Abstract
In response to Roth and Ryckman, I explain in more detail why narratives fashioned with ideal, quantitative covering laws cannot be combined into large-scale covering-law explanations and specify further reasons for supposing that history can be conceived as dynamically nonlinear. I also appeal to an episode in the history of science to examine the idea that dynamical complexity is local in historical space and time and to suggest that such complexity does not pose a unique problem for historical narration. Finally, I suggest that Roth and Ryckman's critique of the use of nonlinear dynamical concepts in historical explanation must extend to explanations employing concepts from linear science. I conclude that their warning against the incoherence of scientism is not convincing.