According to externalist reliabilism and dogmatic foundationalism, it's possible to gain knowledge through a perceptual experience without being in a position to know that the experience is reliable. As a result, both of these views face the problem of making knowledge of perceptual reliability too easy, for they permit deducing perceptual reliability from particular perceptual experience without already knowing that these experiences are trustworthy. Ernest Sosa advocates a two-stage solution to the problem. At the first stage, a rich body of perceptual animal knowledge is acquired. At the second stage, perceptual knowledge becomes reflective after deducing perceptual reliability from the initial body of perceptual animal knowledge. I defend the alternative approach of rejecting both externalist reliabilism and dogmatic foundationalism. According to the alternative view, perceptual knowledge and knowledge of perceptual reliability require each other. Such a cognitive structure seems viciously circular. I propose that the appearance of vicious circularity dissipates when the relationship in question is viewed, not as one of temporal priority, but instead as synchronic mutual dependence. At a given time, one cannot have perceptual knowledge without knowledge of perceptual reliability, and vice versa. Such mutual dependence, I argue, is benign.