It is a long-standing observation that cross-linguistic syntactic variation sometimes seems to correlate with variation in the inflectional morphology. An attractive consequence of this observation is that it may provide the basis for genuine explanations of certain aspects of syntactic variation and thereby replace accounts in which the variation is simply due to random differences in parameter settings. However, in many instances in which such a relation between the syntax and morphology has been postulated, occasional counterexamples to the generalization can be found and they shed doubts on its validity. In this paper, it is argued that, once we consider such counterexamples not just from a synchronic but also from a diachronic perspective, we are not necessarily forced to abandon the basic intuition behind the generalization. Taking data concerning the order of arguments ('free' vs. fixed) and the distribution of subjects and adjuncts in subject-verb inversion contexts as illustrations, it will be shown that apparently problematic cases can often be dealt with if we adopt a specific conception of how the syntax is related to inflectional morphology and of how grammars change.