Computationalism and consciousness with Nic Rouleau, Jaan Aru, and Borjan Milinkovic.

[00:00:00]say a very brief intro, Pokey, and then you can take a moment to think what you want to add.

[00:00:05]So, the key question that has bothered us and that brought me and Pokey together is the following that on the one hand we have people who are sure that AI is conscious.

[00:00:20]And I personally I don't think there's too much scientific evidence for that.

[00:00:24]On the other hand, we have people who say, "No, only biology can be conscious."

[00:00:29]Um and we think and perhaps you guys also think that, you know, the true answer is somewhere in the middle that, you know, biology as it is today, it's way more complex than anything that large language models can instantiate, but it's still some types of computation that we can and should scientifically understand. So, our paper and the presentation was was kind of inspired by this idea that, you know, we should go forward, try to pinpoint the specific computations that underlie consciousness, and unfortunately not too many people are actually interested in that. People are screaming here, "AI is conscious." Or screaming, "They are never conscious." But, you know, there is this important scientific question that we want to tackle, and I guess you guys are also interested in.

[00:01:24]>> Yeah, I I think generally there there isn't much more to add there. Um I I I do agree with Jan that this was our main call where there was not any focus on particularly some middle ground between the two and uh a broader definition of why particular artificial systems right now can or cannot um have some level of sentience or consciousness.

[00:01:57]And on the other hand with um biology and neurobiology there was not enough uh uh consideration or argument as to why biological systems might have or be like the unique vessels of of some conscious or sentient experience.

[00:02:18]Um another thing that I also believe was I'm dissatisfied with our own paper about is distinguishing between intelligence and sentience or consciousness which I would have liked to have been able to go into slightly deeper but um of course word limits uh are a problem.

[00:02:42]>> a next paper, Bobby.

[00:02:43]>> Yeah, there'll be a next paper.

[00:02:45]>> You're dealing with it next paper.

[00:02:48]>> Yeah. But also one other thing to add is for me I have a slightly like optimistic perspective on computation coming from a neuroscience perspective.

[00:03:02]A lot of neuroscientists who are maybe in the field of trying to understand artificial systems want to just um potentially claim that what the brain is doing is not computation. While for me I want to claim that maybe we need to like democratize the term computation away from just Turing computation.

[00:03:27]Turing holds this like monolithic uh you know, kind of ownership over computation or Church-Turing thesis and for me I'd rather democratize the term computation cuz I like it.

[00:03:41]And I think it can be broadened to include biophysical systems.

[00:03:47]So that's that's kind of the underlying drive for me is to develop a different formalism to computation that is constrained by what physics allows.

[00:04:02]And I'm not the first maybe person to think of this and there's a lot of things, but from our perspective, we kind of have this notion of multi-scalarity and and and truly parallel computations happening across scales, the way they might be integrated in a way and and things like that.

[00:04:18]Um so I I I And really the focus was on on [clears throat] on sentience or subjective felt experience rather than intelligence.

[00:04:29]And yeah, I I I still think that there is more to flesh out there even for us and yeah.

[00:04:38]>> Interesting. Yeah, thanks. That's a that's a great that's a great introduction. Um yeah, I have a I have a bunch of things to ask about. Um Nick, do you want to go first or or shall I?

[00:04:46]>> Sure, yeah.

[00:04:48]Um yeah, it's fascinating. I I'd love to know more about this. So, my first question would be I guess kind of foundational, which is do you do you think um first of all, do you think that consciousness is substrate dependent?

[00:05:03]And whether or not you do, um do you think that other cog- cognitive functions like learning, problem solving, decision making, these sort of easier things, do you think those are also substrate dependent or not?

[00:05:20]>> Mhm.

[00:05:21]I can I can try and answer from my perspective.

[00:05:26]I think I think consciousness might depend on being able to So, I'm trying to see how this fits into the same uh terminology, but yes to an extent, it needs to be substrate dependent, but not in the essence that it completely is and that was the the what I or what we really attempted to do in the paper is try and pick out features of substrate dependence that might need to be included and this is kind of the the kind of phrase I I I mentioned at the end. It's like the computation that physics allows or that is constrained by the physics. So, so I do think that there is some um part to play, some causal part to play um in the morphology of the system and the kind of features of of activity that it might allow.

[00:06:26]Um and I do think that these computations need to be related to the physical material system in some sense.

[00:06:38]Absolutely.

[00:06:40]Um but whether it is uh 100% substrate dependent on on biology, I would off the cuff claim no.

[00:06:50]>> As you say.

[00:06:51]>> I mean, that's the same thing for us that there are these people who think that consciousness and other things are completely abstract, comp- completely kind of substrate free and there are others who say, "You know, it has to be biology." But again, I think that there there's the middle ground, which is something that Boki also tried to articulate that some computations they are more likely to happen in certain substrates because the substrate kind of constrains which kinds of computations can happen and uh for example, perhaps in in in simply in in uh large language models, certain computations can't happen that are more likely to happen in in biology. So, we are not kind of we are not saying that it only can happen in brains and in the paper and I guess in the presentation also we say that probably there is a a way to neuromorphic computation, but we are on the same time also saying that substrate matters in the sense that perhaps you cannot get the right types of computations in any substrate.

[00:08:02]So that there are there is also a continuum there, right? I don't know what you think about it.

[00:08:07]>> Yeah. I was In in recent months there's been there have been some papers that have come out um sort of echoing things that people have been saying for almost almost a century now, which is that brains um have many properties that are like analog computers. And during your presentation and um and you during your summary I I can't stop thinking about whether um your interpretation of this this continuum um hinges on like the architecture of the computational system. Like if LLMs were implemented not by um digital computers, but rather by analog computers, or if we had substrates um that were arranged in different circuits that were um more akin to analog computing, um would would that change the calculus for you?

[00:09:03]Like is an analog computer more likely to be conscious than a digital computer?

[00:09:08]And and how does that matter to various materials? What do you think about this sort of thing?

[00:09:12]>> That's great cuz so I I wanted to answer your whether it's related to um whether substrate dependence is necessary for decision making and all of that. That I am not I'm not sure of.

[00:09:23]And then um going now to to what you um just mentioned, there's a lot of So I need to distinguish analog uh computations and analog computing to other stuff. There is a lot of work in um theoretical computer science that is quite scattered, and what analog computation means can be very, very different. Um some of the first kind of analog systems that um that yeah, Shannon um worked on are quite different to the way we think about analog computing later on, and whether we compute on real numbers necessarily does not mean this needs to be this can be a simulation as well.

[00:10:08]Doesn't need to be um uh a kind of uh a a a a physically instantiated computation.

[00:10:17]So, that is vastly different notions of analog computation that I have not yet consolidated in my mind, or there is no consilience. And then there's also like uh you're right about some of the new work um from Earl Miller that is coming out with analog computations being one of the I I guess um premises of or or or or the or the underlying features of consciousness.

[00:10:42]Um and I do like that work. I I I think it's uh really good. And for us, it's kind of um this mix again. Like so, yes, I do think that given the different substrate, I do believe different computations are possible because it is what I term as as a kind of like structural ontological primitive that is necessary for the computation. So, analog computing or some kind of field computations might be one of those, but it's not necessarily like um sufficient for consciousness. For us, we outlined three features, but since then we've also developed much more work into like embodied systems and other features that could add to this biological computationalism we speak of.

[00:11:26]Um yeah, I think that's >> uh so so so Nick, you're right. I mean, it it depends on on the substrate and and if LLMs would be differently instantiated, it would also change the calculus for us. The key that we're after is really finding these properties X, Y, Z that really are crucial for for consciousness, say. And we do think that these properties X, Y, Z are way more complex than simply saying, "Oh, it's a global workspace, for example."

[00:12:01]There is something more interesting. Uh but yeah, it it it is probably So, we don't have anything against LLMs per se. We simply it seems that, you know, this substrate is a very peculiar one where they are currently implemented. Yeah.

[00:12:16]>> I I I should I would also like to add um something else to that. It's um more kind of historical in a sense, and I've always had this assumption that Turing, for example, and Church were mathematicians, not physicists. And in a way, the way the kind of abstract notion of a Turing machine is built on recursion theory, and whether the function um whether all recursive functions are computable in some sense, it was only later that there was this notion that they found the physical instantiation in Boolean logic gates, where the von Neumann machine can be used. And there is an essence already there that the the kind of physical primitives are the ones that instantiate a Turing machine. There is this kind of like implicit circularity there. While for us, it's kind of like coming from the bottom up, the other way around, trying to not just think about this notion as mathematicians, but rather think about it from the physical basis as well. So, given that like physics has some constraints to the way computations might be able to run and the additional constraints of biology, maybe we can begin to understand the particular operations that happen in in in brains or in systems before we then abstract away from them slowly and slowly.

[00:13:42]Um yeah.

[00:13:44]>> So so So could I just confirm a couple of things? First, um you know, if if you've got two systems that have exactly the same behavior. So the input-output uh map is exactly the same, but but the inner architecture is different. So one is built on the you know, multi-scale continuous principles that you were talking about, one is the like classic kind of thing. Uh but the output is exactly the same, let's say.

[00:14:09]Uh and I guess if you if you don't want to stipulate to that that that would be interesting too, but but it seems to me that we can say that you know, it it seems likely that the output can be the same. And in that case, would you say that that their status is actually different even if the output is exactly the same?

[00:14:27]>> Yes.

[00:14:28]>> Yes.

[00:14:29]>> Okay.

[00:14:30]>> Um but I I think it depends on the on the output we're talking about. Um this is also a kind of >> precise, etc. But I think I guess Mike already kind of knows this, you know, that we could talk about the precision of the output and you know, which kinds of tasks, etc., etc. But I think in principle, you know, the first approximation is is yes, yeah.

[00:14:53]>> Okay. Yeah. Yeah, that's that's well, that's yeah, that's I mean, that's an important contradiction to the fact that I mean, basically you're saying that it is it is in fact possible to have a proper zombie that that overtly has has all the right behaviors, but because of the special causal architecture or whatever kind of architecture differences is not in fact doesn't doesn't have the right status. So so that's that's that's of course different than than the standard computationalist view, I think, and it's and it's important. So okay, cool. Um all right, and then uh couple of other other cool questions.

[00:15:26]In talking about um analog versus versus uh a digital computers.

[00:15:34]I want to I want to dig in a little bit about um the role of the observer and and and the perspective. Like, is there really such a thing as a digital computer? I mean, everything is fundamentally made of analog parts. We you know, the observer can sort of impose some some uh some models on it and say that I'm just going to I'm picking a threshold, everything below that is this, above that is that, then you know, maybe maybe we we we tune the materials to make that easier.

[00:16:01]But, what what exactly is a is a digital computer in your view? Is are there any such things and and like where does that where does that status distinction actually come from?

[00:16:10]>> Yeah.

[00:16:11]I I had a similar conversation just recently with a colleague and I I do think I kind of tend to align with you thinking what is a digital computer?

[00:16:20]And really, you're right, like logic gates kind of flip one and zeros given a particular voltage, like for I think Mac laptops, like 3.7 millivolts or something. So, it really is a threshold that's a drawn by the kind of architecture that you've built, but the computation itself happens on those singular flip values in a way. So, I'm guessing like the the digital computer in in essence exists because the computation itself is performed on the single threshold value, whether it's 3.7 or 1.4. 3.7 1.4 is like a zero if it's 1.4 millivolts. 3.1 if it's 3.7, for example. So, I think the the digital computations exist because they're the only states that the system the physical system can take.

[00:17:05]Um >> So, I'm very skeptical of that only because um you know, Josh Bongardner and I have been developing this polycomputing view. I'm I'm skeptical of of the idea that there is the computation. You know, I I I think what we have are observers and and I I and I mean I understand we've all agreed to to sort of primarily observe the system in a particular way and map what we think of as as bits you know, and I and we think this is what the algorithm is doing and so on. But but but I I I think I think a lot of that is by convention of observers and I'm not sure actually that that is the computation the system is really doing from the inside. And I think this ends up actually being important um for for all the these questions of of language models and AIs and I and all of that because we have to remember that while we as observers, you know, it's a little bit like like with a human you can have you can talk to somebody or you can have somebody else who analyzes body posture of voice pitch psychoanalyst, you know, a psychoanalyst, somebody who's analyzing physiology and and these are orthogonal channels of things going on that may have little to nothing to do with what the verbal output is. And I actually have a suspicion that we may have something similar going on with a lot of these systems where we hyperfocus on one channel because we we think we've [music] made it talk and so we're now sort of focusing on the things it says and people ask it do you are you conscious? Do you have you know, do you do you do you are you concerned when we're going to turn you off like all this stuff? But but primarily all of this is going through one particular interface which I think may or may not have much to do with what's actually going on inside and and I think it bottom comes out in this question of is there is there a sense in which there is a the computation that something is doing or is it really very observer dependent when and I I I think it's the latter.

[00:18:55]>> Yeah.

[00:18:57]I I I like that view but I'm not familiar completely with the with the polycomputing idea. Could you just like give me a brief sounds interesting.

[00:19:05]>> Sure. Yeah, the the very simple there's a there's a simple sort of philosophical claim and then there's a bunch of data both from Josh's lab and our lab about it both computational and biological. The The philosophical idea is simply this that that there is no one objective true answer to what computation a process is doing that it's observer relative. That basically there's a set of physical events something is going on. Two observers can can can look at that event and have equally valid model and in fact multiple observers could have equally valid models of what computation if any the thing is doing. So you could look at so and so for example um uh Josh and his postdoc Atusa Parsa had these nice papers where they they looked at uh in this kind of like a vibrational medium you look at you look at the exact same set of events one way and you see an AND gate you look at it a different way and you see an OR gate something like that. Right? And it's the exact same set of events and so you can't ask well what is this thing really doing right? And and and I think I think just philosophically that that makes sense because computation is a formal model that we impose on events and you can s- you can see you can you can perhaps map it differently.

[00:20:18]In biology it becomes really important because I think what's actually going on in multi-scale living living systems is that every level and no level has any kind of ground truth about what everybody else is doing and all of the components of a living system are fun- are simultaneously trying to interpret and and and also hack but but for the purposes of this interpret what everybody else is doing. So you see something going on and you think to me the most useful thing that looks like is a clock. I'm going to I'm going to treat that as a clock and something else look looks and so I'm going to key off of uh you know these these regularities in your timing. Somebody some some other subsystem looks at exactly the same thing and says oh you're making this awesome molecule that I can use and I'm going to do whatever and by the way by looking at how fast it it uh degrades I can also tell something about you and and and what you know whatever. So so there's lots of examples. We have a couple papers with him on this and there's lots of examples of how this plays out. For example, in biology, one of the coolest things about it is that what it means is that is that evolution has a choice. So, when you have a complex system uh and and you and evolution is trying to make changes to it, you have a choice. You can you can change the actual substrate in order to change the functionality in the computations or you can change how the observers interpret it. And the benefit of doing the latter is that if you have strong dependencies, right? So, so there are lots of things like around the, you know, um around the cytoskeleton and my and and energy production machinery where evolution finds it really hard to make changes because so many things depend on it. You change one thing and a ton of stuff is going to break. So, it's kind of locked in. And so, you have those limitations.

[00:21:53]But but instead, if what you do is simply add observers with different perspectives, then you can squeeze lots of new functionality out of exactly the same set of events. You don't need to break any old dependencies, leave them in place, but you're now adding new perspectives where the same thing is doing multiple multiple jobs, right? And we actually Atus and I actually did some work on this. It isn't published yet, but we did some work on this looking at how basically giving evolution the choice of putting in new of tweaking the the computational actual, you know, the the the actual computational of functionality versus adding observers.

[00:22:35]And it's and it's really cool. Evolution is very good at adding new new observers and just saying, "I'm not touching this at all. I'm going to leave it exactly in place, but I can interpret it by looking at it in different ways. I can I can get multiple utility out of it."

[00:22:46]So, so that's so that's polycomputing.

[00:22:48]So, so it's I I think for these for all these questions, it's important uh to to understand when we put labels on these systems, who whose labels they are.

[00:22:58]And and and And especially if we're going to talk about consciousness, and I guess this is my next question is to is to and I'm not too hung up on sharp definitions, but you just to get your idea of when you say consciousness, like what exactly you mean. Because there, if if it's really important to have a first-person perspective, if I since we're talking about consciousness, then I think we have to ask about observers and what does the system itself see? You know, does the system see itself as as continuous versus discrete in addition to how we how we see it?

[00:23:30]>> Cool.

[00:23:31]Should I So I I like this polycomputing idea very much, and I think it aligns with Let me know if it's if I'm incorrect here, but it aligns well with um what we were trying to define in the paper is that there is no privileged scale of computation in the first place.

[00:23:47]Also, I really like this notion of um observers, though I I I kind of would have not [snorts] called them that and generalized, but I guess it aligns exactly something that me and Yiannis are also trying to do further on from this is kind of formalize that idea um in a way where there are kind of events in the universe, physical events in some case, and there are states of the universe or some kind of constraints that can occur, and these could potentially be considered observers from the perspective that you're speaking about, and it is through the interaction of a particular state or an observer and the physical system in which the computation occurs.

[00:24:30]So the computation occurs as a relational quality rather than the state like a state transition um matrix.

[00:24:38]So I I I really >> I'll send this along. Chris Chris Fields has some really good papers on this on on on widening this or this very fundamental concept of observers from quantum theory, but but not talking about microscopic events, talking about like every you know, everything scale free.

[00:24:54]>> Yeah, I I I I I say that that sounds very much up up our alley and what we're looking to do um next. So, that would be that would be great. But, also on the on the consciousness thing, I think for me basically it's felt experience.

[00:25:08]Um and I think the key focus there is the felt.

[00:25:13]Uh there is some kind of notion of being able to at least a partial global signal that can um propagate across a system to be able to within some um time interval uh be able to integrate the quality of what it is like. So, the the the feeling of felt experience rather than subjective experience, which I kind of differentiate um from felt experience, which might be a more complex notion that we're usually used to defining from philosophy and neuroscience and psychology. I'm more thinking about um felt experience from simpler biological systems. And that to me would be a very clean definition of of consciousness. I don't know about Yian. I think he's along the same line.

[00:26:08]>> I I agree. That's why we're working together. But, I think going to Mike's question, I think we kind of we still both look or think that there is some perspective, right? Also, to the to have this experience, you there is some perspective. Um so, it's very interesting that you bring it up. And Mike, I don't know if you have thought further about it, if you if you think about consciousness from from from that perspective, then what is the observer then or or or what are the observers in in this framework of of yours, Mike?

[00:26:44]>> Well, so so I I And this is so I guess the only the interesting thing I can say here, I think um is is the following.

[00:26:50]And I already I was already bugging Nick about a couple weeks ago.

[00:26:53]Um, I was thinking about the following, you know, uh when people talk about consciousness, they almost uniformly focus on what what Boki just said, which is felt experience. And [snorts] what I think is happening there is that we're we're focusing on the input or the read side of consciousness. So, here I am. I I what what what am I what am I getting from the world? How what is it? What does it feel like to to to be in whatever perspective I have. But there's a flip side, which I think is the right side, w r i t e, the right side, which is often um I think almost uniformly neglected. And I think the way to to get at it is I draw I draw a little square like a square, a 2 by 2 square of the positions around consciousness. So, one position, right? If you're eliminative materialist, you say there is no there is no there is no actual um what's it like there is no doing of anything. It's just Both of these things are an illusion and it's just sort of you know, chemistry go and physics go on and that's it, right? So, that's one position. The other position uh the opposite position would be some sort of strong interaction interactionist position, which says, "No, I I absolutely There there really are um felt experiences and also there are uh there's the there's the other side. So, I can act in the world, right?

[00:28:13]I have I have actual will that changes what happens in the physical world, right? So, and then and then there's the epiphenomenalist position, which says, "The felt experience exists, but not the action." In other words, you can't actually change you know, it's You you you you do in fact feel things, but it's epiphenomenal. You can't do You can read, but you can't write. It's the world the world evolves according to physical rules, and yes, you do you really do have some kind of conscious states, but but the this idea that you're going to change what happens is is is fictitious, right? So, so so that that so that that's kind of epiphenomenalism.

[00:28:48]But, there isn't a fourth square yet, which and this is why I've been playing with this idea of of flipping the epiphenomenalism. Now, I'm not saying I believe this. I'm just saying you It's important to kind of think about all the four possibilities. Or, the fourth square inverts the epiphenomenalism and focuses on the other side. Not, what is it like to be or what is it like to feel, but what is it to act? So, the idea is So, so the opposite of what's it like to be a bat, what is it to act bat-like? That kind of thing. And it's the Right, it's the it's the do it's the doing. And so, so I really in in my thinking about about consciousness and minds, I think a lot about the responsibility of not just what does it feel like to to receive whatever is going on, but the responsibility of taking the next step, of choosing among multiple options, right? Of of of having to integrate everything that has happened to you before into some coherent like what you you know you you you you I mean, not to get into free will, but but what you don't want to be free from is your own you know, commitments, thoughts of you know, past experience, right? From your own past. Like, being free from your own past is just like what's the that that's not a that's not a good mind. So, so anyway, so that's so that's all. So, I think I think in all of these things, we have to look at the action part as much as the the the the receiving part, I think.

[00:30:07]>> I um I would agree with that. So, maybe our associations with felt are different or then I'm not using felt correctly because I even during the presentation, I think I mentioned this, something to me that differs about maybe both mine and Jan's opinion with the predictive processing or active inference is that actions are usually some kind of subsumed or subservient to the perception. And for me, it's flipped. It's the other way around. I do hold a very similar opinion in regards to that my felt experience doesn't need to be a felt experience of some incoming in a way. It's more the felt experience of self-generated patterns of activity that allows you to distinguish self from other in in some sense. Um that might be one credit, but that would be the primary focus there. So, in my respect as well, that kind of self-generated pattern in order to distinguish self from other emerges from this like need to move an entire body body. So, in needing to like in some sense shift a material lump.

[00:31:18]It to express it in the easiest way. So, yeah, I I would say I'm in the camp of of of action first, which is I'm quite inspired by like Peter Godfrey Smith and then your work as well. Um and I think Fred Kaiser and there's a few of these people working on and I think my thoughts align on it quite a bit. Um Yeah.

[00:31:42]>> I mean, one one thing one thing though to to build off of that, the thing about biology is that long before muscle and nerve showed up, living systems were taking actions in all kinds of spaces that are hard for us to visualize. So, so so metabolic space, transcriptional space, physiological state space. So, so you don't you know, this this this business of embodiment it doesn't have to be you know, you don't have to be moving mass through the three-dimensional world. As a simple chemical system, you still have to navigate the space of all kinds of possibilities. And as we've been studying even even small like really small molecular networks can already do associative conditioning and habituation and things like that. So, this stuff starts very early on and I think like you know, I I I think we have to ask them what is the if if if if we think that movement in 3D space and the integrated control of your body and all of that is fundamental to consciousness and and and and I've said this to to both Peter and Fred, I think that we then can port that pretty much all the way down and we can ask you know, yes, okay, because of our own commitment to vision and whatever like we're obsessed with three-dimensional space but but but but living things navigate all kinds of spaces and they've been and and and they've been doing it long before we had muscles. So, we you know, what what is it what is it to be a small molecular network doing probabilistic inference and and causal you know, yeah, you know, learning and doing all of these things.

[00:33:21]>> Is that very >> That's a very good and I think that's something I'm actually in the process of learning from your work, Mike. I'm like for me, I previously bottomed out at conductive tissue.

[00:33:34]And the need for some kind of whole body or partially whole body contractile motion. And now I'm thinking trying to run and see how much I agree with or disagree with it but that's that's been a kind of new phase of learning for me to see how far down we can take it. Um yeah.

[00:33:55]I think it's it's it's it's been a pleasure to kind of extend extend what I thought was the bottoming out level of where I would call uh felt experience to to see how far that extends.

[00:34:07]>> I mean there's a there's also a funky thing there. This is the recent work of of Federico Pegasi in my group where we were looking at which which I you know, people differ on how much they think this is relevant to consciousness but causal emergence metrics like Tononi type of right like phi and those kinds of things. So, we've been looking at those in small gene regulatory networks as we train them and there's something very interesting that happens as you train them the causal emergence goes up.

[00:34:40]And as the causal emergence goes up they get better at learning and so there's this like positive feedback loop right?

[00:34:46]But some of them there are different there are different kinds of networks and some of them as you train them the causal emergence just goes up and up and up and up right? Others while you're interacting with them it shoots up but as soon as you stop stop stimulating them it comes it drops back down and it to me it reminds me I don't know you guys can comment on this just qualitatively it looks like certain networks aren't sophisticated enough to keep themselves awake. It's like they right when you're not pushing on them they just sort of dissolve into the void and they can't hold their hold their stuff together.

[00:35:20]But but but others can others can do it in between and it's like this notion of like you know what's you know what's claw doing in between you prompting it right? Is it you know is there anything?

[00:35:29]Well some of these things do nothing and they sort of disappear and and others don't and you can we now have the math to actually distinguish which which ones are which. So you know I think I think these are extremely minimal models for some of the kind of stuff we're talking about.

[00:35:45]>> [snorts] >> Yeah in the in the language of like IIT you might have something there added on to the sort of canonical theory which is which would be something like like durability or like some lasting property like yes the system is integrated but can it be crystallized in form or function for some period of time so that it it exists for longer right? The intrinsic existence that IIT talks about there's really it's not clear that that existence is being recreated or if it comes in and out of existence or if it exists for long periods of time and they they do have ways to deal with these sorts of things but I do wonder like so when when when you talk about all these things that cells were doing before muscles and nervous tissue existed, um you know, there's all sorts of interactions happening and what you're seeing in your simulations or in the lab with these gene regulatory networks, that's been happening for a very long time. All these circuits may have they may exist on a continuum of um sort of integration for over time. Um and I wonder if we have to revive something like the concept of a homunculus, but extend that out to different systems, not just brains. And by homunculus I I mean in the sort of Hughlings Jackson sense of like a representational space where you have like um parts of the world that are mapped onto other parts.

[00:37:17]And uh you know, I I I I think of the cell as existing in such a way where there are all sorts of events occurring and they they, you know, we can talk about these couplings like, you know, if you have if you have an event that involves two parts and you have a you know, I I think in brain concepts, so you know, we can think about pre- and post-synaptic, but it can just be a pre- and a post- across some interface.

[00:37:44]Um so like, you can think of the post as reading the pre- and it's kind of reading a truth claim about the world.

[00:37:52]And you know, in the abstract sense, we we we sort of assume that the coupling is perfect or that the coupling is, you know, reliable.

[00:38:01]But in the real world, couplings are in flux.

[00:38:06]And so you you may have a reading of, well, the system is off, but in fact, the system is on, it's just not coupled.

[00:38:13]And so there there's all sorts of things happening here in terms of interactions that, you know, just get we're we're we're we're not including alternate options into our models of how things work, but something like a homunculus that could integrate all these events that are occurring inside the cell, for example, in a in a very complex gene regulatory network.

[00:38:34]Something that could piece all that together and map it in such a way where you get a kind of global readout of what's happening in that system.

[00:38:42]Something like that might be useful to try to find in the world in different systems in different unconventional um you know, putative minds.

[00:38:52]>> Yeah. Yeah, we've been looking at So, actually Santosh and I have have looked a little bit in these networks of regions of the network that represent the whole thing and and that kind of like the homunculus kind of idea. Um some of the the thing is some of these networks are are not large and complex at all. Like the smallest network that you can make associative conditioning out of is four four four subunits. Like that's it. They don't have to be They don't have to be very large. And there is, you know, it's what you were saying about durability. Well, one of the funny things is that when you look at the graph of the causal emergence every time we stimulate it, even the ones that fall asleep in between, as you stimulate them, the next the next bump is going to be higher and higher.

[00:39:34]So, they do return to zero, but the bump isn't the same. So, there's they save state so so so somewhere, right? That there's there's stateful in that way.

[00:39:41]And so, even even when the thing falls asleep, it it it also reminds me of um you know, the old the old story of the the the the the people with with um uh brain damage for for they can't form new memories. And so, so I think it was you know, who told the story where where you walk in, shake the guy's hand, you go around, come back in, and he shakes it as if it's brand new. But but one time you stick him with a pin, you know, and he then you come back and he's like, "Yeah, I don't really I don't really shake hands." So, he don't remember the original event, but something something stuck. You know, and and it right? And it reminds me of this is like like like the the the the the thing you're measuring comes down but there's a there's a persistent state that's saved dynamically somewhere. And you know, if you can do that in a in a in a network of under 10 subunits, these crazy things that we make with trillions of parameters and whatnot. I mean, who the hell knows what you know, what what what's going on in there.

[00:40:38]>> So, I wonder if any of us disagree with because the this durability is a kind of working memory but like not in the brain sense in the kind of general sense.

[00:40:48]Does anyone disagree that memory would be a kind of contingency for a conscious system?

[00:40:55]>> Uh I would actually agree. Previously, I I was done some work on another level on that but I would otherwise agree that this is a core uh component. Uh I would also like to use this chance to thank both of you because I I am hearing that I it's dinner time for for me for my family.

[00:41:16]So, I I really love this discussion.

[00:41:18]It's really thoughtful but you know, I have to stick with my with my family. I will watch the recording but I'm truly uh thankful here for your time, Mike and Nick. Thank you.

[00:41:29]>> Oh, thanks. Yeah, good to see you.

[00:41:32]>> See you. Take care.

[00:41:33]>> Um I would I would agree with with the kind of uh So, this continuity you speak of I I I I would call it self-determination in in in a way and I I think that if I'm thinking about the causal emergence measure you're using, I'm not sure if it's the same one or if it's still based on effective information.

[00:41:53]Um but if it if it is, I I I think one great extension to that would be actually within the measure itself to have a notion of of of self-determination or or kind of this self-actuation.

[00:42:06]And that's something that um I've been currently working on with a different measure from effective information and from causal emergence, but it's on the same idea picking out kind of macroscopic observables that in some sense have um a self-determining capacity.

[00:42:22]And I I definitely would have would agree that uh I guess memory or or some kind of information storage is required or is a contingency to to conscious experience.

[00:42:35]Yeah.

[00:42:36]>> Is Is there any objective Do you think there's any objective claim that's being made when we talk about self versus non-self, or is that entirely a subjective claim? Like if a supervising non-object could view everything that's happening including, you know, things that we consider to be subjects and things that we consider to be non-subjects, do you think that that distin- that that distinction could ever be made um from the outside, so to speak, or can that kind of distinction only be made from the perspective of a of a subject?

[00:43:14]>> That is a honestly a great question, and I'm not ready enough to answer it like on the spot. Um I don't know.

[00:43:23]>> The reason I bring this up is because, you know, I I tend to I tend to view self and non-self in like a kind of cybernetic um from a cybernetic perspective. It's really difficult to dissociate them when you're including them in a in a kind of feedback loop. If If in fact you think environment is driving um organism, but organism is also directly driving environment and vice versa, um like it's very difficult to separate them. They look like, at least in that in that sort of frame, they they kind of look like one thing.

[00:43:57]And yet um cells distinguish self and non-self um when when you ask a person to distinguish themselves from their environment. I mean, there's certain brain injuries that will actually make that distinction go away perceptually, but generally people will say this, you know, this object is not part of me and so on.

[00:44:17]>> Yeah.

[00:44:18]>> But I've always wondered like from an outside perspective, you know, the miasma just kind of floating around, could is there any distinguishing factor that can actually separate um a subject from its environment or is that just illusory?

[00:44:35]>> I I I might have a better answer now. Just I'm I'm thinking about it in the context of autonomy, which is something that I'm I'm primarily focused on in my uh theoretical thinking and and I I align with some other work done on autonomy, where there is a notion of constitutive autonomy, which is the idea that your border between self and the environment is driven by the kind of recurrent processing within the self, within the system itself. And there is a notion of interactive autonomy, which is autonomy from the environment. So, this is like two slightly different measures.

[00:45:09]And in this sense, I I I think that there is a there is a clear and kind of operational definition of how you can consider a dynamical subsystem within a larger dynamical universe as um a self from the outside as well as from the inside. Inside, it would be high constitutive autonomy. Outside, it would be uh low interaction, so higher autonomy from the from the interactive environment. Um one other thing that I I I I I I should know and I wonder what you what you think about this is um something that I've been inspired by.

[00:45:46]I think Humberto Maturana said it in one of his papers, where he says, "A being comes into existence when a world is severed in two." And I love this idea because it kind of defines an environment relative to every agent. So, there is a kind of there is a degree of autonomy or there is a degree of agency given a particular scale from an environment as well as the fact that the environment is induced by the agent which has the autonomous border itself. So, what might be an environment or a kind of like working environment for a cell is very different to a working environment for a heart or for for a brain, for example.

[00:46:30]And there is some kind of integrative quality across these scales. But, yeah, I think that was an extension on your point, but I'm I'm also interested to hear what you guys think of of that, the induction of an environment.

[00:46:47]>> Yeah, I mean, so so so I have a I have a a framework where the boundary between self and world is defined by the size of the goals. So, it's the the this notion of the cognitive light cone, which is it's it's not how far your senses reach and how far your effectors reach. It's the size in some space. It's the size of the biggest goal you can pursue. So, the size What are What are the states in the world that you actively care about managing? And the and the and the the size of the states that you are concerned with is tells me whether, you know, for example, in space time it tells me whether you're a bacterium that only cares about the the local sugar concentration versus a dog that, you know, doesn't care about what happens 3 weeks from now, but does have like a bigger a bigger area that right and things like that. And experimentally where this comes up is is is two places.

[00:47:38]One is um our [clears throat] our story on on what happens with cancer and what happens when individual cells disconnect from the electrical connection network of the rest of the body and their cognitive light cone shrinks to the point where they're just amoebas again and as far as they're concerned the rest of the body is just external environment, whereas before and and what has happened there is that prior to that, they were part of an electrical network that shared memories. It had It had this like a collective memory thing going on where their goal state was this enormous abstract thing. We're building a limb, you know, with with five fingers. And no individual cell knows what a finger is, but the collective absolutely does. And and and and the collective is able to support a a a standing bioelectrical pattern that tells them what the heck they should be working on. But once they disconnect from that, it's gone.

[00:48:25]>> [music] >> The the the boundary be So, from that sense, the cancer cells are not more selfish. They are They just have smaller selves. So, so their selves have contracted and and you know, the rest of the body is just external environment.

[00:48:37]But the But the other the the other side of that um that's that's really instructive is is the embryonic side. So, you look at a blastoderm and there's, you know, I don't know, 100,000 cells or something.

[00:48:48]And you look at that and you say, "Oh, there's an embryo." Well, you have to ask, what is there one of? What What are we What are we counting when we say there's one embryo? I mean, it's a you know, there are hundreds of thousands of cells, molecular networks inside of that. Like, what is there one of?

[00:49:01]And I think what you could say is that what there's one of is a kind of um shared delusion almost. It's a It's a It's a shared vision of uh we are all supposed to be taking this journey in anatomical space where we're far away. We're a single cell right or, you know, a a flat disc right now. But we're supposed to be this this complex thing that's going to be a gastrula and a neurula and whatever. And so, everybody's in agreement that that's where they're going. And one of the cool things you can do is uh and and uh you you know, I I I used to do this in duck embryos. You can You can They're They're nice and flat.

[00:49:35]And you can take a little needle and you put some scratches in that blastoderm and separate them into islands. And when you do that, every island doesn't feel the rest of it and says, "Well, I'm an embryo and I'm going to self-organize into an embryo."

[00:49:46]And you find out a couple things from that. First, that the number of individuals in an embryo is not fixed.

[00:49:52]It's not genetically determined. It's anywhere from zero to, you know, half a dozen or more, right? And it's a dynamic process of of self you know, pulling yourself together and figuring out where do I end and the outside world begins.

[00:50:04]And then and then the other cool part is that if you you you make the scratches but you let them heal, then you end up with conjoined multiples, so twins, triplets, and so on in the same blastoderm. And then you get some really cool stuff because the cells in the middle of of two have to figure out am I the right side of this guy or am I the left side of this guy? And you actually get and so this explains for the first time who I who I did this in like you know, '96 or something.

[00:50:29]It explains for the first time why conjoined twins, a lot of times in human conjoined twins, one of the twins has laterality defects. Their their heart might be on the wrong side or whatever.

[00:50:38]And it's precisely because of this. When you're sitting next to another twin, there's a lot of uncertainty about what what what am I part of and and that that am I the left side or the right side or what? And figuring out where that boundary is is is not trivial. And so to get back to kind of your original point, Nick, I think that there are again multiple perspectives here. I think we as external observers can try to put some labels on things, but I think for significant I think I think this agents are interesting to the degree that they have their own opinion on the matter. And so [snorts] it's right, so when when there is a when the system itself is is trying to define into in inside from outside in some in some way, whether that's a single cell boundary or an embryo boundary or whatever, I think its opinion carries a lot of weight too. And to the extent that it has strong self models and self opinions about that, it becomes a much more interesting agent where it's not just, you know, external observers that that can that can do this kind of um uh you know, you know, estimation of what what we see. How many individuals do we see?

[00:51:44]Yeah.

[00:51:45]>> Yeah, it's really interesting. And like I know you can even do this stuff statistically like if you have if you have large data sets and you're looking for patterns in the data set that might exist, depending upon how you apply, you know, something like a factor analysis, um you can end up with a very large set of distinct um factors, or you could force cluster things into two or three or four, depending upon how you set up the the statistical model.

[00:52:13]Um and I and I wonder, based upon both your comments, like it seems like boundaries are really important and like the word divide is doing a lot of work. Like what what what does divide mean in like to divide a world?

[00:52:26]And I wonder if um what this just means is uh like there's a kind of you can think of every object in the universe as having a relationship with every other object. And some objects are hidden from other objects either by distance or by um you know, like charge mismatches or you know, all sorts of contingencies that make them less likely to interact.

[00:52:52]And so you could actually seriate every object in the universe with every other.

[00:52:56]You could say like this particle has more interaction with this one, less with this one, less with this one, less with this one. You could just do that.

[00:53:05]And you could create a kind of gradient of separation, let's say.

[00:53:11]Um and then, you know, that's just the most extreme form, but you could do that with biological systems, chemical systems, physical systems where you're just looking at relationships between parts, let's say. I mean, maybe I'm putting the cart before the horse here and and calling things parts and then looking for a way to separate things, but you could um you could do this kind of operation where you're you're looking at the subset of the whole and then kind of defining relationships between um stuff and seeing how it breaks apart. And I think you're right. Like in the end, depending upon what lens you apply or what algorithm you apply or, you know, what kind of sorting mechanism is like brought to bear on this large set of things, you're going to get different divisions. Like it's the there I don't know that there is an answer in terms of like, "Well, there are actually two parts here." or "There are three parts."

[00:54:06]I There are there are potentially many answers to that question.

[00:54:10]>> I I I would I would tend to agree with that cuz I generalized the notion of observers as I as I mentioned before and and in my case, a if we're looking at a system itself and how it um develops agency or or autonomy in some sense, I would consider the environment to be the observer that constrains that particular system. So, the coarse-graining of the of the system itself is relevant and relational to the environment, which as something that um induces constraints on it, is to me an observer-like quality.

[00:54:47]Um and I've been thinking about how to define this observer for a long time because, you know, the way you think about it from classical physics is the fact that this observer is some kind of human with some perception in in essence, but that's not how we think about it when we look at a um quantum physical level or just in general in the like standard model of physics, the an observer is a relational quality from from from the outside.

[00:55:12]Yeah, so for me, the kind of macroscopic grain at which some um agency's maximized uh has to do with the environment that induces it. And this kind of goes back to Mike's comment, actually, that it depends on the on the size of the of the state space in which it is interacting in. And to me, that just speaks to maybe the emergent complexity, which is um induced given a particular agent and how it interacts with an environment.

[00:55:41]So, yeah.

[00:55:42]>> That's really interesting. That that connects with something Mike and I have talked about before um in conversations.

[00:55:48]It's a kind of paradox which is in order to get one, you actually need two. Like in order to in order to have a an an object that's distinguishable from the world to say there's one, you actually need to start with two. It's it's it's it's a very odd thing because if it's relational, if you need the environment, then you actually can't have one until you have two. It's just very it's very backward. Um >> I I but that's that division of inside I I spoke of with that like poetic comment by Maturana. It's it's exact I completely agree with that too. Yeah.

[00:56:23]>> Yeah. Yeah. Yeah. Chris Chris Fields and I have this this thing coming soon on the the symmetry between agent and environment and and how systems offload computation to the environment. And then basically it it ends up being extremely symmetrical to say like who who's [snorts] doing you know who who's doing who's doing the work here. It becomes it becomes really interesting. So, yeah.

[00:56:47]Cool. Very cool.

[00:56:50]Okay. Uh yeah. Thanks very much guys. Uh yeah, really really interesting as usual and I think So so So Bookie, you you said you were working on some new stuff.

[00:56:58]Send me whatever whatever you guys have, please send it along. I'll send you I'll send you some stuff. Uh >> I would I would love to. That's that's great.

[00:57:05]>> I I think we we align on on on many things and and I like this