Can innate linguistic knowledge be eliminated from theories of SLA?

The extent to which it is possible to eliminate the assumption of innate linguistic knowledge from a theory of second language acquisition is considered briefly. The agreement-with-subject possibilities of copula be discussed by O’Grady (this issue) are reconsidered, and it is suggested that there are some problematic data for a non-linguistic linear processor. It is also argued that some emergentist accounts of the acquisition of plural marking and agreement make assumptions that are tantamount to innate linguistic knowledge. Finally, it is proposed that an appropriate model of input is an important component of a nativist theory of second language acquisition.