In this paper, I argue that, in his Theory of Moral Sentiments, Adam Smith conflates two different meanings of ‘self-command’, which is particularly puzzling because of the central role of this virtue in his theory. The first is the matrix of rational action, the
one described in Part III of the TMS and learned in ‘the great school of self-command’. The second is the particular moral virtue of self-command. Distinguishing between these two meanings allows us, on the one hand, to solve some apparent paradoxes of the text; and, on the other,
to identify various features of both the practical reason and deontological ethical traditions that are present in Smith's sentimentalism, enriching his phenomenological account of moral actions.