And yet you've still failed to even present an argument as to why it cannot be both. — Isaac

It’s certainly possible that the Standard Model is deficient and fails to uncover what ordinary perception shows, but I think that less likely than what is shown in perception being a product of the perception and not a way the mind-independent causes are.

Much in the same way that Fitch’s paradox shows that the knowability and non-omniscience principles are incompatible, direct realism and scientific realism are incompatible: if the mind-independent world is as the Standard Model says it is then it isn’t as we ordinarily perceive it to be and vice-versa.

So pick your poison: either indirect realism or scientific instrumentalism.

But then if you want to say that scientific realism is incompatible with indirect realism as well then unless there’s some third alternative we have to reject scientific realism in favour of scientific instrumentalism.

And if we reject scientific realism then we lose all empirical evidence of there being a mind-independent world at all, and so we lose any kind of realism and are left with something like phenomenalism or idealism.

Direct realism doesn’t appear to work under any scenario.

How do we know the difference between our experience of the world and the way the world is independent of our experience? You must have had some experience to even make this claim, so there must be some experience that has informed you how the world is independent if your experience. — Harry Hindu

My experience doesn’t show me the nature of the world independent of experience, the Standard Model and other scientific theories do.

There are two philosophical points here. The first is that, since the "unseen" world causes what we see, we can and have used those causes to grasp the nature of that unseen world. Science did what Kant imagined to be impossible. — Banno

Scientific realism isn’t a given, and even if it were true, the world as described by the Standard Model is very unlike the world as seen in everyday experience, and so something like Kant’s transcendental idealism (just less extreme) is suggested. This is the position I tend to take.

The reply to him is simply that since such a world is utterly outside of what we can comprehend, it cannot act as the cause of what we experience. — Banno

That doesn’t follow. A dog cannot comprehend the physical causes of his experiences but his experiences nonetheless are caused by these physical things. We don’t need to be able to make sense of some cause for it to be a cause.

Recall the scene in A Beautiful Mind where Nash asks a passing stranger if they can see the representative from the Nobel Foundation. — Banno

I would not agree these ever could be the same kind of experiential event. An hallucinatory experience is private to the subject. There is no verification of a subject's hallucinatory reporting, while a veridical experience can in principle be verified since they can report on a public environment. — Richard B

From the SEP article:

The argument from hallucination relies on the possibility of hallucinations as understood above. Such hallucinations are not like real drug-induced hallucinations or hallucinations suffered by those with certain mental disorders. They are rather supposed to be merely possible events. For example, suppose you are now having a veridical perception of a snow-covered churchyard. The assumption that hallucinations are possible means that you could have an experience which is subjectively indistinguishable—that is, indistinguishable by you, “from the inside”—from a veridical perception of a snow-covered churchyard, but where there is in fact no churchyard presented or there to be perceived.

…

However, as noted above, from a phenomenological point of view, hallucinations too seem as though they are direct presentations of ordinary objects: from the subject’s perspective a hallucination as of an F cannot be distinguished from a veridical experience of an F. This is why it seems so plausible to think of them as fundamentally the same.

That we can determine whether or not an experience is an hallucination by trying to verify them with other people has no bearing on the Common Kind Claim.

And to take it out of the hypothetical and into what actually happens; the schizophrenic hears voices and these voices are just “in the head”, and so the indirect realist’s general claim that there is something like “sense data” that can be the direct object of perception should be understandable. They just claim that this is what happens in the case of a veridical perception as well. Veridical perceptions just differ in that they have some shared external cause that causes the same kind of experience in other people. And so, again, your objections here do not undermine, or even address, the indirect realist’s claim.

You made a further claim about what does inform our intellectual considerations. You did not merely claim that the external object does not inform our intellectual considerations directly. You made the claim that it does not tout court, and that something else does. — Isaac

Yes. Just like it is not history that informs us about history but the words written in the textbook that do. But it’s still about history.

Yes, by saying that it “can” be called into question I meant to suggest that it was depending on the various interpretations. Sorry if that wasn’t clear.

I'm disputing your claim that indirectness prevents aboutness. — Isaac

I wouldn’t make such a claim. To borrow bongo’s example, a history textbook is about history, but it doesn’t provide direct access. The indirect realist’s claim just amounts to the claim that when reading about history we’re just reading words, which is true.

You seem to think that the hidden states' steps (light scattering, retinal stimulation, occipital modelling...) mean that the connection is indirect, but the steps that the visual image takes to our response (hippocampus re-firing, working memory channel, sensorimotor inference, proprioceptive cascade...) are direct. Why? Both processes seem to have steps. There are a number of steps between object and model. There are a number of steps between model and response. Why are the latter steps direct and the former indirect? — Isaac

It might be worth us looking into what is actually meant by “direct perception”. The SEP article is a good place to start:

Putting the pieces together, our ordinary conception of perceptual experience involves:

Direct Realist Presentation: perceptual experiences are direct perceptual presentations of ordinary objects.

And to better explain what I have said above:

Clearly, there are differences between these categories, but from a phenomenological point of view, these experiences seem the same in at least this sense: for any veridical perception of an ordinary object, we can imagine a corresponding illusion or hallucination which cannot be told apart or distinguished, by introspection, from the veridical perception.

…

Thus, a veridical, illusory, and hallucinatory experience, all alike in being experiences (as) of a churchyard covered in white snow, are not merely superficially similar, they are fundamentally the same: these experiences have the same nature, fundamentally the same kind of experiential event is occurring in each case. Any differences between them are external to their nature as experiences (e.g., to do with how they are caused).

…

The two central arguments have a similar structure which we can capture as follows:

A. In an illusory/hallucinatory experience, a subject is not directly presented with an ordinary object.
B. The same account of experience must apply to veridical experiences as applies to illusory/hallucinatory experiences.

Therefore,

C. Subjects are never directly presented with ordinary objects.

(C) contradicts Direct Realist Presentation, and thus our ordinary conception of perceptual experience.

If it’s not direct then it’s indirect (unless there’s some third alternative?).

Is there a Cartesian theatre when I hallucinate? Is there a Cartesian theatre when I am presented with an illusion? Is there a Cartesian theatre when I dream? Is there a Cartesian theatre when I think about the pain I feel?

It's just a strawman. The grammar of ordinary conversation doesn't dictate the (meta-)physics of perception. You need something better than "saying that we see the mental image of an apple sounds stupid, so indirect realism is false", or whatever it is you're trying to say.

But then what's indirect about it? — bongo fury

It is this visual and auditory imagery that informs our intellectual considerations, not whatever distal causes are responsible for such imagery.

Oh, so it's a picture, after all? — bongo fury

They don't see a picture. They see an apple. And the apple is mental imagery. Your mistake is in conflating two different domains of discourse. It's like saying that Frodo carried the One Ring to Mordor, that the One Ring is a fiction, and so that Frodo carried a fiction to Mordor. It's just bad grammar.

Literally they obviously don't. They 'see and hear things'. — bongo fury

That's just playing word games.

What matters is that the visual and auditory imagery that occurs in the case of veridical perception is no more "direct access" to their external cause than the visual and auditory imagery that occurs in the case of hallucinatory perception. The difference is that in veridical perception the visual and auditory imagery is causally covariant with some external stimulus, and maybe also that the character of this visual and auditory imagery is isomorphic with the nature of the external stimulus. That's indirect realism. You cannot simply dismiss this by appealing to some strawman Cartesian theatre grammar.

What kind of thing is it, if not an actual voice, and now apparently not a mental image either? — bongo fury

I didn’t say it’s not a mental image. When a schizophrenic hears voices those voices are just “mental imagery” but it’s bad grammar to then describe this as “hearing mental imagery.”

So a question back to you: do you accept that schizophrenics see and hear things that aren’t there? If so, what is it that they see and hear?

Obviously the point for me is the usual one, of whether or not seeing an apple is a case of seeing a picture of the apple — bongo fury

Does the schizophrenic who sees people who aren’t there see a picture of people (who aren’t there)?

Your Cartesian theatre grammar is a strawman.

But the book itself: is it directly about the historical events, or only indirectly? — bongo fury

I don't know what "directly" means in this context.

Regardless, as I said above, reading a history textbook doesn't give us direct access to history, and similarly (according to indirect realism) seeing an apple doesn't give us direct access to the causal world. But that's not to say that nothing in history "corresponds" to what is said in the history textbook, or that nothing in the causal world "corresponds" to what is seen. So I don't understand the point you are trying to make. You just appear to be trying to play some kind of word game.

So the true factual literature "my dog has fleas" isn't about an actual dog?

Or is it that actual things are the same kind of things as made-up things? — bongo fury

Reading a history textbox doesn't give us direct access to history.

And, I suppose: the kind of thing that we read about in true factual literature is the same kind of thing that we read about in fiction? — bongo fury

Yes.

But if it didn’t challenge scientific realism, then there wouldn’t even be a metaphysical question. — Wayfarer

You appear to conflate two difference senses of "realism". In the context of the phrase "scientific realism" it's contrasted with "scientific instrumentalism". Scientific realism says that scientific theories are "true" in the sense that the world is as the theories say, whereas scientific instrumentalism says that our scientific theories are just useful or not.

The kind of realism that can be called into question by quantum mechanics is that of counterfactual definiteness, which asserts that there are objects and that they have properties even before they are measured.

The bit where someone explains what indirect realism is still needs work. Seems unclear. — Banno

Something like: the kind of thing that we hear in the case of a veridical perception is the same kind of thing that we hear in the case of an hallucinatory perception (e.g. the schizophrenic who hears voices). The difference between a veridical perception and an hallucinatory perception is that in the case of a veridical perception the thing that we hear is causally covariant with some external stimulus (and, if a representationalist kind of indirect realism, that the character of the thing that we hear is isomorphic with the nature of this external stimulus).

I'm inclined to agree as per this particular argument. However the sentiment behind the argument, the rejection of radical scepticism by showing that it undermines itself, remains. Neo was evicted from his pod, and hence there is a world in which there is a pod. For the brain in the vat, there is a vat. The phenomenalist conclusion ... fails because the pod and the vat are not just "theoretical constructs". — Banno

The brain-in-a-vat and other such hypotheses are just analogies. The underlying principle is best exemplified by Kant's transcendental idealism. There is indeed something that is the cause of experience, but given the logical possibility of such things as the brain-in-a-vat hypothesis, it is not a given that everyday experiences show us the cause of experience. The causal world might be very unlike what is seen. And that includes being very unlike the material world as is understood in modern physics. So it's not that we could just be some brain-in-a-vat, it's that we could just be some conscious thing in some otherwise ineffable noumena.

At the very least this might warrant skepticism (in the weaker sense of understanding that we might be wrong, not in the stronger sense of believing that we're likely wrong). However, it might not warrant phenomenalism. That we can't know that the causal world is like the world we experience isn't that the causal world isn't like the world we experience. Something more than just the skepticial hypothesis is required to defend phenomenalism.

Regarding this latter point, let's say that there are two possible worlds, one which is phenomenalist (e.g. Kant's transcendental idealism obtains) and one which is direct realist. Unless such a scenario necessarily entails that the character of their inhabitants' experiences differ, skepticism is warranted in both the phenomenalist world and the direct realist world (none of the inhabitants can know which of the two worlds they live in), but phenomenalism is false in the direct realist world (and direct realism is false in the phenomenalist world).

So what we need to discuss is the assumption made in this hypothesis. Would the character of one's experiences in a phenomenalist world differ from the character of one's experiences in a direct realist world? If so, what character would we expect in each, and which of these (if either) is the character of our experiences?

Functional integration and the mind, Jakob Hohwy

There is much more to be said about these broad kinds of models, for which I refer to the papers by Eliasmith, Grush, and Friston & Stephan.

One important and, probably, unfashionable thing that this theory tells us about the mind is that perception is indirect. As Gregory puts this Helmholtzian notion:

"For von Helmholtz, human perception is but indirectly related to objects, being inferred from fragmentary and often hardly relevant data signaled by the eyes, so requiring inferences from knowledge of the world to make sense of the sensory signals. (1997, p. 1122)"

What we perceive is the brain’s best hypothesis, as embodied in a high-level generative model, about the causes in the outer world.

But I don't even know why I even need to quote this. The previous quote of Friston, especially about seeing red, is enough. It's quite clearly a form of indirect realism. It just replaces the usual notion of "raw sense data" with something else.

I don't need to understand cognitive science to understand that it's clearly a form of indirect realism. I speak English and what they say there is quite straightforward non-technical English.

Do I need to properly understand cognitive science to understand the philosophical implications of what Friston and Hobson say here?

This paper considers the Cartesian theatre as a metaphor for the virtual reality models that the brain uses to make inferences about the world. This treatment derives from our attempts to understand dreaming and waking consciousness in terms of free energy minimization. The idea here is that the Cartesian theatre is not observed by an internal (homuncular) audience but furnishes a theatre in which fictive narratives and fantasies can be rehearsed and tested against sensory evidence.

...

These facts have a powerful bearing upon our assumptions about how consciousness is engendered by the brain. We are forced to conclude that we live in something like a theatre and, while it is certainly not Cartesian, it does have properties that lend themselves to the sort of neurobiological and cognitive specification that we attempt to demonstrate in this paper.

Finally, associating consciousness with inference gets to the heart of the hard problem, in the sense that inferring that something is red is distinct from receiving selective visual sensations (visual data) with the appropriate wavelength composition. Furthermore, you can only see your own red that is an integral part of your virtual reality model. You cannot see someone else’s red or another red because they are entailed by another model or hypothesis. In short, you cannot see my red — you can only infer that I can see red. In one sense, tying consciousness to active inference tells one immediately that consciousness is quintessentially private. Indeed, it is so private that other people are just hypotheses in your virtual reality model.

What is it you understand by 'generative model posits'? What definition of generative model posit are you using? — Isaac

I'm not using any, I'm repeating what they're saying. Another quote:

But thus constructed qualia, we argue, are of a piece (modulo that added certainty, more on which later) with other inferred variables such as dogs, cats, heatwaves, and vicars. This gives our story its slightly more realist tinge. Qualia – just like dogs and cats – are part of the inferred suite of hidden causes (i.e., experiential hypotheses) that best explain and predict the evolving flux of energies across our sensory surfaces.

Do you want to argue that this theory says that dogs and cats don't exist?

The authors identified the models associated with them seeming to have qualia, but they do not actually have qualia. — Isaac

That's not what they say. They say that qualia isn't what most people think it is:

If the term ‘qualia’ is constrained to pick out some kind of raw experiential data, then qualia are an illusion, and we only think (infer) that such states exist, But in another sense, this is a way of being a revisionary kind of qualia realist, since colors, sights, and sounds are revealed as generative model posits pretty much on a par with representations of dogs, cats, and vicars.

It's right there in the quote. They're saying that qualia – colours, sights, and sounds – aren't "raw experiential data" but are "generative model posits".

And again:

[our story] identifies qualia with distinctive mid-level sensory states known with high systemic (and 100% agentive) certainty.

And I'll add something else from the conclusion:

We see red because we infer a strangely certain and peculiarly independent dimension of ‘looking red’ as part of the mundane process of predicting the world.

We see red, not because the external world stimulus has some property of being red, but because of something going on "in the head". They just make sense of what goes on "in the head" as being a Bayesian inference rather than as the occurrence of "raw experiential data".

Other papers have identified the schizophrenic's 'demon voices' with failures of backwards acting suppressive models in the autidory system.

None of these papers are saying that the phenomena is actually happening as it reported. We don't actually talk to god, we don't actually have out-of-body experiences, we don't actually hear demons, and we don't actually see qualia. — Isaac

This is misleading. The schizophrenic does hear voices¹, she's just wrong to interpret these voices as belonging to some demon (or person, or something external to herself). Qualitative experiences don't depend on there being some "corresponding" external world stimulus, and we have words like "red" and "painful" that refer to features of these qualitative experiences, e.g. the voices the schizophrenic hears have a tone, a pitch, a pace, a volume, etc.

¹https://www.ncbi.nlm.nih.gov/books/NBK557633

The fact that you've had to change the wording of the quote to make it match your conclusion says it all. — Isaac

I didn't change it to match my conclusion. I changed it so that the grammar flows better. But if you prefer:

I don't know. I don't understand what relevance other papers have to what this paper is explicitly saying? It quite clearly says "[our story] identifies qualia with distinctive mid-level sensory states known with high systemic (and 100% agentive) certainty".

If you want to argue that this account is problematic then you need to speak to the authors, not to me.

So the papers identifying out-of-body experiences with certain activity in the parietal and premotor cortices, is saying that we do actually have out-of-body experiences?

Or, is it saying that we feel like we have out of body experiences (but don't really) because of the modeling assumptions of those regions? — Isaac

I don't know. I don't understand what relevance other papers have to what this paper is explicitly saying? It quite clearly says "qualia [are] distinctive mid-level sensory states known with high systemic ... certainty".

If you want to argue that this account is problematic then you need to speak to the authors, not to me.

The claim is not that we actually have qualia, it is explaining why we might think we have qualia when thinking about perception. — Isaac

The paper concludes "[our story] is realist in that it identifies qualia with distinctive mid-level sensory states known with high systemic (and 100% agentive) certainty."

It's quite clearly accepting that we have qualia, it's just arguing that qualia is something other than the "raw data" that some think it to be.

So, again, nothing in either this paper on the "hard problem" of consciousness or the other paper on the responsiveness to sensory impressions supports your claim that being red is (only) an external world property.

Do you have privileged and unfettered access to everything that happens in your brain? — Isaac

And going back to this, no, I don't, but I do know how to speak English and use colour-terms like "red" and "blue", and so it is a mistake to think that some in-depth scientific and mathematical account of brain activity and sensory experience determines what we mean by such words.

What I do have access to is qualia, and it is this qualia that directs my use of colour terms like "red" and "blue".

If you want to translate it into what 'red' refers to it would be the trivially true statement that, for some, 'red' refers to the quale 'red' when they are verbally reporting their meta-theory of perception. — Isaac

That's what I've been trying to explain to you all along. When I say that I see a blue dress and you say that you see a red dress, you and I have different qualia (however you want to make sense of qualia), and that in this context the colour terms "red" and "blue" refer to this qualia.

The paper says:

"Our claim is that when the brain estimates that a suite of mid-level re-codings, couched in terms of features such as redness, roundness, loudness, pulsatingness etc. etc., as highly certain, it can simultaneously compute that this vivid set of (perhaps 100% agent-certain) contents is consistent with multiple ways the real world might be."

It quite clearly uses the words "redness", "roundness", "loudness", "pulsatingness", etc. to refer to "mid-level re-codings", not to any property of the external stimuli, and distinguishes this "vivid set of ... contents" from "[the] ways the real world might be."

Here's a Bayesian model

P(S1,…,Sn)=∏i=1np(Si|parents(Si))

What colour is it? — Isaac

If that’s what they mean by quaila being a Bayesian model then they’re demonstrably wrong because nothing like that happens when I see red, or indeed when I see anything. Perception/experience isn’t maths. So it’s as the first paper I referenced said: this Markov blanket theory says nothing about the philosophical hard problem of consciousness, it’s just a mathematical description of an organism’s functional response to stimuli.

Regardless, they’re the ones saying that redness is a Bayesian model, not me. My point is only that by their own account of perception redness isn’t a property of some external stimulus, contrary to your claims.

I think we would both agree that the sense data is exactly the same whether you call it a image of a duck or an image of a rabbit. — Richard B

I don't know. Perhaps we see different colours. Perhaps it's darker for me. Perhaps it's smaller for me. Perhaps your sense data is vertically or horizontally mirrored to mine. Perhaps all the lines are straight for me.

Regardless, even if in this case our sense data is the same, direct realism wouldn't follow. What you appear to be doing here is denying the antecedent. That two people having a different character experience shows that the mind-independent object isn't present in at least one of their experiences isn't that two people having the same character experience shows that the mind-independent object is present in both of their experiences. You might as well say that because everyone who sees the dress as black and blue sees the dress as black and blue then a mind-independent black and blue dress is present in the experience, ignoring the fact that there are (or can be) people who see the dress as white and gold.

Is it a matter of temperament? Indirect realism just suits you better? — Tate

I choose to argue for indirect realism because it's easier than arguing for idealism. What I actually believe is irrelevant.

Why do you think it's warranted? — Tate

I'm undecided actually. I just find it simpler to argue for indirect realism than for idealism. There's at least some common ground with the direct realist that makes for fruitful discussion.

I mean, if this is true, then how do you know about mind independent objects? — Tate

They're inferred as it can be considered the best explanation for the occurrence and regularity of experience. Of course, some don't think this inference warranted and so opt for idealism instead. But I'm not sure I'm willing to commit to that.

Why do you have confidence the standard model if you learned about it through your senses? — Tate

Why wouldn't I?

1. Difficult to understand how a scientist would observe a subject’s phenomenal character of experience since it is private to the subject.

2. Assuming that 1. Is achievable, how can a scientist compare it if mind independent objects are not directly accessible according to phenomenology — Richard B

We're able to ask people what they see. For example, there was the infamous photo of a dress that some people saw as white and gold and some as black and blue. Of course, we can never look through someone's eyes to see what they see, and it's entirely possible that everyone saw the same colours and yet described what they see differently, but I think it more reasonable to just accept that different people saw different colours.

And this itself, I think, is evidence against direct realism. Given that the character of each group's experiences were different, for at least one group either a) the properties of the mind-independent object weren't present in their experience or b) the phenomenal character of their experience was not a property of the mind-independent object (or both). Unless you want to argue that the mind-independent object was in some sort of superposition of being both white and gold and black and blue, with each group having direct access to one "version"? But that seems like quite the reach.

You agree with phenomenalism because of subatomic particles? — Tate

I think that the science of the Standard Model shows that the character of our experiences and the nature of the mind-independent world is very different.