G.E. Moore, answered from the texts and cited to the page.
When I look at a red surface, what I am directly aware of is a red sense-datum. The status of this sense-datum — whether it is identical with a part of the surface, or with my consciousness, or with some mediating mental object — is a separate question I have worried at across the corpus without settling.1234 The position is that something is directly given in perception, and that something is the sense-datum; the relation between the sense-datum and the physical object that causes the perception is the philosophical problem.
The paradox of analysis: if "X = Y" is a correct analysis, why is it informative? If "a brother is a male sibling" is a correct analysis of the concept of brother, the analysans ("male sibling") must mean the same as the analysandum ("brother") — but if they mean the same, the analysis is not informative; it is a trivial identity statement.
If they do not mean the same, the analysis is wrong. The paradox is structural; my engagement with it across the Schilpp Reply and the late lectures preserves the difficulty without resolving it. Moore's Paradox: "It is raining but I don't believe it is raining" is logically consistent (the conjunction can be true) yet pragmatically self-defeating to assert (the act of assertion involves expressing belief). The paradox is well-known in the philosophy of language and the epistemology of assertion.5
My sense-datum can, therefore, be identical with this surface only on condition that it is identical with his sense-datum; and, since his sense-datum sensibly appears to him to have qualities incompatible with those which mine sensibly appears to me to have, his sense-datum can be identical with mine only on condition that the sense-datum in question either has not got the qualities which it sensibly appears to me…Chapter 7: A Defence of Common Sense, p. 131
And, if so, what is it that I am knowing about the sense-datum itself? This is the question to which, as it seems to me, no philosopher has hitherto suggested an answer which comes anywhere near to being certainly true. There seem to me to be three, and only three, alternative types of answer possible; and to any answer yet suggested, of any of these types, there seem to me to be very grave objections.II A Defence of Common Sense , p. 56
I therefore define the term in such a way that it is an open question whether the sense-datum which I now see in looking at my hand and which is a sense-datum of my hand is or is not identical with that part of its surface which I am now actually seeing. That what I I know 'This is hand, seems to many parts (e.g.II A Defence of Common Sense , p. 55
For my judgment, whatever it is, is true. And yet, if this presented object is not identical with this part of the surface of this inkstand, it certainly is not a part of an inkstand at all; since there is no other part, either of this inkstand or of any other, with which it could possibly be supposed to be identical.Philosophical Studies, pp. 239–240
But to say that 1 feel doubt as to this, is to say that it is fosstble that it is identical. And, tj it is identical, then I am both feeling sure of and doubting the very same proposition at the same time. 1 do not say, of course, that 1 am doing this. 1 only say that, so far as I can see, I don't know that I'm not. But it may well be thought that to say the latter is as bad as to say the former.A Reply to My Critics, pp. 103–104
The sense-data theory was the dominant philosophy of perception in early-twentieth-century British philosophy. Austin's Sense and Sensibilia (1962, posthumous) attacked the framework across the board; the contemporary direct-realist tradition has largely abandoned sense-data. The paradox of analysis remains a live problem in metaphysics and philosophy of language. Moore's Paradox is a standard topic in the epistemology of assertion. Baldwin's Chapters VII-VIII give the systematic contemporary exposition of all three.