Saying X rather than Y

This is why I like I.A. Richards: the way he is so attentive to the psychology of communication, or rather how the necessities of communication require certain things from one's psychology.
More fundamentally, it is his attentiveness to particular openings in the process of communication for reform, for precision, when that space is currently in complete disarray--precisely because we do not see how communication bears upon psychology, and posit ourselves as very in control of an internal situation that is already much more external, out in the world and negotiating its demands. Watch him explain the following diagram and you'll get what I mean.

S = Selection
E = Encoding
T = Transmission

R = Reception
D = Decoding
DV = Development

[SOURCE]→S→E→T⇒[SIGNAL]⇒R→D→DV→[DESTINATION]

Richards explains this in the following way:

Let me stress first the cyclic mutual dependence, the complexity of feed forward and feedback between S, E, and T.
What is selected is commonly selected in order to be encoded and transmitted. One selection is, indeed, only doubtfully distinguishable from another, except through some encodement. Hence the old question: How do I know what I mean till I see what I say? This encodement need not be what is transmitted. Many people, in writing and in speaking, use, here and there at least, private codings, ellipses, substitutes, schemata, which would be useless to others; they compose in these, translating thence into public language for transmission. Most people recognize the process describable as "making up one's mind to say X rather than Y--without either X or Y as yet being put into words." (We may, of course, differ deeply on what account we would prefer to give of this process.) The point here, however, is that priority of X over Y may be, among other things, determined by encodement problems that have not yet explicitly arisen.

Richards, in other words, doesn't necessarily use the diagram to explain communication. He does this, but also in order to poke holes in the account--eventually he will argue we need more names for "encoding," which is not just one particular activity. Or perhaps it is better to explain what Richard does as poking holes in what the diagram says but using it still, nevertheless, to try and say something about communication. He uses the useful elements, but subordinates the use to a critical activity--without getting rid of the usefulness.
In this case, he is able to pull out an entire space of problems between selection and encoding, and by the time he is done, one both has a great sense of the particular phenomenon he is getting at here--the transition between the two stages--and think that we need a lot of reform of that area. We need some way to deal with "encodement problems that have not explicitly arisen"--that is, we need to make them more explicit. We get, then, a sense of the immensity of the problem and the need for reform, but at the same time, because we see the problem so precisely (it is the problem of trying, in the process of saying X rather than Y, to see whether X or Y can be put better in words, or rather of putting something in words precisely because putting something else in words might be too hard), we feel that we really can carry out that reform--that we have many, many tools to deal with it.