Suppose we create a mechanical brain that we believe is functionally equivalent to a normal working brain. For those who think science can explain consciousness, how would we scientifically determine whether the mechanical brain is conscious or not? — RogueAI

Good question. Indeed the same question applies to other humans too. How do we scientifically determine if another human is conscious, without begging any questions?

EDIT: I'm still replying to posts from page 6. Nicky's unfortunate suicide should hasten the process of me catching up. It would be weird to reply to his posts without him being able to respond.

The issue of awareness in relation to biology and how it is produced by brains will fascinate and always elude neuroscientists for years to come.

The isolation of thought to certain brain regions is the age old fallacy of old scientific ways which still pervade modern science and hold it back in its various fields including neuroscience although not so much physics and the last incarnation of quantum physics.

The problem neuroscientists face is replicating results which science always tries to do however the region always changes in the brain therefore scientists are unable to isolate thoughts and so consciousness itself

OK, science geeks, how do we determine whether an AI is conscious? What do we do? What tests do we give it? — RogueAI

Another good question. It seems to me we'd have to assume a hypothesis first that makes a prediction which we then test. How do we choose a hypothesis? All the (serious) current theories of consciousness make predictions that are compatible with observation. How do we choose between them? I think we have to do it conceptually at the moment - check for things like consistency, fidelity to definitions, entailing the fewest/least-serious problems etc.

Specifically with regard to AI, it has been argued that we might be able to interrogate it to see if it understands the concept of consciousness, to see if can introspect its own awareness. I'm not sure about that.

"Why did you do that?" - list of motives
"Why is the sky blue?" - physical cause of 'blueness'
"Why did the chicken cross the road?" - surprising answer (or non answer) designed to amuse
"Why do humans have noses?" - evolutionary (or developmental) advantages of the nose...

"Why do we have consciousness?" - ...

... what's the kind of answer that goes there? — Isaac

In this thread, it's like your second example: "Why is the sky blue?" - physical cause of 'blueness'

I want to know the physical 'cause' of consciousness, if anyone thinks there can be a plausible account of this. But I'm happy if people want to use a different verb than 'cause', because 'cause' implies a duality that some might not want.

Why should we accept that definition for machine consciousness? It's not the same thing as qualia. You just created an arbitrary definition and assigned it to 'consciousness'. It doesn't answer the question of whether a machine can have qualia. — Marchesk

Indeed. This happens a lot, even academics do it. Functionalists sometimes end up saying 'but that's just what I mean by consciousness'. Which is fine, but then they're not talking about consciousness as we know it, Jim.

Well if nature is fundamentally physical, then subjective experience doesn't conceptually fit. The biological level is still function and structure. — Marchesk

That's interesting. I frequently wonder what the word 'physical' means. I think it may be 'whatever has structure which does stuff (function).' And I agree that subjective experience doesn't really fit, as I said in the OP, "Why can't all that (structure and function) happen anyway without consciousness?"

As a panpsychist I go much further, and assert that any behaviour at all, including the behaviour of atoms, is valuable for the mind of the atom. Everything happens because of consciousness. I've been toying with the idea that all causation is actually psychological.
— bert1

I would class this understanding along with such other non-physicalist explanations of reality as Tegmark's mathematical universe hypothesis. They are metaphysical approaches and, so, there is no empirical way of testing them. They are not facts, they are ways of thinking about something. As I see it, they are not useful ways of thinking, but that is certainly opinion, not fact. — T Clark

'Metaphysical' yes, although perhaps 'conceptual' might be a better word. Opinion is contrasted with knowledge rather than fact. Panpsychism might be a fact, but one that I don't know empirically. I might know it conceptually though.

If the ultimate nature of matter is mental (i.e., idealism is true), doesn't that blow neuroscience out of the water? Isn't the whole point of neuroscience based on the assumption that mind and consciousness are produced by a physical brain? — RogueAI

Yes, which is to say that philosophy, to borrow @fdrake's metaphor, is 'upstream' of science, perhaps. When assessing a claim we could go through the following steps:

1) Is it logically possible? (this is the headwaters, the spring of the stream). If no, it's false. If yes, proceed to:
2) Is it conceptually possible? (this is a wee burn perhaps, as the local Picts say where I live). If no, it's false. If yes, proceed to:
3) Is it physically possible? (we should probably widen this out to a full-on river here, as there's a lot to this stage). I'm not sure this is totally separate from the previous stage. New concepts might emerge as a result of physical investigations, which make things conceptually possible that weren't previously conceived of. This has not happened yet in the science of consciousness, IMO, to the extent that it is now conceivable that consciousness is emergent in some way from brain structure and function.
4) there are other kinds of possibility after this... technological, practical etc, which aren't relevant here.

Debates on consciousness are at the divide between the conceptual and the physical here, I suggest. Disagreement welcome.

My question is: how would we scientifically go from there? How would science "nail down" the question of whether X is conscious or not? What tests could we perform, that would give us conclusive proof of consciousness (or lack thereof)
— RogueAI
I find this question really good and challenging!!!!
The steps are the following
1. identify a sensory system that feeds data of which the system can be conscious of.
2.Test the ability of the system to produce an array of important mind properties
3. Verify a mechanism that brings online sensory input and relevant mind properties.(conscious state)
4. evaluate the outcome (in behavior and actions) — Nickolasgaspar

I wasn't gong to quote Nickolasgaspar but I will here as he offered a response to RogueAI's question.

Predictably, this is a purely functionlist analysis. This requires either assuming a functionalist definition of consciousness to make it work, or, repeating (begging) the question, as in #3. RogueAI is precisely asking how we 'verify a mechanism that brings online sensory input and relevant mind properties.(conscious state).'

But at the bottom of it, the fact is that the subject of experience - you and I - are not reducible to objects - which is what neuroreductionism, as a philosophical attitude, tends to do. — Wayfarer

The problem with this type of reductionism is that nothing ever gets 'reduced to...' in any absolute way. The attempt at reduction always leads off into unintelligibility as the inevitability of infinite regress is approached. So the attempt to reduce the material world of inanimate objects, in this way gets swallowed by quantum uncertainty, leading to the unintelligibility of symmetry-breaking and related concepts. Likewise, the attempt to reduce an organized being (a living being), to an inanimate object, so that it might be reduced in the way of physics, is futile because of the two incompatible ways of understanding the source of organization in the material body. The form of organization required to understand the living being is not compatible with the form produced by QFT, and so unintelligibility results from the attempt at reduction.

This problem is nothing new. Though it is framed in modern terms it is as old as philosophy itself. And, in the past it has been demonstrated that a well formed dualism provides an adequate resolution.

Good question. Indeed the same question applies to other humans too. How do we scientifically determine if another human is conscious, without begging any questions? — bert1

Because I want to have my devil's advocate cake and eat it too... Assuming you don't want to beg questions, you also need to suppress the intuition that observable behaviours aren't sufficient to justify inferring something is conscious. So, y'know, purposive behaviour, ability to adapt to new scenarios, attempts to communicate, appearance of sensations and emotions - the kind of things we'd expect from a human agent. The more it quacks like a duck, the more likely it is that it's a duck.

I think that suffices to justify inferring the tested organism had "wide" content, but not sufficient to justify that it has "narrow" content - the part of an organism's state with qualia which are somehow "internal" and "private" or "caused" by their intentions alone.

But I don't know how to justify someone else having "narrow" content since everything observable seems to be "wide" content, when you take others' self reports as a form of behaviour anyway. Like p-zombies can say "I see the traffic light has red, green and yellow lights" or "Ouch" without, allegedly, the qualia. A p-zombie can behave as a qualia-haver in any way, AFAIK that's part of the point.

I don't think there is a way of inferring someone, besides you, has narrow content. Since anything the test subject does influences the tester, and thus can be construed as environmentally caused (making the content wide). And we're in a strange place if believing that an organism has wide content entails one should believe that it has narrow (phenomenal) content while seeking to preserve an observable distinction.

@Isaac - does this make you more happy?

So, y'know, purposive behaviour, ability to adapt to new scenarios, attempts to communicate, appearance of sensations and emotions - the kind of things we'd expect from a human agent. The more it quacks like a duck, the more likely it is that it's a duck.

@fdrake That's interesting. Those premises form the basis of the argument by analogy, or the abductive argument. No science necessary. An armchair philosopher who had never touched a Bunsen burner could make that argument. You could also make the same argument, but weaker, for rocks.

I'm not exactly trying to defend the ideas, just gatekeeping how they're argued against. I'm also not strongly committed to what I've written. — fdrake

That's fair enough.

I think this is dealt with by "consciousness cannot be explained in physicalist/functional terms (see prior arguments)". So it turns on the prior arguments. — fdrake

Can we have a reason why it cannot be (not 'isn't', or 'would prefer not to' - 'cannot' is big word here)? Coming back to satisfaction. I'm satisfied with the explanation in physicalist/functional terms. So how do we accommodate that? Am I wrong to be satisfied (if so where is my error, what criteria of satisfaction should I be employing)? Are some people satisfied by explanations others find incomplete (a psychological issue)?

I think for Chalmers the bridge is one of conception otherwise. — fdrake

This is my interpretation of the issue, but I can't get past the fact that it comes down do something more psychological than philosophical. Why can't Chalmers conceptualise it. I can (or at least, I think I can). Is there something wrong with Chalmers? We seem to be talking about someone's capabilities (I can/can't conceptualise it) not about the bridge itself (which would be more like 'no one can conceptualise it').

As Patricia Churchland put it

One set of reasons for dooming the reductionist research strategy is summed up thus: "I simply cannot imagine that seeing blue or the feeling of pain, for example, could consist in some pattern of activity of neurons in the brain," or, more bluntly, "I cannot imagine how you can get awareness out of meat." There is sometimes considerable filler between the "it's unimaginable" premise and the "it's impossible" conclusion, but so far as I can tell, the filler is typically dust which cloaks the fallacious core of the argument.

Chalmers is arguing against the "necessarily" part by tweaking/analysing/finagling the relevant concept of necessity. — fdrake

I don't really understand this. Are you saying he's making a point about what it is to be a necessary explicator for consciousness? Does he have an answer to that question, or a reason to doubt?

does this make you more happy? — fdrake

Things very rarely do. If you sweetened it with a bottle of Talisker...

In this thread, it's like your second example: "Why is the sky blue?" - physical cause of 'blueness'

I want to know the physical 'cause' of consciousness, if anyone thinks there can be a plausible account of this. — bert1

OK so taking like for like...

"Why is the sky blue?" (according to Nasa)

Sunlight reaches Earth's atmosphere and is scattered in all directions by all the gases and particles in the air. Blue light is scattered more than the other colors because it travels as shorter, smaller waves. This is why we see a blue sky most of the time.

Why does blue light travel a shorter waves? Why are shorter waves scattered further? Why should it be blue light thus treated and not red light or yellow? Why should gases scatter light at all? How is it that photons (energy) even interact with gases (particles) and why do wavelengths (shapes) have any impact on eyes (cells)?

I could go on, but I don't. I'm satisfied with the answer given, as it seems are you. Can you explain why you're not satisfied with any of the many answers given to the question of how/why we're conscious?

The point he’s trying to make is that while cognitive science is adequate for the explanation of the various functions of consciousness, it can’t show how to bridge the explanatory gap between those accounts and the felt nature of first-person experience. — Wayfarer

Yes, I understand that's what he's saying. Isn't that rejecting scientific explanations for conscious experience? That's how I would characterize it.

OK, science geeks, how do we determine whether an AI is conscious? What do we do? What tests do we give it?
— RogueAI

Another good question. — bert1

How do we know other people are conscious? What standards do we use? Apply those same standards to AI.

[edit] I see you made that same suggestion in a previous post.

Panpsychism might be a fact, but one that I don't know empirically. — bert1

We could have a long and fruitless discussion about this, but I suggest we don't.

Can we have a reason why it cannot be (not 'isn't', or 'would prefer not to' - 'cannot' is big word here)? — Isaac

That's not the negation of the version of physicalism/functionalism which gets attacked. If the position is "consciousness is necessarily explainable by physical/functional accounts", the negation of that is "consciousness is possibly not explainable by physical/functional accounts". Someone who takes the position of physicalism, under that guise, is actually the person who has a "cannot" statement involved in their thesis. The qualist/mysterian in that sense has a much lower burden of proof. They don't need to demonstrate necessity, they need to find one counter example. The physicalist/functionalist has to establish necessary.

If you buy that framing of the debate, anyway.

I don't really understand this. Are you saying he's making a point about what it is to be a necessary explicator for consciousness? Does he have an answer to that question, or a reason to doubt? — Isaac

If you look at the above paragraph about necessity, and accept the framing that the physicalist/functionalist is committed to a claim of metaphysical necessity, that requires an account of metaphysical necessity. Chalmers looks at that account. As far as I understand his view, he equates metaphysical possibility with conceivability - or at least takes conceivability as a sufficient condition for metaphysical possibility. Metaphysical necessity is the same as not possibly not true. If you take conception, or the other arguments like Mary's room/inverted qualia/ and all that, as sufficient for establishing metaphysical possibility, then that is actually a negation of the physicalist position. If you grant that it could be true that phenomenal consciousness isn't explainable by physical/functional processes, then if Chalmers is right, that suffices to show that physicalism is false. Because metaphysical necessity, if it really held, would require that not to be a possibility in the relevant sense of the word.

So you get into arguments about what the relevant notion of possibility/necessity is in this debate. And which notions of necessity entail which others - like what @bert1was organising in their OP. I think that's the place in the debate such dependencies between modal concepts (possibility senses) intervene.

As to the Patricia Churchland quote, you can address it like I did above. You can argue quite convincingly that it's the physicalist who needs to demonstrate that their position cannot be imagined to be false without entailing an internal contradiction.

How do we know other people are conscious? What standards do we use? Apply those same standards to AI. — T Clark

:up:

I'll also bump my pet theme. So often we don't seem to have much of a grip on what is supposed to be meant by 'consciousness.'

This is the question that I think is ill-formed. — Isaac

I can see that. I'm not sure I fully agree; I'll need to think some more. I might have made progress by the time the topic comes around next.

I don't feel like I have experiences in the sense that some proponents of the idea feel. — Isaac

This, though, I'm not sure how to read. This is probably where I locate our disconnect. I feel like you may be making a distinction I can't grasp. Maybe. I wish I could be more specific, but I'm just confused.

So often we don't seem to have much of a grip on what is supposed to be meant by 'consciousness.' — plaque flag

That is painfully true, as evidenced by just about every related discussion here on the forum. Be that as it may, with current issues about AI, it looks like it's going from an interesting philosophical problem to a practical political and social one.

fdrake That's interesting. Those premises form the basis of the argument by analogy, or the abductive argument. No science necessary. An armchair philosopher who had never touched a Bunsen burner could make that argument. You could also make the same argument, but weaker, for rocks. — bert1

True! I think it would also apply to brainstuff and brain behaviour though. You see (neurons arranged in some way), it makes it more likely the organism is conscious and/or has qualia. The relevant distinction there for a person attempting to infer a test subject has phenomenal consciousness is that observables about the test subject are "wide" content, precisely because they partially derive from environmental interaction. If you want to infer the presence of narrow content (phenomenal consciousness) from this, you'd need something more.

Part of why p-zombies are said to work, no? You can fix any physical fact about the world, and they're don't have phenomenal consciousness. All the observables are the same, they can give you the same self reports as a person with qualia, but precisely because their narrow content is "intrinsic", "internal", "first person only" etc, you can't demonstrate that they do have narrow content.

I suppose it's possible to walk the path; there are some physical observables (behaviour etc) which provide sufficient justification for claiming that a test subject has narrow content - the thing is it would always be return that the subject would have narrow content as a p-zombie is stipulated to be able to emulate any physical aspect of a human. The fork in the road is that there are non-physical observables which suffice for that justification - but I've no idea what they could be.

That is painfully true, as evidenced by just about every related discussion here on the forum. Be that as it may, with current issues about AI, it looks like it's going from an interesting philosophical problem to a practical political and social one. — T Clark

And yet isn't it fundamentally an experiential question? Is studying the nature of consciousness equivalent to actually charting the boundaries of consciousness? Or is it just a lot of talking about consciousness? Personally, I believe the boundaries have to be studied with severe existential commitment, otherwise, it is mostly just words.

It's not that the difficulty of locating consciousness among the neuro-signaling forces us to look for it in something else--that is, in some other sort of special substrate or ineffable ether or extra-physical realm. The anti-materialist claim is compatible with another, quite materially grounded approach. Like meanings and purposes, consciousness may not be something 'there' in any typical sense of being materially or energetically embodied, and yet may still be materially causally relevant. — Terence Deacon, Incomplete Nature

Like numbers, and natural laws.

it looks like it's going from an interesting philosophical problem to a practical political and social one. — T Clark

I think artificial intelligence will prove or at least threaten to be a mirror for us.

Like meanings and purposes, consciousness may not be something 'there' in any typical sense of being materially or energetically embodied, and yet may still be materially causally relevant. — Terence Deacon, Incomplete Nature

Why not 'materially' embodied ? Why not there as a dance is there ? 'Materially casually relevant' is hard to make sense of otherwise.

Like numbers, and natural laws. — Wayfarer

My objection is that we don't know exactly (do not agree, at the least) what numbers as opposed to numerals are. 'Law' is obviously a metaphor here, though its wax has cooled. We 'project' functions on measurements, postulate patterns from which we can derive implied observations, etc.

Why not there like a dance is there ? — plaque flag

Why not indeed? Have you encountered that book, Incomplete Nature? It's been discussed here a bit over the years. Cheatsheet here. Next on my list.

Why not indeed? — Wayfarer

It seems obvious: we witness dances, we don't witness meanings, purposes and consciousness.

If the position is "consciousness is necessarily explainable by physical/functional accounts", the negation of that is "consciousness is possibly not explainable by physical/functional accounts". — fdrake

Indeed. Which is why I referred you to the title of the OP. The position is "Neuroscience cannot explain consciousness"

I don't see anyone arguing that consciousness is necessarily explainable by physical/functional accounts - not even Dennett, or Churchland, or the like. Only that such an explanation is sufficient. Hence my focus on the arguments about the (in)sufficiency of those explanations, because that's the only ground I see for disputing them.

If you buy that framing of the debate, anyway. — fdrake

I suppose this means I don't...

As far as I understand his view, he equates metaphysical possibility with conceivability - or at least takes conceivability as a sufficient condition for metaphysical possibility. Metaphysical necessity is the same as not possibly not true. If you take conception, or the other arguments like Mary's room/inverted qualia/ and all that, as sufficient for establishing metaphysical possibility, then that is actually a negation of the physicalist position. If you grant that it could be true that phenomenal consciousness isn't explainable by physical/functional processes, then if Chalmers is right, that suffices to show that physicalism is false. — fdrake

Yeah, that's definitely how I understand Chalmers' argument too. I think the link between conceivability and possibility is, as Churchland put it, incredibly weak, for the following reasons;

1. I think I can conceive it. On what grounds could Chalmers argue against that to maintain his claim that his personal failure to do so has some metaphysical implication? The argument becomes one of either mental capability, or a random accusation that I'm lying and I can't really conceive it.

2. The experience of quantum mechanics should be quite sufficient to give us very good grounds for discarding the idea the 'the way the world is' is at all readily conceivable by us. The way the world is seems to be, if anything, consistently fiendishly complicated and mind-blowing at every turn.

I might have made progress by the time the topic comes around next. — Dawnstorm

Then I look forward to hearing your thoughts on the next one. There's usually one a month, they come along between the 'direct/indirect realism' one and the 'about this God chap...' one.

We have a pretty clear physicalist understanding of how, for example, a material object can become hot; by agitation of the molecules.Do you have an equivalently clear physicalist understanding of how matter can become conscious?