How can "non-conscious parts of the brain" have experience like conciousness have?

So basically what you are saying is that the underlying layers, when writing that sentence, make up some nonsense which they cannot know... which turns out to be true coincidentally. Seems less plausible than the conscious mind writing that sentence directly.

...and if it be nonsense it could be that only for somebody's consciousness, isn't it?

@ttnphns They don't. The OP's point is that in order for those parts of the brain to generate that sentence, the would have to have experience, but they can't, so they wouldn't generate that sentence. What I'm saying is, why can't they generate false statements about having experience?

@vonjd makes a good point. There would be a probabilistic argument against the position you are presenting. That some people have false memories does not mean that we all do all the time.

@vonjd It need not be "nonsense"; the layers that generate actions may be using data taken from elsewhere: conscious parts of the brain, culture, etc. And, if so, that it matches to reality is not coincidence. But the statement can still be generated without conscious will.

Can there be a reality without a consciousness?

I would agree for most statements but I just don't see why (and how) layers that basically don't know what they are talking about should generate accurate statements about consciousness.

@vonjd Why do you assume they don't know what they're talking about? I gave you two possible routes of data flow to these centers in my previous comment.

Chelonian, pardon please, but tell me how a physical system or machine can assert, judge, know, understand? All that is impossible without consciousness.

@vonjd Just to even further articulate what Chelonian is saying so you don’t overlook it... There’s no reason to assume that the underlying layers cannot see and understand the consciousness. If this is the case, the underlying layers would have an understanding of what they would define consciousness as: the experiential manifestation of their actions in the consciousness. Such layers could legitimately say that they, who operate a conscious being, perceive the world from a first person perspective, and can act according to it. (It is not the consciousness, but the layers below, that write).

@ttnphns how can a program, through machine action, know how to ‘best’ act in certain circumstances? Such an action requires judgement and assertion (of a different sort than you would normally describe, but real nonetheless).

@ttnphns How do you know that? Why can't, for example, artificial intelligences do just those things--without consciousness? [The rest would have to go in a chat, thanks]

@Kieran, that is crucial point where I think you and Chelonian are mistaken. A program "judges", "asserts" etc. not itself by per the human consciousness of its inventor. It is he/she who has translated their conscious project into machine algo, more or less successfully, but that did not make the program or the machine conscious. The reason why it is so is this: only consciousness can deny (negate). Denial lies at the core of assertion, judgement etc. Machine or neurons can't deny.

@ttnphns judgement does not actually require denial, it requires the value prescribed to certain ideas to be of greater value than other ideas. If a machine learning program could not do this, it would never progress towards its goal, because its goal has a direction of progression, along which certain ideas are more highly valued

@Kieran, to my mind "greater than", "equal to" etc. certainly requires denial, or attenuation of self-identity. Also, when you reassign a variable a new value, you deny not a value but actually the variable: transition from Y=i to Y=ii is possible only due to denial of Y. Only consciousness can deny self-identity of things. See also philosophy.stackexchange.com/a/45193/28067. That's how I think, I may be wrong or may change my mind one day.

@ttnphns I think you're being naive about what AI can do. Take a look at Alpha Zero in particular; specifically: (1) self trained chess engine using no human knowledge (outside of just what the rules of chess are) (2) uses two distinct fundamental mechanisms... a policy network that recommends moves, and a value network that judges move quality. Both networks are trained entirely by self play in the domain (playing chess against itself); the "zero" in Alpha Zero refers to the fact that it has no human knowledge about how to play.

@HWalters, I admit readily that machines can be much more clever than humans, can learn from zero level etc. What I've expressed is a different point. Efficacy cannot be a criterion to ascribe having or not having consciousness. The fundamental property of consciousness is an ability to question, or doubt, for oneself.

(cont.) A machine can effectively simulate the behaviour which will look as if it is asking itself or is doubting, but that resemblance will be an analogy, not homology. Because machine cannot conceive of, say, 2<>2, that is, think that 2 is not identical to itself. Machine just won't work if it is instructed "suddenly 2 can be nonequal to itself", but consciousness just needs this capacity to neant (take to nothing) self-identities in order to exist.

According to Heidegger, human is the only entity which existence implies asking about own being in the world. When machines start producing questions such as why am I here/living they probably have accuired consciousness (?)

@ttnphns "can learn from zero level" ...this. The narrow point I was making involved zero level learning; A0 makes "judgments" that are intentionally isolated from its programmers (which is how I read "not itself by per the human consciousness of its inventor").

@HWalters, pretty good. But I would yet claim these are pseudojudgements, they become "judgements" only per human consciousness of the perciever observing the machine output behaviour. Insects are too, look very "smart" from outside. Do we easily say have consiuosness like we do?

@ttnphns You wrote:

> Denial lies at the core of assertion, judgement etc. Machine or neurons can't deny.

So, since we have neurons to think with, wouldn't that imply that we, too, don't have consciousness...unless you assume some non-physical aspect to our minds? Is that what you think?

@Chelonian Consciousness is by definition the non-physical aspect, sure. That neurons think for us is a naturalistic scientific theory and have nothing to do with philosophy proper (in my view).

Hmm, then let me rephrase that: Do you think we are thinking with something non-physical in addition to our neurons? Like a supernatural soul or something like that?

@ttnphns Maybe should have tagged you on that last question so you were alerted to my reply, so doing it now.

@Chelonian We don't think or perceive "with" something (which itself is outside of consciousness or soul or whatever to call it), we just think and perceive. Within a philosophical thinking, that is suffice to me. (Sorry, I have to go for today. Cheers to you, and thanks!)

OK, thanks, but it seems I'm unable to understand you.