In Bertrand Russell’s The Principles of Mathematics and related works, the notion of a proposition plays an important role; it is by analyzing propositions, showing what kinds of constituents they have, that Russell arrives at his core logical concepts. At this time, his conception of proposition contains both a conventional and an unconventional part. The former is the view that propositions are the ultimate truth-bearers; the latter is the view that the constituents of propositions are “worldly” entities. In the latter (...) respect, Russellian propositions are akin to states-of-affairs on some robust understanding of these entities. The idea of Russellian propositions is well known, at least in outline. Not so well known is his treatment of truth, which nevertheless grows directly out of this notion of proposition. For the early Russell, the import of truth is primarily metaphysical, rather than semantic; reversing the usual direction of explanation, he holds that truth is explanatory of what is the case rather than vice versa. That is, what properties a thing has and what relations it bears to other things is determined, metaphysically speaking, by there being a suitable array of true and false propositions. In the present paper, this doctrine is examined for its content and motivation. To show that it plays a genuine role in Russell’s early metaphysics and logic, I examine its consequences for the possibility of truth-definitions and the problem of the unity of the proposition. I shall draw a few somewhat tentative conclusions about where Russell stood vis-à-vis his metaphysics of propositions, suggesting a possible source of dissatisfaction that may have played a role in his eventual rejection of his early notion of proposition. (shrink)