On the dyad:
You've reformulated my position as "perceptual act → correct perceptual act" and then objected that the intentional target of perception isn't a perceptual act. But that wasn't my claim. My dyad is: the act as performed, measured against a normative standard fixed by the world — the wall's stable reflectance properties, normal illumination conditions, etc. The second term isn't "another perceptual act"; it's the worldly conditions that determine what a successful perceptual act would disclose. So the dyad is act vs. world-anchored norm, not act vs. act.

You then propose your own dyad: perceptual act → world-state. And you say this is IR, because "the perceptual act discloses world-state without being world-state, and therefore intermediates between subject and world-state."

But look at what you've done. You've defined "intermediation" as: any cognitive act that discloses its target without being identical to its target. On that definition, every cognitive act is an intermediary — understanding intermediates between subject and meaning, memory intermediates between subject and past event, reasoning intermediates between subject and logical truth. You've made "indirect" trivially true of all cognition, which is exactly my objection.

The question was never whether the perceptual act is identical to the world-state. Obviously it isn't — an act of seeing a white wall is not the same thing as the wall's being white. The question is whether the act interposes an object between subject and world, or whether it constitutes the subject's openness to the world. You keep sliding from "the act is not identical to its target" to "therefore the act produces an intermediary entity." That inference is what I deny.

On "unlike types":
You claim perception is special because the two terms of the dyad are of unlike type — perceptual acts are not facts about world-states — whereas in other cases the terms are of like type. But I don't think this holds.

An interpretation is a mental act. The speaker's intended meaning is not a mental act of the listener — it's what the speaker meant, which is normatively fixed by their communicative intentions and the conventions of language. A recalled event is a present mental episode. The actual past event is a concrete historical occurrence that no longer exists. In both cases, the cognitive act and its target are of fundamentally unlike type — one is a present mental episode, the other is something in the world (a meaning, a past event, a logical relation) that the act aims to disclose.

You make these look "like" each other by using loose language: interpretation "is" meaning, memory "resembles" its target. But by that same loose standard, a perceptual act "is" a disclosure of world-state — just possibly an inaccurate one. If you tighten the standard, all the dyads involve unlike types. If you loosen it, none of them do — including perception.

On strong epistemic mediation:
You propose that perception involves “radical” multiple realizability—two possible realizers that share no properties whatsoever. A hallucinated apple and a real apple, you say, share no properties. But that’s overstated: hallucination and veridical perception share plenty of relevant properties (phenomenal character, inferential role, behavioral upshot). The difference is in fulfillment by the world, not in a total lack of shared properties. And in any case, the same kind of “radical” gap shows up wherever cognition can go wrong: a confabulated memory vs an actual past event, a delusional interpretation vs a speaker’s intended meaning, a fallacious inference vs a valid entailment. If your criterion tracks the mere possibility of empty vs fulfilled acts, it will generalize across cognition, not isolate perception as uniquely indirect.

On your response to my first objection:
You say: multiple realization requires an intermediary, the transformation must be "housed somehow," the subject must be aware of a "signal," and therefore the subject is aware of an intermediary.
But this just reasserts the conclusion. That the system transforms its input doesn't entail that the subject is aware of the transformation as an entity. I am causally mediated by my optic nerve, my visual cortex, and countless neural processes — these transformations are "housed" in my nervous system But I am not aware of my optic nerve. The processing occurs; I am not aware of the processing. I am aware of the world through the processing. You need an argument that the subject's awareness takes the transformation as its object, and you haven't provided one — you've simply inferred it from the existence of the transformation.

This is, at bottom, the same inference I've been resisting throughout: from "the system processes" to "the subject is aware of something processed." The first is a claim about subpersonal mechanism. The second is a claim about personal-level awareness. They are not the same claim, and the second does not follow from the first.

Right — phenomenal character is necessary for awareness of the apple. But necessity (or counterfactual dependence) is not mediation. I can’t see the apple without my eyes, but my eyes aren’t what I see. Phenomenal character is what my awareness of the apple consists in — the mode of perceiving — not a second object I perceive on my way to the apple. — Esse Quam Videri

I'm not saying that the colour red is an object, just as I wouldn't say that my headache is an object. I'm saying that I see the colour red, that the colour red is a mental phenomenon, and that seeing the colour red (usually) mediates seeing 700nm light and/or a surface that reflects 700nm light.

You might say that "the colour red is what my awareness of 700nm light and/or a surface that reflects 700nm light consists in", but notice that there's no "X is what my awareness of the colour red (or a headache) consists in". That's why my perception of the colour red (or a headache) is direct/unmediated/immediate in a way that my perception of 700nm light and/or a surface that reflects 700nm light isn't. This sense of direct/unmediated/immediate perception is the sense that the naive realist claims about our perception of apples — that there is no "X is what my awareness of apples consists in", only apples as literal constituents of first-person experience — and is the view that the indirect realist is opposing.

Apologies for my long silence. I needed to reflect and find some clarity. All quotations are from the message in the "reply to" link.

Then what would experience be of? If the objects you witness aren't part of your experience, and yet there are also no images in your mind that could be part of your experiences, where are you getting them? Here, image can simply mean "the image" of hte apple when you cast your eyes to it; it need not be mediated. I just want some story that doens't require an apple to be in your experience.

So we agree that the apple isn't part of our experience. It's not much, but it is something. Suppose I understand seeing something as relationship between the subject who sees and an object which is seen. Then the demand that the apple be in my eye is a misunderstanding of what "see" means. I think that is down to thinking of introspection as, in some way, a paradigm of how the senses work. In that case, perception and hearing are suspect, just because they work at a distance from their objects. For me, it is introspection that is suspect, just because it cannot be wrong and therefore cannot be right. I think the model of perception (as involving a subject and an object that is distinct from the subject) collapses in introspection. Hence I regard "I am in pain" as not a proposition like "I see an apple"'. I go with Wittgenstein in thinking of it as an expression, not a statement.
A picture of an apple is not an apple. "Picture" and "apple" are distinct objects. We can say that a picture of an apple contains an image of an apple. That's what the concept of an image is for - to articulate the way in which the picture is a picture of an apple. So it is not follow from the fact that I perceive an apple that there is an actual apple in my mind; in the case of perception, of course, there isn't an image of an apple - only apple-appropriate behaviour. But that's sufficient. If the experience is thought of as some sort of copy or model, it is needles reduplication.

I don't even understand how that could be the case. To me, it(the scientific story)'s a full analysis of what actually happens when we cast our eyes about us. I refuse, on grounds of consistency/incoherence, to call it Direct. There's nothing further needed imo. It's just slightly uncomfortable for those of us who require that the apple is in our eye.

Well, there are grounds for calling the scientific story "indirect" and grounds for call it "direct". I think the relationship is more complicated than that. "I see an apple" has what is called "success logic". It is only true I do see an apple. It is like "I won the race", that is, it is about outcomes, not processes. The running of the race stands to the winning of the race in the same logical relationship is the scientific story stands to "I see an apple".
When I said it was partial, I only meant that we do not yet understand the complete process, because our understanding of the brain is as yet, in its infancy. I wasn't suggesting that it would always be partial. But note, the scientific story has no place for the experience of seeing an apple - though it may well find correlates in the way that it has found correlates to the experience of pain.

The 'vulgar' ways of talking are heuristic/pragmatic/easier to parse but that doesn't make them right. They can just be wrong, but helpful.

Just to be clear, I don't think "I see an apple" is anywhere near being any kind of theory. It is where theory might start, but only as the question - no particular answer is implied. See above on success logic. It follows, I think, as @Banno suggests, or at least, as I interpret him as suggestng, "Direct realism" as a theory of perception is coined as a reaction to indirect realism.