I don't think much depends here on whether psychology or psychiatry is a science. What matters is whether they are doing good or not for patients. Supposing we think they are, a commonsense criterion seems to be that even if patients do not accept the diagnosis or the treatment, things change if they are contemplating harm to themselves or others, as the jargon has it. The jargon persists and is more than jargon because it has a real role; it is a criterion. A second criterion or perhaps test might be this. Can the patients get through the day or week to their own satisfaction, or look after themselves OK? If I have a friend who never gets up, doesn't bathe, won't look me in the eye, can't work and doesn't eat, and has gone down to eighty pounds, then I feel justified in intervening, because of this second criterion, over the stated wishes of the patient. Something is wrong, and I have an obligation, or some responsibility anyway, to help, especially if I am a psychiatrist. Most people in trouble do want help, I think, even if we are too dim to find the way. "Normative" ethics means ethics which is not-metaethical. I think you may be misusing the word; what you may mean is a highly directive ethics that conflicts with the patient's own ethics. The patient too has a normative ethics - his own. And what if psychiatry is misguided or "pseudo-medical" and does harm? If that is true, then we should avoid it, of course. But is it true? It seems to me it is a bit improbable that the profession as a whole is phony. It is not monolithic enough for that, for one thing. One can always switch psychiatrists.
You have described a fascinating phenomenon that I think is remarkably common, though I don't agree that it always happens It certainly happens frequently in my experience. Perhaps we both have very bright colleagues whom we happen to know very well, and can anticipate what they will say! I am delighted to see "the phenomenon" so well described. However, in the form you present it, I think most philosophers and psychologists would say that the question you ask is a psychological one, not a philosophical one, and that no doubt it is amenable to empirical research. Still, it does prompt a philosophical thought or two. I am put in mind of Wittgenstein's observation that 'In philosophy it is not enough to learn in every case what is to be said about a subject, but also how one must speak about it. We are always having to begin by learning the method of tackling it.' Perhaps when you ask a colleague about your problem, you have to decide not just what to say but how to say it, and that is enough for your mind to turn up the answer. This is true in philosophy, but perhaps if the problem is one whose shape is completely obscure to you, it is in effect a philosophical problem for you, at least temporarily. Then you see the way through the confusion to what is actually going on. I wonder what sort of experiments a psychologist might suggest to answer your question, 'What principle could explain the phenomenon?' I am sure there are bad philosophical principles that some people would drag in, such as this: you already knew the answer in a previous life (Plato). I would go very gingerly with answers like that.
My answer is a little different from Olilver's. Why do so many scholars and intellectuals think that language is necessary for thought? Answer: Because it really is easier to think about definite rather than indefinite things. But indefinite and formless things also have to be thought about. It takes more of an effort of course to think in a pathfinding sort of a way about something new, and one may or may not be thinking "in" language, whatever that means (muttering to oneself, sub-vocally?) If one is trying to come to an understanding of some hard and new logical or mathematical matter, it may be more like shaping forms in ones mind, and then moving them, and less like chattering in French. If one insists on calling "shaping forms", or whatever the metaphor is "a kind of language", then of course the claim is drained of any content, and with that of any interest. People of say that mathematics is a language, or a "language", something like a language. But it has a function and a status very different from those of a language.
You might want to take a look at some of the recent work of John Searle, such as The Rediscovery of the Mind. Searle argues there that "The notion of an unconscious mental states implies accessibility to consciousness. We have no notion of the unconscious except as that which is potentially conscious." (p. 152). While this isn't quite an endorsement of what you call the "transparency thesis," I believe it might be seen as a quasi-descendant of the view you describe.
As an aside, the name "the transparency thesis" is nowadays often used by philosophers of mind to refer to quite a different phenomenon, namely, that when we introspect our experience, we don't seem to be able to attend directly to it; rather, our attention seems to slide right through to the objects of our experience.
If I dream that I am in Lisbon, it does not follow that I am, and I may not be. Nor does it follow that I am not, of course. But if I say that I dreamed that I was in Lisbon last night, this may be one way of saying that I was not in Lisbon. If then I say that I dreamed (in a "ground-floor" dream) that I dreamed something (in a second-level dream, so to speak), it seems to follow that I did not; I did not produce a dream within my dream. And it seems to me hard to see how I could. For the contrast is between dreams and reality, and there is no reality for the second-level dreams to return to. There are just sequences of images and dreamed descriptions of them. If I dream that I have a dream, the second thing I have is not a dream within a dream, but just part of the ground-floor dream.
Depression used to be classified in two forms: endogenous ("originating from within") and reactive. There was an obvious point to this way of classifying things, but a different way has been suggested recently. The newer way is to distinguish between cases in which depression, the medical condition, is present, with its bodily neurophysiological causes, and cases in which only symptoms of depression are present, and which lack the underlying neurophysiological cause. The symptoms might however be produced by some environmental cause, a depressing incident, a personal loss or a major setback in life, without the marked and persisting change in the biochemistry of the brain which is believed to underly the medical condition. The significant difference between the two kinds of condition seems to be that changes in the environment typically don’t touch the medical condition. That may even be a defining characteristic of the condition, as well the usual things doctors look for, such as failure of concentration, changes in eating habits, sleep disturbances, and so on. (Grief, perhaps, can be severe enough to be a borderline case, or develop into real depression.)
Whether to treat with medication should be based on this distinction, or one like it. When the condition is serious enough to block the happy and satisfying life that we could otherwise reasonably expect, then we should perhaps be thinking about treating it with medication. If we see a friend or family member going through the darkness and dangers of severe depression, we will want to do what we can, and we all have an obligation to help, especially if we are medical professionals.
There is by the way also some interesting empirical evidence that people suffering from depression are actually more accurate in their judgments of situations with possible negative outcomes than people without the condition. From this point of view it might look as though the depression is the result of the more accurate perception! On the other hand, depression causes inaccurate and falsely negative mis-readings of facial expressions.
The world situation is disturbing. But if someone has real depression, can’t work, can’t sleep properly, can’t form ideas or concentrate, then this may have to do with an organic problem. What is being removed by medication is this very clear and noticeable condition, with all its miseries. It would be a trick and a tricky philosophical proposition if treatment just removed the perceptions and judgments of the world situation. But they may not actually be the cause of the distress, though it can seem as though they are. The situation is rather that proper medication may be able to prevent irrelevant and unnecessary distress, and even to make it possible to act on the view one has of the world situation.
"How can I hear my voice when I'm not speaking?" is your question. If we reserve the word "hearing" for what the ears do, and "the voice" for what the mouth speaks - not unreasonable, I think - then your question becomes, "How can I "hear" my "voice" when I'm not speaking?" i.e. "How can I undergo something which seems rather like hearing ("hear", in an extended or perhaps metaphorical sense) something which seems rather like my voice ("my voice", taken in an extended or perhaps metaphorical sense)? The important thing is to try to get clear about what the metaphors are metaphors on, if I can put it this way. ("The sun" is a metaphor on Juliet.) It is more than just imagination, because I can imagine a voice speaking, and hear it, without imagining that I hear it. That is, the minute the voice "speaks", I "hear" it, without an added act of imagination of the auditory (or rather, "atidory") kind. That is a puzzle: direct realism about inner "voices"! To the wider question, though, it is helpful to consider how astonishingly varied and numerous the metaphors are, even for one little subject. "He is de-spirited, rat-ty, a clown, a lounge lizard, a drag" and so on and on and on. Do we feel the "drag"? Do we feel the "rattiness"? Or do we judge it? Each of these metaphors trails its own distinctive cloud of epistemology. And so it is with the mind. Talk about the mind is often highly metaphorical, though not, on that account, by itself, false. If he is said to be ratty, the fact that we use a metaphor does not mean that what we mean is not true. Nor is it just metaphorically true, if this is taken to mean that "true" here is a metaphor.
There are objective scientific tests which show that we don't all see colours the same, such as the Ishihara test for colour vision. Most people don't even see the same "colours" out of both eyes. For many people the left eye might see things more saturated than the right.
The question should also perhaps be refined a bit. Shouldn't it be formulated as whether we see things (objects, surfaces, volumes etc.) in the same colours? "Do we see colours the same?" as it stands seems to mean, "Do you see red as I see red?" But this presupposes that we are both seeing red, and then the question seems to ask whether we see it the same way, for example with the same degree of saturation or exactly as blue.