Cool. Thanks.

I only know what you've written here about IIT. I'm not at all sure what it is a theory of.

Is it to be implemented on a digital computer? You mentioned logic gates.

That would be a non-starter, as far as modelling the brain is concerned, for the reasons outlined above, and more. For example there are wave-like phenomena involving large groups of neurons, also what I think is called back-propagation, with impulses travelling both upstream and down.

But there's no consciousness associated with your liver,

I still maintain that the relevant system is the entire body. The brain is not modular like a man-made machine (see above) and neither is the body. The brain relies on a supply of blood, and the liver plays a major role in providing that.

Is it to be implemented on a digital computer? You mentioned logic gates. — Daemon

The idea is to be able to make predictions about whether a system is conscious. I'm not sure how they would test it, though.

The brain is not modular like a man-made machine (see above) and neither is the body. The brain relies on a supply of blood, and the liver plays a major role in providing that. — Daemon

I guess the assumption behind excluding most of the body is that it powers the system which produces consciousness without participating in consciousness.

I think the whole thing is scientifically naive. I was thinking about parts of the body that don't participate in consciousness, I thought of hair. But try stroking your hair.

Consciousness is embodied.

IIT, based on the scholarpedia page.
In formulating the axioms, Tononi uses these criteria:
1. About experience itself; — frank

Scott Aaronson debunkificated this — fishfry

For assertion 1. the philosophical question is what is x if anything at all. Since experience is private there is no way to answer that except for claiming that experience-in-itself exists as a Platonic concept and as a corresponding linguistic proxy.
In astronomy there are the analogously fuzzy notions of dark matter and dark energy which are postulated to reify their effects on galaxy clusters and on theoretical universal expansion. Neither can be directly seen and identified as objects but physicists can justify supposing that they necessarily exist.

Totoni's phi would be a basis for an objective measure of something-or-another that he labels as experience/consciousness. It is not a measure of my mentality before my first cup of coffee but what it might do is to define totoni-ness, an entirely different thing with hopefully some connection to what is generally thought of by others. Whether that is meaningful or just useful would depend on physical implementation of measuring and classifying phi's for various living and inanimate subjects. If the quantification of a cat's phi lies somewhat between Totoni's and a sunflower's then he will have achieved some success.

think the whole thing is scientifically naive. I was thinking about parts of the body that don't participate in consciousness, I thought of hair. But try stroking your hair.

Consciousness is embodied. — Daemon

It's not obvious to me that consciousness requires a system in possession of a liver. That it needs a power source, yes. Since there's a certain amount of closed causation, intuition says there's some work involved.

Does it need a liver to filter and provide digestive enzymes? I don't know.

Downstream we may realize it does, but I think we can start with the assumption that we don't need it. Since we can change out your liver without altering your consciousness, maybe you don't.

Consciousness is embodied. — Daemon

What does this tell us?

Neither can be directly seen and identified as objects but physicists can justify supposing that they necessarily exist. — magritte

Dark energy is at the root of the present crisis in cosmology. The crisis (having to do with conflicting measurements of the universe's expansion) promises to increase our understanding of it.

Yay for linguistic proxies.

Scott Aaronson debunkificated this a while back.

Aaronson demonstrated that the IIT significantly doesn't match some of our intuitions about consciousness, so in my opinion IIT isn't correct, but IIT still matches very well some other of those intuitions. It may be used for example to recognize that a group of humans doesn't constitute a conscious entity greater than a single human.

In my opinion IIT is a great attempt at solving consciousness, surprisingly serious approach to solve a non-trivial philosophical problem. I'd really like to see it tweaked to make better cassifiers of conscious entities.

I think defining consciousness using only the flow of information is lacking. For starters I'd include that conscious entity needs to recognize patterns in this information.

Consciousness may not require a liver, let's say a lobster is conscious and doesn't have a liver, but the lobster or human does need to have the equipment to remain alive and...conscious. If the human hadn't had a liver to start with, it wouldn't have become conscious.

I do think the idea of the brain and body not being modular is an important one. They haven't been designed in the way we design machines. I don't know how much bearing this has on the maths side of it, that is completely beyond me. But if they are really equating neurons to logic gates, the maths is just meaningless I think. The actual mechanisms are so much more messy, plastic, multifaceted than binary logic gates.

Brains and bodies don't work by processing information. The brain works through things like neurotransmitters, neuromodulators, electrical impulses, wave-like phenomena. Calling all this "information processing" doesn't tell us anything more about what is happening.

"Patterns" is another troublesome concept @original2. Can we think about a relatively simple biological process to see why. Bacteria can swim towards a desirable stimulus (let's say some sugar) or away from a toxic chemical. It seems they would need to recognise a kind of pattern in the increasing or decreasing concentration of the chemical. But in fact we know every detail of the chemical process that achieves the directional swimming, and there's nothing left for "pattern recognition" to do.

The same is true of the more complex processes in the brain. It works through things like electrical impulses and so on, not pattern recognition.

Pattern recognition is something a person does, not their brain.

What does the embodied mind tell us @frank? I suppose it tells us that Tononi isn't seeing the whole picture?

Consciousness may not require a liver, let's say a lobster is conscious and doesn't have a liver, but the lobster or human does need to have the equipment to remain alive and...conscious. If the human hadn't had a liver to start with, it wouldn't have become conscious. — Daemon

Sure. You need some way to keep the brain alive. We can take over the functions of the heart, lungs, and kidneys with machinery. Hospitals do it everyday. The patient can be wide awake while being supported in this way. So whether the body is modular or not, whether a human needs a liver, I think that's a side issue.

The actual mechanisms are so much more messy, plastic, multifaceted than binary logic gates. — Daemon

But it's not like some sort of mystical fuzz. Is it?

Bacteria can swim towards a desirable stimulus (let's say some sugar) or away from a toxic chemical. It seems they would need to recognise a kind of pattern in the increasing or decreasing concentration of the chemical. — Daemon

I don't see why it would need to recognize a pattern.

We can take over the functions of the heart, lungs, and kidneys with machinery. Hospitals do it everyday. The patient can be wide awake while being supported in this way. So whether the body is modular or not, whether a human needs a liver, I think that's a side issue. — frank

There's a point here which I haven't yet properly expressed or thought through (which makes it interesting, to me anyway).

When the hospital takes over vital functions they are taking over something that's already in operation, that already has to be in operation for the person to be in a position to be conscious at all. We can't make the whole thing from scratch, using machinery.

And the brain can't be isolated from the rest of the body, it's enmeshed with the rest of the body. There's no sense in thinking about it operating in isolation, it would have nothing to operate with.

Or to look at it from another angle, feeling is primary.

But it's not like some sort of mystical fuzz. Is it? — frank

Why should it be? Why introduce the idea of mystical fuzz? Seems to me this is to accept the categorisations of cartesian dualism.

I don't see why it would need to recognize a pattern. — frank

And I don't see why the brain would need to recognise a pattern in information as @original 12 has suggested. A person can recognise a pattern, a brain can't (it's the homunculus fallacy).

When the hospital takes over vital functions they are taking over something that's already in operation, that already has to be in operation for the person to be in a position to be conscious at all. We can't make the whole thing from scratch, using machinery. — Daemon

I understand what you're saying. The body comes as a package.

If we want to know which parts drives blood pressure, we can pick out the heart and kidneys. We know your foot isn't part of it.

But for you consciousness is different from that kind of function? We can't identify a body part that produces it?

And the brain can't be isolated from the rest of the body, it's enmeshed with the rest of the body. — Daemon

The brain actually is isolated by the blood brain barrier. The CNS has its own private immune system. At 2 in the morning while studying A&P, it might occur to a student that the CNS looks like an alien that invaded a tetrapod.

But I digress. :cool:

But for you consciousness is different from that kind of function? We can't identify a body part that produces it? — frank

We can identify it in an abstract sense, but not in a practical sense, as we can with a manmade machine.

We have "brainoids" now, grown from adult human skin cells. But unless they are connected to sense organs, and yes, things like feet, they can't do what real brains do. There isn't anything for them to be conscious of.

I see what you're saying. The IIT approach would be like starting with the assumptions that we digest food and that some parts of the body are causing that.

We would then start with specifying the parameters of digestion:. food goes in, food breaks down, the body keeps the good parts and throws away what it doesn't want.

Then we would hypothesize for testing.

It's ok that digestion is inextricably linked to other body functions. An "abstract" division, as you say, is enough for our purposes.

Is that satisfactory? Or is there some reason consciousness should be looked at radically differently to digestion?

An "abstract" division, as you say, is enough for our purposes. — frank

I still owe most of my knowledge of IIT to you, but from what I understand, the purpose is to quantify what is required to achieve consciousness. But it seems they are abstracting an arbitrary aspect of the biological machinery, and quantifying that. They aren't taking account of all the brain stuff that isn't just neurons firing (the neuromodulators and so on), they are pretending that neurons are logic gates (which they aren't), and they aren't taking account of the essential involvement of the body beyond the brain.

Do you know of responses to such criticisms? Should I read the Scholarpedia page, or are you planning to post more?

I think defining consciousness using only the flow of information is lacking. For starters I'd include that conscious entity needs to recognize patterns in this information. — original2

A conscious entity would need to interpret the information flow. But what does interpret mean? In the broadest sense even a rock interprets the information flow in its form and position.

According to Fritjof Capra: "cognition is a reaction to a disturbance in a state." And it would seem everything is a system in a state.

It is not a measure of my mentality before my first cup of coffee — magritte

My understanding is that Tononi's phi is intended to be exactly that. But I only linked to the Aaronson article and haven't paid much attention to IIT, and can't comment authoritatively.

The brain actually is isolated by the blood brain barrier. The CNS has its own private immune system. — frank

It's not entirely isolated though. The blood brain barrier is a filter isn't it, not a seal.

It's not entirely isolated though. The blood brain barrier is a filter isn't it, not a seal. — Daemon

Yes, the bloodbrain barrier is a filter. I realize the CNS is not a parasite. It just kind of looks like one from a certain angle.

I will be moving on regarding IIT, just need a minute to sit down. :)

I will be moving on regarding IIT, just need a minute to sit down. :) — frank

Can I just say how much I'm enjoying the discussion @Frank, I really appreciate you posting the summaries and will wait patiently until more arrives.

IIT, originated by Giulio Tonini, is an attempt to specify the system requirements for consciousness. — frank

I appreciate that Tononi began with abstract mathematical Information as fundamental, and derived human-like Consciousness as an emergent phenomenon. That bypasses the distractions of worrying about the feelings of subatomic particles. :smile:

Can I just say how much I'm enjoying the discussion Frank, I really appreciate you posting the summaries and will wait patiently until more arrives. — Daemon

That's so cool! Thanks!

So we talked about the postulate that covers the intrinsic character of experience: the spec being that the system must have a certain amount of closed causation (a cause-effect space).

For composition, the postulate is that the system has to be structured:

"Composition
The system must be structured: subsets of the elements constituting the system, composed in various combinations, also have cause-effect power within the system. Thus, if a system ABC is constituted of elements A, B, and C, any subset of elements (its power set), including A, B, C; AB, AC, BC; as well as the entire system, ABC, can compose a mechanism having cause-effect power. Composition allows for elementary (first-order) elements to form distinct higher-order mechanisms, and for multiple mechanisms to form a structure." --Tononi in the Scholarpedia article.

IOW, the intuition is that if the system is structureless like water, it can't produce a structured experience.

I appreciate that Tononi began with abstract mathematical Information as fundamental, and derived human-like Consciousness as an emergent phenomenon. That bypasses the distractions of worrying about the feelings of subatomic particles. — Gnomon

I think its kind of the opposite. He starts with phenomenal consciousness and is trying to derive a system that would produce it.

A conscious entity would need to interpret the information flow.

Yes, but not any interpretation suffices for consciousness in my opinion. Any information transformation could be arguably seen as interpretation and trivial information transformations are what for example inanimate objects do with information thrown at them.

My gut tells me that if an entity is able to match information to patterns, it is a mark of consciousness, though not necessarily a big one. By pattern i mean some generalized, meta-information that describes information succintly.

An animal being able to differentiate between predator and prey for example would have baked in (perhaps learned, but not necessarily) information patterns that allows it to do that.

An entity being capable of finding new patterns in the information would be in a certain way better that one that isn't capable to do that, but I'd argue that it would be useful to call them both conscious. It would be useful to have terms for both of those categories though.

P.S. Is there an automatic way to quote other posts in the style most people do that here? It eludes my perception.

P.S. Is there an automatic way to quote other posts in the style most people do that here? It eludes my perception. — original2

If you highlight the text a quote option appears.

My gut tells me that if an entity is able to match information to patterns, it is a mark of consciousness, though not necessarily a big one. By pattern i mean some generalized, meta-information that describes information succintly. — original2

That is pretty close. I would say if information can be integrated and symbolized, and physical form is a symbol, imo. It has to start somewhere, and this way it starts at the beginning.

We should let Frank finish his excellent summary and perhaps discuss later. Anyhow, welcome to the forum.

Composition allows for elementary (first-order) elements to form distinct higher-order mechanisms, and for multiple mechanisms to form a structure. — frank

Does the theory ever address the question of what constitutes a "distinct" mechanism (without a human being making the call)? Without that, the theory doesn't get off the ground, or we have panpsychism, which doesn't explain anything.

Does the theory ever address the question of what constitutes a "distinct" mechanism (without a human being making the call)? Without that, the theory doesn't get off the ground, or we have panpsychism, which doesn't explain anything. — Daemon

Their approach has been to try to determine which parts of the brain are actually involved in consciousness. It's not the whole brain.

Remember, they're building a hypothesis which could be tested.

We can count your disapproval of their system boundaries as a potential flaw, but it's ok to consider their hypothesis as it is.

I've read somewhere that they accept that a thermostat is conscious. A thermostat but not the whole brain? And the whole body is involved in consciousness!

What's the hypothesis and how would it be tested?

Why is it ok to consider their hypothesis as it is, when it seems to be fatally flawed from the outset?