In order for the concept, 'recognition', to play a critical role in social theory, it must be possible to draw a distinction between due recognition and failures of recognition. Some recognition theorists, including Axel Honneth, argue that this distinction can be preserved only if we presuppose that due recognition involves a rational response to "evaluative qualities" that can be rightly perceived in the context of social interaction. This paper points out a problem facing recent defenses of this "perception model" and (...) proposes a solution. I begin by making explicit three criteria broadly shared by recognition theorists: an adequate theory of recognition must be (a) critical, capable of drawing a distinction between due recognition and recognitive failure; (b) minimally idealist, able to accommodate the mediating role of subjectivity in human receptivity to value; and (c) historicist, should not rule out a priori that the values involved in recognitive practice are historically emergent. I then argue that Honneth's version of the perception model cannot fulfill these criteria simultaneously. Drawing upon an Aristotelian notion of evaluative perception and working through an example from Zadie Smith's On Beauty, I defend an alternative view that is, I argue, better equipped to meet these three criteria. (shrink)