Narrativism, as represented by Hayden White and Frank Ankersmit, can fruitfully be analyzed as an inversion of two brands of positivism. First, narrativist epistemology can be regarded as an inversion of empiricism. Its thesis that narratives function as metaphors which do not possess a cognitive content is built on an empiricist, "picture view" of knowledge. Moreover, all the non-cognitive aspects attributed to narrative as such are dependent on this picture theory of knowledge and a picture theory of representation. Most of the epistemological characteristics that White and Ankersmit attribute to historical narratives therefore share the problems of this picture theory. The article's second thesis is that the theories of narrative explanation can also fruitfully be analyzed as inversions of positivist covering-law theory. Ankersmit's brand of narrativism is the most radical in this respect because it posits an opposition between narrative and causal modes of comprehension while simultaneously eliminating causality from narrativist historical understanding. White's brand of narrativism is more of a hybrid than is Ankersmit's as far as its theory of explanation is concerned; nevertheless, it can also be fruitfully interpreted as an inversion of covering-law theory, replacing it by an indefinite multitude of explanatory strategies. Most of the striking characteristics of both White's and Ankersmit's narrativism presuppose positivism in these two senses, especially their claim that historical narratives have a metaphorical structure and therefore no truth-value. These claims are hard to reconcile with the factual characteristics of debates by historians; this problem can be tracked down to the absence in "metaphorical" narrativism of a conceptual connection between historical narratives and historical research.