 So, uh, hello, I'm, I'll, I'll, I'll, I'll start a site for us. So, hello, I'm my, hi, everyone. Welcome to the active inference lab and to the first active inference guest stream. I'm here with our special guests, Majid, Benny, and we're just really honored and excited to be hearing your presentation. We're going to be having a about 30 minute presentation on whatever you'd like to be speaking on. And then I'm going to be collecting questions from the live chat and then relaying them to you. And it was already just fun to talk to you right before we started. So thanks again, Majid, and take it away. Go for it. All right. Hello. Thank you very much. Yeah, I'm very excited too. And I thank you very much for giving me the chance to present a part of my work. So, uh, I'm basically beginning to speak, uh, what I've done, um, a part of my work, uh, on a free energy principle. So I will try to share my screen rights now because I'm a great believer in extended mind and I barely could do anything without having my screen shared. So you, you most be able to see my screen. Yep. Looks perfect. Thank you. Perfect. So, uh, uh, I'm basically going to present, uh, a part of my work in a recent paper in cinemas, critical analysis of Markovian monism. And the structure of this talk is quite, uh, straightforward. So, uh, I'm not, uh, I briefly, uh, very briefly and with a very broad branch, we'll speak about free energy principle and Markovian monism. So, uh, I gloss over that, uh, technical stuff because, uh, for one thing, uh, I'm a philosopher and I'm not very well versed in these technical aspects. And for everything I'm going to present the core of my work, which is about scientific models. Uh, so, uh, after that I speak briefly about scientific models and scientific modeling in philosophy of science and coming from the background of philosophy of science. This is what, uh, it is the main part of my contribution to this literature free energy principle to look at it through the lens of philosophy of science and the issue of scientific representation. Um, then I'm going to, uh, describe Markovian models as some sort of scientific models and scrutinize the relationship with their target model in the real world and see, um, whether any of these, uh, descriptions of the relationship between Markovian models and their target system could support Markovian monism. And briefly at the end, I will also evaluate, um, freestyle and, uh, colleagues attempt at developing Markovian monism along the lines of a structural realism. So, um, so most probably you know more than a free energy principle and predictive processing. Uh, the idea has been reintroduced to, uh, science part. So the roots of this idea seems to be quite old, but, uh, it has been reintroduced to restore to science by Carl Briston and Carl Dixon. It is all based on simple and subtle idea. Uh, so, uh, generally, uh, it provides a framework that explains how a biological system exposed to random and unpredictable collection in its external environment can restrict itself to occupying a limited number of the state and therefore, therefore, survive in some recognizable form. So, uh, in order to be able to resist the dispersing of force of the environment, the organism have to, in a nutshell, have to keep the amount of its free energy low. And, uh, predictive processing is some sort of a theory of application indicating that when applied to the brain of some kind of system, sort of ourselves in all free energy minimizing free energy. Uh, could be a specified in terms of, uh, minimization of prediction error. So there are these generative models in the brain and through a process of a top-down process, the brain or the organism. And they were to, uh, apply these models to reality or at least to check them against the reality, against the evidence that it receives from the external world. Uh, and when there's discrepancy, it endeavors to, uh, uh, somehow revise the models, update its model or at least make changes to the environment. So there is some sort of action, either, uh, cover mental action or overt action, but active inference is the other side of minimization of prediction error. And, uh, Princeton and Codding's mainly Princeton developed free energy principle into a fascinating theory of consciousness in terms of temporal richness or temporal depth of generative model. So not only some organism can minimize their free energy, um, and optimize the representation of the environment, they could also somehow form some sort of a foresight or prediction about the possible outcomes of their action in the environment. And, uh, it seems that this capacity is somehow tied with, uh, with conscious properties or consciousness. And of course, there is this idea of Mark of the Lankens, uh, do the appeal introduced them, I think in 1980s and there are some sorts of Bayesian networks, uh, which could help us to make a distinction between conditional independent according to conditional independence between internal and external spaces. So, uh, the literature itself, uh, free energy principle has been quite flourishing, uh, quite a number of scientists and philosophers, uh, have contributed to developing this literature. And amongst the philosophers, uh, just to mention a few, uh, great people who contributed, such as, uh, Jaco Paul, Andy Clark, Sean Gallagher, and, uh, even people who are coming from a background of philosophy of science, such as the Stepan Hartman have contributed to this, uh, rich literature. Uh, what, what's interesting is that the free energy principle has not been developed into some sort of, uh, metaphysical thesis about cognition or consciousness or brain. So, uh, as far as I can see, uh, this, uh, uh, paper, which is the target paper of my paper, for instance, from Cartesian duality to Markovian monism is the subtitle of Christon and Meese and Hobson, uh, as far as I can see, it is, uh, the main work, the main paper, uh, that's actually contributed to, uh, as fading out the metaphysical, uh, implications of the free energy principle. It is the main paper, as far as I can see, that tried to develop free energy principle into some sort of metaphysical, uh, thesis of consciousness. And I think that for this reason alone, the paper deserves, uh, attention and, uh, I'm a great admirer of this work. I usually do not engage with, uh, uh, reviewing or criticizing single papers, but this paper attracted my attention and interest because I think that, uh, it is, so even before this, uh, people such as, for example, Christon and Hobson themselves tried to develop free energy principle into a theory of consciousness. That provides a point to, um, Cartesian dualism, uh, souls to try to do so. But, uh, as far as I can see here, for the first time, free energy principle has been developed into, um, a very, uh, well-defined, uh, metaphysical thesis, which is monism. Uh, so, uh, uh, the main point here is that I'm more sympathetic to the goal of this paper, to the objective of this paper that at times the paper may indicate. And what I'm going to say, uh, is mainly, uh, aimed to invoke, provoke the free energy principles to develop this Markovian monism into a slightly, uh, better, uh, I quote, a well-supported, uh, account of consciousness, metaphysical account of consciousness. And I'm going to look at this issue through the lens of philosophy of science to see the relation between Markovian blankets and, uh, their environment as, as some sorts of relation between scientific models, because Markovian models are some sorts of scientific models. And I'm going to, uh, argue for, uh, provide some argumentations for that, uh, to draw a connection between, uh, Markovian models on the one hand and the target system on the other. And so, whole consciousness that appears through this relationship, uh, could be, uh, put into some sorts of, uh, monism, which is a metaphysical thesis. I'm going to look at this issue through the lens of philosophy of science, uh, and unfold the implication, a spell of the implication for Markovian monism. So the thesis itself, Markovian monism, uh, is quite a straightforward. I have highlighted the part of it for you here. So the thesis is, is that fundamentally there is only one type of thing and only one type of irreversible property. And Markovian blankets form the calculus of belief, which is not only a nice scientific, mathematical model of consciousness, but also provides a clear, metaphysical statement. And I can shed light on current philosophical debates. So, um, the highlighted parts, so that it is, this is the statement of Markovian monism. And the red part indicates that, uh, how forcefully the theory have been, uh, a spell told in a metaphysical term. So there is only one type of thing is a strong metaphysical, uh, cases. And this is something that has been also stated in, uh, prison and colleagues paper. Um, so, uh, and I, here I highlighted another part of the same phrase, which indicates that, uh, this statement of consciousness on the basis of Markovian models is not just a nice scientific or mathematical model of consciousness, but also provides a metaphysical statement. So the main question here is that how could, uh, how come that we could build, how come that we could read off a metaphysical thesis on the basis of our scientific models of the phenomena, scientific, mathematical models of phenomena. Uh, so I'm going to, uh, um, ask this question, as I say, through the lens of philosophy of science to see what's the relation between, I think that this basic question of philosophy of science, what's the relation between scientific models and their target system. They are significantly on the claim of Markovian monism, which tries to develop some sort of a strong metaphysical thesis on the basis of this efficient modeling tool, scientific modeling tool, which are Markov blankets. Um, so this is an insight from philosophy of science and the insight has been, so I think that originally it has its roots in works of Ronald Geary and, uh, uh, Paul Churchland, for example, uh, and a number of, uh, and, and Boston for us and himself, a number of important philosophers of science who have been developing this idea in a model based. Approach to scientific theories in, uh, 1980s and it has been recently, uh, surfaced, uh, resurfaced in works of Michael Weisberg and, uh, Roman Fried, uh, and, uh, God's free as means. So, and the general idea is that scientific practice is basically a consist of model making, making scientific models. And the thing is that, so we have, uh, when, when engaging in scientific practice, we just do not directly engage in, uh, reading of the ideas from the feature of war, uh, reading of the theories. The theories do not just latch on to the war directly. There is some sort of representation and the representational relation between scientific theories, scientific models, and the world is in, in, in direct. So scientific theories do not represent the features of the work directly. Um, the main thing here is that, uh, quite a few, uh, few main factors interfere through the process of scientific representation. So to name only few, for example, scientists interest and explanatory, uh, goals and intentions, uh, play a role in how scientists are constructing their scientific models and how use them to represent the features of reality, uh, by scientific models. And that's, uh, well, free, for example, among other people explains this process of indirect representation, uh, is quite sophisticated. It includes several stages such as denotation or designation, exemplification, King of and imputation. So, uh, the main point here, the take home point here is that the relationship between scientific models on the one hand or reality on the other hand or the target system on the other hand is quite, uh, sophisticated. It is never straightforward. And there are three ways of, so Mark Mark of blankets, Mark of models are just efficient scientific tools in the modular toolbox that allow the modular to represent the features of the world in indirectly. So, uh, I'm going to speak about three ways of, uh, understanding Markovian models. Um, I'm beginning, so, uh, as the sketch that I showed earlier in the case, I'm beginning with, uh, interpreting Markovian models as physical or material models first and as fictional models and finally as mathematical models. And this order has been chosen deliberately because I'm going from the least likely reading of Markovian models. I'm going to are the most likely interpretation of Markovian models. So I'm convinced that Markovian models could be understood as some sort of mathematical models. But I also consider other options because there are some sorts of textual evidence to support them. So for example, we could, it is impossible to understand Markov blankets or Markovian models as some sort of physical models, material models. For example, for instance, in 2019 paper that Markov blanket is not some static statistical device for which we could or we use it to observe or model the world. It is a necessary attributes of the universe that can be carved into teams. So, at least to the first approximation, we try to read this phrase or the second phrase from the same paper. It seems that for free energy principle theories for briefly, for example, Markov blankets are actually features of the reality themselves. There are physical things. There are real, real things out there. And so I think that it is not only Princeton is advocating this reading. I think that also I have seen similar phrases in, for example, recent med singers introduction to predictive processing. And also in Allen and Princeton's paper is sentence 20 and 60, I think. So this is, there is some textual basis for understanding Markovian but Markov blankets as physical models. But I think that this indication, this reference to Markov blankets as physical things has to be taken with a grain of salt. I do not think that there are actually such things. So if we, for example, open the brain or open an organic system, I do not think that we can find anything resembling a Markovian blanket. There is no blanket. There is no such physical thing. But I can understand if a terrorist could conceive of the target system as if it really includes this representational device. The point is that even if we can say that Markov blankets Markovian models are features of reality or physical things that exist in the world, even if we consider that it doesn't follow that reality itself shares the ontological and material features of Markov blankets. It doesn't follow that there is the same self same ontological category that includes both Markov blankets and their target system. And there's an example that could demonstrate that. So, for example, there is such thing as a hydraulic home model of economics. So hydraulic home is a physical object. And for hydraulic model to work, some properties of economic system must be identified with hydraulic properties of the pump. It could be, for example, exemplified through the process of imputation. We could ascribe or attribute some of the properties of the hydraulic home to the physical object, sorry, to ascribe some of the properties of hydraulic home to the economical system that is supposed to be represented. But it doesn't follow that the economical system and hydraulic home, which is a physical object can be put under the same ontological category or material category. So, for example, hydraulic home, which is a physical object is made of, for example, a steel or plastic or some other material thing. It doesn't follow that economics shares this material ontive features. So, I simply put the materialist interpretation of Markov blankets aside and proceed to consider Markov blankets as fictional models. So, and the main reason for this is that I assume that if we, for example, explore the actual brain of organism, we couldn't find anything resembling Markov blankets inside the brain. So, we could interpret them as some sort of fictional objects. And so, and the fictionalist, and again, there are textual support for this, for example, Chris and colleagues say that if we look into the behavior of creatures through the lenses of Markovian models, it will look as if it is acting, the creature is acting to minimize its particular surplus. So, this acid phrase, Luke has a phrase, I think that is very important. In case that they do not really think that there is such thing as Markov blankets in the creature, but it looks as if there is such thing. And it has been this interpretation, this fictionalist, insterimentalist interpretation is gaining momentum, as far as I can see. So, there are a number of other people such as Maxwell-Romsted and Vannes that are developing some sort of fictionalist and insterimentalist interpretation of pre-energy principle. But the thing is that if we concept this fictionalist interpretation of Markov blankets or pre-energy principle, it follows that, so if Markov blankets are useful fictions, we couldn't draw any sort of metaphysical conclusion from them. So, from a fiction. So, for example, if you are studying Greek mythology and assuming that Greek mythology is some sort of fiction, you cannot derive any conclusion about the nature of deity from what you read in a story from Greek mythology. There are no metaphysical conclusion to be derived from fictions and stories, and this is precisely the point in invoking a fictionalist approach to scientific models. So, people are using fictions, assuming that scientific models are fictions, are useful insteriments to just get rid of the metaphysical questions about the nature of fiction models or the relationship with reality. To make a longer story short, this fictionalist interpretation doesn't lead to any sort of metaphysical conclusion about the nature of consciousness or the nature of mind or anything of that sort. So, this brings us to the last interpretation of Markov blankets. And it comes in terms of, I don't have much time, I try to write it up quickly. So, the final interpretation is to interpret Markov blankets in terms of mathematical models. And I think that there is good textual growth. So, it is, at least to my mind, to some extent, for someone who is coming from a background of philosophy of science, it seems rather obvious that Markov blankets are just mathematical modeling tools. So, they are mathematical tools in the context of Bayesian networks to represent the conditional independence between internal and external spaces. And there is textual evidence for that. Again, I have highlighted part of that, which indicates that their enterprise aims to describe a construct from information theory. And mathematical physics that helps them to define the properties of living system and their cognition and give them some sort of meaningful interpretation by invoking mathematical objects that could be used to consolidate this interpretation. So, I think that there is some sort of enlightenment inside, in modern philosophy of science, in the wake of works of Patrick Soops and Basavamparasan in 1960s and 1970s. So, people's attention has been drawn to the importance of mathematical models, set theory and model theory and that sort of stuff, for modeling natural phenomena. So, when doing science, scientists, so the rise of modern, in some sense, the rise of modern science has been a result of invoking these sort of mathematical approaches that allow us to represent and model natural phenomena of, you know, of precision and rigor and rigor. But the thing is that if we assume that Markovian models are some sort of mathematical tools that can be used to represent the features of reality, the thing is that we couldn't develop any sort of mathematical conclusion about the nature of reality on the basis of application of these mathematical tools. Look, if we assume that Markovian models are mathematical tools and if we want to identify their nature with the nature of reality, the only conclusion that follows is that the reality is some sort of mathematical object, is some sort of mathematical entity. And interestingly enough, as I will argue, this is precisely the same sort of conclusion that Princeton and colleagues reached in the end of their paper. So, at some sense, they are saying that they are, they intend to defend some sort of reductive materialist or reductive physicalist. But what happens is that they come to the conclusion that because these two ways of describing the nature of reality, this division on the basis of Markovian models, both to external and internal spaces, because none of these interpretations, none of these geometries that they develop, none of these intrinsic and extrinsic perspective could be supported. Under the assumption that neither perspective is privileged, one would have to conclude that reality is fundamentally ontologically neutral. And they assume that by speaking of ontological neutral nature of reality, they are trying to build open the mathematical nature of Markovian models to make concession on some sort of monism which identifies the nature of reality as neither physical nor mental. But I would say that it would be quite a lot to read into the use of mathematical models. So, of course, mathematical tools are neither, in some sense, are neither mental nor physical if you are not an idealist or a realist, especially if you are not a physicalist or idealist of mathematical entities. But it doesn't follow that the target systems that could be used on the basis of using these mathematical tools itself, the reality, the target system that is going to be used by invoking this sort of mathematical tools itself is neither mental nor physical. So, this is the conclusion of the paper and for coming to that conclusion, I have also discussed some sort of a structural realist interpretation of the paper. Again, it is about the relation between mathematical tools and the physical reality that they represent. So, again, Freestyle and colleagues would be happy to make concession on some sort of a structural realist interpretation of the Markovian monism. But in the context of a structural realism too, a structural realist as some of the leading structural realists such as a Stephen French and James Leibman are quite explicit about this clear. They do not make their ontological commitments on the basis of the mathematical framework that they use. So, they use mathematical framework, they use model theory, they use set theory because they believe that this semantic framework, model-theoretic approach or set-theoretic approach to science, via its structuralist loyalties on its sleeves, but it doesn't mean that ontology itself could be read off from the features of these mathematical tools that they use. So, they are very clear that the job of the mathematical tools that they use, so orthodox and structural realists are using mathematical tools such as set theory and model theory and category theory. But they are very clear that the job of these mathematical tools is mainly representational, it is not constitutional, it is not ontological. On the other hand, it is the physical reality that is supposed to be represented that has some sort of ontological constitutional rule. And with regard to Markovian monism, if we want to derive any sort of, the use of the mere application of Markovian models doesn't lead to any sort of metaphysical conclusion because Markovian models are themselves mainly representational tools. They help us to represent features of the reality that is the target system, the brain or consciousness or organism or the environment or whatever. It doesn't itself, it could not be viewed upon the features of this mathematical model itself to derive any sort of ontological or metaphysical conclusion of the features of the target system. So thank you very much for listening and I hope that what I said makes some sense at least to some of you, but although I could not be completely sure. So thank you very much. Awesome. Thank you for that fascinating presentation. We already have a couple questions in the chat. So if anyone else wants to ask questions, we'll have a good amount of time now just to talk a little bit. And I'll just read down the questions that were asked in the chat and a few things I wrote down. So the first question in the chat was, as far as different styles of monism or different threads in the history and the philosophy of monism, how does a Dawkins based monism, so thinking about like memes, biological or non biological mimetics, that's kind of a Dawkins monism. How does that parallel or overlap with Fristonian monism? So in other words, how does Friston's monism or the Markovian monism relate or overlap with other approaches to monist type theories? Well, thank you very much for the question. I'm not very knowledgeable about the different types of monism. So I know about some quite classical theories of monism. I know a few things about, for example, a Rosalian monism or the form of monism that has been developed by Charles Sanders Peirce, a side problem. For example, the classical forms that are developed by Espinosa. So I'm not very well versed. I'm not very knowledgeable about this kind of Dawkinsian monism. I barely know anything about memes and I'm not sure that we could identify memes as real ontological entities. And I have already argued that I have some sort of misgivings about Friston and Wies and Hobson's version of monism. So I do not want to give a readymade answer. I would say that if Dawkinsian monism resembles Fristonian monism with regard to the fact that it invokes memes, which are, as far as I can guess, some sort of theoretical tools as features of reality, it could be liable to the same kind of objections that I had as Peirvedet here against the Fristonian kind of monism. Thanks for the clear answer. And I guess the follow up on that is where do you situate this Markovian monism in relationship to these classical kinds of monism that you just mentioned? Is that discussed in the paper or how would you frame that? So, all right. Thank you very much. Again, it is a fascinating question. So before this, I have developed some sort of a structural realist theories of consciousness too that have been basically based. So my own approach to free energy principle is associated with some sort of a structural realism too. So again, I myself have invoked some sort of mathematical structure, embodied informational structures to describe the structure of the mind and consciousness. And my basic argumentation there was that this structure underpin the diversity between different icons of consciousness in terms of free energy principle, Georg Nordkopf's resting state activity theory, and for example, integrated information theory of consciousness. So I myself have developed some sort of a structural realist, but with regard to the metaphysical aspects with regard to. So I see monism specifically as some sort of metaphysical approach. And also, I have been very cautious with regard to stating a metaphysical stance. So the version of a structural realism that I have defended was some sort of epistemic structural realism. It wasn't still realism, but it was epistemic in the sense that I was ready to make commitment to the fact that to the extent that our scientific or scientific knowledge is informing us the future of reality is a structural. But we couldn't make assertions about the nature of reality itself. So whether it is a structural or not, or whether it is the same stuff, the self-same stuff that is there, or there are several kinds of substances. So I profess agnosticism about the features of reality in itself. So the version of a structural realism that I defended was epistemic. And I have no stance on the nature of reality in itself, whether are monies or dualist types or any other kind of approach. That's really interesting. Seems pretty fair to be agnostic on the ultimate structure of reality. And you had a quote on one of the slides that the structural realism is taking the pressure off of strong ontological commitments. So how would you summarize this kind of best of both worlds structural realism approach? Are we just, is it a buffet? Are we just taking what we like or what are we committing to with epistemic structural realism? Very, very fair. So I think that it is partly an issue of motivations. So when I'm stating a structural realism, for example, people such as John Goral built open arguments from history of science to argue that. So it takes the pressure off ontological pressure in the sense that it doesn't make commitments to the intrinsic nature of entities. It doesn't make commitment to objects. It is committed mainly to the relations. So and I think that I'm defending some sort of epistemic form of a structural realism because as I say, it is completely agnostic about the features of reality itself. So, for example, people such as John Goral are defending a structuralism on the basis of history of sciences with very clear examples, for example, from the history of optics to show that there is some sort of continuity between theories of optics. Although it seems that there are radical, theoretical sheets in the history of science. So some of the terms, some of the theoretical entities that were supposed to be the subject of ontological commitments before are eliminated from ontology today. But it seems that there is some sort of continuity at the level of a structure of theories at the level of form. So they are not making come and in this sense they are indeed taking ontological pressure from their theory. However, there are other forms of, for example, antique structural realism, which itself comes in various strives. For example, James Lagerman is defending some sort of animinative approach to structural realism. I think that for good reasons from recent made particle physics. But I'm a slightly skeptical that that sort of structural realism is taking any sort of, in terms of ontological commitments, I'm skeptical that it is less conversant that the orthodox forms of scientific realism. I'm sure that there is a perfect match between animinative structural realism and what we know about, what modern physics, many particle physics is representing about the pictures of the board. But with regard to metaphysics, for example, ontological structural realism claims that all that there is is a structure, which is a strong metaphysical claim. So it eliminates entities, objects from the ontology, which is quite ambitious and a strong ontological commitment. So, and the thing is that I do not think that the version of a structural realism that Princeton and colleagues have defended is epistemic, because they are defending some sort of metaphysical thesis here. Monism is some sort of metaphysical thesis. So although structural realism was supposed to be ontologically modest in the sense that at least, for example, Warhol's version doesn't make commitment to the nature of the world itself. The metaphysical versions of structural realism, such as James Layman, are including, are harboring a strong metaphysical thesis about the features of reality that they are claiming to be completely a structural. And I think that in this sense, Markovian monism that is defended by Christophe and others is not metaphysically modest at all. And I think that it is one reason. So I didn't, and thank you very much for bringing this point to my attention. I didn't spell this point clearly in the paper because I didn't spend many time to discuss a structural realism in this specific paper. But I think that, you know, that I look at it from here, I can see that it could be the main motive behind my mind that encouraged me to be a bit critical about this formal monism. That is presented in this paper, wonderful paper by Christon and colleagues. Cool. Cool response. Here's another question from the chat, kind of keeping with this theme of physics and realism and scientific communities that are evolving their models. You mentioned optics. So let's think about particle physics. Cambridge Breathes wrote, Can the author describe in his personal belief the main difference between a mathematical Markov blanketness of an electron when it's behaving as a particle versus when it's behaving as a wave? So how do we think about quantum mechanics and quantum phenomena when we're talking about these different realism ideas? To be quite honest, I have no idea. Do you think that it is mainly because of my ignorance of modern physics. So I have some general conception of what happens at the level of my many particle physics. But how to state that in terms of Markov, and I can imagine how could we state that in terms of Markovian blankets, but I cannot see how the distinction between a particle being in a wave state or being an entity being in a wave state or particle state could be related to this issue of Markovian models. I assume that Markovian models. So one thing that I know about Markovian models is that they are so powerful. They are such powerful modeling tools that they could be used almost to model everything that there is to be modeled. Well, and I suspect and it is only some sort of wild guess that we could model the behavior of an entity either in its wave state or as far as we try to provide some sort of probabilistic representation of the behavior of the entity to model the conditional independence between different states of the entity. We could involve Markovian blankets, Markov blankets to do so, but I think that at the same time, I miss the core of the question. I think just maybe guessing from the context, the electron is something that we might want to say it ontologically exists or some might think it ontologically exists. Yet it can display this sort of bimodal behavior when it can, depending on the double slit and the context of the observer, it can be acting. Yeah, great. So, all right. So how does Markovian blanket does to do with that? So, as I say, this kind of approach to the behavior of entities in many parts of physics was what motivated people such as Steven French and James Ladyman to develop forms of a structural realism. So it is not only an issue of an electron. The thing is that when we look at this, the behavior of entities at the level of at the fundamental level, if I don't want to use a fundamental at the level of many particle physics, they should do all behavior. There is some sort of package that includes individual objects. And it is the interpretation, for example, of the electron that allows us to identify that as some sort of individual objects. And there's a package that doesn't allow us to model it as an individual object. So we have to conceive of that as both individual objects and non-individual objects. And this seems to recover with your ontology of modern physics. And a structural realism provides an answer to that saying that, all right, let's consider the commonality. Let's consider the common structures that underpin both formulation and the behavior of electrons, for example, are stated in terms of several sorts of statistics that are developed from Maxwell's or, for example, both Einstein statistics and other sorts of statistics. So I assume that we could also invoke some sort of a marker of blankets to model the behavior of these kind of entities at the level of sub-particle physics. But does the question mean to indicate that there would be some sort of ontological gain to do so? The thing is that even in philosophy of physics, the discussion of a structural realism, whether we could invoke some sort of representation of mathematical tools or we have to just rely on the geometrical or physical statement of the situation. So I think that these are sophisticated issues in the philosophy of physics, and I'm hardly an expert in that field. But I'm not very clear what obvious advantage would state the regimentation of states of sub-particles in terms of marker blankets would do to the extent that metaphysical instances are at question. So here are a few general questions. What does it mean to be ontologically neutral? So that was brought up, and I was just wondering what is the alternative or how could it be charged or not neutral? What is that? There are several ways of reminding ontologically neutral. So for example, a way that I prefer is to, as I say, progress ontological agnosticism to not make any commitments about the natural reality. But I do not think that it is this sort of ontological neutrality that is at issue in Princeton and colleagues paper. I think that they want to be ontologically neutral in the sense that, so this is the tradition of division in the philosophy of mind between physicalism, between dualism for example, physicalism. People who think that the nature of mind is physical and people who think that mind or at least mental properties are different from physical properties. So I think that Markovian monism wants to be ontologically neutral in the sense that it doesn't want to make commitment to either mentalism or physicalism. It doesn't want to identify consciousness either in physical terms or mental terms. And in this sense, it tries to remind ontologically neutral, which I think it is quite cool. So as I say, I'm a great admirer of this project because I think that it is the first significant attempt that tries to develop re-energy principle into a remarkable metaphysical thesis. And I think that if re-energy principle theories could really show that the theory could support or could be construed along the lines of neutral monism to show that it really supports a stance that doesn't commit us either to conceiving of the nature of mind as a mental state or as a mental stuff or mental properties or as physical. It will be a real achievement, I think. But at the same time, I think that just depending on the mathematical framework of Markovian models doesn't provide sufficient purchase for doing that. It doesn't provide a sufficient basis for boosting a well-supported form of monism. Very cool. So it sounds like there's the eternal debate of whether mind leads to matter or matter leads to mind. Is there one type of thing, two kinds of things? And it would be awesome for free energy principle to step into that gap. But what we've seen so far with the interpretation of the mathematical construct of the Markov blanket does not adequately resolve this metaphysical question. So what kind of analyses or research or measurement or calculation? I mean, I'm a scientist. So I'm thinking about measurement and experiments. But as a philosopher, what would reduce our uncertainty about where the FEP fits into all of us? So far it's been from a philosopher to try to prescribe or endorse anything to the scientists. The thing is that I feel that a cool thing about cognitive sciences and neurosciences is that I think that philosophic ideas could really be useful to scientists in this school. But the nature of this collaboration, the nature of what happens here is so sophisticated that one can hardly... So I'm told that prediction is pretty hard, especially about the future. So I couldn't have any idea or any prescription of what can we do to make this approach more fruitful. The only thing that we could, I think, we could do is precisely the thing that we are doing. Just trying our best and collaborating. So I think that this is why this active inference lab is a cool idea because it allows people from different backgrounds come and share their ideas. And we will see whether anything significant will emerge from all of that or not. But I think that this is the only thing that we could do. It has been the only thing that we have been doing for quite a while. And in fact, well, it was far. Cool. Very... It's almost like there's the mind-matter dualism and then there's the science and the humanities or science and philosophy. And so is one going to envelop the other? Is there really only one and the other one is kind of a subfield? And instead of playing into that dichotomized debate, which is polarized, it's almost like we're taking an ontologically neutral stance here and highlighting the importance of coming together across different contexts and having that conversation. So my question on that would be, let's just say that there was somebody who were curious and wanted to jump in to learn more. What kind of research or skills or mentorship might help this person on their journey, whether they're from a philosophical side or from a scientific side? What do you think would be some key tips for somebody who's just starting on their FEP journey? Well, the best thing that I've heard about the relationship between these different disciplines come from a student branch who is a structuralist and also a wise philosopher of science, saying that metaphysics is best when informed with science and science is best when... I don't want to say that informed by metaphysics, but that is taking some metaphysical insight into consideration. But on the other hand, I think that the nature of scientific and academic work these days is so... In other words, the scientific activity is so compartmentalized these days. People are doing their different works in different disciplines in their own labs. The nature of the problem is so hard that they have to spend all of their time on a specific problem. So I can hardly also provide any advice of what kind of mentorship suits anyone who wants to work or do some excellent job in free energy principle. What I can say that what's very interesting about free energy principle is that it seems to provide at least some purchase for a discourse, some space for a discourse between scientists and philosophers. So it seems that some of the good scientists, such as for example, Max Veramstad or Alex Konstant or Friestand himself, who are scientists, are completely interested in philosophical discussions of what's happening here. And it seems that what's even more wonderful, it seems that philosophical discussions are affecting the nature of their work in some sense. I think that this is something that you can hardly observe, for example, in the age of the art theories of physics these days. I don't know whether the great physicists are studying philosophy for their amusement or not, but I doubt that they are in this sort of constructive relationship with philosophers. I think that this kind of, and the nature of free energy principle itself is quite unificatory, it brings different disciplines, different ideas from different disciplines together. So I have no clear idea about the mentorship or any kind of a schedule or program, but I can say that people who are interested in free energy principle and studying that have already a chance to get a wider scope, a bit more intellectual scope, with insights from a number of different disciplines, biology, computational neuroscience, cognitive psychology, philosophy of mind, philosophy of biology and that sort of thing. So it seems to come with the discipline as it is, by maybe the theory principle. Cool response. And I liked how you grounded the paper and the presentation in the history and the philosophy of science, because it is a community that enacts the conversation around the ideas. And then when there is not a healthy dialogue, it's kind of like the philosophers are sweeping up after the scientists or vice versa, or they're just sort of spectators to each other. But then we're almost moving towards a fruitful bi-directional conversation that actually, as you're pointing out with the specific people who are taking this into their research approach, it actually is something where the science and the philosophy are co-influencing each other. So that's very interesting and special. It's almost like it's such a big scientific idea and so challenging that we know that we need multiple expertise, we know that we're going to need to be working on it together for a long time. And then that helps set the stage for something that might surprise people because it's not just coming from their camp. So we have history and philosophy, science, and then I guess another thread is math. Mathematics, Markov, Blankets, as you pointed out, came from Pearl in the 1988 work and has been developed a lot. Subsequently, where does set or category theory come into this? Well, very good question. I think that in some sense it depends on your view, your perspective on mathematics. So if you are a set or a foundation of this in mathematics, if you are a mathematician who prays to the God of set theory every night in his loneliness, you naturally think that every kind of mathematical theory is based on set theory and set theory has a general foundation or there are other people who are ready to give this rule to the category theory. But for me, because my engagements with these ideas come from a background of structural realism, the main thing is that it has been going to the wonderful works of soups, of Patrick's soups and Van Prossen, and philosophers of science came to, so rational reconstruction has been a program from the very beginning of the history of philosophy of science. So philosophers, and again, it is why free energy principle is so interesting for me because of its great unificatory scope. And there have been philosophers of science and there are still philosophers of science who think that the main job of philosophy of science is to form some ideas about unification of sciences. And in order to bring about this unification, philosophers taught of using logical, formal frameworks for a state regimenting different theories from different disciplines. So for example, in physics, people are using a special form of mathematics in biology, when they are using mathematics at all, they may use a statistical theories or different forms of mathematics. And they came to this idea that if we use some form of logic, some form of, for example, state theory, model theory, we could have a cool mathematical framework that allows us to regiment different theories from different disciplines and clarify their relationship together and have a well-estated, clear, neatly-estated view of what's happening in the world, assuming that science provides a reliable picture of the world, if we could use model theory or set theory. And the thing is that, and I stated this view in a book that has been published by a springer according to a structure of realism, I begin to have thoughts about the usefulness or at least the realistic impact of using set theory and model theory to represent the features of the world. So I become skeptical in the sense that I thought that this set theory and model theory frameworks are too abstract, and even if we could regiment the structure of sciences into this framework, we still couldn't account for scientific representation because the set theory and category theory do not represent the world. These are scientists who are representing the world using mathematical models. So I try to depart from the usual ways of scientific representation on the basis of set theory and model theory and involve other sorts of mathematical frameworks that could provide a better match, a more congenial tool for accounting for scientific representation. Because I begin to try to introduce some sort of information authority framework, and not only information authority framework, but embodied information authority framework, may precisely the same sort of framework that Princeton and colleagues are using in terms of approximate Bayesianism to regiment the structure of scientific theories, because I came to the idea that it allows us to have a more realistic account if we be able to regiment the structure of theories into this approximate Bayesian framework, for example, instead of set theory and model theory. We could provide a better understanding of the relationship between scientific theories, which are actually embedded in the brain of scientists or at least in the scientific community and their target system in the world, because their set theory and model theory are too abstract to take care of the job. This kind of approximate Bayesianism is something that at least according to our best scientific theories is by default embodied in the brain of organism and in this sense we could have scientific representation, supervene the relationship between the organism and its environment. But it is another story, it is not directly related to this Markovian modelism thing that I developed here. Well, it is super interesting because this embodied and enacted relational approach, we can't leave the scientist out of the picture. So it's almost like we remove too much when we abstract that relationship away. And then it almost hints towards that electron question from earlier, which is, well, if your ontology or your model is just an electron, not even the context, you're going to be lost because you won't understand why certain things are happening to your system of interest. And so how do we pull back rigorously and include enough context setting so that there is an ontology or a local logic that's making sense? So it's just cool like how the historical thread and the philosophical thread, it's not spice or salt on the research, like just adding it up. It's actually something that is intertwined and sets the total scope of the scientific journey. So it's like very cool to hear about that. So if there's any other questions in the live chat, we'll ask. Only with regard to a quick follow up with the last point that you mentioned. So part of it might be the contemporary philosophy of science, at least the dominant approach, which is the structure of realism. Is that precisely as you mentioned, it leaves those human beings out of the picture when it comes to scientific presentation. So there are very a bunch of brilliant philosophers of science, such as Ronald Geary and, for example, Nancy Narcissian, who is fortunately still active, who tried to provide some sort of cognitive model of scientific practice itself, of scientific activity. But for some reasons, and I'm not sure of the reason, it has been pushed aside and it is not the dominant philosophy of science again. So part of my activity in philosophy of science was to bring this sort of cognitive approach back to the foreground. And I think that I was fortunate that I bumped up to this brilliant principle and predictive processing, because in order to account for cognitive models of science, so sometimes we do not just want to sit in your arms around the philosophy of science, but want to support the philosophy of science by some actual sense of theories. And in order to develop this cognitive models of science, cognitive approach to philosophy of science, it seems that free energy principle and predictive coding are fascinating tools because they allow us to provide a very good and well supported picture of the relationship between scientists to our basic, the organism and the community of scientists, which could be specified in terms of, I don't know, free energy principle, theories of communication on the free energy principle, the relationship between community of scientists on the one hand, and their environment, the world on the other. And I think that this picture is much more realistic than the ordinary people, the orthodox picture of scientific representation that is fashionable. Yep. Cool. It's like a yes and with action and inference. So philosophy isn't just the armchair. What does it look like to be embodied and inactive? And then perhaps it looks like what is known by the career name of science, but we know that learning and research happens inside of academia, outside of academia. So it is a more realistic picture. And here's a related question from the chat. Stephen asks, this ability to dissolve the internalist and externalist dualist position is a powerful part of Markovian monism, thus giving an open dynamical system the adaptive open boundaries. Any thoughts on this? Yeah, I completely agree. So in the beginning, I told that I'm a great admirer of Markovian monism. And I think that it is a bigger step forward. And this is precisely because of this reason, because it goes beyond this kind of orthodox dualism between... So here it is described as some sort of dualism between internal spaces and external spaces. And Christian and colleagues put a lot of force and energy into developing this kind of informational geometries. Again, I think it is a wonderful framework to model this idea. So this is precisely, in my opinion, for what it is worth. This is precisely the valuable part of Markovian monism that it tries to use this kind of scientific model that goes beyond this dualism between internal spaces and external spaces to develop some sort of monism, metaphysical thesis. So I completely agree that this is a brilliant idea. So the question is what sort of philosophical argumentation or scientific argumentation could be offered to support this picture? So I think that as it is, the status quo of this theory is not as strong enough to carry the metaphysical weight. The question is, as I said earlier, the paper aims to come with the hope to provoke the free energy principle theories to develop this idea, to develop Markovian monism. So what sort of philosophical or scientific argumentation could be offered to support this insight that has been mentioned by Stephen and Liza at the core of Markovian monism? This is the main question. Awesome. So just in the last few minutes here, this paper came out, you know, one week ago, you probably wrote it several months before. But what are your next research questions or what are some exciting ongoing threads? Like where do you go? Let's say someone read this paper. Where are you taking this frontier of thought? All right. Again, cool idea. So as we have already gathered, my mind is very disorderly and it is because I'm pursuing several lines of research together. So part of my research right now is to develop this kind of model based understanding of Markovian monism to defense of modularity tests. It's not precisely the source of Podorian modularity, but at least the source of modularity that are defended by, for example, Jacob Hoey. And I think that the model based approach provide at least some grounds for defending the source of modularity tests. On the other hand, the other thing is that I feel that there is ample room for metaphysical work on this free energy principle theory of consciousness. So as you may know, there are already, so Markovian monism, for example, offers some sort of metaphysical, really, reading of theory of consciousness. But there are also eugenist, constereous, eugenist interpretations of theory of consciousness under free energy principle. And I think that for some reasons, none of these provide a quiet, good metaphysical interpretation of theory of consciousness under free energy principle. So I'm developing another theory. I'm trying to develop a theory of consciousness under free energy principle that could be construed along the line of strong intention on itself. And it is associated with works of historical works of Prince Baranto and Tinkray more recently. So it is another line of research that I have in mind to associate this theory of consciousness with some sort of intentionality. And then again, I'm engaged with, as the paper may indicate, I'm looking for further grounds for the reasons to develop or at least support this Markovian monism myself. So as I say, this paper is just the first, the negative part of some sort of argument that has to be produced by others or myself. I also am trying to, because I can see the appeal of this Markovian monism, so I try to find reasonable grounds for supporting this theory. I'm not sure where does it go, but it's an awesome approach. A lot of times a critique, it's almost like consequentialism. Someone says, I don't like your outcome. So I'm going to look for all the details or say that it wasn't really supported conclusion. And you're almost like saying, actually, I do see that goal or I see the benefits here and we just need to navigate in a slightly new way. But we're going to be learning and connecting these ideas along the way because it isn't mapped how we're going to get to Markovian monism. And maybe that's not where we're going, but we're working towards it. Brilliant. I just get excited and caught you because it is related to another part of my research, which tries to relate free energy principle to theories of group knowledge, to account for, for example, collective intentionality on the basis of free energy principle. So Princeton and Pierce have produced very fascinating theories of communication and social relation on their free energy principle. And I think that, I think that this theory could be applied to develop the parts of epistemology, social epistemology. And I think that there would be wonderful, wonderful reasons. So it is natural for me to see this kind of scientific back and forth between difference, intellectual, as some sort of collaboration that consequential, consequential as you phrase that masterfully, that hopefully leads to the growth of knowledge in general. Yes, the individual in the armchair, it lends once toward, it lends once almost to neuro reductionism and individualism. But when the person is walking around and at a conference and they're, they're sharing their experience and they're learning and being surprised by, by friends and colleagues, then there really is something that's more like distributed cognition. So I'll really look forward to seeing that. Majid, thanks so much for this appearance. This was epic and special first guest stream for our lab. And you're always welcome. Just let us know if you have any other topics that you want to come on a stream with you or colleagues or students or anything. We're always there to host and hold the space for this kind of work. So again, thanks a lot for the research and I'm sure everyone really appreciated hearing it. Thank you very much for having me and also cool work to do this active inference lab. So good luck with that and very well done. And thank you very much for having me. Great. We'll be in touch. So see you later, Majid.