Six Ways to Know Your Mind
Philosophers often provide accounts of self-knowledge as though we knew our own minds either entirely or predominantly in just one way (Jesse Prinz is a good exception to the rule, though). But let me count the ways (saving the fun ones for the end).
(1.) Self-observation. Here's an old joke: Q.: How do behaviorists greet each other? A.: "You're fine. How am I?" Writers in the behaviorist tradition stressed the importance of observing your behavior to assess your own mental condition. Gilbert Ryle reportedly said, how do I know what I think until I hear what I have to say? Or: I find myself asking the server for the apple pie; I conclude that I must prefer the apple to the cherry. Every philosopher thinks we can do this, of course. Psychologists such as Bem and Nisbett have stressed (counterintuitively) the centrality of this mode of self-knowledge. I think they're right that it's too easily underrated. [Update: Let me fold all "third-person" methods, such as being told about yourself by someone else or applying a textbook theory, into this category, even though, as Pete Mandik points out in the comments, they're not strictly "self-observation".]
(2.) Introspection. We cast our mental eye inward, as it were, and discern our mental goings-on. This "inner eye" metaphor has been much maligned, and obviously there are differences between external perception and introspection -- but at least for some mental states (I think, especially, conscious states), some quasi-sensory introspective mechanism plausibly plays a role, if we treat the analogy to perception with a light touch: We detect in ourselves, relatively directly and non-inferentially, some pre-existing mental state.
(3.) Self-expression. Wittgenstein suggested that we might learn to say "I'm in pain!" as a replacement for crying -- and just as we don't need to do any introspection or behavioral observation to wince or cry when we burn ourselves in the kitchen, so also we might ejaculate "that hurts!" without any prior introspection or behavioral observation. Or a young child might demand: "I want ice cream!" She's not introspecting first, presumably, and detecting a pre-existing desire for ice cream. Nor is she watching her behavior. Yet she is right: She does want ice cream.
(4.) Self-fulfillment. The thought that I am thinking is necessarily true whenever it occurs. So is the thought that I am thinking of a pink elephant, assuming that attributing myself that thought is enough for me to qualify as thinking of a pink elephant. Again, no detection or observation required. I can make this stuff up on the fly and still be right about it. For some reason, Descartes thought this was important. My own view is that it's about as important as the fact that whenever you say, in semaphore, that you are holding two flags, you're right. (Well, okay, let me moderate that a little: For some mental states it may be an interesting fact that the self-ascription implies the existence of the state -- for example if it's not possible to believe that you believe that P without also thereby at least half-believing that P.)
(5.) Simple derivation. I adopt the following inference rule: If P is true, conclude that I believe that P. This rule works surprisingly well (as Alex Byrne has pointed out). Dodgier but still useable as a rule of thumb: From X would be good, conclude that I want X all else being equal.
(6.) Self-shaping. A romantic novice is out on his first date. He blurts out, simply because it sounds good, "I'm the type of guy who buys women flowers". At the same time, perhaps just as he is finishing up that claim, he successfully resolves from then on to be the kind of guy who buys women flowers. No detection, no observation, no self-expression of a previously existing state, no simple self-fulfillment, no derivational rule. His self-ascription is true because he makes it true as he says it. Victoria McGeer and Richard Moran have developed versions of this view, but neither quite as starkly as this example suggests. But I suspect that much of what we say about ourselves (both publicly and in private) is bluster that we find ourselves committed to living up to. This self-shaping model works pretty well for imagery reports too.
Six ways. That's it. Am I missing any? (By the way, so-called "transparency" views, the object of much recent philosophical attention, can be wedded to any combination of 3-6, and are not by themselves positive accounts of self-knowledge.)