Borsley and Stephens (1989/this issue) argue that the analysis of Breton agreement developed by Stump (1984) is untenable for two reasons: first, they argue that the fundamental empirical generalization which his analysis is designed to capture (the ‘Complementarity Principle’) is invalid; second, they argue that contrary to the assumptions underlying Stump's analysis, S cannot be a barrier to government in Breton. Here, Borsley and Stephens' claims are evaluated: it is shown that contrary to their first claim, the Complementarity Principle is a valid generalization for most if not all varieties of Breton; and while the truth of their second claim is acknowledged, it is nevertheless shown that the two principal claims embodied in the 1984 analysis can be maintained in a simple revision of that analysis. The revised account is shown to be fully consistent with the evidence cited by Borsley and Stephens, and is contrasted with the alternative analysis of Breton agreement proposed by Hendrick (1988).