"It has been suggested that the sentences giving rise to the problem of the synthetic a priori are really "postulates about the meanings of words" ar»d, therefore, that they do not express what is synthetic a priori. But if the suggestion is intended literally, then it would seem to betray the confusion between use and mention that we encountered earlier. A postulate about the meaning of the word "red," for example, or a sentence expressing such a postulate, would presumably mention the word "red." It might read, "The word 'red' may be taken to refer to a certain color," or perhaps, "Let the word 'red' be taken to refer to a certain color." But, "Everything that is red is colored," although it uses the words "red" and "colored," does not mention them at all. It is not the case, therefore, that, "Red is a color," refers only to words and the ways in which they are used."

“What we have been saying, of course, is not likely to convince the skeptics and we can hardly claim to have "refuted" them. But our question was not, "Can we refute the skeptics?" Our question was: "Are there positive reasons for being skeptical about the possibility of succeeding in the epistemic enterprise?" The answer seems to be that there are no such reasons. And therefore it is not unreasonable for us to continue.”

“You may have knowledge, then, without having any insight into the epistemic status of what you know. In other words, you may know a proposition h to be true without having any beliefs at all about the fact that h is evident or about what makes h evident for you. You will have some degree of insight into your knowledge of h if you have a true belief about what makes h evident for you. You will have a greater degree of insight if, moreover, you have no false belief to the effect that some other proposition makes h evident for you. And you will have an even greater degree of insight into the status of your knowledge of h, if you also know that e makes h evident for you. But our ordinary knowledge about such things as ships, trees, and houses does not require that we have any beliefs about our epistemic situation.”

“In making their assumptions, epistemologists presuppose that they are rational, beings. This means, in part, that they have certain properties which are such that, if they ask themselves, with respect to any one of these properties, whether or not they have that property, then it will be evident to them that they have it. It means further that they are able to know what they think and believe and that they can recognize inconsistencies.”

“It is true, of course, that most of us have very little knowledge of our psychological makeup and that we are likely to accept oversimplified and false accounts of why it is we think and act as we do. But this fact is quite consistent with what we have said about self-presenting properties. For, although these properties may mislead us about other things, they are not a source of error about themselves.”

“Strictly speaking, "The wine tastes sour to me," and "something looks red to me," do not express what is self-presenting in our sense of this term. For the first statement implies that there is a certain thing—namely, the wine— that I am tasting, and the second statement implies that there is a certain external thing that is appearing red to me. But, "I am tasting wine," and, "There is a certain external thing that is appearing red to me," do not express what is self-presenting. What justifies me in thinking that I am tasting wine is not simply the fact that I am tasting wine, and what justifies me in thinking that a certain thing is appearing red to me is not simply the fact that a certain thing is appearing red to me. To arrive at what is self-presenting in these cases, we must remove the reference to the external thing—to the wine in, "The wine tastes sour to me," and to the appearing thing in, "That thing appears red to me." This, however, is very difficult to do, since our language was not developed for any such philosophical purpose.”