Yesterday I wrote about IIT: “In my view Integrated Information Theory (IIT) is the best theory of consciousness available, because it recognizes that a theory of consciousness can only be the formalization (ideally by mathematical axiomatization) of our previous intuitions. The testing of any theory of consciousness can only be done on a very limited “circle of epistemic trust”: the set of beings so similar to us that we can accept their consciousness as obvious”
In “Sizing up consciousness”, there is a very interesting application on how only integrated information flows produce consciousness (including the interesting case of dreams).
These are strange axiomatic ways of thinking about info and meaning. I like the mathematical breakout moments but I wonder about their significance. They feel to this humanist like an effort to bring scientific gravity to highly debatable theoretical starting points. It is interesting that I never came across these theories in all my intellectual journeys. I am glad I now know about them and want to know more. IIT feels like a pastiche of other philosophies. A good deal of structuralism with its emphasis of differance. It seems like IIT wants it cake and eat it too. Consciousness as a discrete existent and a closed causal materialism. I want that too but that doesn’t make it so. I am also seeing the way IIT has subsequently influenced later ontologies like Harlan and Latour who both to varying degrees insist on the irreducibility of the object, not just of mind but all thing.
Nicely put! I get the sense there's some 'sleight of hand' going on with IIT.
They call their theory axiomatic, but to me, these five basic assumptions seem to be motivated by what they believe to be the basic properties of consciousness. I'm not sure I buy that these five assumptions are the basic properties of consciousness, or even things I should assume to be true.
Even before the classical Greek philosophers Heraclitus wrote that no one can step in the same river twice. Both the one and the river change second by second. You mentioned phenomenology. This method of epistemic activity posits that experience is specific, yes, but how humans experience experiences has universal aspects. Boredom, for example, has a shape and a contour. You recognized it when you used the yellow wavy lines and warned your readers to walk away. You felt unsure of the wisdom of including this information. Was this unsureness about the information information itself? I’m not clear about how information causes physical change. Seems more likely that motives or instincts or reflexes effect change through human instrumentation actlng partly on information but a myriad of other constructs. I recommend going in YouTube and find John Searle’s class on philosophy of mind from maybe 2015 at Berkeley. 25 lectures made of golden information.
Thanks for the comment and recommendation, Terry. You raise some interesting points.
I think someone who advocates for IIT might argue that these 'myriad of other constructs' are information, too. They would argue that information causes physical changes because it is a physical thing—being a physical thing, it can have causal effects on the world.
Just to clarify, I'm not advocating for IIT. In fact, I'm not particularly fond of this theory of consciousness. In this article, I'm simply providing an explanation of how IIT defines information.
Those myriad of other factors include emotions, belief systems, expectations for the future, etc. IIT can hang the moniker “information” on them, but convincing others is another issue. Is depression “information?” Try this out, if information causes changes in the physical world, is information the cause of cancer? In my view, the core logical fallacy is equivocal use of the concept “cause. Information is in the chain of contingencies. But to say information acts on the physical world seems absurd. On this view information itself is physical. Btw, I don’t mean to appear to be arguing against your personal view. As you say, I don’t know what your view is. You are speaking on behalf of someone else. I am interested in texts like this in discerning the author’s “real” position, but I couldn’t find any consistent perspective on the theory.
You would really enjoy listening to Searle and his lectures. His favorite skewer is Dennett (I may have his spelling wrong). I think D passed away not too long ago.
Thanks Terry! I have listened to Searle's lecture series on the philosophy of mind -- it was many years ago. I should have another listen. I remember his philosophy of language series as well.
Yes, Daniel Dennett passed away a few weeks ago. Dennett and Searle definitely disagreed with each other! 🤣.
Although my fascination with consciousness started with philosophy, my research on consciousness is as a scientist, not a philosopher. I think for that reason, I tend to side with the more physicalist views, like those of Dennett, than I do with Searle. But I'm not trying to push my views here -- in fact, I'm trying not to let my views get in the way. I'm more interested in being curious about others ideas, thinking critically about those ideas, and working towards trying to answer some of the key questions we need to answer if we are going to have any hope in understanding consciousness.
Searle’s views on language are seminal in my field, language and literacy, which also embraces science, though one might not think so given the reading wars. That was my motive for listening to his mind lectures. Besides the language series, he has a series on social constructs in consciousness that we accept as physical objects (a red light has a few meanings in addition to the quality of “redness,” one of which is illegal to ignore, one illegal to respond to in most states). I think the bottom line for me is the temptation to diminish the role of motivation, intention, and experience in human behavior. In literacy (cf: Fairclough on discourse analysis) much of the conundrum hinges on how individual consciousness (subjective) becomes shared consciousness (intersubjective) by way of language as a cultural construct. It is very difficult for me to conceive of consciousness as a physical entity. Searle helped me understand how one can sustain a logical conclusion that consciousness is not an illusion and simultaneously avoid supernatural explanations, which I also can’t square with my own sense of reality. I just don’t understand how AI can qualify as conscious in the human sense without reducing important concepts like “information” in absurd ways. When I say absurd, I mean it in a sense Searle would accept.
Thanks Terry! I'm also not a fan of the way information is defined in IIT, and I'm suspicious of their claims that consciousness is an irreducible phenomenon -- this idea leads to panpsychism, a view I find difficult to swallow.
You've convinced me to give Searle's series another listen. I must admit, I first watched this series when I may have been a bit more opinionated about my views 🤣. My memory is that Searle spends a lot of time claiming things about Dennett's theory that are just not true. I should rewatch it, if only to take it in with a less biased view this time around.
He uses Dennett’s name as a whipping post for all of the theorists who posit that consciousness is the sum of physical activity and is therefore physical. He does have an ax to grind with Dennett. I could see where Dennett fans would have a strong negative reaction. But he does the same with Descartes and the ghost. At the time I knew Descartes but nothing about Dennett. I think this was eight or nine years ago. Searle is, well, surly. I’m not sure he would be much help to you in generating a mathematical understanding of consciousness. I think your point about IIT axioms is well taken. Seems like there is a catch 22 at work. We first need to understand consciousness, describe it, and then look at AI. We can’t assume consciousness is thus and so and then look at AI, but this is the situation we are in. Keep in mind, my concern is with teaching and learning. If both AI and consciousness are reduced to physical activity then motivation as emotive isn’t so important. I see motivation as the ball game. Is there any mathematical explanation that accounts for consciousness as a function of emotion? I see such huge potential for AI as a ladder for minority and high poverty kids. Tapping into joy and intrinsic satisfaction with AI could be a game changer.
I'm not sure why if consciousness is reduced to physical activity then motivation as emotive isn't so important. Wouldn't motivation and emotion also be explained by physical processes? I feel like I might be misunderstanding your point here.
As far as a mathematical explanation that accounts for consciousness as a function of emotion, I'd say predictive coding probably fits. It at least aligns with that sort of view. I discuss a little about predictive coding in this week's article on animal dreams. Predictive coding is not a theory of consciousness, per se, but many treat it as so.
Searle would argue that chemistry and neural networks are necessary but in sufficient for explaining motivation. The urge to exert effort to read and write derive from human experience in a sociocultural surround ranging from extrinsic networks of symbols (gold stars, grades) to instrinsic constructs created from participating in cultures large and small. If motivation is the result of physical processes, humans would be less human. Searle relies on analogically proofs rather than empirical arguments in matters where the physicalists are absurd in that their position just doesn’t square with what human experience tells us. Followed to its logical extension the argument that motives and emotives have purely physical roots in chemistry and electricity ignores experience (can’t explain what we see everyday). He gives lots of examples.
I see, you're not arguing that motivation and emotion are completely independent of physical processes. Just that physical processes are not sufficient for them. Necessary but not sufficient. There are neural substrates for motivation and emotions. We can explain these processes with physical processes in the brain. What we might disagree about is whether these physical processes are sufficient for the 'feeling of what it is like' to have a motivation or emotion. I believe as a biological naturalist, Searle would agree with this. Would you agree?
Well written. Thank you. It would be interesting to read more about the IIT theory of time, space, colors... For me it is the best theory in this known universe. I wish you a nice day.
Well, yes, I decided on such a claim from Tononi's discussion at Fondazione Prada (DISCUSSION #5 Consciousness: The Still Open Questions - Two Different Perspectives), when he spoke about "Consciousness and our place in nature" https://www. youtube.com/watch?v=joqY2T4GWHg.
Otherwise... According to IIT theory, consciousness is experience... so I could say even more precisely that studying IIT is the best experience for me as far as theories of experience are concerned.
I am already looking forward to your next writing. I wish you good experiences during the day.
Good post. I hadn't realized the role of causality in IIT. I can't say it improves my view of the "theory" (which is dim enough that I have a hard time calling it a theory). As your other commenters have touched on, there is something hand-wavey and "just so" about it. I'm not impressed by the math, which seems to just be Shannon's entropy equation with some hand-wavey -- and I would guess very hard to quantify -- probabilities.
At best, IIT seems more of a measure or description than an explanation.
I was expecting you'd get into the network aspect and phi. Future post? It's the part of IIT that, I think, might touch on something necessary, a very large, complex network. I think there's more to it than just a big network, but I do think a complex network is required.
The talk about the property of specificity made me think of an old quote that touches on the same thing: You can't step in the same river twice!
Yes, I am planning on writing much more on IIT. Because it is such a controversial idea, I want to make sure to give it the time it needs to steelman it. I want to present the strongest possible version of the theory before addressing its criticisms head-on.
But, yes, I agree, thinking that an integrated network is required does not mean we need to accept all the assumptions and axioms of IIT.
Yesterday I wrote about IIT: “In my view Integrated Information Theory (IIT) is the best theory of consciousness available, because it recognizes that a theory of consciousness can only be the formalization (ideally by mathematical axiomatization) of our previous intuitions. The testing of any theory of consciousness can only be done on a very limited “circle of epistemic trust”: the set of beings so similar to us that we can accept their consciousness as obvious”
https://forum.effectivealtruism.org/posts/FjiND3qJCvC6CtmxG/super-additivity-of-consciousness
In “Sizing up consciousness”, there is a very interesting application on how only integrated information flows produce consciousness (including the interesting case of dreams).
https://www.amazon.es/Sizing-Consciousness-objective-capacity-experience/dp/0198728441
My brain hurts now. I hope you're happy.
Stop making me think so much.
But wait.
Did my thinking, in combination with my unique brain state while reading this article, result in a meaningful difference to me as a system?
I believe it might have!
Haha! Well, it seems so :)
These are strange axiomatic ways of thinking about info and meaning. I like the mathematical breakout moments but I wonder about their significance. They feel to this humanist like an effort to bring scientific gravity to highly debatable theoretical starting points. It is interesting that I never came across these theories in all my intellectual journeys. I am glad I now know about them and want to know more. IIT feels like a pastiche of other philosophies. A good deal of structuralism with its emphasis of differance. It seems like IIT wants it cake and eat it too. Consciousness as a discrete existent and a closed causal materialism. I want that too but that doesn’t make it so. I am also seeing the way IIT has subsequently influenced later ontologies like Harlan and Latour who both to varying degrees insist on the irreducibility of the object, not just of mind but all thing.
Nicely put! I get the sense there's some 'sleight of hand' going on with IIT.
They call their theory axiomatic, but to me, these five basic assumptions seem to be motivated by what they believe to be the basic properties of consciousness. I'm not sure I buy that these five assumptions are the basic properties of consciousness, or even things I should assume to be true.
Even before the classical Greek philosophers Heraclitus wrote that no one can step in the same river twice. Both the one and the river change second by second. You mentioned phenomenology. This method of epistemic activity posits that experience is specific, yes, but how humans experience experiences has universal aspects. Boredom, for example, has a shape and a contour. You recognized it when you used the yellow wavy lines and warned your readers to walk away. You felt unsure of the wisdom of including this information. Was this unsureness about the information information itself? I’m not clear about how information causes physical change. Seems more likely that motives or instincts or reflexes effect change through human instrumentation actlng partly on information but a myriad of other constructs. I recommend going in YouTube and find John Searle’s class on philosophy of mind from maybe 2015 at Berkeley. 25 lectures made of golden information.
Thanks for the comment and recommendation, Terry. You raise some interesting points.
I think someone who advocates for IIT might argue that these 'myriad of other constructs' are information, too. They would argue that information causes physical changes because it is a physical thing—being a physical thing, it can have causal effects on the world.
Just to clarify, I'm not advocating for IIT. In fact, I'm not particularly fond of this theory of consciousness. In this article, I'm simply providing an explanation of how IIT defines information.
Those myriad of other factors include emotions, belief systems, expectations for the future, etc. IIT can hang the moniker “information” on them, but convincing others is another issue. Is depression “information?” Try this out, if information causes changes in the physical world, is information the cause of cancer? In my view, the core logical fallacy is equivocal use of the concept “cause. Information is in the chain of contingencies. But to say information acts on the physical world seems absurd. On this view information itself is physical. Btw, I don’t mean to appear to be arguing against your personal view. As you say, I don’t know what your view is. You are speaking on behalf of someone else. I am interested in texts like this in discerning the author’s “real” position, but I couldn’t find any consistent perspective on the theory.
You would really enjoy listening to Searle and his lectures. His favorite skewer is Dennett (I may have his spelling wrong). I think D passed away not too long ago.
Thanks Terry! I have listened to Searle's lecture series on the philosophy of mind -- it was many years ago. I should have another listen. I remember his philosophy of language series as well.
Yes, Daniel Dennett passed away a few weeks ago. Dennett and Searle definitely disagreed with each other! 🤣.
Although my fascination with consciousness started with philosophy, my research on consciousness is as a scientist, not a philosopher. I think for that reason, I tend to side with the more physicalist views, like those of Dennett, than I do with Searle. But I'm not trying to push my views here -- in fact, I'm trying not to let my views get in the way. I'm more interested in being curious about others ideas, thinking critically about those ideas, and working towards trying to answer some of the key questions we need to answer if we are going to have any hope in understanding consciousness.
Searle’s views on language are seminal in my field, language and literacy, which also embraces science, though one might not think so given the reading wars. That was my motive for listening to his mind lectures. Besides the language series, he has a series on social constructs in consciousness that we accept as physical objects (a red light has a few meanings in addition to the quality of “redness,” one of which is illegal to ignore, one illegal to respond to in most states). I think the bottom line for me is the temptation to diminish the role of motivation, intention, and experience in human behavior. In literacy (cf: Fairclough on discourse analysis) much of the conundrum hinges on how individual consciousness (subjective) becomes shared consciousness (intersubjective) by way of language as a cultural construct. It is very difficult for me to conceive of consciousness as a physical entity. Searle helped me understand how one can sustain a logical conclusion that consciousness is not an illusion and simultaneously avoid supernatural explanations, which I also can’t square with my own sense of reality. I just don’t understand how AI can qualify as conscious in the human sense without reducing important concepts like “information” in absurd ways. When I say absurd, I mean it in a sense Searle would accept.
Thanks Terry! I'm also not a fan of the way information is defined in IIT, and I'm suspicious of their claims that consciousness is an irreducible phenomenon -- this idea leads to panpsychism, a view I find difficult to swallow.
You've convinced me to give Searle's series another listen. I must admit, I first watched this series when I may have been a bit more opinionated about my views 🤣. My memory is that Searle spends a lot of time claiming things about Dennett's theory that are just not true. I should rewatch it, if only to take it in with a less biased view this time around.
He uses Dennett’s name as a whipping post for all of the theorists who posit that consciousness is the sum of physical activity and is therefore physical. He does have an ax to grind with Dennett. I could see where Dennett fans would have a strong negative reaction. But he does the same with Descartes and the ghost. At the time I knew Descartes but nothing about Dennett. I think this was eight or nine years ago. Searle is, well, surly. I’m not sure he would be much help to you in generating a mathematical understanding of consciousness. I think your point about IIT axioms is well taken. Seems like there is a catch 22 at work. We first need to understand consciousness, describe it, and then look at AI. We can’t assume consciousness is thus and so and then look at AI, but this is the situation we are in. Keep in mind, my concern is with teaching and learning. If both AI and consciousness are reduced to physical activity then motivation as emotive isn’t so important. I see motivation as the ball game. Is there any mathematical explanation that accounts for consciousness as a function of emotion? I see such huge potential for AI as a ladder for minority and high poverty kids. Tapping into joy and intrinsic satisfaction with AI could be a game changer.
Ah! I see. Thanks for the heads up!
I'm not sure why if consciousness is reduced to physical activity then motivation as emotive isn't so important. Wouldn't motivation and emotion also be explained by physical processes? I feel like I might be misunderstanding your point here.
As far as a mathematical explanation that accounts for consciousness as a function of emotion, I'd say predictive coding probably fits. It at least aligns with that sort of view. I discuss a little about predictive coding in this week's article on animal dreams. Predictive coding is not a theory of consciousness, per se, but many treat it as so.
Searle would argue that chemistry and neural networks are necessary but in sufficient for explaining motivation. The urge to exert effort to read and write derive from human experience in a sociocultural surround ranging from extrinsic networks of symbols (gold stars, grades) to instrinsic constructs created from participating in cultures large and small. If motivation is the result of physical processes, humans would be less human. Searle relies on analogically proofs rather than empirical arguments in matters where the physicalists are absurd in that their position just doesn’t square with what human experience tells us. Followed to its logical extension the argument that motives and emotives have purely physical roots in chemistry and electricity ignores experience (can’t explain what we see everyday). He gives lots of examples.
I see, you're not arguing that motivation and emotion are completely independent of physical processes. Just that physical processes are not sufficient for them. Necessary but not sufficient. There are neural substrates for motivation and emotions. We can explain these processes with physical processes in the brain. What we might disagree about is whether these physical processes are sufficient for the 'feeling of what it is like' to have a motivation or emotion. I believe as a biological naturalist, Searle would agree with this. Would you agree?
Great conversation here. Number me among those who think Searle makes a lot of sense.
Well written. Thank you. It would be interesting to read more about the IIT theory of time, space, colors... For me it is the best theory in this known universe. I wish you a nice day.
Thank you! I have plans to write more about IIT in the future. The best theory in the know universe, huh!? That's quite the claim 😉
Thank you.
Well, yes, I decided on such a claim from Tononi's discussion at Fondazione Prada (DISCUSSION #5 Consciousness: The Still Open Questions - Two Different Perspectives), when he spoke about "Consciousness and our place in nature" https://www. youtube.com/watch?v=joqY2T4GWHg.
Otherwise... According to IIT theory, consciousness is experience... so I could say even more precisely that studying IIT is the best experience for me as far as theories of experience are concerned.
I am already looking forward to your next writing. I wish you good experiences during the day.
Good post. I hadn't realized the role of causality in IIT. I can't say it improves my view of the "theory" (which is dim enough that I have a hard time calling it a theory). As your other commenters have touched on, there is something hand-wavey and "just so" about it. I'm not impressed by the math, which seems to just be Shannon's entropy equation with some hand-wavey -- and I would guess very hard to quantify -- probabilities.
At best, IIT seems more of a measure or description than an explanation.
I was expecting you'd get into the network aspect and phi. Future post? It's the part of IIT that, I think, might touch on something necessary, a very large, complex network. I think there's more to it than just a big network, but I do think a complex network is required.
The talk about the property of specificity made me think of an old quote that touches on the same thing: You can't step in the same river twice!
I agree with all of this!
Yes, I am planning on writing much more on IIT. Because it is such a controversial idea, I want to make sure to give it the time it needs to steelman it. I want to present the strongest possible version of the theory before addressing its criticisms head-on.
But, yes, I agree, thinking that an integrated network is required does not mean we need to accept all the assumptions and axioms of IIT.