This essay asks whether a pragmatist philosophy of history can make sense of the notion of historical facts. It is tempting to think it cannot, since pragmatists insist, as James puts it, that the trail of the human serpent is over everything. Facts, by contrast, are typically thought of as something untouched by the human serpent, something impervious to what we think and do. I argue, however, that there is a way of understanding facts that is perfectly at home in pragmatist philosophy of history. Drawing on work by Robert Brandom, I propose that facts be interpreted inferentially. On this view, to call something a fact, or to say that the facts make my beliefs true, is simply a shorthand way of saying that a particular sort of relationship exists among certain sentences. I further show that this inferential understanding of facts is fully compatible with the distinctive features of historical inquiry. In particular, it is compatible with history’s irreducibly narrative character, and with the way different narratives can reveal radically different facts. Finally, I use this account of historical facts to respond to a classic criticism of pragmatism: the charge that pragmatism is narcissistic. I argue that pragmatism is narcissistic in only the minimal sense that it cannot countenance theory-neutral givens. But pragmatists can happily grant that there is more to truth than consensus, and that our claims are answerable to facts that everyone can get wrong.