 All right. Welcome everyone. It is September 7th, 2022. We're here in Active Livestream, number 48.1. In the second discussion on the paper, Communication as Socially Extended Active Inference and Ecological Approach to Communicative Behavior. Welcome to the Active Inference Institute. We are a participatory online institute that is communicating, learning, and practicing applied Active Inference. This is a recorded and an archived Livestream, so please provide us with feedback so we can improve our work. All backgrounds and perspectives are welcome and we'll be following video etiquette for Livestream. Head over to ActiveInference.org to learn more about getting involved in different learning groups and projects at the institute. Well, we're here in stream number 48.1, continuing our discussion on Communication, Active Inference, Ecological Approaches, and so on. This is going to be a really exciting discussion and we're appreciative of the first author joining us. We'll jump right in with some introductions and then on to the discussion. So if you're watching live, feel free to write any questions in the live chat. And I know that on our side, we have a lot of things to explore and also, of course, looking forward to learning more and seeing where it goes. So I'm Daniel. I'm a researcher in California and I guess the ecosystem of today is Chaparral because it's very warm where I am. And so I was thinking about how different ecosystems and ecotones might facilitate different kinds of communication in a regional way. And I'll pass it to Dean. Hi, so I'm Dean. I'm in the Rocky Mountains in Western Canada right now and what would you in Fahrenheit? It's about 45 Fahrenheit this morning, but it's going to warm up to about 90. So yeah, there's a bit of an ecology going on around here too. I think it's going to be a really interesting day because I think what we're going to do is take this ecological approach and lens and with the background of everybody here, we're going to really be able to pull up a lot of different interesting things that otherwise we would not pay attention to if we only got stuck on the transmission piece. And I'll pass it down to Blue. Hi, I'm Blue. I am a researcher in sunny, beautiful New Mexico at the other end of the Rocky Mountains from Dean. Like literally if you just go straight down, I'm in the Sangre de Cristos and we are recovering here from a like torrential summer of rain. And I think we're I think like maybe we're finally getting back to some sun. It's starting to get hot now like finally like so I think that the the sun is here to stay for still have a hot month, I guess, but the desert ecosystem and you know, the Chihuahuan Desert is special and fun and always excites me. And I'm interested too in this paper in how it more applies broadly, the communication as compared to the vassal paper, like can we then apply to animals and also to like maybe human animal interactions and what those look like. And maybe not only cooperative communication, but also like competitive communication. And I think like extending this work into those avenues is interesting and exciting. Awesome. Thank you, Blue. So, Remy, welcome. And please take it away. Feel free to introduce yourself and give any context on the paper. All right. So thank you for having me. I'm really pleased to see that we'll discuss our paper. I'm Remy Tzidon. I'm a researcher based in Montreal at UCAM. And yeah, it's sunny. It's a nice day here too. So yeah. The paper was an attempt to to device a view of communication that could be like a starting point for ecological and inactive approaches to language and communication. The problem that we saw is that in most language sciences, like linguistics, in philosophy of language, in mainstream philosophy of language, and in biology of communication, most researchers, most people adopt like automatically a view of communication in which communication is essentially transmitting information or transmitting content or an idea to another being. So usually like in linguistic communication, that would be like I have the belief that it's a sunny day outside. I say it's a sunny day outside. And then by doing this, I am transmitting to you the belief that it's a sunny day outside. So that's the basic view of communication, the transmission view that is more or less adopted by everybody in linguistics and in cognitive science, but a lot of people at least. And that's a view of communication that is problematic from an inactive and ecological perspective, because it relies on the notion of content on the notion of representation to transmit a content, I must have a representation in my head. And then I can transmit it to you so that you can entertain this representation in your head. But as is quite well known, inactive and ecological approaches to cognition reject the notion of representation, the notion of content as a basic building block of cognition. So the transmission view is a problem from an active ecological view. And so what we tried to do was to develop a view of communication from which ecological and inactive approaches could start to understand language and communication without having to rely on the notion of content or representation. So that's like the basic motivation of the paper. That's what we were trying to do. And we thought that active inference and especially its ecological interpretation could be useful for doing that. And yeah, the thing that you mentioned, Blue, yeah, the aim was to devise a view of communication that could be perfectly general, so that could be applied to animal communication, maybe communication between cells at a more small scale level, and eventually maybe communication between groups at a larger scale. So the view, the attempt was to produce a view of communication that is perfectly general, and that can be precise along the way so that we can, for instance, precise the conception of communication to apply it to cooperative communication, for instance, which is the kind of communication that is characteristic of human communicative interaction, and that was treated notably in the Vassil paper, to which you alluded to. So that's the aim of the paper. And so because of its generality, it's a paper that is really in a way superficial, because it really treats a lot of topics really quickly, but that was like a big roadmap for a view of communication that could be used by ecological and inactive researchers. So that was essentially what we tried to do in this paper. Awesome. And just one more follow-up question we often ask would be, how did you come to use active inference? What about the collaboration or your direction led to the fusion of these approaches? Were you studying communication more broadly and came across active inference or vice versa? Yeah. So earlier in my academic career, so to speak, I was really more in a philosophy of mind, sort of more of a philosophical background and not really in cognitive science. So that's where I became acquainted with like issues about content, about meaning, etc. And then I came at UCAM, worked with Pierre Poirier, which is really more centered in philosophy of cognitive science and was really interested in predictive processing theories and eventually in active inference framework because of its link with predictive processing. So coming at UCAM with Pierre Poirier, I really became acquainted with these ideas, with these frameworks. And that's where I linked these new frameworks with my earlier interest in language and communication that came from my background in philosophy of mind. So that's kind of how it happened and talking with Pierre and the community of researchers at UCAM who was really using the predictive processing and active inference frameworks. That's how it all came together. Awesome. Well, many places to pick up, perhaps Dean first, if you'd like to. Well, okay, I'll start with a question. If we're going to take an inactivist approach to this or we're going to look at what communication is through that philosophical approach and take all that active inference forwards in terms of opening up what communication is. Could you talk a little bit about how people who want to argue that there is no content, how could we then describe our experiences? The things that we've played out, the things that we're going to play towards, but mostly what I'm curious about is the things that have been played because we've got a record of that. We don't have to call it content, but I've had a lot of conversations with a lot of people who live in the inactivist thinking community and they come up with, they twist a whole bunch of things around, but they really have a hard time coming up with an argument that says my experiences aren't something, something. So maybe you could talk to us a little bit about that. Yeah, so that's a good question. I think that one of the answers that I could provide is that from my perspective, there are some people in ecological and inactive circles who refuse all reference to content. They see content doesn't exist at all, but that's not my perspective. From my perspective, content is just not at the basic level of cognition. It's like, it's kind of, I don't know if you're familiar with the work of Huto and Mayin. These are radical inactivists and they claim that their content is not like at the fundamental level. It's something that gets added up once there are a particular kind of social normative practices that can give rise to content. So that's my view. I think that we can describe the basic cognitive functioning of organisms and animals and in many cases humans without the use of the notion of content, of the notion of representation. I think that we can do that. But at some point in our social interactions, there are multiple things that can happen that will give rise to content to representation. And that's a complicated story. That's something that I might come back to later on. But the idea is that content is not at the basic level. It's not something that we can first suppose in an theory of communication. It's something that can be explained by a theory of communication. But it's not something that is, that can be taken as something that can explain. It's something that has to be explained, so to speak. I don't know if I'm clear. So we can, I think, use the notion of content to characterize the kind of experience, the kind of phenomena that you allude to. But this notion must not be taken as granted. That's the view that I take. And I think there are some ways to explain how content arrives on the scene. But it's not something that can be taken as granted in a theory of communication. That's what I'm critiquing in this paper. It's like in traditional transmission view of communication, people presuppose that there is content, that there is representation. And communication is defined as the transmitting of this content, of this representation. What I'm saying is that representation is not something that can be used as this basal explanatory notion. It's something we have to describe communication without the use of this notion. And then with various tools of social interaction, and particularly social normative practices, we can explain how content can then arise. So I can, if you want, give more detail on that front. But that's my two cents. So content exists. It's not just at the basic level of communication. Awesome. That reminds me of, and to stay on our philosophy, wavelength, the difference between the explanandum and the explanands. So just quoting Wikipedia in explanandum is a sentence describing a phenomena that is to be explained. And the explanands are the sentences adduced as explanations of that phenomena. And so it's a very interesting difference amongst these theoretical views and framings, which I'm sure will be able to unpack a lot of implications of that content as a phenomena or descriptive category, sign or symbol, we could have many ways of even describing what that is. However, under the transmission view, you're suggesting that content is seen as what is able to explain different phenomena like communication or cooperative communication. No matter how far out you go in the sub discussions, content is doing something to explain. Whereas in the radical inactivist approach, radical meaning from the root. So with inactivism at the root, that doesn't mean that there couldn't be a branch that summarizes and reflects content. But it's something that is explained through inactive means rather than something that we're going to put like the cart before the horse and then use the semantic content representation view and the transmission associated with it to then try to explain inactivism. Dean. Yeah, I think I I'm on the same page as both of you. I'll take an example in in 2019. I'm walking up the street rue Saint Denis and I'm heading to the Patisserie and I don't have any first hand experience with that. But once I arrived there and I taste one of those those chocolate pastries, now all of a sudden I have content. So it wasn't that I had content first, even when I was looking at the map and walking up from UQAM past all the bars and on a beautiful fall day, I had the content once I experienced something first. But I also can't deny that there was an aspect of that that wasn't limited just to the moment of all of the actions that I took walking up the street. So if we're going to look at an ecology and as I said in my first question, if we're going to use active inference as the lens as the situational analysis method, I don't know how we can drop content. But I do think the timing of when we see it as the important thing or the the central aspect of what communication is. I don't think it I don't think it destroys the fact that we still have it. It's just when does it matter in that in that continuum piece. So yeah. And by the way, thank you for just reminding me of that that trip in to Montreal in 2019 before all the heck broke loose with all these viral things. It was a beautiful time. I also went to Montreal in 2019. Very cool. Lou. Yeah, I am curious about the difference between or maybe there's no difference or how can we clarify the difference between content and context? Yes, great question. What do you think of me? Okay, so content is essentially something that has accuracy conditions. So the idea of content is it's something that can be false or true. It's something that can be or more generally something that can be satisfied or not satisfied. So for instance, usually we'll say that our propositional attitudes have content. So for instance, when I have a belief, my belief has a content and this content can be true or false. For instance, if I believe that it's sunny outside, I have the belief that it is sunny outside and the content it is sunny outside. It can be true or false. And if it's sunny outside, then my belief is true. So that's the basic idea of content is something that is attached to our mental states. It's usually associated with the notion of representation. There are complex debates to be had there. But that's the idea of content. It's something that has conditions of correction in which it can be correct or incorrect. So that's content. And context, I don't know if you mean in general or in the specific way in which we use it in the paper. How do you use it in the paper? Yeah, so context is a set of elements that are relevant to the joint action. So that's how we use it. A context of action in particular in the paper, we're talking about joint action. So when two people or more people are performing a joint action, they're doing something together, there are going to be some elements, some events, some objects of the environment that will be relevant on which some action has to be done in order to advance the joint action. So the context is like this structured, special temporally structured set of elements on which action has to be taken to progress the joint action. So that's the notion of context that we use in the paper. I don't know if that clarifies something. Well, so it implies like the set of elements is therefore also content. Is that true? In the paper, we don't think that it is. My describing it as a set of elements is a contentful description because I'm talking about it and there's a content to the sentence of me who's saying that there's a set of elements. But the set itself, the thesis that we're pushing forward in the paper is that it can appear in the field of affordances of an organism without it having to represent it. So that's what we're going to, in ecological psychology, there's work on the notion of a field of affordance. So it's like the set, the structured set of affordances to which an organism is sensible. And it's like there are some elements that will drive its action differentially depending on various internal factors. And this sensibility of an organism to its environment doesn't have to be characterized in terms of content if we believe the ecological psychologists to develop this notion. So this set of elements appears in a field of affordances. And these organisms who are sensible to their field of affordances don't have to entertain a representation of this set of elements. They're just sensible to their field of affordances. And that's how they respond to these sets of elements of their environment. So maybe when you talk about content, maybe that is what needs to actually be clarified. So content is specifically some kind of semantic representation or semiotic representation. So it's something that is explicit. But can you give me an example of what would be an element in a set of elements? How is that also not content? Yeah, these are obviously complex questions. But typically, affordance is a relation between an element of the environment and an action capacity of an organism. And the thesis that is taken by an ecological psychologist, not everybody will agree with that. But is that the perception of affordances does not need the mediation of a representation. You can directly be sensible to a certain pattern of energy, for instance, in the energy array around you. And you respond to this pattern of energy by, for instance, taking the affordance that is specified by this pattern of energy. And you don't have, it's not necessary to have a representation of the affordance to which you're sensible. It's something that you can directly perceive. And that's like a central idea of ecological psychology. Some people will disagree with that, will say that you cannot be sensible to an affordance if you don't represent it. So that's obviously something that you can argue for. But there's a lot of ecological psychologists to believe that you can be directly sensible to affordances of your environment just by responding to the energy array that is in your environment without needing to represent this affordance. So in doing that, they say that we don't need the notion of content or representation to characterize the way in which an organism is responsive to the affordances of its environment. And that's what we presuppose in this paper. But obviously, there are disagreements to be had there. There are people, obviously, you think that a different organism is to act on its environment. It has to be to have a representation of its environment in order to act correctly. So that's obviously something that can be defended to. Great. Thank you so much. No, that's super helpful. Thanks. It makes me think of a tug of war game, or just two people pulling on a rope. They're engaged in joint action, which is how action and activism are placed at the center of the account of context. One individual pulling the rope. They might be thinking or thinking that they think any number of things. But through the field of affordances being shaped through the regularities and the, I think, a raise of energy that you described in the niche, that is going to alter the field of affordances of someone on the other side of the rope. And that could be mediated through their proprioception, or maybe the rope is now not graspable, but it was graspable previously. None of that requires a symbolic or even necessarily semantic component to be transmitted along with the rope pole. Maybe they had a pre-play and they said two tugs means this and three means that. And that can also reflect planning to communicate. However, in the act of communication, as we're exploring, we're putting action and the embodiment in the environment as primacy, and then seeing certain types of cognitive agents as being able to develop representation like properties on top of some physical regularities in the niche. Dean? Remy, what would you think of this idea if you, if I were to, we haven't defined ecology here in this conversation, but if someone like me came along to you and said, can we define it as both a process, ecology as process and ecology as a product, would you push back on that ability to see it two ways? And then I guess on the general level, to Blue's question, then can we see it as content and context, this idea, this placeholder of ecology? If we keep it on that general level, can we have our cake and eat it too? Yeah, so ecology here really refers to something specific. It's not ecology in general, like in the branch of biology. In philosophy of cognitive science, when we talk about ecological psychology, it's a particular movement that began with the work of Gibson in the 50s, 60s and 70s. And it's an approach to psychology that states that the cognition of animals is the matter of its direct relation with its environment, and is a matter of the kind of skills that are exerted by this animal in its environment. And the central notion of ecological psychology is the notion of affordances. An affordance is a possibility of action. It's a relation between an element of the environment and a skill that can be exercised on this element of the environment. And the thesis of Gibson and all the ecological psychologists that followed after that is that organisms and animals primarily perceive affordances in their environment. They don't perceive objects, they don't perceive events or properties. What they perceive first and foremost is possibilities for action. And they are pushed and pulled by their environment in various ways and responsive to their environment to take some affordances and don't not take some other affordances. And that's how that's like in a constant interaction with their environment, organisms will respond to these affordances and do some things and will be attracted by some affordances, will be pushed by some other affordances, and that's how an organism interacts with its environment. So that's ecology here really refers to that body of ideas that came from the work of Gibson in ecological psychology. There are obviously links with ecology more broadly as a branch of biology, but it's really specifically like this set of ideas linked to ecological psychology and Gibson's work. Do you think that that what you just described the possibility could be for people like yourself both a product and a process? Is that possibility something that can be categorized two different ways at once? That's a good question. You mean like... Is it a category error to say it's only a process? That would be another way of asking the same questions. I think that ecological psychologists would more prefer the description of the interaction of an organism with its environment in terms of process because the organisms is always changing through its interaction with its environment. And obviously the organism is also changing the environment culturally as is described in active inference accounts of niche construction. So there's always an evolving interaction in which the organisms gain skills and the environment is transformed. So the field of affordances is like never static, so it can never really be described as a product. I don't know if that answers a bit your question. Well, are there products within what you just described? So for example, if I'm an organism and I've had experiences and I can now use those experiences to continue my process, are those things that I've experienced, do they just have to stay in the same category as now processed as in past or have they solidified? Because I can't go back and change my history per se. I can build off of it, but those are pretty absolute now. This sentence that I'm getting out is going to have a record that I can't erase. So are those things products now in the enacting of whatever I've decided I want to focus on in terms of possibility? Maybe, I'll just say something quick. Maybe when you take like a definite time slice of an evolving organism, you can say that at that point there's a product that can be defined. You can say, okay, at that point there's this set of abilities, this set of skills, this set of experiences that can be attributed to this animal. So there's a product at that point. But if you take the organism's evolution, like its entirety, there is always changing. So it's always a process. In truth, there's always a process. But if you freeze frame, maybe you can say, okay, at that point there's this and that is a product of the interaction. But I don't know if I answered correctly your question. But that would be my my idea. Yeah. You are answering and it's great. I'm just trying to get in my mind, if we're going to talk about relativity, and I know that I step into a river and step out and then step into it a second time, I'm not stepping into the same river. So what gives me the affordance to realize that difference? Is it a product? Is that first time I went in now a product? Because I can't go back and change it. I can move off of that and see the relative distance or the relative time. But I don't know how we do processing without product. That's I guess what I'm asking. One quick thought on that is that product doesn't necessarily mean singular final static. There are working products that can be versioned, which digital files have helped us see. And then also, we could say that the past is produced. But also it's reproduced when it's remembered, when it's brought back together. And even in the directions you point to towards the end of the paper with what kinds of empirical neuro physiological accounts are helping bolster the ecological claim. The realization that memories are not being accessed as static files, but are recreated contextually and relived is an interesting angle. Yes, Blue? So the recreation of memories contextually in dreams is really interesting because it's the recreation and also recombination in a different context. And just before we leave context entirely, I think I have a hard time with all of this because even a field of affordances has content. I can go fly a kite if it's windy outside. I have a set of actions that I can choose from. And I think because we have to talk about it in a way that's semantic, it's difficult to experience my field of affordances because I'm a human. And not only that, like I'm a very verbal human. So I think that there's like that. But wait, even the field of affordances has content. But I realized that it's the mental representation aspect that is the distinction between content and context. And I just want to point to the quantum paper where we had the quantum reference frames and Chris Field's paper on contextuality that like we did on the live stream, I think, number 40 and live stream, number 17. And just it's not possible to have a full contextual alignment here. And it makes me think of in the paper how you were describing like there's this person on one end of the behavior of the tug of war. And then there's another person on the other end of the tug of war. And then there's actually the tug of war itself. And I just wonder about like what is the, so person A has their context, person B has their context. What is the context of the tug of war itself? Like it's another entity. Anytime you have this like interaction, and this is something that I've said to Daniel before, like you have a relationship with a person, you have a set of communication with a person, like that's its own thing, like that has to be like loved, protected, nurtured. Like you have to keep it up. It has to like it has to undergo like auto poisons. It has to self assemble, right? Like it's its own entity. And I wonder about modeling in that way. Like if I model my communication, like there's me and the person and then the interaction or the collective behavior or whatever it may be. I don't know. I just it's I've been thinking about it. And yeah, that's all. It's like a view from the bottom or inside or however of communication where there's two entities in a partially aligned reference frame. And then there's also a view where it is just one thing, albeit maybe with partial information encapsulation, but it still becomes one thing. And then is that one thing atomic, even though it contains divisions and partitions as well? Remy on that or anywhere you want to go? Yeah, I think blue, your interventions that were really are really, really interesting. So for the first thing that you said, I think I know that it's hard to, to, to, to like remove our, our, our habit of describing things in terms of content, because we have language and we describe things in terms of content. But it's something that has been remarked by really early by a phenomenologist, like Merleau Panty, and I dig our that our primary contact with the world is not some like a description. It's not like a representation. It's not like a reflexive consciousness. It's a pre reflexive consciousness. And in your day to day interaction with your environment, you're not using this kind of language like description to experience your environment. Like if you're playing a sport, and like, for instance, with your playing soccer, and you, you see like a pass that you can make, you're not reflexively thinking that, oh, you could make that pass. It's like a push or a pull in the field, in an attractor field kind of thing. And that, that's how we experience. And at the second where us humans, especially, we, we begin to think about our experience, then we cannot help but describe it in terms of content, in term of a representation that we entertain in a linguistically with a linguistic means, but in our pre reflexive engagement with the environment, that's the thesis of the phenomenologists and ecological psychology. It's really just a push and a pull from the environment that drives us in different directions. And we don't have this kind of reflexive consciousness, this kind of representation on which we rely to describe and understand it with these linguistic means. So I think that's really the difficulty that we have of like, getting to this pre reflexive consciousness when we are so used to, at the second that we describe something, we are describing it in terms of content. So we have, it's hard to like say, oh, there's no content when we talk about it, we are invoking content automatically. So yeah, there's a difficulty there. And about the autonomy of the dynamic of an interaction, that's something that's really, really interesting. And there's been a great work on this by Ezekiel DiPaolo. I don't know if you're familiar, he's working, he's an activist researcher, he's working on language too. And I'm, I have been really inspired by his work. And he's describing situations where like, there are two people interacting. And these two people have goals that are aligned. But because of the particular dynamic of the interaction, these goals don't come to fruition. For instance, when you're walking in a corridor in somewhere, and you're facing somebody, and both of you want to like, fluidly like pass each other, but you're both trying to go for on one side and then on one side, and you're blocking each other, you're trying to achieve the same goal, which is individually passing past each other. But because of a dynamic that arrives from your interaction, you're not, you cannot achieve this goal. So there's really like this autonomous dynamic at the interaction that is really important. And I don't think that obviously that's something that we wanted to put in the paper, because we're talking about the interactive dynamic, dynamic systems. But I don't think that I've succeeded in translating this in the way we think of shared field of affordances, because I, I, I define shared field of affordances as the subset that is common to both individual fields of affordances. But I think that neglects this part, this autonomy of the field of affordances that is shared in, in the interaction. I don't know if I'm being clear, but I define the shared field of affordances as a product of the combination of the two individual fields of affordances. But it seems that there's a lack of autonomy in the shared field that I don't know how to account for. That makes me think of the ant case, where one nest might be able to grab a single seed and carry it home. So that would be compatible with a view on communication as shaping the field of affordances, because it modifies the capacity for another nest mate to pick up that seed because it's been moved, doesn't need to have any representational content, although especially when we're linguistically describing it does. So I definitely want to come back to that. Then there's this seed grabbing ability, which is predicated on a lower level affordance, like the ability to grasp with the mandibles. And that lower level affordance enables true joint action, like multiple nest mates working together to drag something bigger home. But dragging something bigger isn't an affordance that even exists at the single nest mate level. So there's like shared affordances and joint action at the level of what we can both do. And then there's also this very interesting higher level view where perhaps novel affordances become accessible, but then whom or to what or when do these special affordances or maybe nested system affordances come into play? And how will we mediate the handoff or the way that those different levels of affordances come into relevance or out of relevance? Blue? So in terms of that modeling the field of affordances, it changes through time. And this is why I was thinking about a competition, like in a tug of war between two people, both of them have the affordance to win or there could be a tie. But it's a potential opportunity for either one to win, but only one can be the winner. And so then one person winning by default makes the other person lose. And so this field of affordances changing through time is really interesting and I think very broadly applicable for modeling all kinds of processes. Yeah, that's true. And the idea about the collective affordances, there's been some interesting work on that and clearly some organisms like ants don't perceive collective affordances. They perceive if they perceive something, they perceive their individual affordances and at the collective level, there's the taking of this higher level affordances, this collective affordances, affordance, but they don't perceive individually probably. I don't think they do perceive individually the affordance, but some organisms such as humans, I think because we have maybe a deeper generative model or something, we are able to perceive some collective affordances because we're sensible to them. For instance, if I'm moving with a friend, I can perceive that a couch is unmovable by me, but I can perceive it as movable by us. And then I can say like, oh, we could move this couch together because I perceive it as something that can be acted upon, not just by me, but with somebody else. And then I have a perception it appears in my individual field of affordance as a collective affordance, but that's different from the ants who achieve a collective action, but without having the perception in their field of individual field of affordances of this collective affordance. That's awesome. It makes me think of the human capacity to have counterfactuals and speculative realism on teams and projects like, hey, everybody here, we could do X. So there's the ability to communicate semantically and representationally where needed plausible narratives or little submeshes, new things can be suggested and arise that are specifically about shared or collective affordances. And then additionally, there are special names for when we take individual level affordances that have consequences on other levels of decision-making, like there might be a multi-sig wallet in cryptography or a voting system in democracy where an individual's affordance is like to pull the lever or to fill in a box or to sign a transaction. And then it's being engaged in specifically because there's knowledge, if 100 people do this, then this affordance is going to happen. So it seems like there's many cases where we do our mandible closing, the kind of lower level individual task, also with an explicit knowledge and a communicated knowledge around how someone's individual actions will influence group outcomes that could be like weekly emergent, like merely summating the affordances taken by individuals, as well as strongly emergent or more strongly emergent in the sense that there's nothing different about anyone's pulling lever in voting, but there are phase transitions and dynamics that can be like top-down, Dean? Yeah. So when we have these experiences and they are kind of absolute in the sense that they've happened, they've occurred, and whether we can recall them or not, they seem to have some kind of an effect on this relativity piece and the relationship piece. I don't think it matters whether we see Newton on one side and then Chris Fields and Einstein on the other. The bottom line is that, to what Daniel was just saying, the possible happens when we don't get stuck on what's the gauge of the transmission line, but what does that now allow us to do? So how do we move from the absolutes, the modeling piece to the process piece, the things that are out there that have not happened yet, that are still in that possibility category, and now we're going to use communication, not just to narrow down to the workbench, the ontology, which Daniel and I are our fast friends, except when it comes to the specifics of ontology, the generalities of that and the potential and possibility of that. Maybe you can talk to us a little bit about that from the ecological lens. Yeah, that's a really good question. I think the one interesting thing about like collective and shared affordances is I think they are learned in the same way that we learned individual affordances. So for instance, how do we learn? How do we learn to perceive an affordance? We learn to perceive an affordance when we gain a skill. When we gain a particular skill, we then can perceive in the urban environment the situations where we can apply the skill. So for instance, when we're really young, we can maybe crawl a little bit so we know where surfaces can be crawled upon and we can grab things so we perceive where things can be grabbed. And so we start from this like basic skills, and then we explore environment. We do some things that we were gaining some abilities. And when we do that, probably a little bit by trial and error, but we begin to perceive some higher level affordances because we can do some more complex things. And so our field of affordances depends and becomes more complex as our set of skills becomes more complex. But I think that interacting with others, sorry for my accent, it's not really good. I didn't want to say interacting with the animal others. So when we interact with others, we are also learning to do some particular things. For instance, if I try to pick something up alone, I might not be strong enough to do it. So I don't have the skills. So I don't perceive the affordance. But if in some cases I picked something up and I see somebody else picking it up with me and we succeed in lifting it. And then I have gained the ability to see that with somebody else, I can lift the stick. But it's not a different kind of learning in that case from the kind of learning that I apply when I'm learning individual skills. It's just that interacting with your environment, you acquire skills, sometimes with somebody else, sometimes not. And you understand how you can apply these skills in your environment. And then you navigate your field affordances accordingly. So how you perceive a possibility is about how you have learned to do some things in the past. And then because of this learning, you perceive situations as affording these kinds of actions. They can be individual, but in communication, they are usually about what others can do and what we can do with others. And that's how you perceive the possibilities that can be brought upon, that can be done with the person with which you're interacting. I don't know if that helped, but that's how I see it. From a modeling standpoint, if you move away from the constraints and you look at the emancipation and the freedom piece, it obviously becomes combinatorially explosive very, very quickly because it can go off in many, many directions. But as for example, as Daniel alluded to when we started today's live stream, this could go down many different paths. So he left it open. He left the freedom aspect in. And so I was curious what you think about. So when we do leave that in, when we do move out of the laboratory or we do move away from the mechanics, which you were just describing, and give our attention over to the freedoms piece of it without, without immediately assuming, oh, well, we're going to stop modeling now because this is just going to blow up on us. I was wondering what you were, what your thinking is around that in terms of communication as a platform for the possible? Yes. So the idea of communication is, is that the role it plays in interaction is, is the constraint. It's a constraint of the, on the set of possible action that can be achieved. So the idea is that the, the, what communication achieves, it's, it, it helps to limit the kind, the, the, the scope of affordances that will be salient at a given moment. It's like you will center like the, the, the shared field of affordances around a certain set or a certain affordances and this will limit, this will restrict the, the, the set of affordances that will be considered or that, that can be acted upon by the, the individual participants. So that's the, the, the role that it plays. I don't know if that, that helps, but that's how communication works. I think in, in interaction, that what it does, it, in, in dynamic systems terms, it, the idea was that it helps to create interpersonal synergies in order to achieve the joint goals of the interaction. That was the, the suggestion in, in the paper. So for instance, two people being together in a room, they can be a lot, they can do a lot of things. There's like an almost infinite set of affordances that are open to them. But if one of the people in the room like goes to the couch and say like couch or manifestly tries to lift it, then what is, what is conveyed for, to, to, to take the, the transmission view idiom, what is conveyed or what the effect that it has, it will augment the salience of the, the saliency of the, the affordance of taking the couch. And then this will restrict the set of affordances that are open or that are salient in the shared field. And then the other person will like see, okay, that's what we're doing now. And they're going to take the couch together and lift it and bring it. So, so that's the role of communication is like, there's an all set of affordances open at any moment. And what communication does, if it restricts, it constrains the set of affordances in order to better achieve the, the joint goals of the interaction. That's what we try to do in the, in the paper to understand cooperative communication. You touched on an important difference though, between is and if. So what is, does actually put, put pretty strong, what do we call them? Guard rails constraints. But if I ask what if, I haven't put a constraint on something, I've looked now at the counterfactual and the possible. And I haven't said that the first, the first person that answers back to me is going to necessarily be now the new constraint. The whole point of if is the possible, right? And it is the freeing up. And it's the whiteboard moment. And I'll be know all of this. So again, I understand in the context of this paper, what the is part of it is, in terms of joint action. But is there a, what is there a place in your research here for the if within the same arguments that you're, I believe that there is, but you didn't speak as clearly to them as you did to the is parts, the, the constrained parts. Yes. So the, the, I think what you're talking about here is the, the, the communication about like future or counterfactual situation. That's what, that's what you're talking about. Okay, yeah. So what, what we, we try to do in the paper is we introduce the notion of a simulation of a field of affordances. And that's where we're, we're dancing really closely with like the notion of representation because are there, is it a representation of an affordance or like, that's like a fine, fine line to tread. But so, so there's this notion of a simulation of an affordance field, some sufficiently cognitively complex organisms such as humans and probably some primates can, can produce a simulation of an affordance that will enter in this year or hierarchical affordance competition. So you can like activate an affordance that is not immediate immediately present in order to affect your current field of affordance to, to, to, to for instance, go toward this, this simulation of affordance that's based on the work of Paul Chisek, which is at, at Montreal and his work on affordance competition. And he argues that a lot of, of phenomena linked to cognitive control, so just such as executive function and these kind of things and working memory can be understood as these, these effects of affordances at higher hierarchical levels that will constrain the lower levels of the field of affordances in order, for instance, to act accordingly to this higher level affordance. So the example that Chisek gives is, for instance, suppose that there's a monkey on a, on a branch that likes to see the little berry just, just beside it. So there's the affordance of eating the berry, but suppose that there's an apple, which is a bigger fruit, a bigger price, further down on the, on the branch and the monkey has an understanding that there's this apple further there. He can with this, this simulation of the affordance of eating this apple at the higher level. So like if I do that, if I go there, I will get this apple. This simulation of an affordance at the higher level will constrain with diminish the salience of this immediate solicitation that, and it will help him to like ignore this immediate berry and it will continue on the branch in order to get to the bigger to the apple. So there's like the basic idea of how the simulation of affordances can play like a cognitive control role in order to constrain our behavior according to further goals like in the, in the future, future situation. So that, that's the idea on which it's, we, we speak about it really, really quickly in the paper because it's already a really big paper, but that's an important point. And that's the key to understanding the communication about future and counterfactual situation. We talk about it in the last section. It's the displacement or the absent reference discussion. So the idea of how can we talk about things that are not immediately present in a special temporal sense. And the idea is that we can activate shared simulations of affordances. So I can speak about an affordances that is not immediately present, but it can affect our shared field of affordance and it can drive our behavior toward this affordance, for instance, even if it's not immediately present. But yes, this poses the question of is, are these simulations of affordances representations? I'm not, I'm on defense on this question. But if they are, they are probably like action oriented representations that maybe would be acceptable to an activist and ecological psychologist. Sorry for my really long answer. That's something that, that's something that I've really thought about a lot. Nice. Yeah. Blue want to add anything or there's a few ways to go? No, I'm good. Go ahead. Take it away. All right. Well, one interesting connection with active inference. You mentioned there's nested affordances, including ones that include that are plausible. So that reminded me a lot of the attention and metacognition as mental action, which we explored in live stream 25 with Sandvitz Smith at all's paper, where the same way that at the kernel layer of the partially observable Markov decision process, or however we're doing active modeling, at the most basal or kernel layer, those pies, those policy selections and those ease those affordances have to do with the level of affordance that is embodied, like the actual actions that the eyes are taking in the ocular motor case, or the actual direction that's taken left or right in a teammate's example. And then nested modeling is using an analogous graphical structure where policies of nested levels reflect attention being paid to various lower level lower levels of that hierarchical architecture. So there could be one lever that can go left and right and one that can go up and down. And then there's a higher level decision, like which of these two levers should I be paying attention to and how aware am I of which lever I'm paying attention to, and so on. So I think that paper itself does not present any formalisms, but there's extensive linking to work in active inference and also to really resonant ideas that are in a non slash pre active inference framing that are immediately compatible like nested levels of affordances, the distinction between the primacy of the bodily affordances, but also for certain cognitive systems, the existence of what appear to be cognitive affordances and the desire to model those jointly, perhaps without taking on strong or specific baggage around how we are then constrained to think about the bodily or the cognitive aspects of that modeling blue. And then one other thought was on like shared action, including shared planning. So planning to act together and about how events or activity planning in the future is like an activity that one person can do or groups can do. But it's about preparing a context so that a process can occur. And sometimes all that's needed is to send the calendar events invite. Other times there might need to be a lot of work to be done before the event itself. However, what is the transmission view on planning that might be able to describe. And I think the paper does an excellent job of saying that the transmission view is not refuted by the ecological and inactive angle. You said no transmission, no communication, no communication. It's like, it's an enabling aspect of this action orientation and this focus. Here's some other areas that we had written down from before. One I'm really curious to follow up on is about like language and maybe the role of how we're talking about these issues. What other media or formats have you found to be insightful? Do any inactivists do presentations with silent films or with objects that are mailed to you and you play with or do this? How else can we point at the moon and not have it be confused that we're talking about finger anatomy? Yeah, I think that's hard to talk about academic subjects like this without having an explicitly transmission view, like appearance of communication because these kind of interactions are so far from the basic interactive activities that were like the context of the evolution of communication, which is what I'm trying to explain here. It's not like, you imagine like you imagine like a tribe somewhere in the Pleistocene and like they're not trying to explain academic subjects to each other. They're trying to do things, concrete things together and that's really the context of where communication evolves and what purpose it serves. But I think that it's hard to envision the academic discussions such as the one we're having now that would be like that would be more evidence toward the ecological view than for the transmission view. I don't know if that's clear but clearly like moving a couch together is a better instance of and the communication that takes place in this joint activity is probably a better example of our view than like speaking together about active inference and communication because this was really a specific and really highly complex form of social activity that is more readily described for instance in terms of the transmission view I think. Yeah, nice a lot to say there, Blue. So I think all of those things like even silent films are also representations, right? Like so whether it's a communication like with communication or even pointing to the moon would be like a well maybe that would be different. I don't know. I like this idea of communication as an action that we take and I'm not leaving information transmission behind in the dirt. I don't think it needs to go there I think but I do think that all actions are in some way some kind of information transmission. So even like and Remy you've studied theory of mind so I assume you're familiar with integrated information theory and the emergent properties that information can help us to represent. And so I think like in some way even like eating food like tells the environment that you're hungry and takes away the food and so there's always some level of informational exchange even if it's just trivial information exchange even if it's not meaningful it doesn't create some bigger emergent thing, right? Thanks, Blue. Dean. I think it's an easy again I think it's it's well I won't say it's easy it's Remy did a good job of explaining this and not making it too obtuse. If you use an example of moving a couch the physical manipulation it's not hard to extend that to communication and symbolic manipulation. The manipulation states then the then that sort of reinforces the idea that there is there is a process at hand something that we can observe but I'm still kind of confused about so once that the chair has been lifted or the the words have come out of my I heard an expression because I was watching the John Vervaki thing yesterday and something comes out of my mouth hole now it's it's done it's a product I mean we can read we can resubmit it or reflect on it or whatever but it's there it's now historical and so that transition from process to product I think is something the manipulation by itself doesn't explain we have to kind of delink and see the sort of timelessness aspect of this. Carl talks about deep temporal models and so I don't know that I'm not sure how communication manipulation symbolic manipulation gets explained in these deep temporal senses and I'm not just talking about chiseling it chiseling it into the the the stones the tablets right like there's got to be another piece to this that you've looked into Remy and obviously much deeper than I have what what what is your sense of that once we get past the manipulation processing and start seeing the the products pile up. Yeah so yeah it's hard to answer this question but that's true of everything that is something that happens at some point it's it's it's there and you can't change it so the any event communicative or not or like a rock that falls or something that has that property too no it's once it's happened it's it's something to which we can go back and I don't know if that's so so in the same sense that if you you talk you say something this is something that that has happened and you can it's it's now a product and something to which you you can allude to it has influenced your cognitive system and then you can I I'm I struggle to see what's the question exactly well yeah one of the things that you pointed to right at the very start of your paper was that what these symbols that have been manipulated and now set down in this paper is not the end it's the start of something this product this paper should be the start of a dialogue or a trial log or in this case a quadrilog and it has become something right so so it's the back and forth part that I find really interesting and it's something you spoke to very clearly at the very start of your paper you didn't say okay so if we get stuck in the transmission view when when I stopped talking that's it that's over and now we're going to start something else no you basically said there's a continuum aspect to this it isn't strictly episodic that we have to look at something more than just what Remi says which I think it's a brilliant way to start a paper and an even better way to start a conversation so so there is a product piece to this and there's a process piece to this you put something down and a child will start manipulating it and then they'll start manipulating the symbology that they're applying or affiliating or associating with that so again you've looked at that far deeper than I have obviously because listen to how I ramble right but I wondered what you like how do you take that transition piece now and hold it up to other people who maybe take an inactivist poster or or take more of an instrumentalist approach right because the bottom line is it's hard to deny that it's it's done or it's it's still happening but you focused on the it's still happening so what are your thoughts yeah I think that there's with communication there's a sedimentation the paolo talks about this too like the the symbol that we use the the the words that we speak there there's like this um there's also a speech talk about a first order first order languaging and second order languaging so first order languaging is like the flow the dynamic where every interaction is new everything is always like on the fly and there's always a transformation of the linguistic forms that we use and the second order languaging is like the abstraction where we can say oh that's the same word that has been used that the same sentence that's the same meaning and this second order languaging is like a something that's added up and that that's an abstraction that's something a traditional linguistic has really focused on this second second order languaging which is like the product of language which is like okay the the the the language that the language that we speak is a set of sentences a set of words and that's something that that is identified at the level of the second order languaging but when we're speaking where we're talking it's not at this level that we're doing it when we're acting it's always new we're always reconfiguring we're always transforming and it's only when we abstract from this like basic interaction the some patterns that we see okay at the second order at the second level there is words there's a syntax there's some rules that are followed but it's not the the the the real phenomenon of language is at the first order and that's the so maybe that's the way to see like the real activity is really a process but if we abstract if we look back on this process and we analyze it and we abstract some parts of it then we can see these forms emerging these languages these words that are used repetitively and then we can have like the product of language but in truth language is always the process and that's that the shift the the terminological shift of not talking about language but languaging which is something that is that an activist working on language try to do that really brings forward this this notion of language is always an activity always a process and not like a set of words a set of symbols that are used and these these analysis in terms of repeatable symbols is something that we get when we analyze it like after the fact and we see these patterns that are that are used that we get the second order but when we're doing it and at the fundamental level it's really just a process of continual interaction so that's maybe something to think about and there's something I want to say really quickly about what blue said about the information and the the the way in which information is always present and that that's true that that's what I allude to at the end of the paper when I cite uh Brian scurms when I say yeah there's always information transmission right just moving in my environment there's the information that's transmitted but yeah that's exactly the point and that's the idea is that content is not information information is everywhere everywhere always when we when anything happens in the environment in the universe there's information that is created as transmitted but the idea is that information understood in this in these terms in uncertainty reduction notion of information is not sufficient to explain communication because there's information always everywhere that's that the point that that that people who hold the transmission view think that they can rely on this notion of information to explain to explain communication but it's not a strong enough notion of information they need the notion of content to explain what they want to explain because for instance the information that you think about when I eat when I do wave my arms like that there's information because there's an exchange with my environment but that's not a sufficient to explain communication because this information is everywhere and communication is not everywhere it's a property of certain behaviors that we have to be more specific to analyze thank you blue so the information has to have context right it has to have relevance for the communication for the collective action for the collaboration or something so so there's no like in order for communication like that that contextual I don't know imposition needs to exist I think yeah nice so just a few notes I really like this sedimentation of communication it's a really rich like riverbed or or rock formation natural analogy that's that's very concordant with ecological perspective to reflect how the products of communication are indeed sequentially layered the um pen marks on the page can't be erased we can't undo what was said we can't unrecord we can't untransact on a blockchain all these different aspects and even including the scientific literature being appended to continually so there's a sedimented aspect yet also there's something that is a process like the river that is depositing the sediment and then that made me think about how sometimes communication is descriptive around what was or is and this may map on to the variational free energy inference which is kind of like now casting or even retro casting but once we enter into action and anticipatory or prospective informing and communicating looking forward to reduce uncertainty about action selection is vital for systems to persist and thrive however as we've explored in many discussions sometimes from a really technical angle the introduction of action into model selection and just modeling in general introduces several complexities how the space of possible actions is enumerated and scanned over it's a simpler and a different question to do purely receptive inference on something about the past but to inject one into the present and and embody that presencing requires one to act amidst uncertainty about what the future of the general process will be and the consequences of action which are either slightly or almost entirely unknown so the introduction of action puts us into that realm of expected free energy and highlights all these really interesting dynamics and how the future is not the past but it becomes that way like it gets sedimented but this is the pre sedimentation there's some dust molecules the right next to each other in the river one of them is going to sediment the other one's going to go out to the pacific ocean it may not even be clear in that river which way the different particles are going to go and so that's like where like all the complexity and turbulence can be happening and so it's a very rich distinction and something to really bring back about how the past is sedimented but there's an openness in the action consideration and that's where planning as inference and joint action and shared attention all these features come into play dean yeah i don't know if you're or if you're indirectly referring back to the 47 series of the live stream but there was what we talked about was that people who have looked through an inactivist lens really make fantastic arguments about let's focus on process let's focus on the process apologies if if you're talking about people who work in the instrumentalist hopefully that'll get answered if you're talking about people working in it through the instrumentalist lens they really want to talk about product they really want to say this is a tool and they argue fantastically for that and i think what we tried to do in the 47 was say you're both right you both have fantastic arguments as an agent who's trying to communicate why would we say one is necessary like if we're active in furs why if we're active in furs first and foremost if we're always trying to scan and survey that field of affordances why would we drop one for the other unless that was necessary well even part of affordance the the ability to option and be able to recognize the tool product and recognize the the action in that moment carrying out that that data thread which we can now act further on as Daniel said downstream right so again i don't think anybody would argue that inactivism and that ecological approach is hugely important in order to be able to generate those affordances but maybe maybe there's still people who want to argue know if you just bring your toolkit all of your all of your problems will suddenly dissolve or if you just walk into the room you'll know how to talk to all of these people without really developing that that sense of okay so as you said first order and second order communication right you just walked into the room and then you reflect later on so that the next time you walk into a room maybe you you come at it a little differently each time right so again Daniel i don't know if you were consciously or subconsciously referring back to the last live stream set but that's essentially where i'm kind of that well i've been there for a bit now but i wonder how you can argue that it could only be one and if we're talking about affordances if we're not talking about affordances fine it can be one but if we're talking about affordances and we're optioning out how can it be less than two the menu is not diminished by selecting what you want for that meal you don't rip the rest of the menu out and then also remedy one one uh just insight i've never heard it phrase this way but it's so interesting and hopefully gives many people including us some directions is you said it's hard to discuss academic subjects without the transmission view and that's just such a fundamental point one could take a sort of sociological critique of academic discourse and well the practices and the networks of academics and the genres that they write in the venues and the the paywalls and all of these different sociological features structure academic discourse to be such in such a way that can feel lifted from our first order experience and this actually takes an even more fundamental angle which is yes those sociological features may be what is explained but fundamentally on on that topic or anything what is it about this conceptual or academic mode of synchronous and asynchronous discourse just what is it about what is it on and how can we evolve this and develop this to create an ecology of communication to integrate in a new way for Remy first and then anyone so yeah i think that not just the the academic discussions but all our conceptions of education and learning are influenced by this transmission view also because all the traditional views of of education is the notion that you transmit a certain knowledge a certain view a certain content to like the the student you're the teacher and you have this this content and you're giving it to the that's like the the traditional view of education and there there's been pushback in recent years in the in philosophy of education and trying to understand education and apply more participative methods of learning that that go against that but really the view of education as transmitting knowledge i think is really closely connected to this transmission view of communication and i think that's not a particularly good view of of applying pedagogic pedagogic practices i think that's that's really limited and there's been the the event waited for my for our paper in 2021 to do that but yeah there's been in recent years a lot of work in philosophy of education and trying to push back again this against this transmission view of education and trying to put like having classrooms that are more interactive and in which there's a there's active learning from the from the the students and they're just not just receiving a knowledge but they're gaining skills by by acting in the in the in the in the in the environment a learning environment in which they're agents and not just patients so that I think that's something that's also useful to to to to discuss yes we had some great times discussing instructionism and interactionism in the context of skillful performance and motor commands and the discussion was are descending motor commands representational are they instructions to be carried out or are they interactions that are being engaged in that might not necessitate taking a representationalist view and in the same spirit and strength of the active inference framework that we can talk about patterns across systems here is instructionism and interactionism in a very natural setting of the ecological experience of learning and one can take various critiques or angles on like tabula rasa or blank slate we don't need to endorse the view however what it implies quite directly is that there's a canvas and things are added sedimented on the canvas and one can say that some writing is on the canvas or it's a different shape or those are specifying the metaphor but fundamentally the knowledge accretion model of education the opt-in model as well as the ignorance destructive model of education which is a kind of counterpoint to we're going to add good knowledge and then the shadow of that is we're going to destroy bad knowledge and unlearn both of those are like working in the riverbed are we going to sediment this way or let's modify these sediments it's still seeing the product of education which is a totally reasonable focus for industries and sectors that are focused on the products of education the products of education for the student the certificate and the social recognition and status thereof or just the outcomes of education for the system like amount of people with a certain professional training and what is the river that is sedimenting or de-sedimenting that can get dropped and that's the active turn in education learning by doing as if even a lecture is doing something you're looking you're hearing you're touching so it is quite an exciting area to explore how interactionism in settings that have been considered educational conventionally or unconventionally what the implications are and I'm sure that next week we'll be able to explore many more implications as we kind of have crested one hill and look towards some implications but how do you communicate differently with what you know and have done in philosophy and science of communication? Remy? I think I'm really bad at applying in my own theory of communication I think I give a class at Ucam and I always like I'm angry at myself because I really like just automatically applies this really classical way of transmitting knowledge and I'm always angry at myself because I don't apply enough I think my own ideas about communication but I think I should do more to provide better evidence for my own theory of communication it's funny I'm sure it'll be a lifelong process for all of us because we have priors that are templated on the forms that we know that's what was and what is and what maybe and then there's that expected free energy like active communication perspective and are we going to select only from yesterday's menu or can new dishes and new approaches be added in that we may want to select from are there new modalities or approaches to communication in educational settings or otherwise that are now authorized or enabled by this perspective I'm sure we'll continue to discuss it but let's just say this paper becomes a fundamental reading for a communications department or for a early life learning group like a public school what is different in the short term with the structural constraints and opportunities that exist with someone thinking and acting differently and then over slower timescales the co-evolution of those individual affordances cognitive and bodily affordances and the kinds of systemic alterations to the education architecture dean I know you'll have a thought on that part of what you were speaking to is simply asking as as remi did at the beginning of his paper and then what happens but then not racing to provide the answers it's it's it's essentially opening opening up the space and then and then what happens might mean that you you move out of the classroom and you and you take the classroom migrates to to the lookout on Mount Royal or you or you you simply don't intervene you you kind of let the ecology work itself through now that can be very scary at first because of course you're supposed to be the one hurting the cats but if you let the herd piece of it off and see what cats do when allowed to sort of free think I think I don't know that it provides the the absolute specific answer but it will provide you that joint action opportunity if you if you're really truly looking for affordances as opposed to specifics and I and again I think even in an academic setting I don't think it's I don't think it's too far of a stretch to say affordances matter and we would like to see if people can exercise those things we just want to get too crazy right norms and guidelines or affordances help don't know what else to say but certainly there are things that one would prefer not to see occur in a learning setting right easy to enumerate and then there are the known beneficial affordances or phenomena and then part of interactionism seems to leave that unknown unknown quadrant open and to allow for individual or for collective outcomes that are surprising and maybe not explicitly navigated towards or planned for but how to leave the unknown unknowns open while de-risking unknowns and knowns is a challenge and so again it's natural to see where in the de-risking imperative we de-risk everything and that introduces like second order risks and it closes the door on what may have even been the purpose I mean what's the safest car the one that doesn't move in that one limited local optimization of like reducing the number of high-speed collisions or something like that um yes Remy yeah so I on that topic I think there's uh Brunberg and Riedveld and Kiva Stein talk about an optimal grip on the field of affordance I didn't say it I think but their work was a great inspiration for for my own work so they talk about an optimal grip on a field of affordances and there's like you're always trying to stay in the point where you're there's like the relevant affordances are open to you but you stay you stay at the same time open to switching to other affordances and you're trying to in that process of interaction with your field of affordances continually minimize your your own free energy but there's there's risk in that so there's the the idea the example that they give is the boxer who's trying to maintain a certain distance to its opponent because at that distance if the the boxer is farther it cannot reach with its punches if it's too close maybe there's less impact and maybe it's more open to to being it so there's like this optimal distance to which the boxer has like flexible multiple opportunities that it can that that boxer can take and I think that that's what we're always trying to do we're trying to get to this optimal grip of the the field of affordances in which there's multiple affordances that afford the free energy minimization and that we we want to take that and that's obviously something that can be hard to do but we have through our skillful interaction with the with the environment we learn these kinds of of gradients where where should we go in in our environment in order to find these optimal grips and like you're speaking about car I guess depending on your skill you might want a different car to get this optimal grip on the on the on the road if you're like a fast driver who likes to go to go really fast you might want a faster car but if you go too fast then you're you're getting an accident so there's depending on our skills there's different optimal grips on our field of affordance and something was really afraid to go fast might go with it like release low car and that would be it's that person's optimal grip on on the road that would be the this car would provide it for for this person so that I that made me think about that so some comments on optimal grip thanks for bringing in yellow at all's work it's it's really nice that that was brought in even several years ago into the active inference literature and if we're imagining a tactile grip actual prehension if we're not in a grip strength competition where one might just monotonically prefer more grip if we were like on a playground and we're going to be swinging amongst bars then the optimal grip has to satisfy being able to keep us off the hot lava but also it's not permanent and there's like a level of engagement that is not as strong as we could possibly grasp but gets the job done keeps us in the air and lets us easily and without repetitive stress like grab the next bar so how might that be reflected in some of the formalisms of active inference well for expected and variational free energy there are two terms one is reflecting a divergence minimization and one is keeping the door open with uncertainty about future outcomes so in that balance is like the divergence minimization of the boxer but then also there's something about keeping it open and then it makes me wonder about optimal grip with some of these educational and academic ideas what is the optimal grip in a given context of thermodynamics could it be the optimal grip in a easygoing chill qualitative conversation to know that thermodynamics has to do with things like temperature and now we're going to talk about something adjacent to it that was just one bar on the playground to jump amongst and so what does optimal grip look and feel like and what are the ecologies that support different individuals optimal grip on different concepts so that people can be engaging and interacting in a way that is natural and and seamless for them yeah Remy I think there's a case where the usually talking about the tactile grips there can be an optimal grip in that sense and I think in the communicative context we can also see for instance the the case where we're that we talked about the example of moving furniture together there's this this optimal grip where okay you're holding the couch and you can feel it slipping sometimes and you're like oh damn I must like get a better grip but in that case you have an interactive component and you have a communicative component because you can say when you're feeling your grip slips you can say stop to the other person to okay so to the other person can stop moving and then you can get a better grip on the on the couch so this aspect is that the idea that we tried to try to put in the in the paper when we're talking about cooperative communication is that there can be an optimal grip from a collective from an interactive perspective and communication is a way of achieving that of achieving an optimal grip by a group of individuals on their on their the context of their their joint action and so for example if I'm moving in a couch with somebody and I feel my grip slipping I can say oh stop because it will help it will stop the the couch from moving and then it will help me get a better grip in a literal and a figurative sense on the couch so that we can further advance our our joint action of moving the couch thank you blue so something else like where is an interesting place where we have grip where we don't always think about it is in the in the mental capacity so like when you they talk about overthinking so like you think something you you can underthink under prepare under and then you can also overthink so there's like also this optimal space and just as a practitioner of meditation for a long time when you're meditating on something it's the same so like it's like this over focus like you have to hold it there the object of your meditation but like if you squeeze too tight like if you focus on it too intently it goes away it pops out and lots of things are like I mean even if you literally grip something if you squeeze it too tight it will pop out of your hand just the force and and the the laws of physics make it that way so just wanted to bring that up thank you blue one no on the couch moving example it's quickly become our favorite example we should definitely do some visuals next time it's it's a interesting example because the communication can be representational and linguistic wait I don't have enough room for my right hand to come through this doorway it's very transmission view compatible but also there's a direct continuous time communication through the mechanics of the couch that are synchronously being experienced in real time and mediated through artifacts and through the niche itself linguistic communication has a linear sequential and turn-taking aspect active listening is a practice that from the point of view of the active listener whether they speak or not brings that real time and continuous engagement to the discourse even if they're listening to a recording or whether they're participating or listening in a conversation however it feels like something really important to unpack as we look towards the linguistic direction to differentiate settings where like through digital stigmergy and note-taking or through physical objects like a rowing team or the moving the couch there's a mediation through the environment continuously as opposed to something like a correspondence chess game or a phone call where there's a linear turn-taking sedimentation even though there may still be a continuous unfolding of experience for the agents so that is a very rich area to look at Dean yeah you just you're jumping in and saying things that I want to say so we're obviously the some some sort of simpaticos going on communication-wise what I was going to say is I was going to kind of now I'm going to build off what you said and what Blue said one of the things that we used to always bring to people's attention in terms of their field of affordances is that when you're having to explain something you have to take your time you have to slow down I was working with younger people you have to slow down but the also the other part of it is you have to know when to hit the hand up reaction button in the zoom call you have to time your take for full effect so you have to know both of those uh relatives when we're talking about communication because if you are you can show your passion by constantly interrupting another person and actually do yourself a disfavor so the first thing was always slow down the second thing was always think about when you're putting up your hand think about when it's your turn allow sometimes for there to be a pregnant pause I know Daniel's very gifted at this there's lots and lots of times as the moderator he could just fill in the space but he's really really good at sometimes letting the communication happen when nothing is said so I'm going back to what he was asking earlier is can we signal something by not saying something and I didn't jump in then but I was kind of hoping to be able to explain that before we have to sign off today because that's a perfect example of actually doing the signaling without manipulating the symbols per se interesting stuff does anyone have any final thoughts in this dot one yes please Remy just quickly the the the idea that there's like a skill in when to intervene speak too much speak not enough it's like all this dimension what is called pragmatics in linguistics and there's a really clear example of the notion of an optimal grip again in terms of distance there's an appropriate distance to which you you must stand to your interlocutor and that varies in in in function of culture so if you're really so since some culture and some for some person they can be like really close to you and you're you're uncomfortable for but for them it might be there their optimal grip in this communicative interaction further for others maybe you're you're you must be farther but then if you're too far then you might you're bad though less less have a less good understanding of what this person is saying so there there's this optimal grip also in in the pragmatics of the interaction that yes I think that was a good point optimal grip in space and time and meaning many spaces for a conversation are you intimately familiar and perturbing and six inches away or are you yelling purely strategic impersonal information from across the field a lot of different ecologies a lot of different agents so a lot of different communications um blue or dean if you have any other thoughts otherwise that brings us to a very nice end of 48 one and I'm sure it'll be a great discussion to continue next week just thanks for providing us with a such beefy juicy stuff to talk about I really enjoyed the paper and having you here it's great we do completely concur with me it was a fantastic paper and as I said there's nothing that I find more intriguing when somebody writes something down not as providing all the answers but as opening up all all these questions that we've been able to share today so I'm very I'm very happy that you were able to join us all right thank you all see you next week