It is widely believed that in this world there exist mental objects besides physical entities, such as desires, beliefs, emotions and sensations. These objects can barely be reduced to physical concepts, but we often use them to explain this world. For instance, when Grey runs to the bathroom, we can say Grey has a desire to urinate, and he believes rushing into the bathroom would help to satisfy that desire. Normally we have a decisive intuition about what kind of things are mental, and there is little difference between our intuitions. In other words, we don't hesitate when we classify something as mental or physical, and our judgments are always consistent. However, on the other hand, we usually are incapable of setting precise criteria for classifying things as "mental" or "physical," nor can we define "mental."
It is proposed that mental things are those and only those that we have direct knowledge of it. A knowledge is a direct knowledge if and only is it isn't based on inference or other evidence and beliefs. This proposal sounds reasonable because its criteria makes sense: for instance, the pain I feel is mental because I have direct knowledge about my pain. This knowledge is not inferred or acquired from evidence or belief. This proposal works the same for other desires, beliefs, emotions and sensations.
We can express direct knowledge as a criterion in this way:
Definition A
X is a mental state of S, iff S has direct knowledge about X.
When we define a concept, the definition must be in a sufficient and necessary condition of the object that the concept applies to.
Therefore, if Definition A is correct, then
1. If X is a mental state of S, the any knowledge S holds about X must be a direct knowledge and
2. If S has a direct knowledge about X, then X must be a mental state of S.
Some people considered 2 to be wrong and listed some counterexamples. For instance, when I see something red in front of me, I would have the knowledge that
(a) There is something red in front of me.
According to these guys, (a) would be a direct knowledge because it is not inferred or based on other evidence and beliefs. Therefore, according to Definition A, the red thing in front of me would be my mental state; but that is obviously wrong. They thus conclude that S has a direct knowledge of X is not enough for X to be a mental state of S. Therefore, 2 is incorrect.
I think this counterexample is not a successful one, because (a) is not a direct knowledge. In order for (a) to be a direct knowledge, it must be acquired without any help. But apparently, if we have no knowledge of the existence of the external world, how could we be conscious of external things? Besides, if we can directly possess knowledge about the external world like that, then it would be ridiculous for so many philosophers to trouble over the problems of justification.
Another more convincing counterexample than the previous one is that we have direct knowledge about our present body posture. Given that they are retained, a person doesn't need inferences, evidences or beliefs to know what her body posture is, such as how her knee is bent and where her hands are. In this situation, "my knee is bent" and "I am making a fist with my right hand" would both be direct knowledge. Therefore, given Definition A, my bending knee and my fist would both be my mental states. Nevertheless, they are obviously both physical states. Therefore, Definition A is false.
"A person's knowledge about self body posture is direct?!" My first reaction was "then how could I step on my own foot?" Of course this is not a good objection. With direst knowledge about one's body, one could also accidentally step on one's own foot because one didn't concentrate. A decent objection should sound like this: first, just like the former one, in order to know something about my body, I have first presume that there is an external world, thus any knowledge about my body would at least base on some statement about the existing external world. (It would be no use to reject by introducing some Idealism of ontology. Because in the picture of Idealism, everything is mental state.) In fact, there would be numberless preconditions we take to establish our body-knowledge. Therefore, even though our opponent skip the challenge of the external-world presumption, there are still a bunch of presumption (for example, "I'm not a phantom-limb patient.") he must take on order to grasp the body-knowledge which he expect to possess "directly."
Second, it seems we possess the knowledge about our bodies by some inferences we ain't usually aware of. A more obvious situation would be when one is half awake in bed, one actually don't know whether his knees are bent or not. In order to know whether my knees are bent or not,in this case , I must slightly move my legs and feel them rubbing against each other or against the bed sheets. Although there isn't a serious argument, I believe that our knowledge about our body postures is based on inferences from our perceptional experience. We are barely aware of this because we are used to them.
