Something old, something new.

MacWhinney’s list of proposed mechanisms for language acquisition seems so all-encompassing that it might appear churlish to dwell on a mechanism that’s left out, but on his ultimate list of critical mechanisms there is an indeed an important (and intended) omission: innate, linguisticallyspecialized constraints. Such constraints are indeed mentioned, but if I understand the point of the article correctly, it is to argue that seven cognitive mechanisms, ranging from competition to cue construction, indirect negative evidence and monitoring, collectively obviate the need for any innate machinery that it specialized for language. Why not include room for innate, domain-specific constraints, too? I have little doubt that many of MacWhinney’s Big 7 play a real role in language acquisition. Although there was once a long tradition of treating the mechanisms for language as largely or even wholly separate from other cognitive faculties, there is growing recognition that the language acquisition faculty likely owes something to the rest of cognition. Language, for example, depends on memory, and in many ways the mechanisms of memory that underlie language seem to at the very least resemble mechanisms of memory that are found in cognition more generally (R. Lewis, in preparation; McElree, Foraker & Dyer, 2003) But does overlap in mechanism truly mean that the long-lived logical problem of language acquisition has finally come to an end? Hardly. Saying, for example, that memory matters for language (implicit in MacWhinney’s endorsement of the competition principle) is not the same thing as saying that memory SUFFICES for language. Indirect negative evidence is of use only to an organism that knows which of an infinite number of possible omissions to care about, competition of value only if there is advance constraint on what counts as competitors. As Gibson (1992) argued in a review of earlier work on the Competition Model, as a scientific theory, competition verges on unfalsifiable. As a mechanism for language acquisition, it is only of value in specific cases where a prior constraint suggests to the child that which competes; that’s where J. Child Lang. 31 (2004), 934–935. Printed in the United Kingdom f 2004 Cambridge University Press