The Determiner Wars




A and The: almost certainly the least interesting words in the English language; and quite possibly the least useful as well. In fact, many languages manage perfectly well without them (which is why – for example – Russian learners of English often struggle with them). Nonetheless, these two little words – which linguists call “determiners” – have been at the centre of one of the hardest battles ever fought in the cognitive sciences.

Nobody alive has more citations than Noam Chomsky of MIT. Why? In the late 1950s, Chomsky came up with the idea of “Universal Grammar”. Chomsky argued that the rules of language are too abstract and complex for children to figure out by themselves in just a few short years. Instead, children must be born with some general knowledge of how languages work (i.e. a Universal Grammar). The alternative position, espoused by psychologists such as Jerome Bruner, is that – with the aid of parents and other adults – children are indeed able to figure out the rules of language all by themselves.

For about 30 years, this debate trundled slowly along, proceeding mainly on the basis of armchair theorizing, or the occasional diary that some enterprising academic had kept of her own little darling’s linguistic development. But in the early 1990s, the field was driven forward by some exciting technological developments. The tumbling price of portable tape records and personal computers – linked together by the worldwide web (as everybody called it back then) – made it possible not only to record and transcribe children’s conversations with their parents, but also to share them with fellow academics and – crucially – to search them electronically. Now, at last, it was possible to translate the Chomskyan and Brunerian positions into testable hypotheses.

Enter determiners. According to the Chomskyian position, one piece of knowledge that children are born with is a rule that takes any determiner (e.g. a, the) and (pretty much) any noun (drink, baby, cookie, cat) and combines them together (a drink, the drink, a baby, the baby etc…). If this position is correct, then any noun that a child can stick “a” in front of, she should also be able to stick “the” in front of, and vice versa. The Brunerian position is that children don’t start out with this rule. Instead, they first learn lots of individual determiner+noun combinations (a+drink, the+baby, a+cookie, the+cat) and only later – on the basis of these stored combinations – figure out the general rule.

Thanks to the new technological developments of the early 90s, extracting all the sentences that contained a or the plus a noun from children’s speech was – well – child’s play. And when a team of researchers from the UK cities of Nottingham and Manchester did just that, they found that children did not show evidence of having a Chomksyan general {the/a}+NOUN rule. Rather, each child had a set of nouns that she used only with a (a+drink; a+cookie), and different set of nouns that she used only with the (e.g., the+baby, the+cat). In fact, just one in ten nouns occurred with both the and a (or an, which is essentially just a different way of pronouncing a). Score 1-0 to Bruner.

But the Chomskyans soon hit back, pointing out problems with the analysis. First, Virginia Valian (from the City University of New York) pointed out that the UK team failed to take into account the fact that many nouns were produced only once in a child’s entire recording, and so did not have the opportunity to occur with both the and a. Second, Charles Yang (from the University of Pennsylvania) pointed out that even adults – who everyone agrees have a rule that can add both the and a to any noun – tend, in practice, to use mostly one or the other with any given noun. This isn’t because adults can’t use both the and a with the noun – of course they can – but just because they rarely need to, given the kinds of things that they usually talk about. For example, as Yang notes, “the bathroom is more frequent than a bathroom only because bodily functions are a more constant theme of life than real estate matters”. Children too, reasoned Yang, are perfectly capable of using both the and a with any given noun, it’s just that there are very few nouns for which they ever need to.

Both Valian and Yang produced analyses showing that, when you correct for these problems with the original study, children use about as many different nouns with both the and a as adults do. And since everyone agrees that adults have the general {the/a}+NOUN rule, this suggests that children do too. Score 1-1.

Then, in 2013, Julian Pine – head of the UK team – struck back with a counter-objection: Adults use a lot of very low-frequency nouns that young children never use (e.g., politician, challenge, election). This means that many of adults’ nouns turn up only once in most recordings of adult speech, and so never get the chance to appear with both the and a. Children, on the other hand, tend to use the same handful of nouns (e.g., drink, baby, cookie, cat) over and over again, meaning that these nouns have ample opportunity to appear with both the and a. This means that if children really did have a general {the/a}+NOUN rule, they should use a greater proportion of their nouns with both the and a than adults, simply because they have more opportunity to do so. The way to correct for this problem is to include in the analysis only those nouns that are used by both a given child and his parent. Pine showed that, when this is done, children indeed use fewer nouns with both the and a than do their parents. Score 2-1 to the Brunerians.

Why all this fuss over determiners? Well, determiners are just a battlefield in the wider conflict between the Chomskyan and Brunerian approaches. Pine’s findings support the Brunerian claim that children figure out the rules of language gradually on the basis of what they hear: “Right, since I can say both a+cat and the+cat, and both a+cookie and the+cookie, maybe – just maybe – any noun that goes with the also goes with a, and vice versa!”. The Chomskyan claim is that children didn’t figure out this rule, but were born with it (or, at least, with a rule that combines determiners and nouns; they had to learn the actual words, of course).

That said, it would be premature for either side to declare victory in the determiner wars. While Pine’s analysis might seem watertight now, you can bet that both Valian and Yang will be thoroughly testing it for any cracks. And while it might seem ludicrous to wage war over two of the shortest and most insignificant words in the English language, something undeniably huge is at stake: An answer to the question of how we acquire language: not only our most cherished skill, but the one that is most quintessentially human.

References

Pine, J., Freudenthal, D., Krajewski, G., & Gobet, F. (2013). Do young children have adult-like syntactic categories? Zipf’s law and the case of the determiner Cognition, 127 (3), 345-360 DOI: 10.1016/j.cognition.2013.02.006

Pine JM, & Martindale H (1996). Syntactic categories in the speech of young children: the case of the determiner. Journal of child language, 23 (2), 369-95 PMID: 8936692

Valian V, Solt S, & Stewart J (2009). Abstract categories or limited-scope formulae? The case of children’s determiners. Journal of child language, 36 (4), 743-78 PMID: 19123961

Yang, C. (2013). Who’s afraid of George Kingsley Zipf? Significance: The Magazine of the Royal Statistical Society and the American Statistical Society. Dec. 29-34.

Image via Michelle D. Milliman / Shutterstock.

Ben Ambridge, PhD

Ben Ambridge, PhD, is Reader in Psychology at the University of Liverpool, and author of Psy-Q, which introduces the reader to theories and findings from the psychology literature via a range of interactive games, puzzles, quizzes and tests.
See All Posts By The Author