It’s not hard to see how philosophers have tied themselves into such knots over qualia. They started where anyone with any sense would start: with their strongest and clearest intuitions about their own minds. Those intuitions, alas, form a mutually self-supporting closed circle of doctrines, imprisoning their imaginations in the Cartesian Theater. Even though philosophers have discovered the paradoxes inherent in this closed circle of ideas—that’s why the literature on qualia exists—they haven’t had a whole alternative vision to leap to, and so, trusting their still-strong intuitions, they get dragged back into the paradoxical prison.
Dennett, D. C. (2017). Consciousness Explained. United States: Little, Brown. 369-370.
Dennet points to a lack of a good philosophical framework from which to address qualia. I largely agree with that but that is not a good reason to ignore the concerns that are associated with not having a good philosophical framework. Philosophy goes through paradigm shifts like science and that only happens when the current frameworks are ineffective and people continue to ask questions. I think Dennet also goes into the bucket of philosophers claiming to have answers but actaully providing a meaningles performance, so I double your Dennet with a Dennet
Dennet seems to consider himself a verificationist which is a limited perspective for a philosopher as it basically makes his philosophy a subset of what science can empirically verify. Philosophy has potential value in being broader than science and pointing out where science is in need of a paradigm shift e.g. the empirical verificationist paradigm runs into limits in regards to qualia. Philosophy itself has a paradigm shift on the cards, raised by issues around qualia.
Dennet is probably not a very welcome voice for an HTM believer because Dennet claims people like Jeff are underestimating the diffculties of generally intelligent machines by orders of magnitude. Given that Jeff claims to have the key pieces of the puzzle in place there is an obvious tension there.
@alastair, in the model of TBT I guess the correlation with qualia of red is distributed across thousands of different populations of neurons. The idea that you have a population that maps to “red” does not fit with TBT. If you push your idea to an extreme, imagine you could identify “red” neurons (i.e. that correlate with a particular red experience reported by someone) and you could scoop them out, put them in a dish, and then stimulate them to behave in the same way, where would the red experience be then?
I frankly don’t see where you got that. If you understood Multiple Drafts you would see that HTM is ideally suited to its implementation. I’ll state now that I am not an ‘HTM believer’ as that smacks too much of philosophy and not science. HTM is predicated on the structures in the neocortex and does an excellent job at modeling the spatial and temporal associations that it engenders, which are, if you ‘believe’ Jaynes (and Dennett, and others) the core of human consciousness.
I side with Dennet. Unfortunately most of the fun in philosophy is with the stuff you can’t verify.
Science always runs into problems with anything subjective. As I understand it, qualia is not about the redness, it’s more about the subjective ‘feel’ of red (whatever the hell that means). Definitely not science, or HTM.
HTM (like Siri or the self-driving aids in my car) show how far we’ve come, but they also show how far we still have to go. Siri stumbles on pronouns and simple physics. My car makes the wrong decision way faster than I can make the right one. General AI has been 30 years away for the past 50, and I see no reason to change that now.
In your example of “scooping out” the red neurons, I think it’s clear that the experience of redness would no longer occur to the person who’s just been red-lobotomised : they have no neurons which are activated by that colour, so how could they experience it ? They can’t experience red until their cortex has trained a new set of neurons to respond to that input.
I can’t say I’m really convinced by the idea of having multiple representations of a single object or concept. If you’re going have an invariant representation of something then there are going to be all sorts of problems coordinating them as things change. For example, if I have an invariant representation of a camel (bear with me) as a desert-dwelling mammal with a hump and a long neck etc etc, then what happens when I discover that camels can come with 2 humps ? If my representation of a camel is stored in one place, or across a single network of neurons, then I have only to update that one representation, which sounds fairly straightforward. With thousands of representations, I have to change at least 50% of them to get the “right” answer when I think of a camel. That seems lengthy and cumbersome in comparison. But I guess it’s a different issue anyway, and it’s probably not good to digress on these threads !
This is my understanding as well…but now that we have started talking about it, the question that arises in my mind is ‘how do we feel?’ It has been easy for me to just toss qualia based on Dennetts excellent dismissal of it, but now I am getting bothered about it with all this talk of redness and camels. So when does qualia happen? Does an infant have a sense of qualia? If so, how do you know? If not, when does it happen? Does it happen when the kid first says “Oooooh…pretty!”?
It starts to get to be like the cartoon below, which Dennett included in CE.
If all of the neurons currently associated with redness were to suddenly and simultaneously die or be removed (obviously not likely, but we’ll entertain the hypothetical), then the red-blindness would likely only be temporary. There is likely nothing special about those specific neurons that encode redness. They just happened to be the ones that were in the right place to receive those signals and pass them along. As long as there are still incoming spike trains generated by the red-sensitive cones in the retina, and there are neurons in the vicinity of the terminating axons from those ganglion cells, then the perception of redness would most likely be picked up by and shifted to other combinations of neurons. The association of redness with red objects would need to be relearned by these new connections. The perception might start out a bit noisy and possibly a bit disorienting, but ultimately the sensation of red would most likely return to some semblance of what it was prior to the event.
From my understanding of what people seem to be describing as qualia, it seems unlikely that qualia is some intrinsic property of a direct sensory stimulus. The qualia of red does not come from the perception of red light coming in from the retina. Rather it is part and parcel of the stored representations and associations of all red objects. The feeling of red is closely tied to your feelings about apples, fire-engines and sports cars, or the sensation of hotness. Likewise, the feeling of blue is probably tied to your feelings about open skies, large bodies of water, or the sensation of coldness; green with grass, trees, wide-open pastoral settings… etc. Some of these may be learned associations, but it is also possible that there could be a deeper layer in the brain (amygdala?) that might be more hard-coded to respond to specific sensations with a definitive or visceral emotional response. So, this qualia that you seek may be embedded in structures that aren’t even in the neo-cortex (at least not at first). Our particular interpretation of how these sensations make us feel could be intimately tied up with the responses from the lower brain and our learned associations with specific objects/events.
Barrett states the obvious: At times, we know exactly how we feel: we are angered by blocked goals, saddened by a loss, or afraid of an impending challenge. At such times, we represent our feelings in a precise and differentiated fashion. Knowing how we feel helps to inform us about the significance of the immediate situation, to work out what we should do next, and to indicate what, if anything, we should do about changing how we feel.
In the thousand brains theory, there are many models of each object but each of those model is built using different sensory inputs. So there is a model of camels in the auditory cortex that knows what noises camels make, and another model in the visual cortex which knows what camels look like. When you learn about two humped camels only the model in your visual cortex needs to update.
I think that’s what I was trying to say, but you’ve put it better. The brain will start re-learning ‘red’, but for a while at least, would be unable to register it normally.
My expert knowledge of stroke symptoms (thank you wikipedia : Cerebral achromatopsia - Wikipedia) suggests that colours really are handled in a single location. Knock out those neurons and everything goes grey, at least according to the cases quoted below. It even caused the loss of colour for memories and mental images - so the sensation of redness seems to have completely vanished :
“Mr. I. could hardly bear the changed appearances of people (“like animated gray statues”) any more than he could bear his own changed appearance in the mirror: he shunned social intercourse and found sexual intercourse impossible. He saw people’s flesh, his wife’s flesh, his own flesh, as an abhorrent gray; “flesh-colored” now appeared “rat-colored” to him. This was so even when he closed his eyes, for his preternaturally vivid (“eidetic”) visual imagery was preserved but now without color, and forced on him images, forced him to “see” but see internally with the wrongness of his achromatopsia. He found foods disgusting in their grayish, dead appearance and had to close his eyes to eat. But this did not help very much, for the mental image of a tomato was as black as its appearance.”
I think that’s really interesting, especially the comment about memories losing their colour. It seems as though in this case the qualia can be localised to specific neurons.
From wiki: “Dennett maintains that the difficulties of any such strong AI project would be orders of magnitude greater than those raising concerns have realized.”
You are a believer in some things, it is hard to know what the beliefs (i.e. assumptions) are. Even those we claim may in fact be based on assumptions we are unaware of. This is a good reason to doubt yourself when you make any definitive positive claim.
Right, so the experience is not in the neurons and there is something more complicated going on. Basically the conscious experience is non-local if the neurons in the dish that are beaving in the same way are no longer “red” and the person without those neurons can no longer experience “red” then the redness is something that is more than just the neurons and also dependent on the neurons (but not only those neurons).
I agree with you that the TBT is not compelling yet but maybe Numenta will make it more compelling by showing a working system.
@alastair i agree that article on Mr. I. is cool and thought provoking. It makes one think whether each person’s qualia of perceiving the color red (excluding the feelings associated with the color) is different for each person. Even though our primal instinct/reaction to the color red might be the same in general (e.g. associating it with blood and thus danger) the perception of the color red itself might be different for each person. Hard to prove if this might or might not be the case. Because you can’t objectively describe red without associating it with your experiences and maybe also our survival instincts on the color red. Can’t ask anyone to objectively describe red such that we can compare if our qualia of perceiving red is similar or not. But if I have to guess our perception of red should be similar from one person to another because vision is so important that evolution would have hard coded it in us.
I have a thought that maybe one can’t relearn red because what is the training feedback/signal to train it to relearn red? An aide can try to tell us that something is red to try and train us but training to relearn red using that kind of conscious method doesn’t sound right. It might be possible to relearn sensorimotor stuff if it is damaged because there’s constant sensorimotor feedbacks (i.e. an action causes changes in our sensors and environment paving the way for sequential learning). Even if we relearn to differentiate the color red from other colors will it be the same red as before? Not unlike artificial neural network the loss functions and training method are very important considerations.
I think the whole question of quaila is mostly silly. As long as the sensory apparatus connected to some patch of cortex reliably signals “something” at 640 nm wavelength of light that we all agree is red then that is what red is. The same is true for all of the sensory modalities. For all of us, that wavelength will fall on the same place in a rainbow and will be the same Pantone chip. Trying to convey what what any sensory modality “feels like” is an empty exercise. As long as we can agree on what symbol matches up with what stimulus we have reached the end of usefulness in interpersonal communications regarding discussions about senses “feel like.”
We need to know what functionality qulia gives,…it apparently exists, and in order o exist you must cause something also apparently…I believe that one person that grew up independent of a society would not experience qualia. It needs for the brain to form a model of itself, through beliefs of what happens in other minds. These beliefs are not sentience itself but the recursive model built out of considering evidence induces it and hence experience. Experience is therefore nnot removed from cause and effect but as a result of cause and effect explained, behavior explained. The longer a human is sepseparsted from society the less he is able to qualitatively feel.so I feel because you apparently feel.why couldn’t I simply apparently feel because you apparently feel,because there would be no reason to, I.e. no connection between the two.the evidence might as well cause something else. When brains evolve in a society having qualia is related to having empathy