In this dissertation, I address two distinct, but related questions: (1) Is vision encapsulated from higher-level cognitive content? For example, do higher cognitive states like belief and desire alter the contents of vision? (2) What is the scope of visual content? Is the content of vision restricted to “low-level” properties like shape and color or does vision involve a recognitional component? Regarding the first question, I argue that vision is cognitively penetrable, that what we see depends in part on the particularities of our beliefs, expectations, and goals. Regarding the second question, I argue that we visually represent at least some relatively high-level, abstract properties, such as causal interactions, animacy, and facial categories. Both these positions speak to broader issues concerning the epistemic status of our visual capacities. More specifically, we can no longer understand vision as an entirely non-epistemic capacity, one that merely provides us with a structural description of the environment; rather, the visual system carries ontological commitments and by virtue of these commitments it imposes at least a primitive order on what we see.