Posted by Ben Ambridge
“Do fish have barcodes?” My 3-year-old had just paid his first visit to Pets-At-Home, and clearly had some questions. If you’ve spent any time around children – whether your own or other people’s – you’ll be all too aware of their knack of coming up with the most awkward questions (“Where do babies come from?”, “Why have I got a mouth?”) at the most awkward times (such as when you’ve been trying to leave the house for…the…last…hour).
But even when children’s questions are relatively sensible, they don’t always come out quite right. A particularly common type of mistake that pretty much all English-speaking children seem to make with questions is getting the words in the wrong order. Picture the scene: Your little artist, pen in hand, turning to you for inspiration, asks “*What I can draw?” (the asterisk, as is standard in linguistics, flags this sentence up as a “grammatical error”). They meant, of course, to ask “What can I draw?”; the words just came out in the wrong order.
For child language researchers, these little mistakes are a big deal.
There are two main overarching theories of how children learn language, and the two sides have been engaged in constant battle since (at least!) the 1950s. To illustrate the two theories, let’s take another example of this type of error, which my daughter came out with when she was just two years old: “*What Mummy is doing?”. Aww, how sweet. What she meant to ask, of course, is “What is Mummy doing?”, but the middle two words were back to front, giving “*What Mummy is doing?”. The question is why just about all English-speaking children make similar mistakes. And as with most things child-language related, the two rival theories offer very different explanations.
You’ve probably heard of Noam Chomsky – almost certainly the most famous academic alive; definitely the most highly cited. Well, the first view of how children learn language is very strongly associated with Chomsky; so much so that – at the risk of sidelining the many other researchers involved – I’ll call this theory the Chomskyan view. Glossing over the details, what this theory claims is that children are born with a rule that takes a statement and turns it into a question by (Step 1) moving the question word to the front, then (Step 2) moving the ‘auxiliary’ (here, the word ‘is’):
Statement Step 1 Step 2
Mummy is doing what -> What Mummy is doing? -> What is Mummy doing?
According to the Chomksyan account, children make mistakes like “*What Mummy is doing?” because, for reasons we needn’t go into here, they get stuck after Step 1. But, so the theory goes, the fact that children always make this particular mistake (rather than, for example, asking “*Doing is what Mummy?”) shows that they have the right rules for forming questions (in fact, that they were born with the right rules); they just get stuck sometimes.
That’s the Chomskyan view. The opposing view, which I will unimaginatively call the anti-Chomskyan view, says that language isn’t about being born with rules, but is instead about learning from what you hear. According to the anti-Chomskyan account, children make mistakes like “*What Mummy is doing?” because – again glossing over the details – they hear the combination Mummy + is + doing a lot (“Mummy is doing the delivery”; “Mummy is doing it, sweetie!”; “You should really clear up your own toys, but Mummy is doing it”), and think that you can make a question just by slapping a question word on the start:
What + Mummy is doing ->*What Mummy is doing?
It sounds arcane, I know, but over the years, many experiments have been run to try to figure out which of the two explanations is right, or at least closest to the truth. I know because I’ve done many of them myself, the first all the way back in 2006; the second, a follow-up a few years later; and, finally – thanks to LuCiD – a naturalistic corpus-data study that more or less seals the deal.
The details are technical, but, in short, our experiments support the anti-Chomskyan view (at least, that’s what I argue) because the more common the particular word combination in the language that children hear around them (e.g. Mummy + is + doing), the more common the mistake (e.g. “*What Mummy is doing”) in children’s own speech. This suggests, to my mind, that children make the mistake by learning from the language around them, rather than by getting stuck when applying an in-built rule. That is, I think the anti-Chomskyan view is the correct one.
The results of these studies have, I think, quite a profound implication. If children ask things like “*What Mummy is doing?” because Mummy+is+doing is a common “chunk” in the language they hear, this suggests that – on some level – children must be storing every incoming sentence that they hear. This feels quite counterintuitive – wouldn’t their little heads get full up? – but it really is the only way to explain the finding that children’s language mistakes are so sensitive to the particular sequences of words that they hear. How else would they “know” that Mummy+is+doing is a particularly frequent chunk in the first place (unless you want to assume, even more counterintuitively, that they are constantly updating some kind of internal tally chart)? On the more practical side, we hope that our findings might help researchers come up with better treatments for developmental language disorder (DLD). Children with DLD struggle with many aspects of language but seem to find questions particularly difficult.
Back at Pets-At-Home, it turns out that fish do have barcodes after all; they’re just stuck to the side of the tank rather than the fish themselves (otherwise, the cashier would have nothing to scan). All of this was explained at great length in response to another on-the-face-of-it-baffling-but-actually-quite-sensible question, “How can we buy fish?”.
So next time your little darling comes out with an awkward question, pay close attention – you might learn something that they’ve picked up about your own speech patterns. One thing is clear (and, frankly, quite terrifying): They’re recording everything that you say.