with thought before it -- by which I mean, no one would accuse jumping-without-planning of not being a true jump. I don't think anyone would even make the claim that planning out your jumps makes them objectively better than jumping without thinking first.

It is my contention that language is the same way. When we speak without thinking about it, our utterances do not always adhere to the standard notions of grammatical correctness. We speak in run-on sentences and dangling participles, in "misplaced" modifiers and all of the other things that would send an English teacher scurrying for the red pen. If we think carefully beforehand -- as we usually do when we write -- we can generally avoid these items, but they are very present in our unguarded speech. But, like a jump without thought behind it, I see no reason to consider such speech improper, incorrect, or in any way inferior to its more planned-out alternative.

The notions of grammar and grammaticality are still important to a descriptivist, of course, but here too a jumping metaphor can prove useful. For a linguistic descriptivist, investigating which utterances are grammatical and which are not is as pointless and frivolous as considering which jumps qualify as legitimate acts of jumping. We instead take as granted that anyone trying to jump produces a jump, and anyone trying to speak produces a grammatical utterance. The notion of grammar is used to analyze what all of the utterances have in common, just as a general description of jumping could be produced by observing many people's individual acts of it. But descriptivists do not believe that everything is grammatical, just as surely as no one would conclude that any possible motion is a jump. There are some things that simply no one would do when asked to jump, and there are just as many that no one would do when asked to speak.

And this is perhaps the best test of what is a legitimate instance of a particular kind of action. When someone is asked to jump, and attempts to jump, and considers what they have done to be a jump, we should surely accept that they have in fact jumped. And when someone is prompted to speak, and opens their mouth to speak, and considers what they have said to be a meaningful contribution to the conversation (and not just nonsense noises), we should just as surely accept that utterance as grammatical, whether they planned it all out in their head or not.

In the end, language is just like jumping. When we attempt to speak, we speak -- and what comes out is an utterly genuine act.


  1. It's worth noting Dryer's 4-way distinction between "acceptable" and "grammatical" such that any given utterance can be one of the following:
    1. acceptable & grammatical
    2. acceptable but ungrammatical
    3. unacceptable but grammatical
    4. unacceptable & ungrammatical

  2. I really like this post. I've always thought that language classes focus too much on studying vocabulary from a textbook and not enough on actually speaking. Definitely better to "jump in."

  3. Great post, but I'm not sure about your next-to-last paragraph. You write:

    "When someone is asked to jump, and attempts to jump, and considers what they have done to be a jump, we should surely accept that they have in fact jumped."

    But what if what they did is in fact flexing their left leg while their right leg stayed put and their right foot never left the floor? What if after they did that they swear they have actually jumped, despite the video clearly showing one foot never left the floor? Should we agree with them and consider what they did to be a jump, or should we disagree and insist that for a movement to be a jump both feet must at some point not be touching the floor at the same time?

    My point here is simple: does a descriptivist consider *all* (as in 100%) utterances of a subject (that the subject insists are "OK" to him) to be actually grammatical, or is there also some kind of external criterion (like the jumping definition of having both feet off the ground at the same time) that one can also test the utterances against? Because if only the first alternative is considered, then I don't know how to distinguish descriptivism from language laisser-faire.

    I hope I'm not antagonising you while saying this. I am a descriptivist myself. I know language is dynamic and always in flux (although there's always a tiny prescriptivist in me that laments how people should pay more attention to what they say and how they say it :P), but I'm uneasy in that using this stance I can't find a satisfying definition of grammaticality and ungrammaticality (and "nobody says this" isn't a good definition because we can't possibly ask everyone, except in communities speaking very endangered languages).

    I guess as an engineer I have difficulties with fuzzy definitions.

  4. That's an interesting thought, Christophe. I think I'd like to hear a little bit more about your example before I respond in full. Are you suggesting that the person in question is mistaken about jumping? (That is, if they reviewed the videotape, would they concede that they hadn't jumped after all? And if they saw someone else performing the exact same action, would they say that it wasn't a jump?) Or is this a person whose very definition of jumping is such that flexing a leg counts as a jump? I think you're suggesting the latter, but I want to make sure before I respond.

    While you're answering that question, perhaps you could also flesh out the idea of "language laisser-faire" a little bit. I have a suspicion that my answer to your question is going to hinge on that issue.

  5. Hi Joe,

    Yes, I do mean that the person counts flexing a single leg as jumping.

    As for "language laisser-faire", it's just my way of describing the "anything goes" attitude prescriptivists usually accuse descriptivists of taking when talking about language.

    My issue is basically: what counts as ungrammatical for a descriptivist? To go back to the jumping metaphor: if we are to accept that jumping can be defined as flexing one leg just because one person swears they consider that to be jumping, at which kind of movement do we stop accepting that it's a jump? As I wrote, we can't possibly ask everyone to jump in front of us. How can we know there's no one out there that considers tiptoeing to be jumping, for instance?

    If that sounds a bit like the "slippery slope" prescriptivists sometimes talk about, it's because it's similar. But here I am not discussing about the "quality" of someone's language, but only about the definition of ungrammaticality according to a descriptivist. You write that grammaticality is still important to a descriptivist, but I just cannot find a definition of grammaticality in a descriptivist framework that doesn't devolve into nothingness when I really analyse it fully.

    I hope I've managed to make myself clear. It's a complex issue and I'm not quite sure I'm giving it a good description.

  6. Okay, thanks for clarifying. In response, I would say that I think it's important to look at all data that crosses our path, and not discount something just because it doesn't fit our preconceptions of what we're studying. There are many communication systems such as creoles or sign languages that were dismissed and ignored for years, because they didn't match someone's existing idea of what a language should be. Today those systems are recognized as full-fledged languages, and studying them has greatly increased our understanding of everything that language can be. So I would say it is just good science to keep those prescriptivist blinders off. Preconceptions can and should guide our research, but we should always be open to the idea that our earlier conclusions can be revised. It may be, in the light of new evidence, that we have to revise our definition of jumping!

    As scientists, we also have to accept our inability to test everything, everywhere. No one, in any field, can examine every possible instance of their research topic, so we shouldn't expect students of language to do so either. We base our conclusions on everything we have examined, so that our work always comes with the caveat, "...or so we've seen thus far." I would say that everything in science, from gravity to the gas laws, fits this model. When we have a rich enough evidence bank, we can be reasonably confident in the universality of our conclusions -- but, as I said before, it's important to not take those conclusions as hard facts when potential counter-evidence comes along.

    I'm still not sure I entirely understand the "anything goes" attitude toward language that you've mentioned, but I'll try to answer your question about what grammaticality / ungrammaticality is to a descriptivist. The short answer is simply that everything a person says is grammatical. That's just how language works -- it turns your thoughts into grammatical speech. If you think of language as an input-output machine, the output is always grammatical. (This ignores actual slip-of-the-tongue speech errors, which are a breakdown of the machinery that the speaker generally recognizes and corrects.)

    The reason this may seem like an "anything goes" approach to language is that people do not always share an identical grammar, even when they supposedly speak the same language; what is grammatical for speaker A may be ungrammatical for speaker B, and vice versa. It's thus important to not conflate an individual's grammar with the grammar of a language, such as "English grammar." This latter concept is always an aggregate, a stylized average of what's grammatical for most speakers. If you'll forgive the analogy, this is similar to how two members of the same religion might not believe the exact same things as one another. As a result, any statement about the beliefs of the religion is going to be a generalization that doesn't hold true for all of its members. But if A and B's beliefs are different from one another, that doesn't mean that the notion of belief is meaningless from a descriptivist point of view. We just have to generalize anytime we're describing the patterns of a larger group.

    The reason I object so much to prescriptivism is that it takes the aggregate, the generalized version of a language's grammar, as a set standard and attempts to enforce it on a very heterogenous population of individual speakers. This disenfranchises the speakers whose grammar is far from the standard, and it stifles the very natural process of healthy language change.