Support the show to get full episodes and join the Discord community.
Alicia Juarrero is a philosopher and has been interested in complexity since before it was cool.
In this episode, we discuss many of the topics and ideas in her new book, Context Changes Everything: How Constraints Create Coherence, which makes the thorough case that constraints should be given way more attention when trying to understand complex systems like brains and minds – how they’re organized, how they operate, how they’re formed and maintained, and so on. Modern science, thanks in large part to the success of physics, focuses on a single kind of causation – the kind involved when one billiard ball strikes another billiard ball. But that kind of causation neglects what Alicia argues are the most important features of complex systems the constraints that shape the dynamics and possibility spaces of systems. Much of Alicia’s book describes the wide range of types of constraints we should be paying attention to, and how they interact and mutually influence each other. I highly recommend the book, and you may want to read it before, during, and after our conversation. That’s partly because, if you’re like me, the concepts she discusses still aren’t comfortable to think about the way we’re used to thinking about how things interact. Thinking across levels of organization turns out to be hard. You might also want her book handy because, hang on to your hats, we jump around a lot among those concepts. Context Changes everything comes about 25 years after her previous classic, Dynamics In Action, which we also discuss and which I also recommend if you want more of a primer to her newer more expansive work. Alicia’s work touches on all things complex, from self-organizing systems like whirlpools, to ecologies, businesses, societies, and of course minds and brains.
0:00 – Intro
3:37 – 25 years thinking about constraints
8:45 – Dynamics in Action and eliminativism
13:08 – Efficient and other kinds of causation
19:04 – Complexity via context independent and dependent constraints
25:53 – Enabling and limiting constraints
30:55 – Across scales
36:32 – Temporal constraints
42:58 – A constraint cookbook?
52:12 – Constraints in a mechanistic worldview
53:42 – How to explain using constraints
56:22 – Concepts and multiple realizabillity
59:00 – Kevin Mitchell question
1:08:07 – Mac Shine Question
1:19:07 – 4E
1:21:38 – Dimensionality across levels
1:27:26 – AI and constraints
1:33:08 – AI and life
Alicia 00:00:03 How on earth can I argue for a local science? I’m gonna be laughed out of the room if I argue for a local science,
Paul 00:00:11 But this is kind of what, what Bill Wimsatt hints at in the causal s thicket. Right?
Alicia 00:00:15 Thank you, thank you, thank you. It seems to me that what those folk psychology terms mean represent are what dynamical system theorists would call the constraint regime that holds the dynamics together. The next stage of this is going to be a study of all the different modes of constraint that are all operating simultaneously.
Paul 00:00:47 This is brain inspired. Hey everyone, I’m Paul Alicia Guerrero is a philosopher and has been interested in complexity science for a long time, in her own words, uh, since before complexity was cool. I like that. In this episode, we discuss many of the topics and ideas in her new book. Context changes everything, how constraints create coherence. And the book makes the thorough case that constraints should be given way more attention when trying to understand complex systems like brains and minds, how they’re organized, how they operate, how they’re formed, how they’re maintained, et cetera, et cetera. So, as you probably know, modern science, thanks in large part to this success of physics focuses on or mostly on a single kind of causation. The kind involved when one billiards ball strikes another billiard ball, but that kind of causation neglects what Alicia argues are the most important features of complex systems, the constraints that shape the dynamics and possibility spaces of systems.
Paul 00:01:58 Much of Alicia’s book describes the wide range of types of constraints that we should be paying attention to and how they interact and mutually influence each other. I highly recommend the book, um, and you may wanna read it before, during, and after our conversation. That’s partly because if you’re like me, the concepts that we discuss still aren’t comfortable to think about in terms of normal modes of explanation in science. Thinking across levels of organization turns out to be hard. You might also want the book handy because hang onto your hats. We also jump around a lot among, um, many of the concepts that are in her book. So context changes everything comes almost 25 years, almost a quarter of a century after her previous classic Dynamics and action, uh, which we also discuss and which I also recommend, especially if you want more of a primer, um, to her new, more expansive work. And these ideas touch on all things complex, of course, from self-organizing systems like Whirlpools to ecologies businesses, societies, and of course, minds and brains. Find the show notes at Brain inspired dot coast slash podcast slash 1 74. I’m sure you could tell I very much enjoyed this conversation. We also have some guest questions from friends of the podcast, Kevin Mitchell, who’s also coming up on a near future episode about his new book. And Max Schein from Down Under. Thanks for listening. Here’s e Alicia.
Paul 00:03:37 Context changes. Everything is almost a quarter.
Alicia 00:03:41 It’s Cuban hyperbole. Oh,
Paul 00:03:43 <laugh>. Well, what does that mean? <laugh>
Alicia 00:03:45 Humans tend to exaggerate. I probably should have said context changes most things or a lot of things. <laugh> Cuban hyperbole.
Paul 00:03:53 Well, I was gonna tie it to your previous work dynamics and action, which is now almost a quarter of a century old. Um, and so what you could say about that is that context changes action is a very coarse and terrible summary of, of that classic now work. Thank you. Um, and so I was thinking maybe context changes everything is just kind of an expansion on those ideas. Yeah, it really
Alicia 00:04:16 Is. It really is. Um, after I wrote that one, I thought, well, I’ve got nothing else to say <laugh>, but <laugh>. But you know, there’s been so many issues. And one of the things is people said to me, you know, there, there’s nothing that’s context free. I had used that term in that one.
Paul 00:04:34 Oh, yeah, yeah,
Alicia 00:04:35 Yeah, that’s true. Uh, so what do I call it? Well, what about context? It depends. So I thinking maybe I get gave short shrift to that one, to that that kind of constraint. And so, you know, things start, and of course AI and language models and so on. And I, for e I think came outta the process. So there were a lot of things that I thought once the pandemic started, I thought, you know what, just sit down and do it. You know? And just, just put it all together. Yeah. And Steve, you see if it comes together comes Yeah. And, and I was pleased It did.
Paul 00:05:12 Yeah, me too. Um, I mean, you mentioned the phrase context free, which you have changed now to context dependent in the new book. Independent. Independent. Oh, that’s the context, right? Context independent. Yeah. See, that’s what I was gonna say is, uh, you know, the book, I love the book. It is, um, challenging and one that I will have to revisit over and over because well, no, it’s not bad. So I’m already a, that’s not a bad thing. I’m already a slow reader and, uh, I’m a slow thinker apparently, uh, as well. At least to get any depth. Poor
Alicia 00:05:40 Writer is what I’m I Well, I don’t write very clearly. I don’t know why I’m good in the classroom, but I’m ter I just, it looks still, I spent most of my time just editing. Sure, yeah. But
Paul 00:05:53 Well, that’s, so what I wanna ask you is like, what’s changed in your views since Dynamics in action? And, you know, besides just expanding, you know, you cover like so many different constraints and topics in this book, but what I wanted to comment on before that is one of the reasons why, um, I’m gonna have to go back to it over and over is, uh, because the terminology and is you’re like inventing some of it. And a lot of it is kind of still fu for me, um, still fuzzy in terms of like, what, what it actually like, it’s, it’s, uh, definition and how, like, how I set the boundaries, you know, with these terms and what they mean and how I can use them. Yes, yes. That’s, and so, you know, and we’ll come back to this, but, uh, anyway, that’s why I’m gonna be revisiting a lot. Um, so maybe let, let’s start with, you know, what’s changed in 25 years in your thinking?
Alicia 00:06:43 I think what’s changed is that when I wrote Dynamics in Action, I basically meant it as a brief against analytic action theory.
Paul 00:06:56 Okay. Well, you can you, can you summarize analytic action theory? Well, the
Alicia 00:07:00 Idea that, that either that intentional behavior, intentional action, you know, the book starts with what’s the difference between a wink and a blink, right? The difference being that difference, you should be able to explain away anything that philosophy of mine considers intentionality or intention. So teleology was reduced away. Um, the idea of Contentful intentions were reduced away. And so, and so my thought was because I had published a number of standalone papers about what’s wrong with behaviorist reductions of intention. What’s wrong with, with identity theory, uh, attempts to reduce intentions. And on that complex dynamical systems theory was a very good way of entering this whole field that would solve a lot of the problems or wouldn’t get you into the problems that all these other attempts did. But in the last 20 some years, I keep thinking, you know, most stuff in reality is irreversible, most stuff is context. So it applies not just to action, but it applies a lot more broadly. And so I think I stick my neck out a bit more with respect to general ontology or general cosmology by using the same notion of constraint. Yeah. Um, more broadly than just for action. Yeah.
Paul 00:08:45 Well, I had to revisit dynamics and action. Um, as you know, and as you write about in your, your newer book context changes everything, uh, neuroscience has embraced the dynamical systems perspective. And you talk about some of the studies that, um, you know, some of the people who have been in those studies have been on this podcast. And we, I, on this podcast, we talk a lot about the dynamical systems view, and then going back to dynamics and action, it’s like, oh my God, it’s all there already. Uh, I mean, thank you.
Alicia 00:09:10 Thank you. It’s fascinating to me that somebody like Mark Churchland, whom I use in one of the chapters is the son of Paul and Patricia Churchland who were the
Paul 00:09:24 <laugh> Yeah.
Alicia 00:09:24 Our foremost advocates of eliminative.
Paul 00:09:27 Right? Oh, I didn’t even think of that
Alicia 00:09:29 Thing is as a mine, I’m thinking, oh my goodness. This is, this is this revenge par excellence of dynamical systems theory, because I once met Paul Churchland Yeah. Um, here in dc Dick Schlegel at GW invited him to speak, and I asked him back then, this must have been, this must have been just about when context when Dynamics came up. It was years ago. And I haven’t you ever considered con dynamical systems theory that has nothing to say about neurology or the market, really?
Paul 00:10:05 Yes. Oh, huh.
Alicia 00:10:07 Paul Churchland said that in person, and I’ve never forgotten. I’m thinking No, and the reason I’m was almost sure that was so is because, but way back when I, I was talking a lot to people in operations research and network theory and that kind of thing, and I’m like, you know what? There’s gotta be a network property in the brain that has emergent properties that somehow loops back down and is able to influence behaviors such that these emergent properties get realized in behavior. That’s the difference between an action proper and behavior. There’s gotta be some of that. But no, I think that just traditional notions of causality, traditional notions of ontology that only the tiny little fundamental components are real mm-hmm. <affirmative>. So it was all, let’s look for the neuron
Paul 00:10:59 Right.
Alicia 00:10:59 That does this. Yeah. Yeah. So I find that really, really very rewarding that Yeah. Dynamical systems theory has been vindicated in a sense.
Paul 00:11:09 Yeah. I I wonder what the elder churchlands, they, their daughter is also in neuroscience, by the way. Um, I wonder what they think of, of the dynam. I, I wonder if they’ve come around at all, because the tagline sort of, of Illuminati limitism is that folks psychology, the, the words that we use in folk psychology aren’t real things. Right? And then, and then there’s a, a reductionist way to understand things, and that’s the best way. But the, the dynamic, the dynamical systems viewpoint kind of, um, it, it’s kind of in between that, right? Because it allows for a top down thing, but it doesn’t necessarily say anything about the terms and concepts we use.
Alicia 00:11:44 Well, that’s why I think I stick my neck out in this, in this one where it seems to me that what those folk psychology terms mean represent are what dynamical system theorists would call the constraint regime that holds the dynamics together. Yeah. And so they are capturing an emergent property that is powerful. And that’s where the whole problem of causality kept bugging me. Because thinking Carl Gillette’s book on, on reduction in emergence where he says, you know, nowadays the physicists like Brian Lin, for example, have, have espoused the truth of the reality of emergent properties. But they still don’t want to allow that those emergent properties are causally causally influential on their components. In other words, they cannot change behavior. They cannot change, in a sense, neural architecture in virtue of their emergent properties. Hmm. Yeah. Whereas that’s what I think complex dynamical systems lets you do.
Paul 00:12:56 Um, maybe we could so just segue into causality. Sorry, I’m jumping around
Alicia 00:13:00 A
Paul 00:13:00 Lot. No, that’s great. I do
Alicia 00:13:01 That a lot. That’s great. I do that a lot.
Paul 00:13:03 You’ll have to, I do it too. So we’re gonna be in a real swamp here pretty soon, <laugh>. But, uh, you know, so one of the things in dynamics, in action also, uh, is, you know, you, you’ve railed against this modern notion of efficient causation as the only valid kind of causation. And since you’re just talking about causality, um, maybe we should talk about that. And then I wanna come back to also what you have changed your mind about, uh, perhaps since then. But, um, so, uh, you know, famously there are these four different types of, uh, causations from Aristotle’s days from Aristotle and, uh, an efficient causation, the billiard ball, uh, the thing X hits things y and then that moves things thing y That’s the only valid kind of causation. And a lot of your labor is trying to bring back notions of the other, of formal causality and causality via constraint. Correct? Correct. Um, and so do you think that, and, and from my limited perspective, because I’m reading works like yours and like a, a lot of similar kinds of works that are trying to bring, make teleology not such a bad word, and bring like these constraints to, to bear, um, because of that, I feel like the old story, the, the modern story of efficient causation is the only causation. I feel like that’s dissipating. Do you believe that also, or
Alicia 00:14:17 I think it’s still very present in, in physics, especially with regards to the kind of causality that we would call top down causality. Mm-hmm. <affirmative> that in fact, emergent properties really are not causally powerful. That all the heavy lifting as Van Orton would say, it all happens bottom up. And so it’s really explained through efficient, energetic transfer,
Alicia 00:14:52 The idea of causality as efficient boxes you into the, the poss boxes you into the impossibility of top-down causation. Because if there were top-down causation with that kind of causality, you’d have over determination, you’d have of a violation of the closure of the physical, and you’d have all these problems. And so therefore, knowing that can’t happen, well, but if you think of, but what I found fascinating, that’s one of the reasons why I adopted the notion of constraint, was every time natural scientists, hard scientists got into trouble, I thinking in terms of efficient causality, then immediately the term starts being used as well. It’s factors or conditions or constraints mm-hmm. <affirmative> that are causal. Well, you know, okay, so I’ll, I’ll, I’ll grant you efficient causality as, as, as the term for cause. Right. So let’s call it influenced by constraint. And, and that’s what I was trying to do, because I certainly don’t want to go to an al notion of formal cause I don’t want to go to, you know, the or, or your, you know, pales teleology or even Aristotle’s teleology, which was all built in set up from the beginning.
Alicia 00:16:11 Right. That’s the way Got, that’s why, that’s why Darwin, Darwin bought into this. And so Darwin thought he was very concerned about the notion of evolution, because the idea of new species appearing that that would me mean that it wasn’t all built in from the start. And therefore this developmental process was only an unfurling of preexisting, uh, potential. And what I wanna say is that constraints really create you potential, because I think that’s what you get from PGA gene. I think that’s what PGA gene gets the, he was the good Prego, now that we think about it. <laugh>, he was the one who, he got the Nobel Prize for dissipated structures, I think for that very reason. Hmm.
Paul 00:16:54 But you, you, you mentioned physics, and maybe it’s just in biology, that the notion of, um, of circular causality and, and causality via constraint is more comfortable among biologists than
Alicia 00:17:07 I think the biologists. Biology was the last holdout. They refused to get shoehorned into the model of efficient mechanical causality. And as I was reading your email and reading one, one of the latest issues of science magazine on Australia, you know, they’re talking about the, um, dispersal of, uh, distribution of black beans throughout Australia. And the assumption was, well, this precedes the Westerners who arrived in Australia, but now they’re revisiting the idea that some of, of the songlines of the native Australians have accounts of these plants and the dispersion. So the science magazine article talks about combining not only the genetic, uh, flow of the, of the genes of the black bean plant, but also biogeography and these ethnographic accounts. Mm-hmm. <affirmative>. So I think that you see it now more in eco, in ecology in addition to just straight biology. But you see it, and of course the social scientists have always used other causal factors as their reasons, but of course, that’s what led the heart scientists to say, well, that proves they’re not sciences the social scientist. Right? Sure. Yeah. Because they need to, because they need to appeal to unique causes, to things that don’t scale as a universal law. And because they use top-down factors that are not efficient causes, that shows for sure they’re not science <laugh>
Paul 00:19:04 Before, maybe this can just be a jumping off point, too. Any of these can be jumping off points because we can go any direction. But have you changed your mind, uh, about anything in particular with regard to the story of context and constraints over the last quarter century?
Alicia 00:19:19 No, I’ve just sort of expect, I think I focus more on context independent constraints mm-hmm. <affirmative>, because what I think that does,
Paul 00:19:28 And hang on, let, I’ll, I’ll just say context independence mean, it’s, means it’s a, it’s an external constraint that drives a system, uh, far from equilibrium, like a gradient to gravity, et cetera. Right. Pulls on the system way
Alicia 00:19:40 Gradient particularly. And when you start looking at the different kinds of gradients, either the spatial temporal gradients of gravity, you know, the gravitational <laugh>, right?
Paul 00:19:52 I like that you just illustrated gravity right there. That’s
Alicia 00:19:54 Right. That means, that means that that space time is not homogeneous and uniform. There are in homogeneity in space and time. Um, polarities, we know the importance of polarities all over the place. We have polarities in el electric and magnetic, but also in developmental, anterior, posterior, uh, right and left in the development of, of an embryo. Um, there, uh, we have, um, I think in, in in social systems, we certainly have levels of organization that would give you different, different, uh, points of equilibrium at different stages. Yeah. So I think importance of that, um, boundary conditions, I think as a setting for a context. So basically I think of the, of context independent constraints as those that determine in a sense, the bounds of possibility space. They, they kind of determine the, the, the, the basic fundamental characteristics of possibility space.
Paul 00:21:14 But you need both context, independent and context dependent. And I’ll just, uh, briefly define context. Context dependent constraints are those that which bring things back essentially away from context independence, which means more toward equilibrium. And when you get these mixing and match, maybe not No, I, I know I, I know I just caught myself saying more toward equilibrium, but, um, away from independence, we’ll say maybe that’s a Okay. Correct. Yeah. But when you get these interacting is when, when you can have the ability to develop complex systems. I think
Alicia 00:21:43 The reason I had given context, free context, independent constraints, short shrift back when, is because I don’t think context independent constraints complexify, they don’t, they don’t enable complexity. All you get in principle would be a clump
Paul 00:22:01 <laugh>. Yeah.
Alicia 00:22:02 Correct. They would match at
Paul 00:22:04 The polls. Yeah.
Alicia 00:22:05 And, and clumps and masses and agglomerations have no emergent properties. So context, independent constraints don’t, don’t create complexity. Once you have a, a, a space of possibilities with a certain in homogeneity in it, then I think complex dependent constraints can kick in. My question is, well, why do they show up? Why do context dependent constraints appear? And I’m wondering whether, again, the boundary conditions, if, if, if the clumps are dense enough mm-hmm. <affirmative> and they can’t move much, then do accidental linkages
Paul 00:22:57 The determinism, are you hinting at determinism in the universe?
Alicia 00:23:01 Well, um, even the bumps and jostling of regular brownian motion mm-hmm. <affirmative> that if, if there, I always try to ground what I say in, in heart sciences, and I always end up going back to things like the benard cells and the BBC reaction. Otherwise, I’m very worried that philosophers will be accused of flights of fancy that are not grounded in anything scientific. And so I keep thinking, alright, so you’ve got Bernard cells and you a pan of water heated uniformity from below, and at a certain threshold of, of non-equilibrium, in other words, at a certain gradient, correct. The status quo cannot be maintained. In other words, conduction cannot be maintained. And so the system flips to linking individual water molecules into local perturbation. But if that gets amplified by the conditions, then it becomes a self-reinforcing feedback loop. And that’s, I think, where you see the context dependent constraint kicking in that then the linkages create standalone units. Almost like, isn’t that almost, almost like when you’re playing hangman? You know, we know that instinctively because if we have towards the end of a word that we don’t know what that is, a t and then an I, well, there are two more to go. Well, it’s either gonna be n g or o n because these form like standalone units, because they, the context dependence concatenates and becomes a standalone unit. And now you’ve got, in language, you’d have a pH mm-hmm.
Paul 00:24:57 <affirmative>.
Alicia 00:24:58 Right. But in, in, that’s what a laser is. A laser is laser context dependent constraints of a whole bunch of photon streams getting linked together in a way that they become a standalone entity with emergent properties.
Paul 00:25:17 Yeah. So you’re talking about, and you just mentioned why you use these examples, because they’re less dangerous, I suppose, because they’re in some sense simpler, more, more, I’m gonna say low dimensional, um, because they’re like from physics, so it feels more comfortable. Maybe that’s a way to say it. Um, but, but you do venture into, um, these more, you know, like biological systems, uh, in the book. Um, and, and one of the questions I had for you is, so I, I’m glad that you’re sometimes pausing just a little bit before you say causality or something. Yeah.
Alicia 00:25:47 I always do <laugh>.
Paul 00:25:47 Yeah.
Alicia 00:25:49 Because that’s where immediately a physicist will say, no, that’s not causality.
Paul 00:25:53 Yeah. Right. Right. Well then like, and then I don’t know what’s going on in your head if you’re searching for a different term because there, or you know, alternative types of terms, because, you know, reading your book, and one of the reasons why I need to revisit it multiple times and I look forward to it, is just because things feel so new to me, and the terminology is so new that I, um, that, that I don’t feel like they don’t feel comfortable. Right. So I, I’m trying to think like, how do I apply this term and how does, how does this concept interact with this kind of constraint? And there’s their interlocking independencies, and how do I think about applied to the brain? And so what I’m wondering is, are these gonna feel you’re the expert? How com do these feel super comfortable to you just switching and thinking about these things interacting and the different modalities
Alicia 00:26:38 I wish I had made, I had coined the term enabling constraints. It goes back to, to Patti, Howard, Patti. Mm-hmm. Mm-hmm. <affirmative>, and a whole bunch of these biologists who talked, several of whose books have the same similar title hierarchy theory. Mm-hmm. And it’s because the enabling constraint is able to create a lev, a self-organization that can then also embed in the next higher up level of constraints. So, so I particularly want to use the notion of enabling constraints as constraints that enable the formation of a more complex dynamic.
Paul 00:27:19 And that is, and how does that relate to a limiting constraint? Contrast that with a limiting,
Alicia 00:27:23 That would be the, I think we tend to use the term constraint, especially in physics and in mechanical engineering, especially as limiting as, as, as closing off options. But the biologists tend to use them very often as factors, conditions, um, feedback loops, catalysts. See, the nice thing about the notion of a catalyst, two things, a lot of things <laugh>. One is it spans chemistry and biology mm-hmm.
Paul 00:27:56 <affirmative>. Yeah. Right.
Alicia 00:27:58 And now, and, and even in so and so as a catalyst of change, and we say that of people, so it, it’s used in that sense be, but also because a catalyst is conserved, it does not use up any energy of its own. And so what I like about that notion is that it’s a causal factor. It’s a causal, it’s an influencer
Paul 00:28:24 Makes a difference.
Alicia 00:28:25 Yeah. That is not in the efficient causality mode. So are feedback loops mm-hmm. <affirmative>. And the thing about feedback loop, which I love, is that for eons, for sin, our sin, aris on that never changed. The idea of something creating itself or feeding back on itself to change itself was forbidden, was rebo. Uh, because again, from Aristotle’s notion, then it would have to be before and after itself. Mm. It would’ve to be before itself is caused, it would be have to be after itself is caused. You can’t, oh, no. You can’t have that <laugh>. So the notion of simultaneous causality does not allow for parts interacting to produce holes, holes that have emergent properties, which then simultaneously loop back down and constrain the parts of which the hole is made up. And so that’s why I love the notion of feedback loops and catalyst in particular. But then I think once that whole coalesces, then you’ve got stabilizing constraints because it has to hold together for a persistent length of time. Because unlike neuro equilibrium thermodynamics, the thing doesn’t fall apart. The, the thing about these benard cells is that while the conditions hold, it holds together despite changes at the component level
Paul 00:30:02 Perturbations. Yeah. So it, yeah.
Alicia 00:30:04 Fluctuations, internal fluctuations and perturbation, it has a meta stability, which I also like, which is different from thermo equilibrium of classical near equilibrium thermodynamics while the thing holds. It has a stabilizing power to stabilize the constraint regime as a whole.
Paul 00:30:26 And what is it, you say it has a stabilizing power. That’s the whole what is, what is that
Alicia 00:30:30 Is the dynamic constraint regime that constitutes a Bernard cell. A Bernard cell is not a thing other than the water molecules that make it up, or the constraints that go into establishing the conditions that make it up, the boundary conditions as context independent constraints, and then the linkages of the molecules. Yeah.
Paul 00:30:55 Right. Yeah. Yeah. It’s just so hard to think of across scales, these meteorological. Yes.
Alicia 00:31:02 It’s extremely difficult to think across scale, because we tend to think in terms of efficient cause across scale. Yeah. And all of these biologists will, I always wondered for a long time, why did these bi biologists like Stan Salty, Howard Patti, a lot of them started talking about semiotics and seniors. Why did they ever get into this <laugh>? And, and it finally dawned on me, because the relationship between levels of organization and Herbert Simon knew this mm-hmm. <affirmative>, um,
Paul 00:31:35 The near near com de composability, uh,
Alicia 00:31:37 Relations of constraint. They’re not relations of energetic exchange. Right. So the higher level, it is a form of constraint on the possibility space of the lower level. So it doesn’t let it veer off and, and disin disintegrate. Mm. It’s interesting. The word integrity Yeah. Fits beautifully here because it just holds it together. Yeah. So when I think of top-down constraints, those are limiting because they hold that possibility space in within a dynamic range mm-hmm. <affirmative>. And so the thing hold an organism holds together, our metabolism might change. Uh, you know, a spleen might be removed. A a, you know, a gallbladder might be removed. <laugh> with a constraint regime holds.
Paul 00:32:30 Can you name something that is not an emergent property of something else?
Alicia 00:32:36 <laugh>? Oh, culture.
Paul 00:32:37 Culture is not an, an emergent property culture
Alicia 00:32:39 Is an emergent property. Do you want
Paul 00:32:40 Me to No. Can you name something that’s not an emergency? Oh,
Alicia 00:32:43 Uh, quantity is, is is emergent, but it’s not an interesting emergence.
Paul 00:32:48 Okay. So Oh, that’s fine.
Alicia 00:32:50 No, thank you. That’s a good point. When I think of emergent properties, the emergent properties that I’m, that I, that I’m interested in are qualitatively different properties that affect the behavior of the component level so that that qualitative property continues in effect. Okay. And that’s why homeostasis, I think of as a, an merchant property that has interesting features because it controls the goings on at the lower level, such that homeostasis, homeostatic balance holds. Whereas a whole bunch of, of, um, pile of debris in my backyard,
Paul 00:33:40 It’s still emergent with respect to, ’cause the size, it’s what
Alicia 00:33:45 Size? It’s just bigger. Bigger.
Paul 00:33:46 Well, it’s bigger, but if you throw a bowling ball into it, it would be harder to move than if you just had one piece of debris out there. Right. So is that Well, that’s true,
Alicia 00:33:53 But it does not change the characteristics of any of the
Paul 00:33:58 There’s no top down, there’s no, yeah, there’s
Alicia 00:34:00 No top. There’s, see, and I’m, that’s, that’s what’s been driving my thinking for 20 some years. Yeah. Right. And that is, how can you explain top down causality without falling victim or, or making the mistakes of al uh, all that other stuff. And, and that’s what I think complex dynamical systems gives me. And, and I think, I think the rise of ecology in the last 25 years has helped enormously because we used to think, oh, well, we can kill off the plankton. That’s not a big deal.
Paul 00:34:32 Right? Yeah. Yeah. Correct.
Alicia 00:34:34 And the, and the thing is, the, the understanding that it may mess up the constraint regime that holds the whole thing together. Um, it is what the idea that there is, that there is there, that there are systemic properties
Paul 00:34:50 Yeah.
Alicia 00:34:52 That really affect the components. I think that’s gone a long way to making us aware of top-down in a respectable scientific way.
Paul 00:35:07 Uh, on the other hand, one of the features of complex self-organizing systems is their robustness. So if you did kill off all the mosquitoes, we might be okay.
Alicia 00:35:17 I suspect though, I, I think robustness means that they degrade gracefully.
Paul 00:35:25 I like that,
Alicia 00:35:26 That, that it doesn’t, it doesn’t fall apart the way if you break the axle on a bicycle that, or tric or whatever, that will just, the thing won’t work at all. Right. These, it seems as though complex dynamical systems far from equilibrium, they, they get a bit outta kilter, but they continue to work for a fair amount of time until that then it reaches its level of instability that it, it, that that constraint regime cannot hold
Paul 00:35:56 A new phase change coming up.
Alicia 00:35:57 And you get an entirely new phase change and falls apart. So I suspect that if we killed off all the mosquitoes, we do not know what might enter that, that open niche now. Mm-hmm.
Paul 00:36:11 <affirmative>, right? Yeah. Well, there’s been all sorts of disasters with introducing in, um, non-native species to kill off one thing, and it turns out they’re invasive. Exactly.
Alicia 00:36:21 Exactly. I, I suspect we’ve gotta be careful because of that kind of problem before the thing falls apart, you might have some really nasty surprises that you didn’t expect. <laugh>.
Paul 00:36:32 Yeah. Hey, you mentioned homeostasis and we, uh, again, we’re gonna jump around, but, um, I, I had Henry Yin on, and I’ve had a few other people on who talk about, um, the brain more as a control system than necessarily a complex system. But we need to think about things more in sort of a, the old cybernetics way of a, a control system. However, the, the set point of our thermometer is not set externally. We’re setting our own set point. Right. So homeostasis is
Alicia 00:36:59 In certain balance. Well,
Paul 00:37:01 Well, that’s what I was gonna ask you about. Like, let, how do we, how do you think rather about, uh, is if when we’re setting our internal desired temperature, let’s say that we need to be around 98.6 mm-hmm. <affirmative>, how do you think about that? Where does that set point come from bottom up, top down? Or is it a just the, i
Alicia 00:37:19 I think every not only comes up bottom up, top down, but pass forward,
Paul 00:37:24 Pass forward environment in,
Alicia 00:37:26 We forget temporal constraints as well. Yeah. <laugh>, um, for example, folks up in the Himalayas and in the Andes have a whole different, um, oxygen, uh, hemoglobin
Paul 00:37:43 Capacity Yeah.
Alicia 00:37:44 Capacity and characteristic because of centuries of adaptability and millennia of adaptability. So we forget temporal constraints as well. Yeah. Um, right. It, it seems as though the same temperature in the fall feels colder than it does in the spring, because by the time we get to the spring, we’ve had a whole winter. So I think it’s multi-dimensional, multi-scale temporal temporally. I think it, it, it, we forget time. Yeah. We forget. So we, and all of these complex dynamical systems, I remember Fraser, PGA genes, I’ve, I, I was fortunate enough to meet, meet him a couple of times, and I, I’ve never seen this written down, but he used to use that phrase more than once in person, they carry their history on their back.
Paul 00:38:36 Oh, I like that. Yeah.
Alicia 00:38:37 I love that phrase. In other words, they’re very dynamical architecture and characteristics embody, and that’s where the four E’s come in, embody the conditions under which they were created, evolved, and so on. His examples were snowflakes. Mm-hmm. <affirmative>, they might’ve all started with the same, but over the time that they fell to earth, they, their actual architecture changes to embody the, the, the temperature, the pressure, the, all the atmospheric conditions through which they modified themselves and adapted and until they reached the, to the ground. So I think that’s true for us as well, for for, for body, for body temperature mm-hmm. For body states. I’m, I’m, I’m fascinated by that whole note, inflammation, all the, all these set points mm-hmm. <affirmative> that I think in a sense, those mark the parameters of what you might call the constraint dome <laugh>, you know, we’ve got the proteome and
Paul 00:39:48 The, oh, there it is. There’s the word, the constraint dome constrains.
Alicia 00:39:52 I just, I just, it’s, I was reading proteome and Yeah. Epigenome. And so Well, we know we need a constraint,
Paul 00:39:59 Dom. We do. Well, that’s all in your book. The constraint dome is Yeah. But
Alicia 00:40:03 I don’t, it never occurred to me to use that phrase, which I like, because it’s a whole bunch of them together. Yeah. Um, at once.
Paul 00:40:13 Right, right. Well, that’s the pro that’s, that’s the problem in, in sort of just mentally trying to imagine these things. But real quick, you, you know, you mentioned temporal constraints, but in a sense, all, and this is a question all or most constraints are temporal in nature because a constraint is by definition slower than the process that it’s constrained like a catalyst. Right. Is, um, more, uh, stable than the reaction that it’s catalyzing. And that’s, and that, so it can act as a constraint. Correct.
Alicia 00:40:42 But then the cells which enable the tissues, the cell dynamics, which enable the tissues, that’s faster than the Yeah. Than so, so you have a whole bunch of, so therefore, what’s important are the interfaces between the levels of organization and what are the root, what are the constraints that govern those interfaces? And we would call them what codes, rules, algorithms.
Paul 00:41:12 Laws. Laws.
Alicia 00:41:14 Yeah. But local, but local. Very, yeah. Oftentimes very local <laugh>. Right,
Paul 00:41:19 Right. Well, that’s, so, okay, there’s this, a lot of the physical laws, um, like, like, you know, the notion of a force, right? Yeah. For some reason. Well, when you think about it, it’s, it doesn’t make any sense. But for some reason, when you grow up with it in physics classes and you think, oh, force, I know what a force is. Right. But when you really think about it, it becomes very unnatural, like saying the same word over and over, and you lose its meaning. Are these ideas going to beco, um, seem as natural to me eventually as the concept of a force, which is a, turns out a slippery idea also.
Alicia 00:41:51 I would hope so, but <laugh>, but that’s what I’m always worried about, about, um, neologisms and, and, and, and, and I worried about using the word constraint, but it seemed as though it was,
Paul 00:42:04 What else would you call it? Better
Alicia 00:42:05 Than a cause? Because as if I, if I use the word cause then the assumption that that has to be energetic transfer, that’s almost a, that’s a losing battle. That’s what mm-hmm. <affirmative>. So I don’t know what else to cause call it, but to think in terms of enabling constraints as bottom up, that that, and that within certain boundary, within certain context, independent frameworks, they can generate emergent properties, which in turn operate as second order top down context dependent constraints, if you will. <laugh>, it’s easier to just call ’em constraint regime.
Paul 00:42:46 Yeah. Okay. Regime’s a good word. Yeah.
Alicia 00:42:48 Hold the, the components together so that they continue to satisfy the emergent properties of the whole.
Paul 00:42:58 Do you feel like, we’ll, we will be able to write a constraint cookbook to generate dynamical complex systems that we want to generate eventually,
Alicia 00:43:09 If we think of it as formal systems the way we used to, where you would have a universal law. Yeah. Like Newton’s law’s of motion mm-hmm. <affirmative>. And I think my answer is no, because if I’m emphasizing context dependence, those laws would be what car Popper would call. Well, yeah. Those applies to, to, to a normal science, but evolution to a whole new constraint structure, to a whole new constraint regime, such as the transition, I would say from physics to chemistry, chemistry to biology, biology to psychology first or sociology first, or whatever you wanna call it. And on and on and on. That requires a whole new constraint regime. It’s a, it’s a, it’s a top topological
Paul 00:44:05 Nightmare. <laugh>
Alicia 00:44:06 Nightmare and transmission. You know what’s interesting, Preco Jean detested Cata catastrophe theory, and I never understood why this idea that, uh, uh, Peter Allen, who worked with him for PGA for a long time, you know, Peter Allen’s worked Peter Allen, really nice work. He’s retired now from Cranfield in the, in, in, in England. But he did a lot of work with PGA in Brussels at the free university. And the idea that the transition from independent photon streams to a laser, the transition from a independent water molecules to a benign cell, the transition from a whole bunch of, uh, chemical transformations to a living organism is a, is akin to a, to a bifurcation, to a, a topological bifurcation that you have entirely new coordinates that apply at this new level. That to me, gives reality to emergent properties.
Paul 00:45:20 Oh, ontologically real. Yeah.
Alicia 00:45:22 Oh, I, I wanna say ontologically read.
Paul 00:45:24 Okay. It’s dangerous.
Alicia 00:45:25 I don’t have to worry about about tenure, so I don’t have to worry about, about, about retreating to epistemology when the going gets stuff said, well, well, you know, that’s our models and that’s the way we make sense of, of information. No, uh, overload it. So, no, dammit, I wanna say it. They’re on, they’re ontologically real. That, that the constraint regime that holds a benal together or a culture together or a, uh, laser together are real, they are not stuff, they are not things. I certainly don’t wanna reify them. Hmm. But I don’t see why we don’t allow the dynamical interdependencies, which is all these constraint regimes are. Right. They are, they’re interlocking interdependencies.
Paul 00:46:17 I thought, I thought refi meant make real
Alicia 00:46:20 Oh, uh, info. Well, yeah, sure. Part of it, but, but then only stuff, stuff is real.
Paul 00:46:27 Oh, okay. Okay.
Alicia 00:46:28 So
Paul 00:46:28 Therefore, processes aren’t real
Alicia 00:46:30 Process. That’s why white haired rails against reification, because he needs to emphasize that processes are real, and I wanna say, and the constraints that keep processes within a certain, um, space, face space are also real. They are as real as the face space they’re kept into.
Paul 00:46:53 Which one do you give? Do you give more credence to processes? Or are you a process philosopher, so to speak? That’s
Alicia 00:46:59 A good question. You know, dynamical systems theory seems to place an, if nothing else equal due to the links as the nodes. Yeah. And the question of whether the nodes are condensates, mark Auer uses that term, the con condensates of links of, of, of a lot of, of stable linkages.
Paul 00:47:24 Oh, that’s cool. And we think
Alicia 00:47:25 Of nodes as, as stable condensates, no, as nodes are condensates of stable links or those that, that, that, that happen very often and so on. So it’s hard to say, um, in this new book, I stick my neck out because I did not realize that there’s still a question about exactly what covalent bonds are.
Paul 00:47:49 Oh, right. Yeah. I think when you, when you dig into any established thing, there’s, you realize there go, yeah, there
Alicia 00:47:55 You go. And the intercon conversion of matter and energy certainly says, well, you know, so is a material thing simply the condensate of a bunch of constrained energy forces? Right. Links and constraints, or, or are the forces and the links the manifestation or the behavior of the, then if relativity theory is correct, which I then it doesn’t matter, they’re in a convertible, right? Mm-hmm. But again, they’re very local. See, I wanna, I, I was thinking how on earth can I argue for a local science? I’m gonna be laughed out of the room if I argue for local science,
Paul 00:48:43 But this is kind of what, what Bill Wimsatt hints at in the causal thicket. Right. Thank
Alicia 00:48:47 You. Thank you, thank you, thank you. Oh, I have a sweet spot in my heart for Bill Wimsatt, even though I’ve never met him, because I learned that he, I dunno who told me or how I’d learned about it. He was on the committee that approved my first paper.
Paul 00:49:05 Oh. Oh, it’s
Alicia 00:49:07 In theory, rest on a mistake. And, and so I think that I published that first paper so many years ago.
Paul 00:49:15 Wow. That’s awesome.
Alicia 00:49:16 I think Wimsatt is, you know, and that brings in the notion of explanation, how he’s been arguing that sometimes explanations look upwards, not just downwards. Mm-hmm. <affirmative>. Why? Because if you do believe that there are emergent properties and they are causally constrained, <laugh>
Paul 00:49:37 With the eye roll folks who are resonate. Exactly.
Alicia 00:49:40 Then you have to be able to make reference in any complete explanation to its influence as well as to the influence of the bottom up components, which is the way we tend to, we always look downwards. Yeah. Um, and let me tell you, I think the reason folks like David Chalmers and company all this, this is about psycho, uh, the, the, the, what’s it? Panpsychism
Paul 00:50:10 Panpsychism. Yeah.
Alicia 00:50:12 I think the, what’s driving that is that anytime you find top-down causality, you cannot much deny that’s gotta be mental. Well, let’s put it differently. It’s gotta be top down, and it’s a constraint regime influencing its components. But I don’t want to say that they’re just as chemical properties are emergent from physical, biological properties are emergent from chemical, psychic properties are emergent at a certain level of organization. Yeah. I don’t wanna say that that emergent property way up at the psychic level also exists. I really don’t wanna say the, the, the Bernard cell is thinking
Paul 00:50:59 <laugh>
Alicia 00:50:59 What I, or is conscious or is aware. Yeah. But I do wanna say there is a top down phenomenon, and I think I suspect that a lot of what’s driving the pans psychic,
Paul 00:51:13 Um, what would we call it? Um, fad.
Alicia 00:51:16 I hope so. I cannot imagine that, that anybody saying serious. I think it’s a, I think it’s a, it’s, uh,
Paul 00:51:24 Paradigm. That’s all I can do. Let’s call it paradigm to give it a little more credence than fad. That sounds No, no, no.
Alicia 00:51:29 It was a paradigm way back in.
Paul 00:51:31 Whatever. I don’t wanna get in trouble. Okay. I have to <laugh>
Alicia 00:51:36 Credibility. I don’t wanna, you know what, talking about that, I just spoke at the Theory of Consciousness, uh, conference in Tarina Oh. This past spring. And I suddenly realized I said the same thing. I, I talked about, I talked at that conference in Antwerp 20 some years ago, and I was laughed out of the room 20 some years ago.
Paul 00:51:58 Oh, wow.
Alicia 00:51:59 And I was talking about constraints and dynamical systems and so on, going now, interesting. You get a lot of this talk about constraints there. I mean, people were talking a lot about constraints.
Paul 00:52:11 What, yeah. Well, what is the lesson from that? Because you’ve stuck to your guns and did, did you just brush that off your back? And this is an aside, it’s more of a personal question, you know, because a lot of people face laughter or dismissal. Well,
Alicia 00:52:24 I wasn’t, I wasn’t literally laughed at, but I was ignored. No. And nobody asked any questions, and they just said, it’s, it’s the equivalent. People were polite, but maybe
Paul 00:52:32 They didn’t understand it.
Alicia 00:52:34 Well, I didn’t explain it clearly, but I think it was just, it didn’t, it didn’t, it didn’t fit the, the standard bottom up, um, identity theory or convenience theory, which was the, what was popular at the time, or I never can say hard at the time. Uh, so it was, it was ignored, just as, I hate to say it. Absolutely. You know, bill Tel, who’s, who’s a nice guy, and he’s very
Paul 00:53:04 Good. He’s a, a new mechanic. He’s a mechanisms all the way, uh, uh, proponent,
Alicia 00:53:09 Well, if you look at his mechanics, they’ve sort of changed. And, and he came to Cuba. He and Adele came to Cuba with me in a conference I organized in Havana years ago. And I think his mechanism is not your grandmother’s mechanism.
Paul 00:53:25 That’s true. But he’s adapted, like you just said.
Alicia 00:53:28 I, I, but, but it’s sort of like, come on, bill, just acknowledge complex dynamics.
Paul 00:53:34 Can’t do it. It’s can’t.
Alicia 00:53:36 Well, I mean, it’s hard to, it’s hard to change your mind when you’ve done it so long. Yeah.
Paul 00:53:42 You mentioned explanation a few minutes ago, and Yeah. So in dynamics and action, uh, the, I think the last chapter or the next last chapter, you make a call for the necessity of, um, explaining these sorts of meteorological, um, interactions in terms of hermeneutics or, or narrative story. And, and in this newer book, I, I, I think I searched and I think Herman Ick was in there maybe once or twice, but you also talk, you bring up the idea of order parameters, which are kinda like these hyper parameters of dynamical systems. So I’m wondering where you stand in terms of explanation. My, my original thought was, well, if I could just list all the constraints and the types of constraints and how they’re interacting. Is that an explanation? Do I need a narrative story? Are these order parameter where, what would satisfy me? Or a, uh, an a, um, an a, uh, a reviewer of a paper as an explanation of some cognitive fun, uh, function, for example?
Alicia 00:54:37 Well, well, I, I used the term hermeneutics, um, <laugh> pri,
Paul 00:54:43 I hate another eye roll. Another eye roll. Yeah.
Alicia 00:54:45 I mean, and I don’t like neologism. Um, although that hermeneutics goes back to biblical,
Paul 00:54:52 Right? Well, that’s, yeah. When you look it up, it’s all about the Bible. But your point was to go across levels, you really need to account for these things. And narrative might be the only way to do it.
Alicia 00:55:00 TZ is notion that I borrowed that back then from Cliff tz who almost defined hers. And instead of you, it’s kind of like sailing. You have to attack from the very global to the very local, back to the global, back to the local, and see how all these constraints interact and change each other. And I love also the notion of mutual constraints satisfaction, which is what I think is happening. I think probably narrative does a way better job of accomplishing that than a hempel and Oppenheimer Hempel and Oppenheim deductive numerological model. Samuel.
Paul 00:55:37 Yeah.
Alicia 00:55:38 Yeah. So to that extent, I don’t think I’ve changed my tune. I just, I’m not using the notion of narrative. When somebody nowadays says, Proust Madeleines, you bring up the whole constraint regime of the culture, of his history, of his childhood, of his experience of that to me, is almost like an order parameter. Mm-hmm. <affirmative>, it’s, it’s a parameter in that it, it, it is the intersection of a whole bunch of different independent variables that together give you a whole tapestry of a whole different face possibility space.
Paul 00:56:23 Could you think of a concept as a constraint?
Alicia 00:56:26 I think, I certainly think of a concept of a constrain and a framework, a cognitive framework being a whole, um, a tapestry of concepts, but in such as cognitive, it’s a emotional, I mean, think of, think of olfaction and olfaction does not, uh, go directly to the cortex initially, right? Hmm. But I think a concept ha concept has a lot of characteristics that are important. One is they’re multiply realizable.
Paul 00:57:03 That’s a big theme in your book, multiple realizable.
Alicia 00:57:05 That’s a big theme in my book, because a triangle or triangular as a concept mm-hmm. <affirmative> is multiply realizable. It’s not a one-to-one capture of the individual physical features of one instance. And I think that we know the child is able to, has understood the concept when they’re able to apply that concept to an instance they haven’t seen before, or they were not taught about mm-hmm. <affirmative>. So I think that’s, that’s extremely important. I think what a co a local constraint regime does is it captures a type what, what philosophers used to call types and universals, or a concept or what we would call an emotion. Uh, uh, not this emotion. Now here
Paul 00:57:58 An emotion. Yeah.
Alicia 00:58:00 You know, this type of emotion, right? It’s a bunch of constraint regimes that, that can be identified with a term like fear. Mm-hmm. <affirmative> or anger or pleasure or, um, avoid and a, uh, and, and, um, approach. All, all these are code terms to that extent. The, the, the reductions. They’re code terms, but they do capture something very real. And that is a whole complex constraint regime
Paul 00:58:35 That kind of swirls swirl around this, uh, let’s say fear or something it takes with it, um, you know, fight or flight, et cetera. And it’s almost like a, a process that it, is it still itself kind of condensated to use your term earlier, uh, but still a process and brings with it, carries with it on its back, uh, its history and all of things associated’s
Alicia 00:58:55 History, correct. Correct. Correct. Yeah.
Paul 00:58:59 Um, okay. So I’m <laugh>. So I didn’t tell you this. Um, and I know we’re jumping around. I actually elicited a few guest questions from, uh, previous guests. Uh, the first one I’m gonna play for you, um, and this is kind of, kind of maybe segue into a conversation about, uh, minds and brains and then later ai. The first one I’m gonna play you is from Kevin Mitchell. I’m gonna on the podcast, um, in a couple weeks about his new book. Um, yeah, yeah. Uh, agents, free Agents. Yes. Uh, how Free Will Gave Us Agency. Yes, yes. I’ll have to look it up. Okay. So I’m just gonna play this to you. Make sure that you can hear it.
Speaker 4 00:59:34 Hi, Alicia, this is Kevin Mitchell. I really loved your new book, which very clearly lays out the case that constraints are causes. That is, that the way things are organized is every bit as much of a cause of how they evolve through time as the basic laws and forces of physics. Mm-hmm. <affirmative> in particular, you highlight the really key idea of enabling constraints, especially in hierarchical systems, that organizing the elements at one level can enable new kinds of properties to emerge at higher levels. You say in chapter 16, that emergent properties and powers were not merely waiting to be revealed. They are genuine creations of path dependent, multiple constraints, satisfaction. And I wanted to push back on that a little and suggest that in fact, many of the organizations we see in living organisms and design systems are there because of some functionality they serve a useful purpose. And that many such functional motifs or architectures are in fact waiting to be revealed under this view. A lot of the design work done by evolution is actually an exploration of a kind of platonic design space, followed by retention of useful motifs rather than a real creation de novo. So I wonder what you make of that notion. Thanks.
Paul 01:00:49 All right. So that was a little long one, but um, did you get it all?
Alicia 01:00:52 Yes, I think I got it all. Um, that’s a heck of a question, because I think logically of course there is a platonic, maybe undercurrent, but, but in fact, I, I think I’m more like Aristotle did not believe that Plato’s forms existed in a transcendental realm, and I wanna stay agnostic about that. But I certainly want to say that they are embodied and revealed and acted upon in virtue of the, the confluence of these constraints occurring at a particular moment, at a particular, uh, time locally <laugh>. Um, you know, I, I think I used to think of myself more as a platonist when I remember getting into philosophy, and I got into philosophy because back in the day, a lot of schools allowed you to, if, if you were not a science math major or engineering major, you could substitute a, a symbolic logic course for a math requirement as one of your requirements.
Alicia 01:01:59 I thought anything to get rid of, not to do math, which I am sorry I did. Um, but, but when I took the first philosophy course, it was like discovering a whole new platonic world. It was like, you know, if I’d taken another history course, yeah. You know, it’s another king, another queen, another war, another this, but this was like a whole new floor in a building that I’d never seen before. That’s that, that kind of transformation. Um, but more and more I wanna kind of remain grounded and, and rooted to, to natural phenomena. So I really want to naturalize my arguments and speculating about a platonic transcendental realm.
Paul 01:02:48 Well, he’s said a kind of platonic, a kind. Yeah. So maybe he, he hedged a little bit.
Alicia 01:02:51 Well, if that’s the case, then I’m, I’m, I’m perfectly happy with his book that way. Um, so that, what is the kind, well, in Aristotle’s case, it’s the, it’s the Aristotle believed in forms, it’s just that he believed in embodied forms and, and forms as they were. And so what I want to say is that they are created by this confluence of constraints. Um, my main concern is to make sure that they are taken seriously, such that meaning is taken seriously. In other words, um, the, an embedding organizational level gives meaning to the components. Um, the amount of glucose in a fudge brownie that I can eat and not get sick is given meaning by what my homeostasis metabolism is operating as. Mm-hmm. <affirmative>, uh, well, yeah, you’re okay now, but don’t do it very often kind of thing. So that meaning is the multiple realizability of the ambiguous aspect being disambiguated by context, but, but not, not, and I use the word context very deliberately.
Alicia 01:04:17 I don’t want to use context to mean any old environment. I wanna mean the environment in which a system is embedded, and from which it has drawn some of its context. Its enabling constraints and in terms of which it implements, its, um, stabilizing or top down, uh, constraints. Mm-hmm. <affirmative>. So, so I, I wanna say these are, these are new and the properties are real. So that, um, yes. Super conductivity and, and a laser has brand new properties Yeah. That are real. Did they exist in a preexisting platonic realm? I, I, I don’t think I wanna speculate on that. I’m perfectly happy as long, and to me real means it has causal power. Okay. I mean, it has some kind of, it, it, it makes, it makes a difference if it’s definitely phenomenal, you know, then I don’t, that kind of emergence Yeah. Weak emergence, that kind of emergence doesn’t interest me that much.
Paul 01:05:27 Okay. I, I don’t wanna speak for Kevin, but, um, kind of reformulating his question. You know, a lot of what you and the field and dynamical systems theories talks about is these state spaces. And when you’re on a dynamical, like when you’re on your, a trajectory within the state space, within the Yeah. You can visit different areas within the state space, not all of them. And the, and the constraints constrain where you can visit in the state space. Right. But if you, I wonder if that pla that kind of platonic realm that Kevin is alluding to is, you can think of it as the state space that, um, whether a state space is a real thing, right. On a, a trajectory later we can say that we can define the state space by the trajectory and the, the, the boundary conditions of the possible values that it could have taken. But if it’s not visiting, uh, a particular area of the state space, would that, and it did, would that then be a new thing or would it be visiting a thing for which it already had a capacity for, and thinking in terms of like the functions that Kevin was talking about.
Alicia 01:06:26 I think that’s what normal science is. Normal science is the visiting of areas in possibility space. It hadn’t visited before, but it’s pretty much the same possibility space. I think that’s development.
Paul 01:06:41 Ah, okay.
Alicia 01:06:41 Biology. Mm-hmm. <affirmative>, I think evolution. I think there are times when it’s the entire possibility space that is new. That’s what I wanna call a, an a creative emergent characteristic.
Paul 01:06:57 Is this related to Stuart Kaufman’s adjacent possible idea? Maybe.
Alicia 01:07:00 Maybe. Okay. Whether, whether we go step by step, you know, going Yeah. Each adjacent possible, or in the case of a Bernard cell, it doesn’t seem that it’s a adjacent possible. Right. It seems as though there’s a threshold of instability. Yeah. Beyond which you need to reorganize all the parameters, all the variables, everything acquires a different meaning because it’s in an, and it’s in a qualitatively different space. And it seems to me that that’s what happens, for example, a cell in a Petri dish as opposed to a cell within a tissue. Hmm.
Paul 01:07:40 Okay.
Alicia 01:07:41 It’s a whole, it’s different qualitative phenomenon. And so it’s, and and it’s the, it’s the new possibility space that I wanna say is being created is an emergent quality. Ah, okay. It’s not just a new corner in the existing possibility space. I’m perfectly happy going adjacent possibles. Um, ELA two Kaufman. Yeah.
Paul 01:08:07 This is from Mac Che who’s been on the podcast. He thinks, uh, he, he calls himself a systems neurobiologist and he really tries to think across scales. And a lot of his research is dealing with things like how neuromodulation interacts with, uh, network level things. So these things are hard to think about and to research, uh, obviously. So let me pull up this question.
Speaker 5 01:08:30 Hi, Alicia. This is Max Schein from the University of Sydney. When we think about top-down constraints over bottom-up activity, we often invoke concepts like the self. So I can imagine myself being hungry, and that will help constrain the types of actions that I might take in order to find food and, uh, satiate my hunger. But there are a number of really fascinating situations like the so-called flow state, uh, and also states with pharmacological manipulation when the sense of self can dissolve. Yeah. Yet you still see quite coherent cross scale behavior. People don’t just crumple into a heap. They can still perform music and sports at a high level. Mm-hmm. Mm-hmm. Or have a nice walk through the forest. So my question is, do you view the sense of self as an illusion, or do you think of it just as one of many ways in which a system that’s organized across scales can be organized? Thanks.
Paul 01:09:20 And Mac wants to know whether his question was better than Kevin’s No, I’m just kidding. No, no.
Alicia 01:09:24 <laugh>. No, they’re very good. Both of them. You know, I keep thinking how with Gaga’s experiments with the cut corpus Cossum, that there were a whole bunch of weird things like we discovered about somebody was shown an image to the right eye, another, another image to the left eye, and people confabulated. And so you wonder if in these cases where the self seems to dissolve, um, there’s nevertheless enough connectivity to maintain a coherent structure, um,
Paul 01:10:07 Structure being the self, what’s structure?
Alicia 01:10:10 Structure, I would say a dynamic, the whole dynamic constraint that keeps being you to a lesser example for I think people who sleep walk or who are hypnotized will nevertheless not bump into the coffee table. Hmm. They go around the coffee table.
Paul 01:10:32 But these could be sort of the on the automatic side of behavior, right? Yeah. So there’s automatic and controlled processing. Right.
Alicia 01:10:37 Right. So that, well, but there is control enough that you walk and you don’t, you, you, you, you
Paul 01:10:44 Yeah. Maintain Right. Still well, but maybe the flow state would be, ’cause Mac, uh, so, um, mentioned the flow state when that is like when your automatic processes take over and you, but you still barely have, have some sort of control and you’re in a heightened Yes. You’re trying really hard and it kind of switches, and then you’re watching yourself do it kind of. Yeah.
Alicia 01:11:03 Yeah. And you wonder whether we can, could that be a bit like these optical illusions where you can see one side and not the other? And so, so to what extent do those constraints to, to what extent do those, um, levels of awareness conflict with each other can mutually support each other and reinforce each other or not? Again, I think that, that the next stage of this is going to be a study of all the different modes of constraint that are all operating simultaneously. As I started, as I started writing this, all of a sudden things like entrenchment, like buffering, I have a whole chapter of where I catalog some of some of these buffering and entrenchment and scaffolds. Scaffolds and all of these that seem to all be operating simultaneously. And that certainly a lot of the social scientists have understood very, very well mm-hmm.
Alicia 01:12:08 <affirmative>, you know, in, in, in, in public policy. Oh, look, if we can entrench this rule and regulation, it’ll, it’ll hang on for way longer and, and avoid any kind of attempts to, to remove it, uh, as a regulation that I live in dc. That’s why I think about it. Uh, and so you wonder starting to map things in terms of constraints, in terms of possibility spaces might start to illustrate how all of these work and the possibility of therapeutic inver intervention as a result. Because that’s what all this recent talk about psychedelics. Yeah. No, you wonder to what extent a psychedelic literally disintegrates the constraint regime of the self of, of the pre-established coordinates and parameters and so on. Yeah. The ego, right? Yeah, exactly. The ego. Um, you wonder to what extent that that is a, uh, that is a disruption of the constraint regime more than the disruption of vision or Yeah.
Alicia 01:13:22 Vision too. You get maybe hallucinations, but, but that’s not as interesting as, you know, I go like that and I push my eye. I can hallucinate all these standard halluc, you know, optical illusions. We don’t call ’em hallucination, call ’em optical illusions, right? Yeah. So at what point do we start looking at this, and, and the answer is, that’s gonna take a lot of very interesting research, kind like Ram Chandra’s famous example with the phantom limb and the mirror and so on. You’re going, how is that possible? Unless there is some control regime that is modulating the sense of pain that incorporates, you know, I can imagine being in an automobile accident and my arm mangled, or my ear these descriptions Yeah. Of these IEDs and Oh,
Paul 01:14:17 You chalkboard scraping and things like that.
Alicia 01:14:19 Correct. And so you go, well, here’s everything together. Because if you didn’t have the visual input, if you didn’t have the conceptual understanding of what this means on and on and on, you wonder if the pain would be, would be different. Uh, and so how all of these constraint regimes, um, interact and what are the necessary inputs at each point to maintain that particular constraint regime so that then you have a global sense of self. Um, that I’m not sure we’re all that coherent when every, in, every day,
Paul 01:14:58 I don’t feel coherent, but I Yeah, exactly. My wife would beg to differ. We
Alicia 01:15:01 Aren’t enough Yeah.
Paul 01:15:02 To, to,
Alicia 01:15:03 To be operating. But it’s not like, you know, I heard something horrible happening outside. I wouldn’t react. And I even, even in a sense of flow, I think there’s enough cognitive control that if some horrendous perturbation, Hey, the standard answer is you drive home and you don’t even know if you went through that red light or you don’t because you drive home in that, in that path every day, and you don’t even think about it. Yeah. But certainly if a child ran across your, your front of the, then you’d immediately get out of that flow and, and, and react. So, so there’s how much control, in which case, how can we use it therapeutically for obsessive compulsive disorders, for P T S D, for all of these? I think those are all disease, um,
Paul 01:15:57 Disorders,
Alicia 01:15:58 Disorders of constraint regimes. So
Paul 01:16:02 Regulation
Alicia 01:16:03 Of regulat, it’s a, it’s the regulatory system that’s shot. It’s like the on switch is on all the time, or the, it isn’t modulating it the way it’s supposed to modulate it. So these are regulatory meaning constraint regime, um, issues, not structure, not function. It’s the regulatory constraint regime that modulates these other two that, that we don’t understand that much about. I think O O C D, uh, P T S D, chronic anxiety or depression, all, all of these that are sort of, I think we’ve pretty much taken care. Well, no, we haven’t taken care of it, but we understand how it worked. The low hanging fruit of the, you know. Yeah. Well, you know, the glucose is outta whack that, that caused it. Right. But where stress or epigenetic stress can be inherited several generations down, that’s a regulatory issue then. And, and that’s what methylation does, and that’s what I think a lot of these constraint regimes do.
Paul 01:17:05 So the, there’s this phenomenon sort of piggybacking on what, uh, Mac was asking about. There’s this phenomena in, um, psychedelics, right? You can use them therapeutically, and in some cases, like near death experiences serve as therapeutic events and people come back kind of different, right? Correct. And is this a way to, is one way to imagine that, and maybe even electroshock therapy is another example. I’m not sure if that’s a good example. No,
Alicia 01:17:26 No. I think it, yeah, yeah, yeah.
Paul 01:17:28 So you have all these kind of entrenched, uh, constraints locally in your brain and in your mind and operating things. And then when you, so I’m sure you’re aware of the psychedelic study, that when you, uh, administer psychedelics, often what happens is there’s higher connectivity in your brain, less modularity. Right. And could, could you, maybe this is speculation, but would you think of that as just resetting all of the, um, uh, attractor states and then when they come back up, they can come back in slightly different ways? They might not.
Alicia 01:17:55 Probably. Probably. And it might open up, it might make the interfaces more porous to contemporary experience. So you don’t have to be worried that the explosion is the I e D or that, uh, right. That, that you don’t have to wash your hands a million times in order to, because you’re not Yes. I, I, I think that the, that more and more as we read these kinds of novel forms of therapy that they are, they usually are targeting issues of, of regulation.
Paul 01:18:30 Hmm hmm.
Alicia 01:18:32 Which is really fa to me, fascinating. Yeah. Yeah. To sense, starting to study how exactly that connectome stays, stays whole stays
Paul 01:18:43 Constraint. Dome you mean cons, connect, constraint,
Alicia 01:18:45 Dome, can they all, because all the connectome, the proteonomics and all the rest of them,
Paul 01:18:50 All the ohms, all
Alicia 01:18:51 All the ohms then are modulated by the whole constraint dome of the, of the entire, and but then also includes the constraints of your society. You know, I, I love traveling to Europe. When I travel to Europe, the weight of history is very obvious compared to the United States. You ask, uh, somebody at a restaurant in anywhere in Europe, and how do you prepare this dish? Half of the time it’s the way it’s always been prepared.
Paul 01:19:23 That’s right. <laugh>. Yeah.
Alicia 01:19:25 Whereas in America, if you do it the same way as you did it six months ago, oh, you’re stale, you’re boring, you’re old <laugh>. It’s encouraged the, the constraint regime. Now, that’s not to say that they’re not all sorts of problems for marginalized societies where the constraint regimes there are very clearly kept entrenched very deliberately. Yes, I understand that. But it’s all a question of the constraint coming from the culture as well. That’s why in the four E’s, the whole, in, for example, their emphasis on affordances,
Paul 01:19:52 This is the, so I don’t know if we said it earlier, this is like the inactive embodied, eh, what are the other two? Oh, the four is, anyways, the embodied cognition, correct. Uh, yeah.
Alicia 01:20:03 Embodied cognition, which Andy Clark and Paolo and, and even Merlin. Where about Gwen
Paul 01:20:09 Ex Extended,
Alicia 01:20:10 Extended, embodied enacted.
Paul 01:20:13 There we go. We got three outta four about that.
Alicia 01:20:14 Got three outta. You know, I always say I frankly do not think that a chair affords opportunities for seating for a lot of people. You hear stories of folks from, I don’t know which culture, who were brought to the United States, and they will stand on the toilet before. Oh. Because it didn’t occur. And I don’t think at the Tommy mat affords opportunities for sleeping for those of us who were not raised in a Japanese, I mean, affordances are clearly embodied, but embedded in a, in, in the constraint of the society in which you grew up, or you know about
Paul 01:20:54 Embedded, you just said the fourth e It’s embedded. It
Alicia 01:20:57 Is. Okay. There you go. I, and I think how things get, see, I don’t think you have embeddedness in Newtonian science. Right? Right. Things are just plunked into a featureless space and time in Newtonian science. Yeah.
Paul 01:21:11 Whereas, but that’s the beauty of it too, because it’s sim simplicity, right? It’s correct. Spherical cows. It’s in a vacuum, it’s, yeah. Correct. Correct.
Alicia 01:21:18 But it’s, it’s featureless. Whereas embeddedness is, you’re really embedded, meaning that embedding layer affects you and you are Yeah. A part of that embedding dynamic as well. It goes both ways.
Paul 01:21:38 All right. I wanna ask you, I, because I, and we’re coming up on time and I have a thousand other things to ask you, but let’s stay on brain in mind for just one more question here, and then I have some I wanna ask you about artificial intelligence, um, and, and your thoughts on that. So thinking about dimensionality, um, and going up in these a neurological level, right? So a lot of this modern neuroscience is reducing the dimensionality of your brain activity into these low dimensional manifold structures. And then the dynamics come in when you have a trajectory along that manifold. And a lot of, some people think that that’s, well, maybe that’s the closest thing that we can get to right now in describing, in relating brains to minds, right? Um, and then, but marrying that idea with the idea of capacity, um, and this goes back to Kevin Mitchell’s question. Um, what I want to think is that I’m not a low dimensional mind. I’m a, I have a high capacity, right? Is there a way, like, is there a story I can tell myself that, um, going from the high dimensionality brain activity that I have moving, quote unquote up a level to mind? And that can, that’s a messy statement. I know. Can I increase my dimensionality? Do I have to reduce it?
Alicia 01:22:46 No, I, I think for example, letters form words, I think of a word as a higher dimension. You can do more with those words. You have options available in words and keep going sentences. But that’s certainly true for atoms. Molecules, uh,
Paul 01:23:11 Is the higher dimension, does that ha Do you mean is it a higher s possibility space? Or is it, I guess that’s kind of equated to higher dimension?
Alicia 01:23:19 Yeah. Yeah. See, that’s what I think of. It’s a higher po it’s a possibility space whose coordinates are relational.
Paul 01:23:29 Okay. Yeah.
Alicia 01:23:30 And if that’s the case, yeah, then you have a way higher possibility space. So that’s why, for example, the, the example I used in the book, which is Monte’s, uh, understanding of how these chimps, uh, are able to understand reach,
Paul 01:23:46 Task
Alicia 01:23:47 And reach, because the possibility space they have forged is a task defined possibility space.
Paul 01:24:02 But it’s of a different type,
Alicia 01:24:04 You know? Exactly. The type is simpler in that you have one concept of a type that captures all the, well, you have to look at the color dots, you have to look at the movement of the dots and so on. But it’s the task that is defined and the behaviors def is executed in terms of that task. Mm-hmm. <affirmative>, is that lower dimensional? I don’t, I think it’s way more sophisticated than it would it were just in terms of the, uh, intensity of the flash of light.
Paul 01:24:37 But here, the task is a huge constraint as well. Yes.
Alicia 01:24:39 But it’s a, but it’s a, it’s a, it’s a much more complex, uh, dynamic than would be the physical features alone, right? Hmm.
Paul 01:24:51 Yeah.
Alicia 01:24:52 So it’s sometimes people say that complex systems are a simplified description of what’s going on. Sometimes I think of it more like streamlined.
Paul 01:25:09 Okay. <laugh>,
Alicia 01:25:10 The streamlining gives you characteristics that you wouldn’t have otherwise. And so the organization of that monkey’s brain in response to training, which would be the enabling constraints, Hmm. Alright. It seems to me opens up a world for the monkey that may be captured by the essential characters of the task, but not of the pure physical component alone. So in that sense, I think it, of it as an expansion, chemistry can do more than physics. Biology can do more than chemistry, even though they are nonetheless not violating any of the underlying laws. Right. It’s the organization and the complexity of that organization that opens up possibilities, I think.
Paul 01:26:01 All right. So my mind could be more cool than my brain activity. I could, I could feel better about my mind than my brain. <laugh>
Alicia 01:26:08 Probably. Yeah. Brain gets really, you know, and I think that’s true. We, I think, I think as people age, they get worse at oh, tasks that are just repetitive or, or
Paul 01:26:22 Oh, I thought that’s all they could do are tasks that are repetitive.
Alicia 01:26:24 No, no. But I think understanding connections and understanding nuances and things like that.
Paul 01:26:30 Oh, I’m thinking way aged. You’re thinking like, uh, our age maybe.
Alicia 01:26:35 Well, um, I, I’m thinking, I’m thinking of the difference between, I don’t know, dunno, it’s more like difference between understanding and just pure speed of activity.
Paul 01:26:50 But my grandmother, who’s 99, she cannot learn a new card game. Right? She can play bridge, but she can’t learn a new card game. But that’s 99. Yeah. So maybe there, maybe that’s even a different type <laugh>. Maybe
Alicia 01:27:02 That’s any, and you wonder maybe she’s not interested in any of that.
Paul 01:27:05 Maybe. Come on Grandma
Alicia 01:27:07 <laugh>. Yeah, yeah,
Paul 01:27:09 Yeah. Okay. But I see it’s a different kind of understanding. It’s, I don’t know, would you call that wisdom? Well,
Alicia 01:27:13 Hopefully. I think that’s what philosophers were looking for. Yeah. And the question is what? I don’t think it automatically happens as you get older. An awful lot of old fool around there around. Sure. Yeah. I’m sure are a lot of old fools out there.
Paul 01:27:26 Yeah. Okay. El Alicia, I don’t wanna take all day for you. So I wanna switch to, um, artificial intelligence here. Yes. For a few moments, if you have. Are you, are you good on time for a few more minutes?
Alicia 01:27:35 Yeah.
Paul 01:27:35 Great. This is a lot of fun for me. So, um, no, me
Alicia 01:27:38 Too.
Paul 01:27:38 So I was actually kind of surprised in revisiting dynamics in action and in, and you go into it in a little more detail. I think in context changes everything that you refer to, like neural network Yes, yes. Examples. And in both books you talk about the semantics that come along with, uh, syntax when you put recurrence in these neural networks and, um, you cite Jeffrey Hinton and, um, actually Michael plt, who, or, um, David plt, who was one of the authors on that paper, I did a course with him. Anyway, that’s neither here nor there, but I always
Alicia 01:28:10 Call him Hinton and Shall, because Hinton shoutout shall always, but they’re, that, that, I, well, go ahead. Finish up. Well, anyway, again, I’m Cuban. I shouldn’t even
Paul 01:28:21 Yeah, well, so given what we’ve been talking about and, and, you know, these interlocking systems and how important that is for complexity and, and the constraints, uh, given all that, I’m, uh, slightly surprised and, and you talked about, uh, constraint of closure from, uh, Masimo and me, ma
Alicia 01:28:39 And Ma Moreno and Masi
Paul 01:28:41 Moreno Masimo, which I love that book also. Yeah. That book great. Um, but that is very specific to like biological systems and not just biological systems, but complex systems in the natural world. And it’s not, those types of things that we’ve been talking about are not built into, um, neural networks, or only some of them are because maybe you, you talk about the recurrence and how that can lead to semantics. So anyway, I, I was, uh, slightly surprised that I started reading about like the neural networks because I thought, oh, well, a lot of the things that it seemed to be important, um, in these complex dynamics that give rise to our rich mental worlds and rich ecologies and stuff are missing in the artificial intelligence systems. Or maybe they’re not. And I, I wonder what you think is missing in current ai.
Alicia 01:29:19 I think what’s missing in current AI is, uh, context <laugh>, in other words, um, but that’s what the
Paul 01:29:24 Context
Alicia 01:29:25 Somatic Somatic.
Paul 01:29:26 Oh, okay. Well, I was, because context, I think is it context changes? No, no, that’s your, the title of your book, <laugh>. All you need is, oh, all you need is attention, which is a different form of context. And like the modern transformer, which underlies all these large language models is all about context. Yes. In terms of word,
Alicia 01:29:43 But it does not have somatic
Paul 01:29:45 Okay.
Alicia 01:29:45 Information. Okay. It does not include historical context, you know, well, this is how it’s changed over time. So I think you need, it does not include social except indirectly as it, as they, as the social aspects get revealed in, in language use. Mm-hmm. <affirmative>. Mm-hmm. <affirmative>. Otherwise you don’t have direct somatic interaction. Now, once you have these things, once you have the chip, you have the Nvidia chip in a, in a machine that you kind of raise as a child with all the video, with all the video, with all the auditory, with all that,
Paul 01:30:32 You’re actually, you’re speaking exactly to my previous guest as not, I haven’t released this yet, but he raises artificial agent chicks exactly like newborn chicks to, to look at this developmental trajectory over time. So,
Alicia 01:30:43 Which people did with chimps, remember they tried that of that without turn did not turn out very well. Yeah, yeah. Because at some point it didn’t scale up. Now the question is whether the, but I think that if in fact we have eon we have eons of evolution behind us of the planet, of the, or of live of life and so on, you wonder if you can short circuit any of that and get Right.
Paul 01:31:17 Well, we seem to be partially,
Alicia 01:31:19 Well, I mean, I certainly wanna say a sunflower is aware of the sun.
Paul 01:31:24 Okay.
Alicia 01:31:25 Trumpism is a form of awareness. I don’t care what you call it. Is it self-awareness? No, I don’t think so. I wouldn’t wanna say that. Um, so is the dot example of the, you know, which the elephants is on now, they’re saying crows can do it and so on. Mm-hmm. So is there, and what is it like to be a crow as well as a bat? So at what point do you start saying, well, this is not just, it recognizes its body with something and one, or really recognizes itself mm-hmm. <affirmative> as a self, or is there a self in the, in the chimp that can point to the screen when it sees its image with a thought on its forehead at each point? I would say it’s an emergent property, but certainly, I’ve gotta tell you those hint, I had been playing around with these problems for ages, and I had not sat down to write, uh, dynamics in action until I read Hinton Proud and chais, which time was, oh, I went the fact that they, that that really simple, I mean, we’re talking how many decades ago.
Alicia 01:32:29 Yeah. Uh, neural network was able to, according to chais, I think it was when somebody asked him, how do you explain that it produces the errors of deep dyslexia, not surface dyslexia. And his answer was, I can only explain it by saying it self-organized, a semantic attractor. So semantics would be an emergent property that changes the output. That’s what I, why we would call it strong emergence. That’s what I, why I would call it top down. It changes the probability distribution of what its behavior is such that it produces an output that reflects the semantics of the attractor.
Paul 01:33:09 But so AI does not have closure of constraints. So this is another idea that, um, and you know, it’s not self-generating. It’s, it’s byproducts don’t feed back in that sense. Or, or maybe you have a different outlook
Alicia 01:33:22 Now, now, now,
Paul 01:33:24 Now it doesn’t.
Alicia 01:33:25 Exactly.
Paul 01:33:26 Okay.
Alicia 01:33:26 That’s the question. The question always is not, not what’s practically doable today, but in
Paul 01:33:33 Principle Sure, sure. But I mean, are we gonna have to emulate life then? Are we gonna have to have artificial life to have what, whatever the hell, real artificial, real intelligence, real artificial intelligence is. Right.
Alicia 01:33:46 And that’s again, an answer for the constraint on to how many of these constraints are really not necessary for, for intelligent systems that would assist a, a neurologist or a Sure. Or a, you know, maybe you don’t need experience, you know, emotion or, you know, I mean, I don’t know. Mm-hmm. Um, so, but my, but what I find fascinating is that even without somatic input, without any of this other stuff, the kinds of errors that people talk or they hallucinate, they, it’s so similar to the kinds of errors that you got with, um, uh, the split brain Annika issues. The thing is making up some kind of story that would make all the inputs it has so far coherent. Hmm. Correct. Yeah. Even if it is, well, that’s a typo. <laugh>, I seen one where the output, well, people were, the trainer was saying, no, that’s not right. That’s not right. That’s not right. At one point it said, well, you know, that was just a typo. Excuse me, <laugh>. That’s, that’s amazing.
Paul 01:34:52 Yeah. Yeah. What ideas or concepts do you find yourself like revisiting the most? Um, when, when you’re working, what, what ideas do you have? Do you revisit and are unsure about or think you need to grapple with more?
Alicia 01:35:07 I think the notion of how do you account for the emergent properties characteristics in each particular case, in each particular individual. How do we, how, is there any hope to be able to map this other than say, well, if each one is unique, then well, then that’s the end of science. That’s the, I mean, you, we need some kind of general
Paul 01:35:37 Formal
Alicia 01:35:38 Hmm.
Paul 01:35:39 I never know. And we both winced there. Yeah.
Alicia 01:35:41 Yeah. I don’t know, I don’t know how to phrase it. Is it a back forth, uh, hermeneutic kind of multiple constraints, satisfaction? How does that work? And I love that example that I got from Ling and Kig. I think their book way back went on psychology. You know, imagine you renting out a, an apartment and the apartment has one large wall, and it’s got a window here and so on. And you’ve got, and this is gonna be your bedroom. And so you’re constrained because the bed has to fit against maybe a wall, but then the chest doesn’t fit because the other wall. So, so it’s a question of multiple constraints, satisfaction. So do you turn the chest of drawers into a night table, because that’s the only place it fits. And so you’re adjusting it and it’s constant adjustment. But that’s what you see in dynamical systems, that that’s the kind of meta stability that a, someone riding a bike or someone on a tightrope, you know, one of these people going on a, on a high wire act tightrope, it’s not the kind of equilibrium of your old time near equilibrium thermodynamics. It’s a meta stability where it’s a constant adjustment of constraints.
Alicia 01:36:56 Correct. Yeah. And so how do we, is there any hope to be able to map this in a way that provides some kind of general guidance as to how this is gonna play out in a particular situation? How much do you have to know of the background? How much do you know? Do you have to know of the context to be, you know, and I think, I think that’s one good thing about that Aristotle had said was you don’t, you can’t teach law, medicine and ethics, because those are eminently context dependent compared to science. So, so how do you do it? You can only do it through apprenticeships because the apprentice gives you this. And when I taught class, this is what I do. You know, you get a feel for it. So, so is that what a good coach in a business does? Because they’ve got enough of a feel for it. They might not be able to formalize it into a standard algorithm, but they’ve got a feel for the situation that therefore can provide guidance.
Paul 01:38:05 This is why a coach doesn’t need to be more of an expert than the expert as well. I imagine.
Alicia 01:38:09 You know, you have to have a feel for, for, and, and how do you, how do you formalize the feel mm-hmm. <affirmative> of something And, and that’s what I’m, that’s what a mus, that’s what a master musician does. I just finished reading a book by Adam Knick called The Real Work, and it’s about the mystery of mastery.
Paul 01:38:28 Okay. Oh,
Alicia 01:38:29 It’s fun to read. And so he talks about, you know, how do you, is the master seeing something that’s at a different level? And Hofstetter used to talk about that when he talked about the chess master who’s shown a chess game in progress for a few seconds. And then they’re asked to reproduce the
Paul 01:38:51 Yeah, because they had the field for the board, you know,
Alicia 01:38:53 But he, they would always put the pieces in different locations than what was actually presented to them. Oh, right. But the strategy, the strategic position white and black were in remained the same. Interesting. So what are they seeing that, that the, the newbie doesn’t see? The newbie is worried about, uh, playing the piano, one finger here, one finger
Paul 01:39:17 There, simulating all each move ahead of
Alicia 01:39:19 Time, each, that’s like the physical signal that’s coming in. Obviously, mastery means that you see a level of organization that, in the case of chess strategy, in the case of of, of, um, um, magician is the crusted digitization. It’s the, it’s the magic. And what is it that you’re seeing that then matters? And I think that’s the constraint regime that we’re all trying to figure out. Why is it that the same person that the soldier next to that one didn’t get ptsd? Mm-hmm. <affirmative>, they, they had, they were in the same explosion
Paul 01:39:58 Because they were more of a man.
Alicia 01:40:00 See, that was the old tradition. It had to be an internal, an internal character
Paul 01:40:06 Is sick. Right.
Alicia 01:40:06 Whereas what I’m saying is what are the interdependent constraints that, that make a difference? Uh,
Paul 01:40:13 But you, you stuck, you began that with like, you wondering if there is no hope so, or if there is hope, is there hope? Do you feel optimistic? I,
Alicia 01:40:20 I, look, you’ve gotta assume that we’re all humans. So there’ve gotta be some commonalities somewhere at what, at what scale? At what granularity? It’s hard to say, because you look at families and you see kids from the same family raised by the same friends going to the same school, and they’re so different. So you want, how does that happen? Is that all in birth order? I mean, there was a people lot about that temp order, temporal order has to matter at some point. But to what extent? To what degree? I think that’s what the, what’s fun about this is that there’s a whole new possibility space of research that opens up. Mm-hmm. When you look in terms of constraints, rather than looking in terms of, you know, I, when the people were into the genome, I thought, you know, people are assuming again, oh, one disease per genome looking
Paul 01:41:19 Back. That’s crazy. It’s never
Alicia 01:41:20 Gonna happen. Yeah. I knew that that was the case. Now
Paul 01:41:23 You knew that we I didn’t know that. I mean, yeah. ’cause I had this linear thought. That’s the way I was raised. And so we
Alicia 01:41:29 All were raised that way. We all, that’s what we all assumed. Yeah. And, but now the question is, well, is it gonna be one protein fold, perb per disease, or per dis So then the que So at what point is it? No, it’s the context that makes it fold this way rather than that, and therefore gets this functionality rather than that, even though it’s a same string of amino acids. So, alright, so it’s the confirmation. So what is it in the context that makes it conform a certain way? Well, in the case of development, we know that one cell, in one part of the fertilized egg is the one that, because if it’s in that location, that’s what’s gonna turn into the first neuro crest. Is that pure random, then that’s then that races, that brings randomness there rather than at the quantum level. Yeah.
Paul 01:42:15 Yeah.
Alicia 01:42:15 So how does that work? Who knows? But it’s fun to think about. It opens up a whole new way of looking at, at a whole level of organization that we thought was just illusory or just epistemically convenient. But if we really say, all right, now let’s pretend just for the sake of argument, it’s real. How does this play out? How, what does this mean? What are the implications?
Paul 01:42:39 I mean, it’s not just one level though. God. Now I wanna ask you, you know, what’s beyond consciousness? What’s, you know, if you,
Alicia 01:42:44 It’s turtles all the way up down.
Paul 01:42:45 Oh, I just, that, I just use that metaphor.
Alicia 01:42:47 Every, every, it’s everywhere. Yeah. It’s the whole tangled, you know, it’s an entangled bank. That’s, but it’s not random. It’s, it’s very, it’s, people use the word, still use the word chaos when they refer to things like the Lorenzo tractor to mean, it’s, it’s, it’s this sort. No, it’s just a high level of organization that we didn’t appreciate before. Yeah. That, that lorenzen butterfly is very organized. Are you kidding? But it never repeats the same trajectory twice because of its path dependence and so on. So, so we have to understand what are the constraints that lead to that higher level of organization and how does that work?
Paul 01:43:28 Well, Alicia, in terms of possibility spaces, it just occurred to me that if, if cons, the, the science, the philosophy of constraints is a possibility. Space, your book context changes everything filled the hell outta that space. <laugh>. And I’m glad that we got thank Visit some of those places today. I
Alicia 01:43:46 Thank it.
Paul 01:44:06 I alone produce brain inspired. If you value this podcast, consider supporting it through Patreon to access full versions of all the episodes and to join our Discord community. Or if you wanna learn more about the intersection of neuroscience and ai, consider signing up for my online course, neuro ai, the quest to explain intelligence. Go to brain inspired.co. To learn more, to get in touch with me, emailPaul@brandinspired.co. You’re hearing music by the new year. Find email@example.com. Thank you. Thank you for your support. See you next time.