Simulating the World & Remodeling Philosophy (Session 9)

Hello and welcome to the ninth session of simulating the world and remodeling philosophy with Rezanegar Sani. I'm going to pass the mic to him now. Thank you very much. Thank you, everyone. So I'm going to just straight directly start the fictionalist accounts a little bit, go to the, you know, some of the stuff with regards to fictionalism, then false models and toy models. So my apologies if I don't ask if you have questions or not this time because we are a little bit late and behind in terms of the topics that we should cover.

Anyway, let me start. So you remember that we talked about, with regard to, you know, a little bit of, we talked about the Maynard Smith's account of fictionalism. and imaginary, like for example, an imaginary population of RNA does not exist, we said that as a consequence, then it cannot be similar to any target system of interest. If they are to share properties, then they both must exist. Hence, they cannot be similar.

So what I'm trying to actually do with the fictionalism is just that fictionalism, what you might call to be a kind of anti-realist. So there are three what you might call to be types or creeds with regard to the constitution of models in science. philosophical creeds you can categorize them first in terms of the realist approach by realist

approach i mean scientific realism in which the models do represent a target system and the target system also indexes or corresponds with target properties and entities or objects. Then we have and of course the realist scientific realist approach you might call to be Weisberg, is somehow a representation of a kind of what you might call to be a soft scientific realism. And as you have noticed, those of you who have been reading Michael Weisberg's book,

the concept of similarity plays a central role in the scientific realist account of modeling. Of course, as we talked about a little bit, the concept of similarity is quite also a thorny issue. then uh there is the constructive empiricist account which is more or less in the tradition of what you might call to be logical positivism and a more you know sophisticated account of it would be van frossen account of modeling which you can check

constructivist empiricism in the sense that and of course within the constructivist empiricism theories are in fact classes of models. so the whenever whatever kind of a logical axiomatic system that you form that logical system at some basic level correspond to certain observational protocols or statements or sentences, which is, you know, we know that the germ of this idea coming from Carnap's

logical structure of the world of Bao. And then the third one is fictionalism. Fictionalism in the sense that we talked about last time, make-believe games. Of course, make-believe If games, you can also talk about them not as if that they are absolutely fictional, but they are more in line with that they begin their statements implicitly with such terms as let us suppose, let us imagine, let us assume that this is the case. And this is the case is

actually a would-be. It's a would-be the case, which we pretend as if it was really the case. Of course, in fictionalism, the concept of similarity is no longer playing a central role. So fictionalism, what you might call to be, from a certain standpoint, to be diametric with regard to some of the core concepts, core philosophical concepts, to be diametric to the scientific realist position. And those of you who want to actually read a little bit more about the concept of fictionalism, a new understanding of fictionalism in philosophy of science and modeling, also in comparison to realism and empiricism, I suggest this one.

Let me put it in the sidebar. Sorry. it's a good summary and Leo had also has written you know a number of other you know good papers and these kinds of what you might call to be the philosophical controversies or oppositions

between these three distinctive views of scientific modeling. But yes, this should give you an overall introduction. Nevertheless, so I just wanted to make this brief statement. Why are we talking about fictionalism? because fictionalism, in a sense, does not take the concept of similarity that Weisberg has been advancing as its central tenet. Okay. So, for example, Weisberg considers three other than Maynard Smith kind of account

of fictionalism, he also considers three more detailed accounts of fiction or fictionalism in modeling. One basically based on David Lewis account of model realism or possible You see on David Lewis account, P is true in fiction, capital F, if and only if the counterfactual P would have been true had F been told as known fact is true in every belief world of the author's community. P is true in fiction, capital F

if and only if the counterfactual statement P would have been true had F been told as known fact is true in every belief world of the author's community so a belief world of some community is any possible world where all the overt beliefs of the community are true. So consider a world in which the stories of Sherlock Holmes are told as known facts and where the belief of Sir Arthur Conan Doyle and his compatriots are true. This is a world in which it is true that Sherlock Holmes

lives on Baker Street Hence, it is fictionally true for us. Analogously, for example, Maynard Smith's RNA model description is fictionally true in so far as they are asserted as known fact, and the beliefs of Maynard Smith and his compatriots within the scientific community are true in at least one possible world. Now, of course, this brings a bugbear of the modal realist philosophy, such as Lewis, in the sense that we know for the fact that how David Lewis formulates, at the beginning

of counterfactuals and also possible worlds, the question of possible worlds is that all possible worlds are, in their own sense, actual. Okay? They are, in their possibility, they are actual. It's just that we don't have a causal connection, broadly understood as an epistemological correspondence between such worlds. Now, of course, this, as I mentioned, and that is exactly the, basically, the bugbear of the modal realist idea of the fiction as applied to modeling. In the sense that Lewis' modal realism

leaves us with an incredulous stare as Van Frossen has mentioned it. It is too epistemically costly to accept the existence of non-actual possible worlds of which we can have no epistemic access. Because you see that it's true that from in their own possibility they are actual. but what makes them actual for us is that we should have some sort of epistemic access to such possible worlds such that we can understand their actuality and hence trueness

so this is one I'm sorry I'm just trying to be very brief here so I can get to the nitty-gritty stuff with false models. Another account of fiction is coming from Kendall Walton in 1990. Let me spell it here.

So, Walton's account of fiction is significantly different from Lewis's account of fiction and involves properties, principles of generation, and make-believe statements or systems. In a game of make-believe, there are props about which participants agree to certain convention,

i.e. principles of generation, as how he puts it. When the props are present, they make believe certain a stair states of affairs to be the case for example if we agree that tree stumps are bears or images of green slime on a movie screen are a monster then when we see those props we may believe that there are bears or monsters present respectively in fact we may even have quasi-emotions and certain pragmatic conclusions about them such as quasi-fear or quasi-choice of, for example, you know, taking side with this or that, you know, kind of

affair. So therefore, props and principles of generation generate fictional truths. With regard to Maynard Smith's RNA model, presumably the prop is the model description. And given the conventions amongst population geneticists, we make believe that they are true and they are in the relevant game of make-believe. You see, Lenka, the definition of an impossible world in modal logic is

It's quite actually straightforward. Somehow even from an ordinary sense, essentially an impossible world is a world in which there is self-contradiction. It's an explicitly self-contradictory world. And hence, it is in fact, for a classical model realist like Lewis, we cannot derive the kind of coherency or consistency criteria that we could derive from possible worlds.

No matter if those possible worlds are truly wacky, it's just that they are not self-contradictory. But of course this raises a whole issue in, you know, kind of, paraconsistent logic and stuff with regard to nature of self-contradiction and things, but we don't want to get there. So now with regard to Walton's account of fiction, Weisberg's major objection to it is this. It's, I think, on page 54 of Simulation and Similarity.

If mathematical models are games of make-believe, they don't resemble anything in the physical world because they are scientists scientists mental states thus frig who actually roman frig f-r-i-g-g who applies waltonian fictionalism to the discipline of scientific modeling thus frig has to give us an account of how we can learn about real targets from games of make-believe this is a non-trivial matter because now we are owed an account of how something inside a modeler's head can be compared with the properties of a target

there is a problem here uh but of course i think weisberg has misdescribed it with regard to walton his worry is how can a mental representation be similar to a target system however waltonian fictionalism is not committed to that claim really. Rather the Waltonian fictionalists would claim that for example Maynard Smith is taking a certain attitude to a prop as specifically his make-believing that the differential equations are true of RNA populations.

If this is so it is fictionally true that they are true of some RNA population. In effect the the model doesn't apply outside of the game to some target, rather the target, the population of RNA enters the game. Whatever we learn about it, we'll learn about it in the make-believe game. The more serious objection is sometimes models like Maynard Smith's are explanatory and predictively accurate. However, if modeling is a form of make-believe, then the scientific success is make-believe as well. Namely, it's explanatory and predictive balance.

The predictive and explanatory success due to modeling only occurs in a game. Again, this view also invokes what Van Fossen calls incredulous stare. Now the last fictionalist view that Weisberg talks about is Arnon Levy, the so-called Durei account, D-E space R-E, and his name is A-R-N-O-N space L-E-V-Y. Levi's Durei's account, we make believe of some actual target that it has the properties

that it doesn't have. So, of some population of RNA, Maynard Smith make believes that the model description truly describes it. One problem for this view is that there must be an object which we make believe has some property. However, if there is no such object, then we cannot make believe of it that it has certain properties. Otherwise, it is not a Duray view. The question is then, are there objects which modelers posit, which are non-existent? As one example, sometimes philosophers suggest that population geneticists posit infinite populations and then claim there are no such things.

In response, the DeRay fictionalists could claim there are populations and we might make believe that they are infinite in size. thus it seems that the DeRay view is more credible than either Lewis or Walton view applied to modeling and sorry in that essay see lose fictionalism realism and empiricism and scientific models he actually makes a kind of an analogy with those maps in the Lord of the Rings or any of the novels.

So you have a very specific portion of denotations and geometrical figures, which basically index the relations between entities of the model and some referring devices, usually legends or some stuff on the map. And then that is the main body of your map. And then you see that it's usually surrounded by the sea of monsters, the beyond or something. Those are the fictions. Those are the infinite fictions that allow us about certain kinds of what you might call to be commonalities among these entities and properties within the model itself.

So, independent of these particular proposals, Weisberg provides four arguments against fictionalism about models. the argument from inter-scientist variation, the limited representational capacity of fictions, the inability of fiction view to account for modeling practice, and variation in the face of modeling practice. Suppose we consider, you know, for example, the Maynard Smith equations as a prop for mate belief game. Weisberg writes on page 58,

as proponents of the fiction's account like to emphasize mathematical descriptions are extremely sparse if the mathematical description exhausted the focal properties of the model then models would be correspondingly sparse in their portrayal of fictional scenarios fictions then cease to look at all like real world scenarios militating against the claim that models can be compared to a real target system in a straightforward way. However, as a fictionalist assumes, there's something like Louis's belief world or Walton's principle, hence there is more in the make-believe than just the prop of equations.

Thus, there is no more reason to think that there is more problematic variation here than there is amongst viewers watching movies or reading novels. Sure, there will be differences of interpretation but this is true on one what Weisberg's view too. For example, consider how much debate there is over interpreting parameters in mathematical models. Or, you know, using different kinds of mathematical the structures for describing a model equation. So there is variation in both the sciences and the arts but neither seems particularly objectionable or for

that matter agreeable. Now with regard to the second objection of Weisberg, is that the fictionalists cannot account for different representational capacities of models. He writes page 61, fictionalists regard the Loctavulterra model as an imaginary system composed of a predator population and a prey population. Setting aside for a moment how specific this has to be, are the predators sharks? The model is composed of concrete, discrete organisms that interact with but the equations used to describe the

Loctaw-Volterra model do so in terms of populations this means that no individual organism is represented in the mathematics of organisms According to Weisberg, the model of predation has to be composed of concrete populations of discrete and distinct individuals, as he says it on page 62. But their fictional model must be of individual organisms. Individuals are precisely what the Luchta-Volterra model leaves out. This, however, misconstrues fictionalism. The fictionalists make beliefs that model description is true. Thus the model is silent regarding anything not specified by one and two.

You know, the previous objection. And the common beliefs of the relevant community. They need not have any particular beliefs or make beliefs about the discrete and distinct individuals. A third objection is that fictionalism distorts modeling practice. According to Weisberg, modeling often involves indirect representation. One writes down a set of equations specifying a model, and the model is relevantly similar to the target. However, the fictionalist make-believe of some objects or objects that it has certain properties or some prop is true of it. Thus, it is a form of direct representation, according to the fictionalist view.

First, making believing create an opaque context and thus one can make believe that P without make believing that Q. So if we make believe about some predator and prey, we haven't thereby make believed about sharks, rays, squid, cod, lobster in the Adriatic Sea. Second, one might deny that modeling is a form of indirect representation. For example, one might claim that models are statements with adjustable parameters, which when properly specified, then the model can be true of some target of interest. Models are just abstract or idealized representation and presumably are a species of familiar genus.

Of course, we can study those statements independently of any application if we like, but they in so far, as they are representational, they represent targets in the world. For example, you can think about one of Picasso's three women. Talking about it can be viewed, we can talk about it and we can view it, appreciated independently of whether this depiction abstract or idealize, you know, these so-called, you know, female prostitutes, or we might be interested in that art historical

question, that where, how this form of representation has evolved. And that is more of a generic, the generic idea you know before the adjusting of the parameters of modeling we can we can so fictionalism essentially what you might call to be in a sense in the sense is allow us to talk about the model before in fact through adjusting of the parameters we can talk about the relation between the model and a phenomenon physical phenomenon or a

target system so this was a very very brief idea of you know fiction and there is a massive amount of literature on fictionalism in models. I just wanted to just give you a little bit of, you know, information with regard to, you know, some of the issues that are being raised in fictionalism. And most importantly, as I mentioned in my last comment, that fictionalism, even though kind of like a strong model realist fictionalism falls

into some sort of confusion, but fictionalism also allows us to talk about models in a much more complex way that indirect representation or the scientific realist or constructivist empiricists allow us to talk about them. And of course this view of talking about the models prior to talking about how models are models correlate or map onto a target system, a real world phenomenon, is one of our points

of entry for talking about toy models, the small and big. Because essentially toy models also have this component of fictionalism. So this is one as our point of entry to the domain of toy models. The second one is the concept of false model. You know, we know that all models are in one way or another are false. But what does it mean that we can garner the pragmatics, the advantages of this falsity when we make it explicit?

So before I go to the false models, any question but questions that don't lead me to the rabbit hope of talking more about fictionalism. I assume that there is no question. Is there a way that you can maintain

a type of fictionalist account of talking about models within an uh either the scientific realist position or the constructivist empiricist position yes i think so yes i think so and in fact if you read that uh c lue's uh essay he's actually trying to um come up with a form of hybrid fictionalism in the sense that that component of fictionalism is preserved and can in fact be talked coherently within either the branch of scientific realist approach to modeling or within constructivist empiricist domain yes

I don't think that to commit to the fictionalist thesis basically entail you committing to all of its baggages wholesale. Yes, yes. It's actually a document. Let me actually see if I can. My apologies. I need to be very careful when I drag and drop stuff to the sidebar because I might accidentally put yet another picture of car nap.

It wouldn't be bad, you know, obscene materials. It would be just another picture of car nap. One second, apparently it doesn't work. Let me, one second, one second. Okay. It's being uploaded. There we go. It's on the sidebar, has a Google Drive link. Okay.

Okay. So, we already know that, you know, OK, let me start with this. You see, false models are a species of neutral models.

Now, of course, this means that how can we such thing as neutral models and what what are really neutral models you see the term neutral model is of course is a misnomer if it is taken to suggest that a model is free of biases such as might be induced by acceptance of a given hypothesis such as that the patterns to be found among organisms are products of selection. Any model must make some assumptions and simplifications and or idealization, many of which are in fact problematic.

So the best working hypothesis would be that there are no bias-free models in science. So when we are talking about neutral models is that we are in fact talking about biased models. This observation has a parallel in the question. As Wimsatt poses this question, what variables must be controlled for in an experimental design? Well, there are no general specifications for what variables should be controlled.

Since, one, what variables should be controlled or factored out through appropriate attempts to isolate the system. Two, what variables should be measured. Three, what errors are acceptable. And four, how the experiments should be designed are all functions of the purpose of the experiment. Similarly, one, what models are acceptable, two, what data are relevant to them, and three, what counts as a sufficiently close fit between model and data is a function of the purposes for which the models and data are employed.

Any model, accordingly, implicitly or explicitly, makes idealization and simplification, as we have talked about. It ignores some variables and admits some others. It simples variables in the models and among possibly relevant variables not included in the model universe. These omitted and simplified variables and interactions are sources of bias. In cases where they are important, sometimes certain kinds of variables are systematically ignored in the process of modeling.

Thus in reductionist, for example, modeling, where one seeks to understand the behavior of a system in terms of the interactions of its parts, a variety of model building strategies and heuristics lead us to ignore features of the environment of the system being studied. And even though we move toward the interaction of the system and environment, again, certain forms of interactions between the system and environment are being studied and others are

omitted and ignored. These environmental variables may be left out of the model completely, or if included, treated as constant in a space or in time, or treated as varying in some particularly simple way, such as in a linear or random fashion. In testing the models with the focus of the interrelations among internal factors, environmental variables may be simply ignored or treated in some aggregate way to simplify their analysis.

so you see uh this is also a good lesson for all of us in the sense that we have we usually hear these stuff about you know um basically well these kinds of uh let's call them scientific philosophers or scientific rationalist philosophers. They always simplify stuff. They always miss some important factors. Well, that is not the job of the goddamn philosophy or science to account for any detail. In fact, if we were going to account for any detail, we couldn't make any kind of claim about how things work.

Right? The idea of that kind of all-over-the-place complexity where you have to include every variable, every form of interaction, is in fact toxic to the progress of cognitive science and science in general and the science of cognition. so these are these are i mean i just wanted to make this claim precisely because now we are seeing that you know how the the basically the enterprise of science works is and slowly

uncover the complexity of the world in which we exist is not why these wishy-washy speculative demands on including every detail but by simply factoring out some of them It's what you might call to be unpacking step by step. So, any question before? And by the way, we can have a break, and then we can come back, and then I will start.

yeah well the problem of what to omit is is a big problem not the problem of what to include yes and this is exactly you see it's exactly also uh you kind of an allegory of it is also the frame problem in robotics the problem of what to ignore is important for a fluid navigation of a robot in a very complex environment, not to include every detail about stuff, about possible pathways. To ignore certain features is really the task of scientific navigation.

Sorry Reza, but just to make sure I've understood where you're coming from and you're saying because it like you have a model as complex as the actual universe or a model that is completely intractable and so therefore you suppose for prediction or description, that's where you go. Yes one I think is the question of intractability and the other one is a bigger question of the metaphysical idea of complexity. The world is neither complex nor simple. To in fact posit such dichotomy for the philosophical investigation or the scientific investigation

means that we have a world works. The science of, for example, needs to take into account as its a priori law. But of course that is not how science works because science doesn't work by way of such massive metaphysical claims. But also, yes, the problem of interactability as a pragmatic problem. The idea that navigation only works by way of encapsulation. Encapsulation that is already, we do it with regard to our own concepts. To unpack a judgment, an assertion,

conceptual assertion, we cannot say what it means in every possible context. We can unfold such meaning, such conceptual content, step by step, one context at a time, one task at a time. Well, you have to, I know that you have some stuff in your mind, so let's have a break and then you come back with your challenge. Sounds good. How about a five minute break?

Excellent. Okay. All right. Excellent. Thank you.

Thank you. Thank you.

Thank you. Thank you.

Environmental mismatch. I mean, there are actually some work very, I mean, of course, this has come back to, you know, from, I mean, Adam can talk about it much better than me. But, I mean, you can also compare it with the idea of the compression in information theory, algorithmic information theory.

in the sense that to compress the data that is a task I wouldn't say that it is not the question of exclusion, it is not that it doesn't want to talk about everything, it is not, as I mentioned, it does not bother itself with

those metaphysical question of everything versus nothing or complexity versus simplicity. It's just that to encapsulate information, to compress information, to exclude some variables and interactions is a pragmatic way of how to do the actual scientific to employ and deploy the scientific method. Epistemological tractability. but then yeah just it seems like it's kind of a balancing act because determining what to include inside of a theory or model is the what that's the criterion for what counts is

objectively valid too would you be able to repeat that um the if i i need a little bit of time to think about it but the main gist of what i'm saying is like you're modeling something and you have to choose what to focus on and or what to ignore but the choosing what to ignore is part of what sets the criterion of objectivity within science Yes, yes, yes. However, this does not commit us to any claim whether science actually has a concept of everything or has a concept of nothing. The concept of complexity of the universe or the concept of a fundamental simplicity of the universe. No.

it is just objectivity requires to unfold through epistemological method and this epistemological method should be tractable in order to make the objective claim one of the central points of the scientific enterprise and it seems like if I can respond one more time it seems like this is an ongoing theme that I'm running into at least is that there's this pragmatic concern which is in some way unburdened by the problems of objectivity as such or something like that but well you see here I think that you are

losing pragmatic concern in a very loose pragmatic sense pragmatic concerns of the epistemological scientific method not any sort of pragmatic concern. Maybe Adam can talk a little bit about this with regard to computer science and information theory. Adam, don't be shy. It reminded me of this paper, which I'm almost certain I got from Reza in the first place by Abramsky. It's a beautiful short paper called Two Puzzles of Computation. And

in that he says, why compute at all, right? So by definition, you've got all the information at the start as an input, and then you run some computer program on it, and you get some other piece of information. But you have all the information at the start, what's the point of computing. And you have to have some idea of compression or indexing there, which I think he's drawing attention to. He doesn't answer the puzzle directly within the paper. Which is, I mean, that's why it reminded me of Borges, you know, scientific method, I I forget the title, even though I looked it up five seconds ago.

On exactitude in science, right? They build a map the size of the entire empire. It's useless as a map, right? Because there's no information compression going on, right? You can't use it as a map. You have to walk all over the empire. So, I mean, the only thing that I really was questioning in Reza's account, I actually agree with like all the, but just this calling it pragmatics was the only thing that kind of threw me a little bit right at the end because it's always like, if the whole point is compression for navigability, then it's no longer a kind of a pragmatic concern.

It's the whole point of the enterprise. Yes, yes. No, absolutely. Yes. Okay. Yeah. I concede on this point, yes. And this coming back to try to ameliorate my strong resort to the pragmatic concern with regard to Theo's question. Yes, yes. Cool. Any input from any of you? Yann, no, Jean-Pierre, Justin, Lenka, Marie, Mikey, Cepide, Barrett, well Andrea, no one.

Okay. So, talking about, you know, when I talked a little bit about the neutral model, the The concept of neutral model also comes from evolutionary biology and ecology. Where a neutral model usually means a model without selection, right?

For example, coming back to work of Stephen Jay Gould and Simberloff on random phylogenies. Phylogenetic descent trees in which organizations and extinctions are determined by random variables. Now, these artificial phylogenies, in many respects, resemble those found in nature. They thus reasoned that the similarities between the artificial and natural phylogenies were not products of selection processes operating at that level. Their model did not, as they pointed out, rule out selectionist explanations for especiations

and extinctions at a lower, for example, intra- or interpopulational level. Similarly, for example, the work of Crow and others on neutral mutation theories modeled and evaluated patterns of molecular variability and change on the assumption that selection forces on these variants could be ignored. Similarities between their predicted patterns and what was found in nature led to various versions of the hypothesis that the evolution of various systems or of various kinds of traits was driven not by selection but by various forms of genetic drift. Or for example, Kaufman's work on genetic constraints identified features which because

they were near universal properties of this randomly constructed model genetic control networks are taken to provide baselines for the properties of systems on which selection acts. He argues that these genetic properties will usually survive in spite of selection rather than because of it. Here, Kaufman is using a comparison of models with and without selection to argue that selection is not important. One, of course, must not assume that if the data fit the neutral model then the excluded variables are unimportant.

The research of people like Kaufman shows that selection may be unimportant in some way in producing the phenomenon being modeled, but they do not rule it out entirely. Thus, random phylogenies do not exclude selection as a causal agent in producing individual extinctions through, for example, over-specialization to a temporally unstable niche or speciations through directional selection in different directions producing isolating mechanisms between two different geographically isolated subpopulations of the same species,

because his model simply does not address those questions. Now, again, here I just wanted to very quickly go over this idea of where the concept of a neutral model comes, which I generalized it at the beginning. You can in fact think about the concept of a neutral model as a null hypothesis. As a null hypothesis, it usually establishes that omitted variables do not act in a way

as specific to models under comparison, not that they do not act at all. This is consonant with the earlier claim that the adequacy of models is highly context dependent and that their adequacy for some purposes does not guarantee their adequacy in general. However the use of these models as templates can focus attention specifically on where the models deviate from reality. But only so if the biases of excluding or including variables and interactions between

entities, processes, events, so on and so forth, are made explicit. In that sense, these models can in fact be understood in terms of templates, which allow us to focus attention on where the models deviate from reality, leading to estimations of the magnitudes of variables left out, or to the hypothesis of more detailed mechanisms of how and under what conditions these variables act and are important.

This is a pattern of inference that is both common and important and deserves closer scrutiny. The variety of ways in which this is done is going to be the topic of the rest of our talk today. And then coming back to that idea, you see, you know, kind of the philosophical allegory of this is that, in fact, when we make explicit, like, for example, philosophy of mind, right? For example, I, for example, talk mainly about language, formalisms, the concept of a structure, so on and so forth, with regard to the concept of mind.

I leave out all of the other stuff about unconscious emotions, so on and so forth. You see, we have a tendency in philosophy to think of such moves as pathological, that you are pigeonholing the question of mind into some strict, you know, basically niche stuff. But no, this is exactly what scientists do and what we philosophers should do with the awareness that only in doing so, in constraining the problem to certain aspects and not others, can we actually notice the magnitude of those variables which we have left out.

In starting from the beginning to try to talk about all of these variables, we actually don't understand where our model actually fails, where our philosophies fail, where some other factors that we have ignored might have played a massive role that can actually lead to the revision of our model. and our variables. So this is also a philosophical lesson, a moral philosophical lesson out of this. Don't feel bad if you talk about mind in terms of axiomatics, you know, because that actually you do it with this awareness of

this very fact. You are doing a good job. You are actually opening a pathway toward understanding of possible different variables, the magnitude of them, so on so forth. so we what I just said now we can ask how models can misrepresent as we know already talked about it repeatedly

about the simple observation that most models are oversimplified over approximate fundamentally incomplete and in other ways completely false gives little reason for using shouldn't be taken as a reason for not using them. In fact, their widespread use suggests that there must be other reasons that explicitly false models, misrepresenting models, should in fact be used. It is not enough to say that we cannot deal with the complexities of the real world so

simple models are all that we can work with. For unless they could help us to do something in the task of investigating natural phenomena, there would be no reason for choosing model building over astrology or mystic revelations as a source of knowledge of the natural world. Or does the instrumentalist suggestion that we use them because they are effective tools rather than realistic descriptions of nature gives us much help. For it presupposes that we want to understand, namely, how false models can be effective tools in making predictions and generating robust descriptions and explanations.

I want to suggest now various ways in which false models can, one, lead to detection and estimation of other relevant variables, as we talked about. Two, help to answer questions about more realistic models. And three, lead us to consider other models as ways of asking new questions about the models we already have. And four, in evolutionary or other historical contexts, determine the efficacy of forces which may not be present in the system under investigation, but which may have had a role in producing the form that it has. Of course, this is a classical view of modeling that has been proposed by people like William

Bechtel, William Wimsatt, Carl Craver, and many of these mechanistic philosophers of science recently. So I just want to go over a range of ways in which a model can be false. first one a model may be of only very local applicability

this is a way of being false only if it is more broadly applied out of context okay two a model may be an idealization whose conditions of applicability are never found in nature For example, point masses, the uses of continuous variables for population sizes, etc. But which has a range of cases to which it may be more or less accurately applied as an approximation. And it would be false again. it would be taken anything other than a kind of approximation.

Three, a model may be incomplete, namely leaving out one or more causally relevant variable. Here it is assumed that the included variables are causally relevant and are so in at least roughly the manner described. 4. The incompleteness of the model may lead to a misdescription of the interactions of the variables which are included, producing apparent interactions where there are none, as perious correlations, or apparent independence where there are interactions, as in the spurious context independence, produced by biases in reductionist research strategies.

in these cases it is assumed that the variables identified in the models are at least approximately correctly described five a model may give a totally wrong-headed picture of nature or physical phenomena not only are the interactions are wrong but also a significant number of entities and or their properties do not exist. Actually, a lot of these kinds, this one particularly, a good example of this, are in some quantum physics

or string theory toy models. Okay. Where we are just simply talking about fictional entities. Fictional interactions. None of them are actually, as far as we know, exist. Sixth, a closely related case is that in which a model is purely phenomenological. That is, it is derived solely to give description and or predictions of phenomena without making any claims as to whether the variables in the model exist.

Example of this include the viral equation of a state, the automata theory like in Turing machines as description of neural processing and linear models as curve fitting predictors for extrapolating trends. Seven, a model may simply fail to describe or predict the data correctly. This involves just the basic recognition that it is false and is consistent with any of the preceding states of affairs. But sometimes this may be all that is known currently. Now, any question before, now that I said that, you know, that there are all these ways that a model can be wrong.

Now, what we want to know is that what we can do with this goddamn false models in a scientific way. Before going to that end, I want to hear your ideas. I would like to, it's just a question of clarification, what would be the difference between 5 and seven well you see five or seven so the five one

just to repeat for people who missed the whole list uh how it evolved five a model may give a totally wrong headed picture of nature not only are the interactions wrong but also a significant number of the entities and or their properties do not exist okay seven a model may simply fail to describe or predict the data correctly this involved just the basic recognition that is false and is consistent with any of the preceding the state of affairs. Now, you see, yes, they can be related, but five is essentially is that on these explicit assumptions that all of this stuff that we are talking about

are just make-beliefs, are fictional entities. We are not even trying to pretend that they actually do describe or predict an aspect of the phenomenon whereas seven either we might have some of these fictional entities or we might not have it's just that what regardless of whether we are using with these purely fictional entities wrong stuff or write the stuff we just cannot describe or predict it accurately or with a semblance of fidelity

Yes, yes, absolutely, yes. Essentially all these models, the ways that models can be false, they are false with regard to the basic assumptions of all of the three branches of modeling, philosophical branches of modeling that I mentioned, constructivist empiricism, scientific realism, and fictionalism. Any? Marie? Andrea? Justin?

Can we very quickly repeat the thought? Because it's not coming to your place with me. I missed that one. Just a brief summary, everything else is very clear, but in the fourth I've lost. Sure. The fourth one. The fourth one, the incompleteness of the model may lead to a misdescription of the interactions of the variables which are included, producing apparent interactions where there are none or apparent independence where there are interactions, as in the spurious context independence produced by biases and reductions research strategies. So, for example, you know, coming back,

that is actually a good example of what we talked about earlier on. For example, with regard to a Turing machine. So a Turing machine that's in the first puzzle of Abramsky, Epistemological Closure. It's the idea that from an interactions point of view, now you basically, the Turing machine goes through, it receives a discrete set of input, goes through a state transition, and then it yields a discrete set of output. Now, during the state transition, it does not accept any more input in the classical church, Turing, thesis, any more stuff from the environment.

Now, of course, you can do, include certain specific interactions, as Turing himself did, certain kinds of interactions. And of course, these interactions that you include versus other interactions may fundamentally misrepresent other forms of interactions that are going on to the point that you think that the church-tearing machine, an intractable church-tearing machine, should be able to compute any form of generalized interaction. interaction which is of course it's a big thing or in the in the philosophy of mind and cognitive

science you know uh when we are talking about um uh you know extended mind so we uh think of brain architecture and its estates its internal estates as interaction with some portions of reality But depending on what portions of reality we choose such states or such architecture to be in interaction with, we might end up misrepresenting other portions of reality that the brain or the internal states are truly in interaction with. Reductionism, the same thing. and actually another thing that is really interesting here the way that the false models

are false you can see these that many incompatible philosophical positions in the philosophy of science or in cognitive science like you know for example internalism versus externalism I don't know, all sorts of things, computationalism, mechanistic view versus the inactivist view, so on and so forth. You see that they, in fact, do share a lot of common ways of falsity in this specific sense of modeling. such that, for example, reductionist philosophy of mind

shares, in fact, a lot in terms of this fourth way of a model being false with the inactivist paradigm. Certain interactions are taken to be fundamental such that they misrepresent the entire paradigm of interaction, the entire framework of interaction. Okay, shall I proceed? Okay.

Now, we go over a list of things we can do with false models. from a commonsensical view it might seem paradoxical to claim that the falseness of a model may be essential to its role in producing better models isn't it always better to have a true model than a false one naturally it is but there's never a choice that we are given and it is a choice that only

philosophers could delight in imagining and also scientists for that matter will any false model provide a road to the truth here the answer is just as obviously an emphatic no some models are so wrong or their flaws so difficult to analyze that we are better off looking elsewhere For example, case five and seven, five being a model may give a totally wrongheaded picture of nature or a model completely failed to describe or predict the data correctly. They, for example, if they take into extreme levels, they might have little useful value in providing us with better models.

the primary virtue of a model must have if we are to learn from its failures is that it and the experimental and heuristic tools we have available for analyzing it are structured in such a way that we can localize make explicit its errors and attribute them to some parts, aspects, assumptions, or sub-components of the model. So we are not simply glorifying the falsity of models. This glorification is not coming for free.

we are only interested in this falsity in the biases of the model to the extent that we have either existing tools of localizing the exact errors with regard to the certain aspects of model and making explicit or using different kind of models, you might actually develop new tools that can shed light on the biases. Regardless, the only reason that we are interested in these falsities is precisely because

with regard also in relation to making better models is that our point of emphasis is about how exactly can we shed light on the specific errors now of course as Wimsad mentions uh there is a mythology among philosophers of science for

example the so-called Quine-Douin thesis that this cannot be done that a theory or model meets It's experimental test wholesale and must be taken or rejected as a whole. Kind of like when we say that, you know, what is that saying, sorry, you know, that the best argument, that an argument is as strong as its weakest chain, right? But of course, we can also think about the chain of reasoning.

In a different way, like Charles Sanders' purse, idea that the chain of reasoning is not this kind of monolithic chain, of deductive chain, but it is like a cable made of many, many filaments and smallest wires all packed into one. So just because one fails doesn't mean that the other ones are also failing, right? That is another metaphor. And essentially this metaphor is against the kind of Quine Duem thesis. Not only science but also technology and evolution would be impossible if this were true.

in this and in logically similar cases. Essentially, if everything was going to fail just because of the whole system, if one part of the system fails, one part of the model fails, okay, the whole system should fail or should be left out, there wouldn't be any kind of a species on this planet from an evolutionary perspective. It is about tinkering, you see? Evolution is a tinkering process. It is not a process in which the failure on one part leads for certainty to the complete failure of the entire system.

If something gone wrong with the evolution of your nose, and it had become a trunk, it might actually, well I don't know, I'm not an evolutionary biologist, it might have, yes, created massive consequences that you as a species die, but I'm just making a metaphor. It shouldn't happen like that. Precisely because from an evolutionary perspective and also with regard to the logic, as I mentioned with Peirce versus deductivist chain, but also modeling, the idea of the testing of the model against failure is not a wholesale matter, is not a matter of all or nothing.

It's more of some process of tinkering. only parts of course this has this relies on the idea that if a part fails to what extent that part actually affects other parts to what magnitude you know of course if the magnitude of the influence is high then of course it will lead to a chain reaction throughout the whole system but if the magnitude, the locality of that part is quite limited, it's not going to affect it. And then, so the whole Quinduam thesis on, you know, on this subject matter becomes almost

too rigid from a from either what you might call to be evolutionary naturalist perspective from a logical perspective or from an epistemological modeling perspective question question before i start yes yes yes maybe I missed it but I want to meet by evolutionary perspective in regards to the trunk when we clearly very much influence that with our pre computational thinking by creating beings like fuck for which there is no evolution of reason to exist or do you think this is still

within the evolutionary that you're describing. Yes. No, when I'm talking about the evolutionary, you see the evolutionary, when we are talking about the evolution, we are talking about the Darwinian, modern Darwinian theory of evolution, the theory of biology. In the sense that, of course, you can give a computational interpretation of it or another kind of interpretation of it. But it's essentially the way that the morphogenesis of organisms emerge right we just influenced it so much with our essentially inherently false models and creations that i'm not sure that i can locate that evolution it is not really you see i don't think that darwinian theory is model based the refinement of darwinian theory is model based

Darwin's basically worldview is a theoretical view, a theoretical view that actually undergirds that we, certain models we use and we don't. Of course, to challenge the Darwinian theory, you should be capable of giving us an alternative that is scientifically sound. are we of course there are all these post-Darwinian wishy-washy stuff that we see in the you know kind of the new age kind of philosophies but that absolutely should not

by any means be taken as a kind of what you might call to be a stain upon the theory of Darwin. I actually really, really suggest a very great book on a kind of a very kind of not essentially revolutionary but quite philosophically, what you might say, clear-headed view of Darwin's contribution to biology. Let me just, one second. I'm just trying to get the title for you.

Has anyone read that book? it's a crud cutter rather than a washing detergent does any of you know what the crud cutter is particularly the brand is so good

Okay, let me, I try to just go over some of these lists so we can, we can, next session, finally start talking about the toy models. I mean, let me just read the whole list of what we can do with the false models and then We can talk. So this is essentially a list of functions served by false models in the search for better ones. One, an oversimplified model may act

as a starting point in a series of models of increasing complexity in realism. Just like that kind of allegory of philosophical method that I mentioned to you. You don't want to write a book that is all over the place with regard to the question of mind. You just want to focus on one aspect. And hopefully that, if taken resolutely, understanding its own limitation and restrictions, even provisionally, can lead to an expanded work. And that is usually how philosophers work in the span of their lives. They start with a very modest starting point, and then they branch out.

Two, a known incorrect but otherwise suggestive model may undercut the too-ready acceptance of a preferred hypothesis by suggesting new alternative lines for the explanation of the phenomenon. 3. An incorrect model may suggest new predictive tests or new refinements of an established model or highlight specific features of it as particularly important.

Four, an incomplete model may be used as a template which captures larger or otherwise more obvious effects that can then be factored out to detect phenomena that would otherwise be masked or be too small to be seen. Again, coming back, you know, usually is, when we are talking, for example, you know, coming back to the idea of, you know, when a rationalist talks about mind, sees mind in a Kantian framework. the question of mind the modeling of the mind as a template now of course usually the objections to

such stuff coming from that well you completely forgot to talk about unconscious or emotions or all of this stuff you know real-time interactions and culture so on so forth usually that is not exactly you know sure they are also important but what is missing in fact to these kinds of big views of other variables are in fact is in fact the detection of extremely important and otherwise hidden precisely because they look either too small or too unimportant factors.

You see, that is another way that a model, an incorrect or false model can shed light, lights, not just on big variables that we have excluded, but also on hidden states, on hidden properties, hidden factors, which are otherwise negligible to us, or not even are on radar at this point. Okay? So, five, a model that is incomplete may be used as a template for

estimating the magnitude of parameters that are not included in the model. Again, we talked about this. Six, an oversimplified model may provide a simpler arena for answering questions about properties of more complex models, which also appear in the simpler case, and answers derived here can sometimes be extended to cover the more complex models. Seven, an incorrect simpler model can be used as a reference standard to evaluate causal claims about the effects of variables left out of it, but included in more complete models or in different competing models to determine how

these models fare if these variables are left out among them. Eight, two false models may be used to define the extremes of a continuum or a spectrum of cases in which the real case is presumed to lie but for which the more realistic intermediate models are too complex to analyze or the information available is too incomplete to guide their construction or to determine a choice between them. In defining these extremes, the limiting or the bracketing models simply specify a property

of which the real case is supposed to have an intermediate value. So essentially, we do use sometimes two false models in conjunction to in fact bracket problems, intermediary problems which are tractable 9 a false model may suggest the form of a phenomenological relationship between the variables a specific mathematical functional relationship that gives a best fit to the data but is not derived from an underlying

mechanical model. This phenomenological law gives a way of describing the data and through interpolation or extrapolation, making new predictions, but also because its form is conditioned by an underlying model, may suggest a related mechanical model capable of actually explaining it. Ten, a family of models of the same phenomenon, each of which makes various false assumptions has several distinct uses. A, one may look for results which are true in all of the models and therefore presumably independent of different specific assumptions which vary across

models. These invariant results, so-called robust theorems, are thus more likely trustworthy or true. B, one may similarly determine assumptions that are irrelevant to a given conclusion. C, when a result is true in some models and false in others, one may determine which assumptions or conditions a given result depends upon. Kind of like a detective working here. 11. A model that is incorrect by being incomplete may serve as a limiting case to test the adequacy of new, more complex models.

If the model is correct under special conditions, even if these are seldom or never found in nature, it may nonetheless be an adequacy condition or desidratum of newer models that they reduce to it when appropriate limits are taken. A good example of it is Newton's concept of gravitation versus Einstein. Newton being the special case model with all the limits appropriately applied to it. Twelve, where optimization or adaptive design arguments are involved, an evaluation of systems or behaviors which are not found in nature but which are conceivable.

Alternatives to existing systems can provide explanations for the features of those systems that are found. And thoughts? Maybe Andrea or you can talk about this a little bit more.

Of course, you see conditions of applicability usually in a kind of epistemological sense. They are not purely given. You know, they are not categorical given. So let's put it that way, right? They are applicational givens, right? In the sense that they involve massive amounts of inferences, structural coordination with observations, so on and so forth. But yes, they can be, from an applicational perspective, they can be given. And this is essentially, this is the task of false models to in fact even renegotiate the givenness of the conditions of application. The condition of application being given doesn't mean that it is a categorical given. Surely it is a given, and hence it needs to be renegotiated.

But it is not false or completely dogmatic in a sense that, for example, the myth of the categorical given is. ... No, it is... Well, you see, the question of application is essentially the question of localization,

of the range of application, right? And of course, the localization of the range of application always starts with a specific theoretical assumption. that can be justified within the theoretical structure, construction, or the paradigm of theory, of a specific theory. But of course, either it might be too restrictive or it may be too broad. It's just the adjustment and the determination of the exact local range of application can only be done in the discipline of modeling by applying other models which can renegotiate

this range either by bringing new variables by renegotiating some of the theoretical structure uh you know many many other factors Okay, someone needs to talk about because I don't have anything to say at this point. I think there's a lot happening in this sidebar. i kind of want to put jp to to we're we're uh talking about fictionalism how uh

how is it possible to have falsity under the fictionalist premise that that's still fuzzy for me and maybe um jp has more to add on that I would say it requires a fictionalism understood as a kind of, in a Brandoian sense, a kind of semantic self-consciousness of itself. and of course as how we talked about it to commit to the idea that these false models in one way or another have a fictional status

or fictional what you might call to be fictional framework does not commit us to the entire thesis of fictionalism as is classically understood. But the entire thesis of fictionalism, is that incompatible with the idea of some broad sweeping idea of falsity under that thesis. Yes, I would say that, I mean, depends on what account of fictionalism you take into account, account of fiction you take into account.

Some of them are completely, in fact, commensurable with that, in the sense that all of these falsities, in the sense that all of the, the distinction between the falsities sometimes cannot even be distinguished. I would say that, you know, that is why we talked about this, that, for example, now, of course, this is becoming now complicated here. For example, a modal realist account of fiction, such as David Lewis. Of course, it has different accounts of falsities as well. Right? because they are happening in possible worlds.

And of course, in each of these possible worlds, you have a different account of falsity which might have, you know, garner fundamental insights. But it's the idea that, for example, that account of model realist requires a specific connection, epistemological access from one of these possible worlds with its corresponding falsity to another possible world with its corresponding falsity. I don't think that fictionalism... Okay, maybe I need to think about this. I really need to think about it.

I think that even fictionalism as a whole cannot be understood as a challenge to basically advantages of false models in the way that we have talked about. in the sense that it is just a stance about the role of models with regard to other models or a physical target phenomenon. We do not need, in fact, I would say, to endorse specifically either scientific realist, the fictionalist, or constructivist empiricist.

We can just get the minimal insights of, like, what we were talking about earlier, of each of these branches, and actually talk about them with regard to the criteria of falsities, of what counts as a false model and how these falsities of the models can be advantageous, beneficial for us to make better models without taking side with any of these kinds of philosophical doctrines. I mean, I feel like is falsity even a great word to use here on litograms?

Correctness, incorrectness. And incorrectness, you see the bias. Like there's a mismatch, right, between the description provided by the fiction and the observation within either, within the world or the description in another fiction, right? and there's a mismatch in a particular parameterization or a particular situation. Yes, but of course in a specific fictional world, you can only do certain kinds of things, make certain kinds of relationships and not others.

So even in the fictional world, you have criteria of incorrectness. Sure, sure. I mean, it's not about the internal consistency of the fictional world so much as when you then compare it to some other fictional world or your observation from the world itself. Yes, yes. No, I agree. I agree but but also we can we can say that let's assume that you know we have we create a fictional property and of course this fictional property is defined within a fictional system

and let's assume that this fictional property can in fact be applied to a picture of reality okay Now, of course, we do not yet know what this fictional entity can do in a real domain, in nature, in our theories of nature. All we can at this point talk about first is that whether this fiction, this fictional entity, is actually when we are using it, Is it consistent with its internal domain? To what extent? And again, coming back to your answer, that using this fictional entity deprives us of what other kinds of fictional entities that we could use, according to their own respective internal logic.

All right. Theo, does that get out? But I agree that the question of fictionalism as a whole and the question of incorrect model is yet completely vague at this point to us. It's just that, as I mentioned, we got only, what was it, Theo, what was it? little bit from the Walton idea of fiction. Oh, I'm not sure what you're talking about. The one that we were talking about earlier on, I completely forgot our conversation. I'm scrolling up in there.

Wait, you're not talking about the Lou article. No, no, no, no. Okay, I need to... Sorry, my memory is getting worse. Yeah, there was only one part of the fictionalism that we identified as our point of entry to the toy models. remember anyway let's yeah you were talking about make believe the make believe aspect of fictionalism yes a make believe that allows us to talk about models yes that's the idea

yes yes yes we can even without any regard for what you might call to be the The application to a physical. Yeah, the representational scope of the model itself, maybe just the make-believe, as detached from the embedment into a specific world, taken to be the one. Yes, yes, yes, yes. And I consider this as that also, if taken with a modesty, can count as an enablement, precisely because it allows us to look into the structure of the model, how the components of models with regard to any range of application

or representation or referring capacity work. Yeah, if this is what we're talking about, this is the... Yes, yes, yes, yes, that's the one, yeah. I guess it's just kind of a mystery to me how you preserve those. Well, maybe it's less a mystery to how you preserve those like pragmatic aspects of using fictionalism to talk about models without slipping into what I think is like possibly a looming relativistic, unintelligible version of fiction. Why do you think that making belief games in a logical sense leads to relativism?

Essentially, when we are talking about fictionalism, we are not making any kind of fictions. As I mentioned, we are not making claims about astrology. We are making about modalities, logical modalities and counterfactuals. We are essentially in the domain of the structural logic. I think sorry JP you go first I think for me there's a bit of confusion between I think in this whole discussion between one possible use of fictionalism in order to understand model construction and an ontological purview of fictionalism yes I think that we need to

distinguish these properly yes i am not for ontological and i do and and the whole thing is that well we can actually talk about this so whether fictionalism does in fact have an ontological perspective or not yeah that's a big question but yeah the fictionalism as i understood initially was somewhat... It was an ontological position. I mean, I'm familiar with the term from the philosophy of mathematics, and the thesis is that mathematical entities does not exist, but you can refer to them directly,

like they are denotable. But in order for they are being denotable by words, they must exist in some sense. So what they are saying is that mathematical entities exist in the story of mathematics, not as real entities. So this was the... Yes, because to say that they are real entities, that would be a Platonic realist. You lapse back into Platonism, that's it. Or you can be a nominalist, and the nominalist does not countenance the direct reference. You have to paraphrase direct reference to numbers. Yes. You understand what I'm talking about? Yes, yes, yes.

The idea there, fictionalism for me, came attached to this ontological position that some domain of entities, some specific domain of entities are fictional. So you can be a fictionalist about mathematics or a fictionist about, I don't know, scientific entities like the sellers unobservables or something like that. So for me, it's like an ontological position. So I'm having a hard time understanding. So maybe I should reformulate the question. then if we actually do endorse the pragmatism of fiction, fictional pragmatism or pragmatic fictionalism, how does it actually can be brought back, folded back into these kinds

of incorrect model view. One, and second, how much, to what extent, can the pragmatic fictionalism can be detached from the ontological commitments of fictionalism par excellence? Those are good questions, yes. Theo is getting ready to write a 300,000 paper. So anyway, I have some probably, most probably unconvincing answers to a few of these problems,

but I cannot talk about them precisely until and unless I start the idea of a toy model. And the way that we are going to look at toy models in the next session is by way of two things. One, what we might call to be actually three, well two actually, because two of them look almost identical. One is a toy model as giving us a new understanding, a mode of understanding, as opposed to the model as such. is essentially

it signals a shift from an actual explanation that a model strives to achieve to a possibly so explanation. That's one. Two, we are going to talk about toy models with regard to the metatheoretic assumptions of the models themselves. Namely, a metadiscourse on models.

One. With the understanding that again, and that's why this is more, again, another point of confusion. It is kind of similar to the idea of a false model. In the sense that when we apply false models, exactly we want to make explicit what are the errors of the model itself. Right? to make explicit in a Brandomian sense, it biases. It's theoretical and modeling biases. And exactly to do that requires making toy models

because toy models are explicit metatheoretical accounts. All models are metatheorists. All models are also biased. Whereas it's just that how toy models are is that they are explicitly incorrect and explicitly have such theoretical biases. And of course, this leads to the idea that then how the shape changes from a model to a toy model. When we actually pronounce our theoretical assumptions to be always metatheoretical assumptions.

When we underline the implicit biases of our model, convert them into explicit biases. smell of burning here i think something is on fire It's under your nose. No, I think that's, well, yeah, well, I hope that it's not a cigarette.

Okay, so who's going to present next session? I'm going to drag you at this time. Do you have... Half of you already have given presentations, some of you haven't. Sorry, I have to turn this on. I know that it's not good for recording, but I cannot see anything. That's okay. Do you have specific material that you'd like people to present on? Yes, exactly on the issues of fictionalism and incorrect models. And definitely read that hybrid view by Siliu's paper that I shared on the sidebar.

Actually, I have to have the permission to download it, Josh. You didn't give me permission. Oh, I didn't give you the permission? Okay. okay i will send you the link um i don't know how to give the permission yeah yeah i don't give permission it's just that i don't know how to give one permission and no one said anything we just like cool we'll find it somewhere else oh is is this for the liu paper yes oh yeah that's on like phil papers i think yes yes 16 pages yes or he's on his academia page yeah he's in academia page okay that's that's easy enough okay okay who's going to do it uh let me see justin john pierre marie

i suppose i could do it okay one more lenka thanks him okay we have two excellent super fantastic All right. Should I end the broadcast? Yes. I don't think that it would be any, serves any purpose if I try to say anything more on this front. Anyway, it was fantastic to talk to you all.

And I hope you will have the rest of your great Sunday. Take care, my friends.