Seeing and Knowing
(1969) is an influential book by philosopher Fred Dretske. Dretske explores the epistemology of perception and offers an externalist
picture of the acquisition of perceptual knowledge. With visual experience as his
subject, Dretske posits that there are two fundamental ways of seeing, epistemic
and nonepistemic. Crudely, the distinction is as follows.
Nonepistemic seeing: S sees b
Nonepistemic seeing is the nonconceptual and noncognitive apprehension of your environment – you see an object, but you don't see it as
anything. Note that although this is purely a matter of fundamental sense perception, to be nonepistemically seeing some object you must still visually differentiate it from its environment.
Epistemic seeing: S sees that b is P
Epistemic seeing is both conceptual and cognitive. As the name suggests, it involves beliefs. In the
winter months, I see white shapes fall against my window, and I also acquire the belief that those shapes are snow, not to mention the belief that it is snowing. Other philosophers have dubbed this propositonal seeing.
Dretske was writing at a time when analytic philosophers generally took knowledge to be a shorthand for justified true belief.
Seeing and Knowing rejects that formulation's emphasis on justification, arguing
that the knowledge (true belief) we glean from what we see is not
necessarily or even typically accompanied by internal justifications.
For instance, let's look at the causal link between nonepistemic and epistemic seeing. Specifically, we'll examine the move from (1) seeing an apple on the table to (2) having knowledge that there is an apple on the table. Dretske's major observation here is that in these visual cases, barring any skeptical hypothesis being true, the perceptual experience of (1) can be the basis of our knowledge in (2) without necessarily entailing that knowledge. Furthermore, when (1) does lead to (2), it can do so without relying on extraneous intellectual justification. Let's take a more formal look at what this true belief sans justification entails:
In perceptual cases, S has knowledge of b (sees that b is P) iff:
- b is P.
- S nonepistemically sees b.
- The conditions of S nonepistemically seeing b are such that b would not look the way it does to S unless b were P.
- S, believing the conditions are as (3) describes and not the result of any false assumption, takes b to be P.1
So it is the fact that one sees the apple on the table in this way, under normal
conditions, that allows one to recognize and know that it is there. In these ordinary
circumstances, attendant assurances that he isn't actually looking at a
painted orange don't even enter into the picture.
Dretske's idea of how we know things about what we see relies, then, on his
disjunction of the visual and the conceptual ways of seeing. The ability to pare seeing down to its elementary, nonepistemic component rescues our picture of the world from being irreducibly interpretive, lacking direct connection to the immediate objects of our experience. This gives us room to acknowledge the difference between our unique, individualized perceptions of the world without undermining the objective and universal status of the objects of our perception. Dretske's minimal notion of nonepistemic seeing furthermore allows him to substantiate his claim that sense experience does not always entail belief, and may in fact be independent of it.
I am lifting these requirements from Gerald Doppelt's account of the material in "Dretske's Conception of Perception and Knowledge".