This commentary discusses some aspects of Haider’s model of grammar change that are problematic from the perspective of usage-based approaches to language change. These aspects include (i) the postulated equivalence between intentionality and teleology, (ii) the metaphorical nature of Darwinism when applied to other domains, and (iii) the nature of explanations of language change. With respect to (i), it is argued that equating intentionality with teleology disregards the fact that innovation in grammar is not unprincipled like in genes. With respect (...) to (ii), the question is whether a comparison between as different concepts as human behaviors/brains and genes/populations can be considered as more than a metaphor (however powerful). Finally, with respect to (iii), a number of diachronic-typological studies are discussed that concur to suggest that variation in speakers’ verbal productions is largely adaptive, and therefore selection operates on a skewed pool of variants in which non-adaptive/dysfunctional variants are a minority (if any). (shrink)