Direct Realism and Perception

[EQV]

This is a continuation of the thread by the same name at the previous forum site. The original thread can be found at the link below:

EDIT: unfortunately I am not able to add links to my posts

==============

@Michael - you wrote:

On the monitor

The crux of my point was that these are non sequiturs :

A1. Our minds are “directed” towards X
A2. Therefore, we have direct perception of X

B1. Our cognitive operations are “answerable” to X
B2. Therefore, we have direct perception of X

This is shown to be so when X is “the world beyond the monitor”.

So direct perception must be explained in some terms other than A1 or B1.

On the science of perception

We’re just going around in circles meaning different things by “direct perception”. Given what naive and indirect realists mean by “direct perception”, indirect realism is the scientific view of perception. As above, it’s not about A1 or B1. It’s about phenomenology and the ontological separation between experience (and its qualities) and distal objects (and their properties). That’s all it means for our experience of distal objects to be indirect.

On the structural “mirroring”

It’s “accounted for” by the fact that distal objects and their properties are (usually) causally responsible for first-person phenomenal experience, and if the cause changes then the effect often changes. The colour I see is determined by the wavelength of the light (and my biology), so when the wavelength changes to a sufficient degree the colour I see changes. But differences in biology can entail that a greater degree of change in the wavelength is “hidden” because it does not affect a change in the colour seen, hence this colour being an epistemic intermediary.

On the alleged non sequiturs:

I don’t think I’ve ever argued in the form you’re attributing to me. My claim isn’t “our minds are directed toward X, therefore we have direct perception of X.” It’s that the kind of cognitive directedness we actually exhibit in successful inquiry — where our claims are genuinely answerable to how things are, where we can be wrong about the world and recognize that we’re wrong — already presupposes a mode of cognitive contact with reality that your framework struggles to reconstruct. The question isn’t whether directedness entails direct perception as a deductive consequence. The question is what account of perception best explains the directedness we manifestly have.

You keep insisting that I need to explain perception in “some terms other than” directedness or answerability. But I’d push back: what I need to explain is how knowledge is possible at all, and any account of perception is answerable to that larger question. If your account of perception makes knowledge inexplicable or reconstructible only through a circular inference, that’s a mark against the account, not an irrelevant consideration.

On “just going around in circles”:

I don’t think we’re merely using “direct perception” differently. I think we have a substantive disagreement about what perception is, and the terminological dispute reflects that. You’re operating with a picture on which perception is fundamentally a matter of having phenomenal qualities present to consciousness, and the question is whether those qualities are properties of distal objects (naïve realism) or properties of mental intermediaries (indirect realism). I’m questioning whether that framing — experience as the having of qualities, with the question being whose qualities they are — is the right way to carve the problem in the first place.

On the view I’m advancing, perception isn’t primarily the “having” of qualities at all. It’s an act of awareness in which an object is presented through a complex of sensory content, where that content functions not as the object of awareness but as the medium through which the object is given. The sensory qualities are real, and they are “in” the perceiver in a straightforward biological sense, but they aren’t what perception is of. The distinction isn’t between naïve realism and indirect realism as you’ve framed them; it’s between accounts that treat sensory content as the terminus of awareness and accounts that treat it as the vehicle.

This isn’t just a verbal difference. It generates different predictions about the structure of perceptual experience. On your view, perceptual error is a matter of mental phenomena “erroneously seeming” to be distal objects. But that characterization already presupposes that the natural, default interpretation of perceptual experience — that we’re encountering things in the world — is a kind of systematic illusion that needs to be explained away. On my view, the natural interpretation is basically correct, and perceptual error is a matter of misidentifying which worldly features are being presented, not of mistaking mental items for worldly ones. The phenomenology of error supports this: when I discover that what I took to be a snake is actually a rope, I don’t revise my belief that I was perceiving something in the world. I revise my belief about what worldly thing I was perceiving.

On causal explanation of structural mirroring:

You say the structural correspondence between experience and distal objects is “accounted for” by the fact that distal objects are causally responsible for experience, and that changes in the cause produce changes in the effect. But this is a causal-correlational account, not an account of cognitive access. It tells me that my experience reliably covaries with worldly conditions. It doesn’t tell me how I ever get from “my experience has this structure” to “the world has this structure.”

Consider your own example. You say the color I see is determined by the wavelength of light and my biology, and that differences in biology can mean that changes in wavelength are “hidden.” Agreed on all of that. But now ask: how do you know that? How did you come to know that wavelength and biology jointly determine perceived color, and that some wavelength differences are hidden from certain perceivers? You arrived at that knowledge through inquiry — through physics, through comparative studies of color perception, through reasoning about the relationship between stimulus and response. And that inquiry involved understanding the intelligible relationships between wavelength, biological structure, and perceptual response — not merely having further phenomenal experiences that covary with those relationships.

This is the point I keep pressing because I think it’s genuinely unavoidable: your own account of how perception works — the causal story, the biological mediation, the structural mirroring — is itself a product of understanding, of grasping intelligible relationships that hold in reality. And that understanding cannot itself be reduced to the perception of structured mental intermediaries without triggering the very circularity I’ve been pointing to. You need a mode of cognitive access that gets you to how things actually are, not just to how your mental phenomena are structured, in order to even formulate indirect realism. The question is whether your framework can account for that mode of access or whether it quietly relies on something it officially excludes.

[EQV]

@hypericin

@hypericin

You keep reverting to talk of intentional target. But we previously agreed that intentional target has little place in the discussion. If when reading a translation of Homer, Homer is our intentional target, then there can be no relation between “direct object of perception” and “intentional target”. There can be no perception less direct than our perception of Homer.

The Homer analogy was poorly chosen on my part — it introduced more confusion than clarity, and obscured the point I was trying to make. But the failure of the analogy doesn’t discredit the idea I have been defending, which is that the question of intentionality is unavoidable if we want to settle the DR/IR debate. So I don’t agree that intentionality has no place in the discussion – I think it’s the crux of the debate.

@hypericin

In your breakdown of levels of analysis, which number(s) do you think the brain’s models fall under? I, and I assume most IRists, believe that these models are the objects of phenomenology. And so they straddle 7-10, in their third person, causal description, and 11, in their first person experience.

It sounds like you are advocating for the reduction of phenomenology to computation. I’m very skeptical of this, but I’d be interested to get your perspective before going down that rabbit hole.

@hypericin

Either these models are the objects of perception of level 11, and so IR is true. Or, I’m not sure what you think these models are doing, and why you believe they exist.

As I said in my previous post, I think these models are part of the causal infrastructure that makes perception of distal objects possible. They are the vehicle not the content. They realize intentionality, they don’t terminate it. They are how we see, not what we see.

@hypericin

None of these other requirements (many of which sound more at home in computer graphics than neurology?) are themselves perceived. But qualia are.

As I’ve mentioned before, I don’t think it’s correct to say that qualia are perceived. Objects are perceived, not “objects + qualia”. Qualia are not thematized independently of objects within ordinary perception. When I see a red apple, I don’t see “redness” as an object in its own right alongside the apple. The qualitative character is an aspect of how the apple itself shows up within experience.

You can, of course, reflectively foreground the qualitative character of experience — “what a vivid red!” — but I see that as a second-order act of attention over-and-above the baseline act of perception. Treating the reflective case as revealing the true logical order of the pre-reflective case is a significant philosophical move that needs its own justification.

@hypericin

Qualia are perceived, and are required to perceive distal objects. But distal objects are not required to perceive qualia. Therefore, the perception of qualia is logically prior to the perception of distal objects.

Your argument runs:

(1) Qualia are perceived.

(2) Qualia are required to perceive distal objects.

(3) Distal objects are not required to perceive qualia.

(4) Therefore, the perception of qualia is logically prior to the perception of distal objects.

Premise (3) depends on premise (1), and I’ve already explained why I reject (1). That said, it might be worth saying a word about hallucination.

In my view, hallucination occurs when the perceptual system misfires causing an experience that purports to disclose a worldly object but fails to do so. Notice that we can only identify hallucination as hallucination against the background of a norm set by veridical perception. The hallucinator experiences something as though it were a public, spatially located, persisting object; the experience has the intentional structure of world-directedness built into it, even though in this case nothing in the world answers to it.

So even when perception breaks down, it breaks down in the mode of world-disclosure. The directedness toward mind-independent objects is so fundamental to the structure of perceptual experience that even its failures take that form. So the phenomenology of hallucination doesn’t demonstrate the logical priority of qualia, because even here it the qualia are presented as an aspect of something else – only this “something else” doesn’t exist.

This is why building a theory of perception on the hallucinatory case gets things backwards. Hallucination is intelligible only as a privation — a falling-short of what perceptual experience is constitutively structured to achieve. It’s parasitic on the veridical case in the same way a counterfeit is parasitic on genuine currency: you can’t explain what a counterfeit bill is without first understanding what genuine currency is and does, and the existence of counterfeits doesn’t show that all currency is "really just worthless paper.”

I think the indirect realist faces an underappreciated problem here. For the IR, veridical perception and hallucination have the same proximate object — a mental model or structured quale-complex — and the difference is merely that in the veridical case this model happens to have an appropriate causal history.

But then the directedness of experience toward the world becomes entirely epiphenomenal to the analysis: it drops out as a mere seeming, a user-illusion generated by the model, doing no real explanatory work. It seems like the IR is committed to saying that what the experience is of, in the most fundamental sense, is the same in both cases — the model — and the world only enters the picture as a distal cause.

But that makes it deeply mysterious why perceptual experience has the phenomenological structure it does — why it presents itself as disclosing a public, mind-independent world rather than as the inspection of an internal display. The DR has a straightforward explanation: experience has that structure because, in the normal case, that’s exactly what it’s doing.

@hypericin

cientific understanding is never identical with what it attempts to grasp, it is always an approximation. This is why science never ends, it is constantly being revised. And there is seldom a single universal understanding. 100 different physicists might have 100 different understandings of color charge, at various degrees of accuracy. Some may be entirely wrong. Only numerical distinctness between reality and understanding can support this diversity.

Yes, understanding is always revisable, approximate, and diverse. But partial understanding of X is not the same as possessing a numerically distinct copy of X that happens to resemble it. It is an incomplete grasp of X itself.

This talk of ‘grasping the same intelligibility’ across mind and world isn’t some extravagant metaphysical thesis – it’s the minimal condition for knowledge to be possible at all. This becomes clear when you try to deny it.

If understanding and reality are always numerically distinct, then assessing whether a model corresponds to reality requires cognitive access to reality independent of the model. But on IR, all cognitive access runs through numerically distinct models. So checking correspondence just produces another model needing its own check — a vicious regress. Unless you stop somewhere and say: “Here, our cognitive activity genuinely reaches reality itself.” But that stopping point is just the identity I’m pointing to.

This is why your own formulation presupposes something akin to my position. Every time you say perception “captures some structural features of the object,” you’re helping yourself to a cognitive grasp of the object that can’t itself be just another representation awaiting further verification without falling into regress.

The diversity and fallibility you cite are fully accommodated on my view — each knower grasps the same intelligible structure partially, from a different angle, with different degrees of adequacy. Error is the failure of an intended cognitive act to reach its target, not the possession of a low-fidelity model. On your account, it’s unclear how error differs from success, since in both cases all the subject has is a model, and whether it “matches” is precisely the question that can’t be settled from within the model itself.

The same applies to perception. That different organisms perceive differently and fallibly doesn’t entail that none perceives the object itself — it entails that perception is a partial, perspectival, species-specific disclosure of something real, not a sealed-off confrontation with an intermediary.

[Wayfarer]

This discussion has been carried over from here

[AmadeusD]

@EQV

But that makes it deeply mysterious why perceptual experience has the phenomenological structure it does — why it presents itself as disclosing a public, mind-independent world rather than as the inspection of an internal display. The DR has a straightforward explanation: experience has that structure because, in the normal case, that’s exactly what it’s doing.

Some pretty serious issues here imo.

There is a pure non sequitur between
“Why does experience feel the way it does?” and a conclusion about that, because its the easiest explanation. I think that’s lazy and likely my biggest gripe with DR thinking. It just stops at the easiest possible train station to get off at. Totally unsatisfying, and does not explain much of experience, to me.

You’re right, it is mysterious. That’s why there’s questions. How is it possible I can experience an apple that isn’t in my mind, where I experience everything I experience?
DR-ing wont do it to answer that question, though, because there is nothing claimed by DR which gets the apple into my experience. It sits outside my mind, and there is stays until the DRist comes and literally hand-waves it into experience. Some find this satisfying - I think its both unsatisfying, unphilosophical and unable to explain the facts of what eyes, rods and cones do. That is an hurdle that must be overcome for any theory - This is where Michael’s biggest pressure is as I see it and it is the one thing no one has attempted to address beyond hand waving in years on TPF.

How does the apple get into my mind? “You look at it” begs the question, so wont do. You need to explain. (not you, you - whomever).

[Michael]

@EQV

On the view I’m advancing, perception isn’t primarily the “having” of qualities at all. It’s an act of awareness in which an object is presented through a complex of sensory content, where that content functions not as the object of awareness but as the medium through which the object is given. The sensory qualities are real, and they are “in” the perceiver in a straightforward biological sense, but they aren’t what perception is of. The distinction isn’t between naïve realism and indirect realism as you’ve framed them; it’s between accounts that treat sensory content as the terminus of awareness and accounts that treat it as the vehicle.

I say that direct perception of pain mediates indirect perception of the fire burning my skin.

You say that pain is the medium through which you have direct perception of the fire burning your skin.

This is a distinction without a difference. If sensory content is a medium as you say and if sensory content and its qualities are ontologically separate from mind-independent objects and their properties as you say then indirect realism holds. Whether to call pain an object of perception or a mode of perception is irrelevant. The epistemological problem is the same regardless; perhaps mind-independent objects and their properties are nothing like sensory content and its qualities, and perhaps some of our sensations are not caused by “appropriate” mind-independent objects, even if they seem to be — like with phantom pains.

[hypericin]

Hey there, sorry for the long pause. Something else took my attention, and also frankly I needed a vacation from the topic.

@EQV

If understanding and reality are always numerically distinct, then assessing whether a model corresponds to reality requires cognitive access to reality independent of the model. But on IR, all cognitive access runs through numerically distinct models. So checking correspondence just produces another model needing its own check — a vicious regress. Unless you stop somewhere and say: “Here, our cognitive activity genuinely reaches reality itself.” But that stopping point is just the identity I’m pointing to.

I think every DRist falls prey to this misunderstanding. How can a brain “capture structural features of the object”? There is one and only one way: the brain builds a model of the structural features. That’s it. There is nothing else a brain can do. You keep wanting to say that the brain directly reaches out and grasps its object. But this is very vague. How can a brain do that, operationally? It cannot.

Your concern has been that this places the perceiver behind a sealed wall of models. Seeing is believing, but if seeing itself is a model, then we never grasp reality itself? This is correct, and is exactly what allows for radical skepticism. Skepticism arises because we do not have direct contact with reality, and so the models we construct in reality’s stead can be (radically) multiply realized. Skepticism is a result of the structure of perception, and would not be an issue if we could somehow grasp reality directly.

Rather than this impossibility, at some point we assume that if multiple models are in agreement, they are accurate. If I see a rock, feel it, if others see it and feel it, that there is a rock is overwhelmingly likely. But never certain.

@EQV

But the failure of the analogy doesn’t discredit the idea I have been defending, which is that the question of intentionality is unavoidable if we want to settle the DR/IR debate. So I don’t agree that intentionality has no place in the discussion – I think it’s the crux of the debate.

We can intentionally target the objects of intention. We can intentionally target perceptions themselves. We can intentionally target Homer. So what can intentionality tell us?

You keep saying that distal objects “are the target of perception”. But this is not always the case, qualia are as well in some contexts. When we smell a flower, listen to music, the qualities of perception itself are the intentional target. So, what is the significance of objects being the target of perception, some of the time?

@EQV

It sounds like you are advocating for the reduction of phenomenology to computation. I’m very skeptical of this, but I’d be interested to get your perspective before going down that rabbit hole.

I do. Either phenomenology is free standing, or it is reducible to computation, or it is reducible to physics. I find computation to be by far the most attractive of these options. But as you say, this is a rabbit hole we had better not descend unless we have to.

@EQV

You can, of course, reflectively foreground the qualitative character of experience — “what a vivid red!” — but I see that as a second-order act of attention over-and-above the baseline act of perception.

I thought you had agreed earlier that perception of qualia is not a introspective act. There is no intermediary step between hearing a chime and perceiving the sound of a chime.

@EQV

Premise (3) depends on premise (1), and I’ve already explained why I reject (1).

Rejecting my argument on the basis of rejecting the premise Qualia are perceived feels quite thin. Most everyone would agree with this premise. Even if you believe that this perception is reflexive, this does not contradict the premise.

@EQV

But that makes it deeply mysterious why perceptual experience has the phenomenological structure it does — why it presents itself as disclosing a public, mind-independent world rather than as the inspection of an internal display

I’m not seeing this deep mystery. The function of perception is to disclose information about the external world. It is how we interface with the world. What function would it serve if it were presented as an internal display?

Whereas, I have not seen you address the typical hallucination argument. Hallucination and veridical perception can be in principle identical. Yet, in one case it is perception of the world, in one case it is not. There is therefore no room for us to perceive the world directly. If the two experiences are indistinguishable, that one of them is of the world cannot be a part of the experience itself.