The inferential character of self-knowledge

Do I have to figure out what I'm thinking, or do I simply know? The opening gives the familiar distinction: I supposedly know my own thoughts directly, while I infer someone else's thoughts from what they do. If someone seems upset, I might guess what they believe or want, but that guess could be wrong. The common-sense view described here treats my access to my own thoughts as different: direct and incapable of error.

Two terms matter. Noninferential means not reached by drawing an inference from clues. Infallible means unable to be mistaken. The psychologists will challenge both claims, so don't let the unfamiliar vocabulary hide the basic dispute: perhaps we work out our own thoughts too, and perhaps we can get them wrong.

The evidence comes from children. In certain experiments they describe simple phenomena correctly but misdescribe their own thoughts about those phenomena. The psychologists take this to suggest that children can have the relevant thoughts without being good at identifying them. On their account, thoughts help explain our actions, and we infer those thoughts rather than directly observe them.

My first question is whether difficulty describing a thought necessarily means difficulty knowing it. A child could know what they think but lack the words to communicate it clearly. That's a possible objection to the interpretation of the experiments, not a fact the writer has already established. It matters because reporting a thought and having access to it aren't obviously the same task.

Next comes a problem for the psychologists' view: if we're constantly making inferences about our own thinking, why doesn't it feel that way? Their answer is expertise. Think of the chess example. A beginner might inspect the pieces and work through possible moves to judge a position. An expert may look at the board and immediately recognize that a position is strong or weak. The absence of a noticeable calculation doesn't establish that no inferential work occurred.

Apply that explanation to self-knowledge. We get so practiced at making inferences about our thinking that they happen incredibly quickly. We don't notice the process, so it feels as though we're simply seeing our thoughts directly. From there, it's easy to assume we couldn't be wrong about what we think. The psychologists are explaining the impression of direct, infallible access; they aren't conceding that the access really is direct or infallible.

The final paragraph heads off a troubling implication. Must I observe my own behavior from the outside, just as I observe another person's behavior, to figure out what I'm thinking? Not necessarily. The proposed evidence includes internal sensations and emotions. These are described as cognitive activities that aren't themselves thoughts, but that can provide clues to thoughts.

That distinction also explains why inferential access to my own thinking needn't mean I have exactly the same evidence as another observer. I have access to these private internal clues; someone else doesn't. Their frequent occurrence gives me plenty of practice making quick, reliable inferences. Their privacy prevents someone else from using those same clues to contradict my inference. Both features help explain why my conclusions feel immediate and beyond correction, without establishing that they actually cannot be wrong.

The overall movement is from a familiar belief, to experimental evidence against it, to an account of why the belief persists. We can follow the psychologists' explanation without treating every objection to their experiments as settled or claiming that another person necessarily knows our minds better than we do.