This paper raises specific puzzles for the Strong Minimalist Thesis (SMT) based on certain crosslinguistic patterns. I do so by pointing out that the SMT entails two undesirable consequences: first, the SMT assumes that the Borer-Chomsky Conjecture is true; in other words, that all syntactic variation across languages is due to lexical differences. Second, it assumes that there can be no ordering restrictions on Merge, because they would imply the existence of an independent linguistically proprietary entity. I first present crosslinguistic evidence from case and agreement that the Borer-Chomsky Conjecture alone is not sufficient to account for syntactic variation. I then present evidence for the existence of ordering restrictions on Merge, based on a cartographic distinction between high and low complementizers. I argue that both of these patterns are purely syntactic, in that they are independent of Merge. I conclude that these independent problems raise puzzles for saltationist theories of language evolution.