Correcting your child’s grammar

Back in 1971, a guy named M. D. S. Braine published [“On two types of models of the internalization of grammars” in D. J. Slobin (ed) The ontogenesis of grammar, pp. 153-186, Academic Press] a rather interesting observation:

… I have occasionally made an extensive effort to change the syntax of my two children through correction. … the interchanges went somewhat as follows: “Want other one spoon, Daddy?” — “You mean you want THE OTHER SPOON” — “Yes, I want the other one spoon please daddy.” — “Can you say ‘the other spoon’?” — “Other … one … spoon” …

Doubtless (as he mentions) that parents had noticed this centuries ago, but Braine was trying to use the fact to learn something about language.   It’s an interesting fact and even a surprising fact: you’d never suspect it before you have kids and it can easily get forgotten later because older kids do respond fairly well to that kind of correction.

Years later, in a frequently referenced (but sloppily argued) paper by G. F. Marcus [Cognition 46 (1993) 53-85], Braine’s observation is treated this way:

Of course, even if parents provide feedback, it might turn out empirically that children do not use it.   For example, Zwicky (1970) … points out that “six months of frequent corrections by her parents had no noticeable effect”.  Braine (1971) even more clearly illustrates that feedback may be ineffective:

[Above quote from Braine, 1971]  … “Want other one spoon, Daddy?”…

These examples suggest that it is the child’s underlying linguistic system, rather than negative evidence, which forces children to change their grammars. …

Now, does it really?   Let’s look at that last statement carefully.   Where did the phrase “underlying linguistic system” come from?  There is nothing in Braine’s daughter’s speech (i.e. the actual data) that makes it so.  That phrase comes from a theory that Marcus buys into.   He says this:

Suppose that a child must learn that a certain sentence, B, is ungrammatical.  The child might learn this in two ways.  First, something external to the child … could tell the child that B is ungrammatical.  If the parent does not provide an explicit denial of sentence B, the only alternative is that the child has a mechanism…  That mechanism, innate or learned, must be internal to the child…

There are a pair of logical leaps here that are really pretty dubious when you think about them carefully:

  1. That whatever teaches the child is either  internal or external, but not a mixture of both.   Think about hunger as an analogy:  what makes you come to dinner?   It can be a mixture of hunger (an internal cause) and the smell of food (an external cause).
  2. An “explicit denial” from a parent is the only external thing that matters.  What about the possibility that when you ask for milk in an ungrammatical sentence, the milk takes several seconds longer to arrive?  What about an implicit denial?

Now, back up and look again:

These examples [children not responding to adult corrections] suggest that it is the child’s underlying linguistic system, rather than negative evidence, which forces children to change their grammars. …

There’s something odd about that paragraph.   It talks about what forces children to change their grammars.  But, as far as we can see from the evidence presented, the children are not changing their grammars.   Of course, children will eventually stop saying things like “other one spoon”, but when?  The important moment is the one when they finally get it right, not months before.    But, from what we read, Braine is talking about the long periods of frustration rather than the moment of accomplishment.

Suppose we liken learning grammar to mathematics.  Give children (or adults) math problems that are too hard.  For instance, give calculus problems to people who really only understand arithmetic.   They will do it wrong, and you can correct them as much as you want but the corrections will not help.  It’ll be very much like a 2-year old learning grammar (though not as cute).  But, if they are also learning algebra and trigonometry and such, the day will eventually come when calculus “clicks” and they figure out how to do it.  That moment and the few days before are the time to study.

It is certainly possible that one of your corrections (number 317, perhaps) hits fertile ground at a time when they have learned the foundations underneath calculus, when they almost understand what they are doing.  It is possible that your correction causes an “Aha!” moment and makes them finally understand it.  Or, it is also possible that the “Aha!” moment comes because of the memory of your corrections.  So, it could be a combination of an internal maturation along with external corrections.

But, the point is, that even if one of your corrections was crucial to a child/student learning grammar/calculus, most of your corrections will seem to be failures.  All but one, in fact.  To people paying attention to the failures of correction (like Braine and Marcus), it can look like all corrections are useless.

I’m not trying to make the argument that this is the way it happens.  Rather, I am claiming that Braine and Marcus were looking at the wrong moment in the child’s development so that Braine’s interesting observation may not actually mean what they think it does.  They have failed to imagine several possibilities and therefore the argument is not nearly air tight.  Logic is a tough customer: if you get your arguments nailed down tightly, you can trust your life to the result, but if you leave some loose ends in the argument, then the conclusion means nothing.  So, do these examples “suggest” anything?  No.  It’s really Marcus making the suggestion.

So, here I am, correcting a linguist’s logic.  I wonder if my correction will have an effect?