Pretty much every kind of human (and, for that matter, animal) learning shows frequency effects: the more we hear or see something, the better we learn it, remember it, and even like it. But in the domain of children’s language acquisition, both the existence and meaningfulness of frequency effects have proved controversial, particularly because they have implications for the (in)famous nature-nurture debate. In this target article, we argue that frequency effects can be found absolutely everywhere in language acquisition, from the level of abstract strings to the level of abstract syntactic cues. In fact, high frequency items are not only early-acquired and resistant to errors (when children are attempting to produce them), but also cause errors, when children use them in place of lower-frequency targets.
What does all this mean in terms of theory? Well, we argue that while frequency effects are often taken as evidence for constructivist/usage-based accounts, they are not necessarily incompatible with nativist/UG accounts in principle. However, because these accounts draw a sharp distinction between the lexicon and grammar, for instance assuming that even infants’ grammatical rules are formulated in terms of syntactic categories and phrases rather than individual lexical items, they do not straightforwardly explain frequency effects that cut across these levels of representation.
There are commentaries too; nine of them, in fact. While most of them are generally supportive, many point out that the real trick is going to be disentangling the effects of frequency from those of other factors (e.g., serial position, communicative intent) with which frequency frequently interacts. In our response, we acknowledge that this disentangling work has only just begun, but conclude that – nevertheless – frequency effects are real, and are therefore something that any serious theory of language acquisition – of whatever theoretical stripe – must explain.