In Smith’s naturalistic account of morals, the impartial spectator might be seen as problematic: If this spectator were interpreted as an ideal observer, an omniscient judge, or as the representative of God speaking to us through our conscience, it would be inconsistent with Smith’s naturalistic ambitions. We propose a reading of the Theory of Moral Sentiments according to which the impartial spectator is a normal person, one who has learned what morally proper emotional responses to particular situations are for all those affected by them. Moral disagreement among normal persons triggers the engagement in sympathetic processes. In the course of the sympathetic process, agents involved in a situation exchange standpoints with unconcerned bystanders so that both the agents and the bystanders become spectators of themselves. They also share factual information about the respective situation and about their respective prejudices and convictions: Factual ignorance, prejudices, and personal convictions can be sources of bias or partiality. The parties involved in a sympathetic process reach an agreement about the proper response for a person affected; this agreement is justified in virtue of being achieved through a sympathetic process. The participants in a sympathetic process learn to become more impartial than they were before. Impartiality is a virtue—like, say, patience is a virtue; impartiality is a virtue that people cultivate in sympathetic processes. The cultivation of better impartiality is a matter of learning by communicating and doing. This learning never reaches its ideal end; our impartiality can always be improved. But not being perfectly impartial does not mean not being impartial at all. Smith’s claim is that after each successful sympathetic process, by and large, people are able to make moral judgments better aligned with those that an impartial spectator would do.