Simulating the World & Remodeling Philosophy (Session 3)

Hello and welcome to the third session of simulating the world and remodeling philosophy. I'm going to pass the mic to him now. Thank you everyone. So I remember that last session we basically discussed a little bit more. This session will be more reading and then at the end some discussions but before moving forward if you remember Mikey had a response and let's first listen to Mikey and after that you know any kind of

questions that you have from the last session please you know ask them and and then i will start okay mikey uh it's pulling up just a second my things don't worry don't worry i hear some background noise Yeah, I do too. I'm not sure whose mic is on. It might be Lenka's. All right. Okay, Mikey.

All right. You're ready. It's almost. So I wrote a piece responding to a certain kind of part into simulation similarity. And I wanted to, I'm not going to recite the piece, these little parts I've moved on from, but there's a few parts I want to reflect on some questions coming from that I'd like to consider. So, and this was going back to the first questions I asked in the first seminar, going into like

the predictive policing models and where that came from. So I was investigating a little bit on the base of that model, which originally came from an actually earthquake or earthquake aftershock prediction model. They essentially took that model and just put new data into it. But this time it was like with human elements to analyze crime rates within the city of San Diego, breaking into sections of 500 squares, and then using it to predict like criminal hotspots where they would then deploy police and then in response be able to send more police out and the what they kind of found was essentially

that it was just validating you know what they've already done because kind of of course it would because it's already sending police to the point places they're already going so it's just this feedback loop. And so I was using the idea of Lacoste-Schules to kind of analyze the model itself and the validity of it. And, you know, looking at how, what the scope was, how it was deployed, and then also just, you know, how it actually like rolled out into the real world. And so kind of my questions are going into the applications of other models. And seeing how they can be like, certain things can be pulled from different models and then laid upon like other models or other scenarios.

where in this one, a lot of the bias was able to just be validated, validated, instead of it actually being like using as a model to investigate the concepts or to look at the theories in it. So yeah, that's kind of like my big question is just being able to, how do we use models in those scenarios to investigate other models and expand further on and not make it something where it's just like self-validating. And then also in terms of specialization, where the police, when they're putting this model out, you know, essentially you're saying like, this is really complicated mathematics that we're using.

So, you know, just kind of using this language as being like, you would not understand this unless you're a specialist. And a lot of times it's kind of like these models and the scientific theories that kind of go into these concepts, especially in social and political practices. The language is kind of put forward as a way that you want to get this, let that specialist handle this. And the way I kind of see a lot of this is that this practice of like within reason and and in scientific theory and model theory could actually be used by anyone and how to approach that like systematically and just basically finding false ability in different ideas to be able to fully investigate

and be critical within certain ideas, particularly within power structures or things that are presented to us as givens based on specialization. So yeah, I mean, there's, yeah, so those are kind of like first like basic thoughts and things I wanna explore more and looking to kind of like expand more in this class. Super, okay, excellent, thank you so much, fantastic. With regards to the first question, in the sense that yes, models are, you know, a discrete model that is being deployed, whether for, not just for prediction, sometimes for describing a

phenomenon, sometimes for even explaining the mechanisms that are responsible for giving rise to that phenomenon. If they are treated as discrete models, yes, they fall in what you call self-verification criteria. But the thing is that this of course comes back to what we talked about with regard to the epistemological criteria of how models should be understood within a broader domain, within a broader epistemological realm that allows us to investigate models without settling for the criteria of their application, the criteria of their representation, their

fidelity criteria, so on and so forth. One way to go on and investigate models is by looking into the structure of any model. And relocate that structure, reinvestigate that structure within the context of established theory. You know, that's one way to go about it. Now another basically way to do this, this kind of investigative work, rather than just simply treating model as its own self-verification

in a kind of circular manner is by way of toy models, which Adam also brought up with regard to coding and programming last session. Essentially toy models, of course, toy models also can be very different. But we will talk about them down the line that you might say that you see every model structure every model has a core. This core is an interpretive structure of a certain kind of phenomenon. Now it is a structure in a kind of quasi-isomorphic

sense in the sense that we call it the structure precisely because we have seen which we can explain or we can map this structure by way of a theoretical framework. So a structure is in fact a theoretical entity. Models just crystallize this structure and make it applicable to a certain kind of phenomenon now if we go into we look at a model we see that it has the model structure has certain kinds of what you might call to be information with regard to the kind of a structure that it wants to interpret the kind of structures I

want to represent so on and so forth but the thing is that here inside the model itself inside the model itself in the same vein that you talked about it we cannot in fact accurately assess the kind of theoretical assumptions which are encapsulated in the model. A theory, you can think about this like this. So you get theory, which is a very expansive system of structuration or structures, their relations,

so on and so forth on different classes and scales. And a model is a kind of crystallization, a kind of narrowing down of the scope of this theory. So obviously, when we are using a model, there are so many theoretical assumptions that are only functioning in our model in a hidden manner. As long as we just pay attention to the model, we cannot even talk about these assumptions, because these assumptions are implicit, are implicit in the structure of our model. to broaden the scope, and hence if we are talking about some sort of epistemological detective work, if we are really going to talk about model coherently, it's a structure,

we have to reconnect it back to its theoretical context, moving to a broader realm and bringing some of these hidden assumptions that are operative at the level of the model but nevertheless precisely because model is a very narrowed down version of it. We cannot talk about them coherently or explicitly at that level. So this is one level. But there is also even a larger domain that even at the level of theory we have certain metatheoretical assumptions. You see, let me water it down to this kind of formula.

So the theory tells you what you ought to do given such and such circumstances, given such and such laws will produce such and such results and not others okay not others is what you might call to be the realm of meta theory even a larger realm than the theory itself for for example equations of motion for newton what you might call to be theoretical assumptions Okay, and orary is a concrete model. It's essentially, you know, that kind of sphere and the elliptical motions of celestial bodies

which you can turn the handle and they go around, you know, around a particular celestial body. So this is your model. If you want to talk about it coherently, you need to go look at the kind of theory that is responsible for crystallization of these models. And this is what you might call to be Newtonian equations of motions. But Newtonian equations of motions obviously index or include certain laws by virtue of the theoretical structure itself.

They don't, for example, talk about some sort of chaotic phenomenon in the trajectory of, for example, the movement of a celestial body. So they allow you to talk about certain things and not others. And of course these are quite implicit within, for example, Newton's theory of motion. Not only implicit, sometimes they are in fact unconscious, like unconscious entities, just simply Newtonian theory cannot imagine what would be a chaotic theory of motion at a celestial level.

So you see, in order for us to actually move from this talk about the model coherently, not only we should recontextualize and reconnect back the structure of the model to the structure of the theory which is responsible for it but also moving from a theory to a broader expanse of meta theory basically laying it down and comparing it with other kind of competitive theories you know theories that might actually challenge newtonian equations of motion

either by giving something more or by challenging some of its theoretical assumptions so this is this is the kind of an epistemological route that ought to be followed uh for us to get out of that kind of discrete self verificationist way of looking at a model. So this is a kind of an epistemological lesson. Now, your second question was about what you might call to be socio-political and cultural implications of this epistemological route. Well, I don't think that this question is very easy to answer and I don't think that to say

simply that just because a specialist do it there might be some sort of sinister or hidden sordid agenda involved in application of these models or saying that basically we need to put ultimate trust in the kind of a specialist who deal with these models the way that you were talking about. I don't think that either of these choices are socio-politically sound, I actually say that yes, there should be a balance between

an epistemological trust in the edifice of science, but also this brings back the idea of education and scientific education at the level of the public discourse, again, to the foreground as long as people are not even willing to talk about hardcore epistemological constraints and scientific procedures and just simply make some sort of cultural commentary about them as we see today in philosophy social sciences so forth then that would be the pathology that would be just natural pathology there should be obviously there should be a some sort of modest epistemological trust. The person who actually advocated this was

James Ladyman when it came to one of his lectures at Oxford University. When he brought this issue of trust in the epistemological acuity of a specialist he brought this issue up when he was asked about you know creationism you know well yeah you need in order to I mean creationism is just like you know total bullshit but if you don't have an epistemological trust in evolutionary biology the kind of stuff these computational or evolutionary scientists

do to make their theories and come with facts then obviously you can you are basically you are prone to saying that oh well maybe actually creationism you know is not that bad you know it might have some grain of truth to it well it absolutely doesn't have any grain of truth to it so that's one But I would say that this needs to be a modest thing. Yes, a kind of fanatic trust in a specialist and what they do create blind spots, social culture of blind spots. Because as you say, such a specialist might actually fundamentally be misguided. In fact, there are so many examples of this in the history of social sciences.

people who come up with models purely based on theories which are utterly incoherent, utterly dubious, so on and so forth. So that's one. And the second one, which is basically a task of a socio-political enterprise, is the idea of the scientific education. It's It's just that we cannot talk about such stuff, whether trusting scientists or not trusting a specialized scientist, if we don't have a level, a robust level of scientific education ourselves. Then obviously if we don't have it, then pathologization, this kind of bipolar pathology of mistrust

and absolute trust will inevitably happen. Yeah, a couple things to think about on that is, you know, in terms of like the kind of specialists, it's like the, even though it's like, you can have the trust in those who are highly educated and work and specialize in a certain field, having the, you know, education or like working towards like a general practice of a critical thought and critical reason and looking at things like uh model theory it it focuses more on um issues that like you don't need to be a specialist but you can find the foundation of an issue uh that something's

being presented and like for instance this model theory it's like there could be you know the the specialists in the police departments the specialists with like certain data um and the specialists even like an earthquake that like built this model they can each have that but when they're all put together from an outside view of using things like model theory or just critical reason, you can find issues in the whole, the structure and the patchwork of that model. And I've noticed too, like part of this, I'll share my paper here in a little bit just on PDF, but a lot of the discussion around predictive policing has more so in just like police practices and you know, the moral issues and that kind of polarization and but no one i haven't seen much investigation

on the actual like algorithm using and the application of like does this earthquake algorithm actually really apply to this but is there any much validity in actually using this and right well that's that's basically again comes back to the first question the idea of the epistemological called detective work that surely you cannot just, for example, this also happens in, I mean social Darwinism is a great example, models of social Darwinist dynamics, whether couch in terms of game theoretic or rational choice theory, which are kind of on the same level. the thing is that okay so inside your first level the modeling and stuff well

sure things you know cohere up they sound all good you know the math is solid so on so forth but then the whole point is that how can you in fact apply Darwinian theory of natural selection which is essentially about biological laws into a different scale, namely society. That very application of a theory, namely Darwinian theory, to some sort of socialist structure, that requires metatheoretical assumptions. And that's why you have to go to an upper level to investigate it, that whether it is actually a

sound idea of theoretical application or not. And of course, when you look into history of biology and the great works that have been done at the intersection between physics, biology, and social sciences, the thing is that there is absolutely no way that you can extend or overextend what you might call to be a natural law to social rules, or what you might call to be a natural theory to the social structure.

I mean, yes, you can do it. You can do it. But essentially, you cannot get the same conclusions that you could get from Darwinian natural selection inside a different scale, namely the social scale. This is when you move from one scale to a different scale, constraints are added. and these constraints transform even the application of that theory, even the nature of that theory. So this is really important. Question, questions.

Heckling. i was supposed to ask some questions today i think yes okay yeah i have some um throw it uh i think they are kind of metafilosophical in scope. First one is I would like some specification regarding the reasons one can sustain as Weisberg does the priority of the mathematical models in the sense that

computational ones would be a subset of those yeah and the reasons one can sustain as you you do the priority of the computational models understanding the mathematical model as a subset of the of these right okay this is this is one very simple I mean I don't know if it's very simple but anyways is more more of a understanding I would like to understand more about this sure second one is something about ontological commitment historically I would say one of the reasons some people like for instance

Wittgenstein would sustain the possibility of a kind of reduction of mathematics to something like computational computation like an algorithm something like this one of the reasons historically was to dispense with the infinity and i was asking myself if the position that takes the computational model as the larger set would be committed to finitism to mathematical finitism because of the algorithmic nature of the computational structure so this is a fine

normally we would think of it as a finite structure like a finite number of steps and you can construct like mathematics in a somewhat intuitionistic way like a number of steps will be made by a empirical individual and but this commits intuition is to a finitist position all right yes yeah yeah okay and the third one is something that maybe is a question for later, maybe, I think it is, because actually I was influenced by seeing in the bibliography the presence of Nelson Goodman's ways of world making. So I was thinking about when you dispense

with bold realism, the idea of direct access. Naive realism, yes, okay. Yeah, bold realism. Right. I ask, doesn't the concept of model gain a much more general remit in the sense that every theory you can fathom would be thought of as a mediated access not anymore as a direct description so in a sense the question would be what is not a model if you uh dispense with the

bald realist thesis of direct access right right essentially are you saying that so what basically once you dispense with this kind of old or naive realist approach, the distinction between model and theory collapses or blurs. Yes, this is what I'm asking. I'm asking what's your position on this. I'm not sure about it. Right, right. Okay, yes. Well, the first question, I mean the first and two questions were, I think that they were on the same spectrum. So I would just try to come up with the broad answer to this. I don't think that the mathematical models are subsets of computational models. I would say that mathematics is a subset of computation. Now this is an entirely

different thing. Computational models essentially might have different patterns of computation or hidden theoretical assumptions about what counts as computation, right? So you can have in fact different kinds of computational models. And by no means at this point in theoretical computer science, we can just say that well, the Turing finite to save machine is a theory of computation. No, in fact, that is actually a model of computation, a specific model. And in fact, Turing also talked about this in one of his later pieces, that he can imagine

different paradigms of computation, different pieces about what counts as computation. By virtue of that, we cannot say that mathematics, mathematical models are subsets of computational models, not only because we might have different paradigms of how going about doing mathematics, but also because we do in fact have at this point different theories of what counts as computation. So I would say that rather than saying that mathematical model is a subset of computational model rather than it being a kind of, you know, kind of switching around

the priority, it actually creates confusion. It just doesn't make sense because at the level of the model, insofar as any model is beholden to the theoretical oedipus, theoretical framework in which the model structure has been couched, has been distilled, as long as we don't know about these theories, we cannot make such formulas. It would be just like vague. What does it mean? It really doesn't mean anything. I would say that it is far better to go on a kind of a different scale of the ontological status of mathematics with regard to computation and then at that point I would say yes I think that mathematics can be understood as a theory of computation precisely because

computation is what you might call to be in the broadest possible sense a way of of how you can move back and forth computational authoritarianism between mathematical structure, the question of mathematical structure, the question of logical proof, logic, and the question of programs. Now the thing here, when I'm talking about computation, it no longer means the kind of canonical Turing thesis. It is a very, very, what you might

call to be theoretical understanding of computation in which we have simply what you might call come across some sort of deep correspondence between proofs and structures and the computational programs responsible for this move back and forth but here again at this point you see a kind of a Godelian or even a Carnapian problem arises in the sense that, okay, we say that structures, namely mathematics, proofs, namely logic, can be thought in terms of programs. Let's

not talk about programs at this point in terms of a finite state machine a lot Turing and Church, but talk about computation simply as a view of the processes a view of processes that basically both mathematics and logic can be thought in terms of two manifestations of a very divine vision or underground vision of processes what that's how you know they're both involved in process proofs it's essentially a proof since the time of gerhard Genson, we know that any proof can be actually thought about two processes interacting with

one another, the proof of A and the proof of not A. The same thing about mathematical structures, certain kinds of processes, geometrical, algebraic, so on and so forth. Now here, of course even at this loose talk of computation we will eventually force to a deeper level to talk about meta processes what counts as a process a theory of what counts as a process so when I'm talking about computation by no means I mean it's a kind of classical church curing thesis but nevertheless Nevertheless, Church was, I would say, Church and Turing pointed toward this way, even though

their theory of computation was in the classical sense that they formulated initially was restricted. Throughout the past few decades, we see that this model, this Church-Turing model has been loosened to a theory of processes where you can have different qualitative levels of what counts as computation, interaction between processes. So that's all I mean by computation in a kind of a very abstract theoretical computer science talk, not by any means in the way that applied computer science talks about finite state machine in the universal theory machine so on so forth

okay so you think this does not this this uh this sense of talking about computation does not commit us to finitism in that same sense because it's a theoretical sense yes and precisely because you have to move it enables us to move the scales of different different constitutional system, different paradigms of the understanding. Yeah, yeah. Kind of like a Godel's incompleteness theorem or later Carnap's constitutional system, moving from one axiomatic computational system to another axiomatic computational system in which you can talk about proof of those axioms in a new framework. And hence from the second level again to the third level up to infinity.

Okay, I see. Yeah. Now, I just forgot about the third question. The third one was the collapse between model and theory when you, once you... Oh, a bold realism. Bold realism, yeah. So you're not a bold realist. I mean, the naive realist position. In the naive realist position anymore, is there a collapse between models and theory and everything you say will be, in some sense, a picture, but the picture that is mediated is not anymore, it's not a direct description anymore. Right, right, right, absolutely, yes. Yes, I absolutely do think that once we dispense

with naive realism, namely a thesis which purports that we have a direct access to the description of reality, we end up inevitably in the second position, namely mediated access. Every access to reality from now on is going to be mediated. Now, however, that does not by itself collapse the distinction between a theory and a model. precisely because even though models are implicit theoretical entities, they are not theory to court. They are not simply theories as such. You see, yes, they are both purport to be that every axis

that we are ever going to have with regard to reality is going to be mediated. But of course they have different kinds of mediation. And yes, at the level of pure mediation there is no distinction between them. But the kind of inferences, the kind of qualities and systems that they create for this kind of mediation, they are fundamentally different. A theory is not a model. Yeah, I was actually waiting for something like this. It's like there is a taking direct access as a criterion.

Okay, they are similar, but they are differentiated by other criteria. Yes, yes, yes. And actually you started to answer this question when you were answering the last question, when you were talking about a theoretical internal way to differentiate between a model and a theory, in the sense that a model is kind of a crystallization. I think this points to something which can be a distinction. and this distinction is to be traced, is to be made inside the theory, not by comparing a direct discourse with reality and a model, which is, you know, this is mediated, this is not mediated,

not that way, in a different way. Right, yes, absolutely. So let's imagine a diagram. So we have boxes here, lower boxes. It's a kind of a tree diagram. So you cannot compare, simply compare two models with one another. Because models as my key somehow intimate, there are like this kind of self-verificationist entities. Essentially, when you are in the business of the model, every assumption is there. there so it's just all about that we have when you can week the board with a model yes the data might be wrong you might have misapplied the model but

these are like what you might call to be human model of fault it's not the fault of the model in the model itself is not fault yet that level at the level of discrete model the model itself a singular model but then of course the model can be wrong itself but there is no way to talk about whether the model is wrong or right at the level of the model itself and there is no also a way to talk about this model versus that model versus another model so on so forth so so these are boxes model one model two model three model four there is absolutely no connection between them so imagine arrows coming out of these boxes going

to a different level a higher level which is the realm of a theory so at the level of theory which is a higher level a broader level we can talk about within this theory which of these models are actually right and which are them are false we can compare the models at the first level at the bottom level but then this is not enough imagine that at the level of theory we can go to a higher box a rectangular box the highest level a meta theory where we can think about that the theoretical assumptions which we take for granted nevertheless are hidden

yet are responsible for us to making these models at the bottom level in accordance with the set theory. The theory cannot exactly like how discrete models couldn't talk about themselves, couldn't reveal their secrets to us, whether they are right or wrong, a theory itself cannot talk about itself, about its axioms, about its stuff. We need to go to a higher hierarchy. That's a metatheory where we can, and of course there are many infinite hierarchies of metatheory where you can refine, bring forward to the light theoretical assumptions

which were there but nevertheless were hidden. Some of them might be wrong, some of them might be right, so on and so forth. Yeah, yeah, yeah. Actually, yeah. You answered my next question. That would be the difference between the use of… I'm thinking, actually, I don't know if you're familiar with this book. I was reading What is a Philosophical System by Jules Villemin. Oh, yes. I'm a big fan of him. He's superb. Yeah, but it seems to me that he ends up being in a place where it's like each philosophical system is like a closed system.

Axeomatic system, yes. You can't compare and you can't choose. And yeah, it seems to me... well you see women comes from women comes from a do him and coin you see the thing about so especially the the vision of philosophical system that he endorses by the way Joe women is a French philosopher of science was a sorry was a French philosopher of science and he sees philosophical systems a kind of like these monads, these kind of self-contained spheres of axiomatic systems, which of course

they can be expanded, but only according to the kind of axioms they use in order to index certain kinds of formal properties, so on and so forth, which can be mapped onto some domain of discourse. Now, the thing is that this vision of Jules Bouiman is very, very compatible with the early car map, kind of these kind of enclosed axiomatic systems where you don't have any criterion of comparing these, you know, basically what you might call to be philosophical systems.

And of course that leads to some sort of liberal pluralism. But the thing is that I think Carnap came up with a solution. You know, Carnap exactly endorsed the same kind of overarching multiplicity of these self-enclosed axiomatic formal systems in his earlier work, but then he listened, he attended Godel's incompleteness theorem where basically a discrete axiomatic

system cannot verify its own truth you see that's all point a philosophical system that is made of a certain axioms and itself and close can never actually verify its own truths it cannot say the kind of statements I'm making this kind of system are actually true or false. It cannot prove them consistently. Now in order for, so that was a blow to Carnap and that's how he moved from his earlier work to his later work to imagine exactly like a Godel incompleteness theorem, particularly the first incompleteness theorem, a kind of level of meta discourse, meta logic or meta

language where you can go to a different axiomatic system in which all of the other axiomatic systems can be compared and talked about, they can be assessed. But even that meta-theoretical system requires a further meta-theoretical system such that from a higher point of view or vantage point which can talk about the proofs, the statements and the axioms of a lower system, so on and so forth. Yeah, yeah. It's like isomorphic with the idea that you can't really compare two models but you have

to jump into the theory level and then you can compare and then you have to jump to the the meta theory level and so on and so forth. Yes, yes, yes, absolutely. Yeah, okay. Yeah, I think I'm done with my questions. I mean, you can proceed. Fantastic questions, fantastic questions. Thank you for your answers, fantastic answers. Okay, so any more questions about all of this stuff that we've talked about? Ian, you are so silent today. What? No, I'm good for now. I'm just putting things together. Questions later.

Excellent. Any person? Marie actually had something came up on the sidebar. I just couldn't read it. Sorry, my sight is so weak. My microphone is on. Hi. I think we moved away quite a lot. I just wanted to slow down a little bit to ask you to maybe clarify for me what you meant with experts, because from what it sounded on, it's like we're creating perhaps a dangerous boundary if no one who's not an expert on that particular model is allowed to have any kind of inquiry. And isn't it

note a little bit on the part of experts to also make that expertise accessible. What value does an expertise have if it's only locked in within that field? Maybe I understood what you meant, but I find it very dangerous to use a lack of expertise as a form of deterring anyone from thinking outside of the expertise. How do we have a progress if we don't start somewhere? Well, you see, I think the problem is a little bit topsy-turvy here. In the sense that, right, you know, obviously the lack of expertise is not a vice. It can be turned into a virtue.

But of course, that's why I brought up the idea of a kind of scientific and systematic education around these kinds of topics. However, we need to be realistic at this point. I mean, I really genuinely don't think that a person who has zero scientific knowledge can talk about science at this point without any kind of epistemological inquiry, philosophical inquiry in a systematic way that is also acquainted with modern sciences. Essentially, we get down the road, what you might call to be a kind of, you know, again, I'm sounding, you know, even I'm Marxist, but any comment I make about Marxist sounds

up as if I'm like anti-Marxist. But really, this is, isn't it really the whole point of the orthodox Marxists today, that They say that, oh well, there are these stuff, you know, and these scientists are these shady, dark people doing sinister stuff behind our back and we don't know they might be actually real subsumption there, capitalist has hijacked their science and so on and so forth. But no, there is absolutely a way to go through them in the sense that science, scientific procedure is not responsible.

It's responsible to make it accessible to the public. Yes, I agree with that, with what you said. But the whole point is that the scientific procedure is not something that abide by our normative prescriptions about how society should be, how this should be. This is the whole point, that when a scientist talks about stuff, she's only interested in the scientific procedure. There is no... it might be, it might be, but let's pretend that it is just a pure scientist, a pure scientist.

Following the formal procedures of science, well, here again another problem arises. What if these scientists are in fact being politically bought, socially bought, culturally hijacked so they produce certain kinds of procedures and not others well we don't know that we don't know that as long as we are illiterate we don't have the capacity to rise above to a certain kind of scientific method of inquiry where we can actually look into these problems but also those scientists have somehow found a method to make accessible their procedures to

the public. There is no, I would say, non-pathological way out of this scenario. We can just talk about all sorts of stuff, they can just talk about all other kinds of stuff, There is no incommensurability. I would say that Marx would have said if he was alive that this happens when proletaria fails at his task of cognitive revolution. And when proletaria fails at the idea of cognitive revolution, science becomes fundamentally detached and in its detachment it may very well can

be hijacked by some sort of capitalist or other kinds of agendas, political agendas which are not scientific and merely ideological. But I would actually put the blame here at the people. People should ask these questions. People should be able to come to the level of the adequate or sufficient or necessary understanding to look at these problems. When a Marxist basically dismisses the idea of philosophy of science, and philosophy of

science essentially this midterm between the two, between the kind of general cognition and the scientific cognition, when orthodox Marxists dismiss with the idea of philosophy of science, then what do we expect? And yes, Joven, not everyone needs to be scientists, but at the very least we can do in fact, try to acquaint ourselves with how science works. And that's exactly what philosophy of science is. Expect scientists to create a kind of a cognitive trickle-down economy, giving us a little bit tidbits of their stuff. fact let's look at it today in popular science look at all the pop sci science that's isn't

a trickle-down economy from scientists so the scientist tries to water down this stuff give it to some other journalists and that journalist makes it accessible and then you get the worst kind of ideological science ever existed about a string theory quantum theory so on and so forth you get like the ultimate mishmash of stuff it's really i i do think that this is this is a real uh a problem in the sense that people have failed to live up to the cognitive revolution well who's going to be in charge of the level of experts and if it becomes an elitist club that

only one who is at the top of the expertise get to decide who they engage with then I think we're in a very risky situation because we're just going to end up in a place where I could say look you don't know anything about art and I won't talk to you. And then... Yes, yes, sure, sure, this happens right now but let me tell you that there was a better time at the end of the 19th century where scientists didn't have this attitude of saying that like Ludwig Boltzmann or Einstein, the greatest scientist, I mean you cannot really, just simply you cannot find better scientists than these two, or Newton. They talk to philosophers, they talk to psychologists, they talk to artists and they actually create a cultural revolution

rather than just mere scientific revolution. So if this has happened before, why can it not happen today? Obviously the reason that it's not happening is precisely because of the sociopolitical framework in which both philosophy, general inquiry and scientific inquiry are embedded, but also because of paradigms of education, purely market-based, purely profit-based, so on and so forth. um i'm i'm just sorry i'm just like looking at um the sidebar um um we're all heckling you at the

sidebar yes yes yes well yeah go on go on please please well yeah i was i just said you know the freeing of intellectual property has kind of through like the sort of technology of the present it's made this all slightly more possible but in the past a lot of this stuff has been limited and also like i have a lot of skepticism swords like social collectivity and kind of this apparatus that we can you know if you look at the past 20th century and sort of the failures of kind of like bureaucratics you know soviet communism it's like there was an anti-science to that bureaucracy in and of itself and like all these systems of social of of of collectivity within

whether you're sort of like a marxist or a post-marxist or whatever like i just have this great skepticism that it's going to yield this sort of like universal um basic interest and sort of like so when you're talking about i forgot the term it was it was uh um it it had to do with the the proletarian the cognitive revolution of the proletarian like i i guess i have morris was alive he would have said i would say he would have said that yeah but one of the things that i i guess i'm just a little bit more skeptical about this sort of like as as we kind of see ourselves kind of morphing into this more kind of like machinic practice of of of endless intellectual

property and our ability to even like atomize these ideas by our own self-determination like I can't even potentially see now how we would be able to systematically implement these ideas in a framework that would yield some kind of scientific universality. And I think one of the issues that I'm also seeing, and one thing you pointed out that I thought was interesting is by constraining with science the sort of possibilities of maybe our outcomes and sort of like creating new commons based on sort of science, like

you can potentially do that when it has to do with things like perhaps like air quality or sort of like things that happen because we're kind of destroying our underwater ecosystems, levels are rising, we have fungal blooms in the ocean that are creating these very high carbon levels. And if there's some sort of like collectivity that's kind of shifting what's going on right now, that becomes, I guess you can find new scientific universals throughout that. But I guess I just have a lot more skepticism on kind of like, I guess maybe I have a lot of faith in sort of like the individual agent and maybe sort of like skepticism and sort of the systemization of kind of like...

Right, right. Okay, now I see. That's really good that you are coming from a different perspective. I'm coming from a rhetorical perspective. No, actually I think that you are coming from a kind of individualist, a logically individualist perspective. A little bit, a little bit. Yes. I wouldn't say... quite familiar with this kind of line of thought. And to be honest with you, for a very long time I dismissed that kind of credo of thinking, being a Marxist. But to be honest with you, the more I have looked into these problems, the more I have thought, I have become pessimistic about what can be done and what ought not to be done.

I would say at this point, the pessimism should not just be levied against universalism of any sort or proletarian collectivity. I would say that the idea of individuality is even more fragile than the idea of collectivity. Oh, yeah. You see? That's one of the problems that we're having today as sort of this infinite apparatuses through, I guess, I don't like saying light capitalism, but whatever you want to say, which kind of give this kind of like endless culture of self, of customization and representation. Yes, yes. I agree with that, and that's one of my criticisms of capitalism where it's at right now.

But that would almost be a post-Marxist critique, right? Yeah, actually it would be more in tandem with some of the more sophisticated communitarian philosophers. But the thing is that you see science, this might actually strike you very, very weirdly. I do believe in Heidegger's idea that science does not think. Scientists think. What? Yes, science does not think. Absolutely science does not think. To think means not to discover just facts, but to be capable of differentiating between facts and values. Hume, essential problem of philosophy.

How can science do that? Look at Knight Brice Tyson and so many other scientists who are basking in the light of higher sciences, and they are constantly confusing facts with values. How can we actually derive values from facts? This is the premium question of philosophy, the question of thinking. no I agree that's um that's that no who said what I didn't say I said it okay you have to explain yourself now no no no I like all the these questions are super pertinent but this is you

know I think a lot of the I think Laurel addresses a lot of these questions which is why I'm doing a lot of work on him but yeah I mean his whole thesis is that science can think and what he means by sciences it's a distribution of labor which is what Mikey was talking about where where science is no longer simply you know a scientist but how can we make science equivalent to art to etc to etc but still maintain the objective rigor of a science and so this of Of course it's more like artistic project, more theoretical, more critical project, but I think it's one way that he's doing it. Yeah, no, no, absolutely. And to be honest with you, in order for us to in fact think coherently about the political

political problems of scientific manipulation or common sense for that matter, we just cannot go on and talk about expert versus non-experts. And I actually like Adam's sidebar comment, you know, one to the last. I don't think that this is, at this point, this is really something that is fruitless discourse, fruitless critique. The more fruitful question is to talk about the idea of how technique actually popularizes science and redistributes it. How basically what we talked

about the re-understanding of what science is as a kind of a systematic discourse. Philosophy, I would say is science, absolutely the science, but not an empirical science. So there are all these kinds of stuff and this is why I absolutely do believe that philosophy has a role to play so as science and also a kind of a mediated tissue between the two. Well, We don't know yet this mediated tissue, how it would look like. I would say that philosophy of science for the most part is not really a proper mediated

tissue in the sense that if philosophy is about fact value distinction, about the true ambitious cogito or cognitions, theoretical and practical, and also axiological, and science for the most part is about explaining, describing facts, then we need to have a better tissue, a mediating tissue between the two. We just don't know. And as long as we don't even think about such problems, no matter how much we complain, oh, these experts do stuff, other people doing other kinds of stuff, we are not going to go anywhere. I really want to ask about your inherent Platonism in this entire view, but I think we should

talk more about models. Yes, yes. Okay, let's have... I need to have some water. Let's have a quick break, and then I will start reading no question asked at that point, and then we will open it to questions at the end. Sounds good. How long a break do you guys want? I don't know. Five minutes? Five minutes is good to me. Sounds good. OK.

Thank you. Thank you.

Thank you. Thank you.

Thank you. Thank you.

Thank you. Thank you.

One thing that I just look at it's Barrett and Alan talking about this. Yeah, sure, this is of course, it's a kind of a Althusserian line of the critique of science and ideology. Well I would say to overextend this picture that everything that science does might, might

be kind of already subsumed within the capitalist mode of production and thus some sort of ideology and so on and so forth. I just generally don't see the merit of this discussion precisely because that is already philosophically presupposing that capitalism is a totality under which every productive, cognitive relation has been subsumed. To even pose this as a problem, you require to in fact pose as the premise of your argument the totality of capitalism. The totality of capitalism, if understood as a totalized and achieved totality in which every possible

mode of production or social relation has been subsumed is already a metaphysical reality and metaphysical illusion and by that it's no longer in fact a coherent either political thought or scientific thought. One of the greatest communitarian thinkers, he used to be in fact a kind of orthodox communitarian thinkers, Rob Lucas, after he got disenchanted by this idea of capitalism as an achieved or completed totality, which you can talk about, you know, everything might or possibly be ideologized or, you know, subsumed within this kind of a specific

mode of production of social relations but nevertheless that idea is already a kind of a metaphysical wishy-washy picture, a kind of something that doesn't really, it's not scientific, in fact, yeah, I think that Rob Lucas talked about this problem really, really nice, a very kind of a Marxist tradition to show that essentially if we think like that literally then there is absolutely no way to talk about anything because the metaphysical

illusion takes the upper hand ultimately and why do we even think about such things as as an escape, as how things should be rectified, so on and so forth. And that paper is called by Rob Lucas, I don't know whether it has been published yet or not, it is called How to Educate a Child. an absolutely fantastic paper I highly suggest it I don't know about whether Ellul could be considered to be communitarianist or not no I don't think so but nevertheless you see Barrett I think that many this

It's kind of a communitarian, orthodox communitarian vision, kind of like, you know, the idea that, okay, so you have some sort of completed totality, like the engine of history, like capitalism. So everything that happens within this time is in fact an index of such time. Everything that is happening is molded by this time, so on and so forth. You see, this is not really, I would say, the particular or the exclusive characteristics of orthodox communitarian thinking. In fact, you can go and think about no reaction. No reaction is in fact, has much more to do with communitarian thinking than any other

stuff. The whole idea of the floating island is the idea that, okay, well, basically this whole world has been molded by multiculturalism and all this shady stuff going on in the world, and okay, let's make a commune, but let's not call it the commune, it's called a floating, no reactionary island. But the thing is that first of all, this kind of prescriptions never actually substantiate, precisely because it has nothing to do whether everything has been totalized by some sort of higher order or not.

It's just the idea that as long as you posit such a metaphysical totality, your utopia, Whether it's a non-reactionary island or whether it's a commune or an atopia, a negative space in a Foucaultian sense, atopia, would be parasitic upon that logic of metaphysical illusion of totality and not an escape from it. I agree.

I well I had one question that I was kind of thinking about in sort of like you know because I I'm very interested in everything you have to say it's like totally different than where I'm sort of where I've been thinking for a little while but the sort of kind of like I guess going back to the systems that that inform our sort of our sort of versions of scientific rationality I think like this sort of latent totalitarianism of scientific rationality rationality when it's kind of when it's kind of prefaced by a certain system like you know let's say you know we kind of talk about the story and i i heard you say something at an earlier uh seminar is like you know our our science now under sort of neoliberalism is is almost more

mythological than that from coming out of like baroque theological catholicism i i heard you say that. I'm always thinking about the systems. It's like using the method of scientific rationality. It's like, okay, well, if you're creating a commons condition where we're all supposed to jump off a cliff because scientific rationality tells us to do that. But I understand there's an inherent need for us to analogize these things and sort of so i'm not discounting science i used to be somebody who was like a magic person and i'm not anymore right right you know and i so i understand i definitely understand that and and so but i

guess i'm just a little bit skeptical of the sort of like that of this kind of thing because you see a lot in today when we talk about sort of like the atomized society which is i live in america We have this thing. Yeah. OK. Where we have this kind of this dilemma where we have people. I see this constantly who will defend science on one end and then they'll they'll they'll sort of also on the other end defend this kind of like, I guess I don't want to say post rash. I don't want to say post-rationality, and they kind of conflate the two and they become this sort of interchangeable thing due to the convenience of an ideological system that

predates them. And so I guess I'm very, very skeptical of these things as well. Yes, yes. Essentially, I would say that both the kind of extreme universalists of traditional Marxism or even can't and the kind of this kind of individualistic sense where basically ultimately all individuals should take their cues from science about how to live a good life are just pathologies and philosophy was invented precisely to thwart such pathologies of course it failed but this doesn't mean that we cannot reimagine a new task for philosophy reinvent that task the idea of a good life requires you to

to think at least about three domains or realms or categories of thinking one theory or theoretical thoughts practical thoughts and then axiological thoughts value thoughts we can you can even add another fourth dimension aesthetic thoughts yeah and as long as you don't have this for then everything that you do will become in one way or another pigeon-holed ultimately pathological and exclusive I see that Artemis asked a very very good question here about this idea that Okay, so all of this stuff,

then why are we actually talking about models? Well, I think Jean-Pierre brought his, by way of her question, sorry, his question, that the idea that models, first of all, the first things that they do is that they make this evident that any sort of access to reality is going to be mediated. Now, when it comes to mediation of reality, okay, let's say the ultimate question is to renegotiate our positions in the world.

The preenial question of all thoughts, who are we? What are we doing here? You know, this is renegotiation of our position in the world. And by extension, the renegotiation of who we are. Okay. If that's the case, if we already know that there is no such a thing by way of scientific revolution, there is no such a thing as an immediate access to reality, and we have basically bunch of models and then some theoretical frameworks, then this question becomes even more convoluted in the sense that exactly how can we talk about coherently about our

a specific mediation with a specific sectors of reality. This is the question of modeling in the broadest possible sense. The question of a specific forms of mediation with a specific forms of reality, namely particular phenomena at hand. This, I would say, is the ultimate question of modeling. Okay shall I start?

Okay, so as I mentioned, you know, that there is a, to understand what models are, we should understand what models are not, you know, in a determinate sense. Of course, in the history of science, the talk of modeling goes back to a few particular figures. As I mentioned in the first session, if I remember correctly, I mentioned that in the

early history of philosophy of science, there is no mention of models. Sure, there are some mention of models, but models are quite vague entities, you know. They literally don't have any distinction between models and theory in any sense. It's just that you can use these two words interchangeably at that point. But later on people think about these two things very, very differently along the kind of discussion that I had with Jean-Pierre and also Barrett, Mikey and also Adam in the previous session. One of these figures who kind of highlights this distinction, this essential contrast

between models and theories, but in a rudimentary sense, is Patrick Supes, S-U-P-P-E-S, and his colleagues. They were, at that time, when they were writing their papers, they were endorsing something that is called a semantic view. And according to the semantic view, theories, physical, psychological and biological theories,

are sets of models and not sets of axiomatic sentences. So what is axiomatic sentence? You see, with the rise of logical positivism and logical empiricism, you know, the Vienna circle in analytic philosophy and philosophy of science, early on, so there is this idea if, and please do watch those of you who have access to archive, you know, the previous session precisely because this session is just simply a sequel to the previous one where we talk about logical empiricism. So logical empiricism, early on the proponents of it

believe that basically what counts as a description of reality are some sort of pictures. What What are these pictures? These are axiomatic sentences, elementary self-evident sentences which map on a one-to-one basis or corresponds to facts of reality, facts of experience, empirical experience. And then if we believe in a set of axiomatic sentences which represent or picture, not

represent, picture, facts of experience, then we can create a kind of a higher hierarchy or system of further axiomatic sentences derived from the first set of axioms on top of them, ever more hierarchies of sentences and systematize them, cohere them, and that's how basically we created theory. So basically these first, the problems of the semantic view of modeling, which was like the forerunners of the model, scientific modeling and model sciences, were believers of something different in the sense that theories are not in fact sets of axiomatic sentences which

map onto facts of experience, but they are sets of models in the sense that we have been somehow talked about models very loosely. And there is, of course, a good deal of variety among even proponents of the semantic view as to the nature of scientific models. Soupis and his colleagues, for example, have argued that scientific models are simply logicians' set theoretic models. Later philosophers in this tradition, the semantic view, sought to describe models in terms closer

to those actually used by scientists conceiving of models as either as sets of trajectories to a status space like for example van frozen or as systems that would be concrete if they were real. Now proponents, generally proponents of the semantic view of modeling are primarily giving accounts of the structure of theories. Yet, insofar as these accounts make implicit claims about the practice of theorizing, they treat this practice as univocal, focusing exclusively on model-based representation.

If we set aside the question of whether all theories are best reconstructed as sets of models, we can still ask how scientists go about theory construction, construct a specific theory, and whether or not this theory construction explicitly depends on models. Now in arguing that modeling is just one kind, albeit an important kind, of theorizing, we We can argue that some theoretical practices explicitly depend on the construction and analysis of models, while others do not.

Modeling is the indirect theoretical investigation, this is indirect in boldface, modeling is is the indirect theoretical investigation of a real-world phenomenon using the model. This happens in three stages or three phases. In the first stage, a theorist constructs a model. In the second, she analyzes, refines and further articulates the properties and dynamics of the model. Finally, in the third stage, she assesses the relationships between the model and the world if such an assessment is appropriate.

it if the model is sufficiently similar to the phenomenon in the world that you wanted to model then the analysis of the model is also indirectly an analysis of the properties of that real world phenomena. Therefore modeling involves indirect representation, an analysis of real-world phenomena via the mediation of models, hence no direct access, but also a very specific mediation, mediation via models rather than mediation by way of direct theoretic representation.

Now of course this is not the only way to go on and about making or constructing the theory. Phenomena can also be represented and analyzed without the mediation of a model. This kind of non-model based form of theorizing is what is usually called abstract direct representation, ADR. Now similarities and differences in these two forms of theorizing are best appreciated by looking at examples in details. So the models you can think about, the Bay model, San Francisco Bay model, Volterra's

model, the Schelling's model of segregation. Now as in contrast, and I briefly explained them and described them early on in the first session. Now with regard to what counts an abstract direct representation, we need to have a little bit of a good grasp of you know what what kind of theories are not just sets of models. And if I remember correctly, I mentioned as an example of

an abstract direct representation ADR approach, Mendeleev's table of elements. So the story of But Mendeleev's construction of the periodic system has a very modest beginning. When Mendeleev was assigned to teach courses on inorganic chemistry at St. Petersburg University, he found out that there was no good inorganic chemistry textbook available. text lacked an organized or coherent structure from which to characterize the known elements

and inorganic reactions. So in order to deepen his and his students' understanding of the elements, he wanted to essentially develop a classification system, a taxonomic system, that elucidated their underlying properties. This would allow for a more systematic understanding of the properties of each element, the reactions each element could participate in, and trends underlying these properties and their corresponding reactions. So, Mendeleev essentially, again, faced a daunting theoretical challenge. Samples of the pure elements have many chemically important properties any of which could form the basis of a classification system.

Essentially, precisely because there were so many properties and their respective reactions, it would be really hard to make a non-arbitrary classificatory system. One might, of course, sort elements by color, density, conductivity, melting points, or by their affinity to react with various reagents. But in the end, however, sorry, one second, you know, Mendeliev decided to focus his attention on finding trends in the properties of valency,

isomorphism, and above all, atomic weight, abstracting away from all of the other details about properties and their corresponding reactions. Atomic weight is a familiar concept. What valency and what 19th century chemists call isomorphism may not be. are said to be isomorphic when families of salts containing chemically similar but distinct metals form similar crystal shapes. Valency, on the other hand, refers to the combining ratio of an element.

For example, carbon is tetravalent, meaning that it can combine with four equivalents of hydrogen. Mendelio's first step was to organize the elements by atomic weight. This allowed him to form a one-dimensional ordering of the elements, which served as the initial organizational device, classificatory device, but did not reveal any information about the elements underlying the structure or unity. Then next step, valency and isomorphism.

Mendeliev at this point tried to find out dimensions along which to organize dimensions, other dimensions along which to organize the elements. Like in the modern chemistry sense, we can think of his next step as trying to figure out where each period or row of the periodic table ended. In some accounts, Mendeleev is said to have put the names and properties of elements on cards and played chemical solterres on long train journeys until he found a satisfactory ordering of the known elements. In 1869, Mendeleev analysis ordering of the elements according to their weight and properties.

This ordering, which later became known as the periodic table of the elements, organized the elements in order of atomic weight and then in columns of groups in virtue of their chemical properties. When the elements were properly ordered, Mendeleev discussed, one could see the periodic dependence of elemental properties under atomic weight. This principle, which Mendeleev called the preiotic law, is one of the bedrock principles which organize the entire domain of chemistry. It is still recognized as one of the most basic patterns among chemical phenomena, although we now think of it as a preiotic dependence of elemental properties on atomic number or

effective nuclear charge. So Mandeliev's theoretical achievements are sometimes overlooked because of the suspicion that the priodic table is just and merely a taxonomic device, a classificatory system. It makes certain trends explicit, but it has been argued that the table does not actually explain anything about what these properties are. It has also been argued that Mendeleev articulated an important classification system, but not a theory as such. For example, a few philosophers of chemistry have argued that what Mendeleev discovered

was an ordered domain, and that orderings of domains are themselves suggestive of several different sorts of lines of further research, but not themselves, theories. Now, this is somehow actually a misinterpretation of Mendeleev's ADR approach. The first reason is because the remarkable predictions that Mendeleev was able to make on the basis of the periodic system. you know, predicting what would be the next element that could be ever discovered, you know?

For example, in 1869, he noted that there were gaps in his table of three elements. On the basis of information about chemical trends encoded on the table, he hypothesized the existence of what he called eca aluminum or eca silicon and eca boron the properties of these novel elements are listed in table one and just a few years later the elements gallium escandium and germanium were discovered and as indicated on table one their properties were in remarkable, almost precise agreement with Mendeliev's predictions.

So Mendeliev's predictions might look like trivial exercises making inferences about missing books on the shelf. Like you know, you go into a library and you know, by virtue of the topics and how they have writers have been classified, you think that, oh, some books might be missing on this shelf or these empty slots. This, however, underestimates the significance of the achievement, Mendeliev's achievement. Mendeliev had no empirical knowledge that there were any empty slots to be filled. His task was

Therefore, not as simple as interpolating the properties of unknown elements on the basis of known elements. He first needed to hypothesize the existence of the missing elements by, in fact, analyzing the theoretical structure he had created in the guise of a taxonomic system, namely the Priotics Table. Then he was able to use the trends posited by the periodic table to make predictions about the properties of the missing elements. This prediction was theoretical and not merely

classificatory. So to quote, actually let me... so basically without going into much details when we think about Mendelius periodic table and for example Volterra's Adriatic seeds model of predator and prey, we can see that the central contrast between their theoretical style involves their approach to representing and analyzing real world phenomena.

Mendeleev created and studied a representation of real elemental properties while Volterra created and studied representations of mathematical models that were similar to real phenomena. The difference is the essential contrast between modeling and ADR, approach, abstract direct representation. To clarify this contrast, we must of course take a closer look at the various stages of

both modeling and ADR. As a construction and analysis of models are key steps of modeling, we must first consider some properties of scientific models. So essentially what I'm going to do from now on, looking at precise stages by which a modeler makes a scientific model. And through these explication of the kind of stages that go into the model construction, We can distinguish a modeler from a mere theorist.

Now, questions? Adam, Theo, Mari, everyone else who has been silent. So is the explicitness of the model one of the key criteria here?

Like is that one of the things that... Because I mean, we talk about, like Weisberg talks about having a structure and an interpretation. trying to keep parts of the model. But it seems to me like a theory has a structure regardless. Yes, no, absolutely, absolutely. We will get into this when we are looking a little bit into the work of Joseph Sned. Yes, theory has a structure, and the class, the structure of a model is, in fact, a subset of the structure of the structure of a theory. But the thing is that what really here is the distinguishing factor is that the way

this structure is being encapsulated, the theoretical structure is being encapsulated within certain constraints with regard to applicability, representational criteria, description of such a structure so on so forth. Questions?

Okay. So as Weisberg has mentioned, there are of course many, many characterizations of scientific models that have been offered in the philosophical literature. And of course, to that end, we should be capable of coming up with certain kind of classification of what counts as a scientific model in which the majority of this literature can be included

but also some can be excluded by virtue of not really capturing the necessary stages of modeling practice. So, first, I would like to give a quick description of the nature of models and their relationship to the world. with regard to also, you know, the kind of literature about modeling that has been written. Then I would like to consider the role of the theorist's intention in the evaluation of the model-world relationship.

These features have been discussed in connection with models of experiment and models of data, and the inferential conception of scientific representation, but not, for the most part, played a major role in the discussion of scientific theories and theorization. So, again, as I mentioned in the first session, models are abstract or physical structures that can potentially represent real-world phenomena. Abstract ones are kind of like, for example, Schelling's model of segregation, a concrete one is like a fruit fly or an orari.

Many different things can serve as models, including physical constructed scale models, model organisms, and mathematical objects such as sets of trajectories through a status space, which are usually used in order to predict the evolution of a dynamic system over time. however for now we are going to restrict our attention to abstract mathematical models precisely because from a historical perspective in the in the practice of modeling they have always been the most primary forms of model-based theorizing.

So when employing mathematical models, one studies the model by studying representations of the models, which we can call model descriptions, and I talked about this, that model descriptions you can think about this as these kinds of sets of equations by which you say that well this is this is the model yet like look to Volterra's models or you know some other model like Schilling's computer algorithm for the segregation model so this is this is a really a model

description and a model description is not actually the same as the model structure the model structure is the core of the model structure is the core of the model it contains information implicit and explicit about what this This abstraction is not only that but of what this is an abstraction, of what kind of real phenomenon this is an abstraction. So you see, here is a broader domain of information and of course to talk coherently about it

need to have certain kind of constraints about the kind of phenomenon you are trying to model as well as the kind of representational constraints that allow you to talk about this phenomenon coherently with regard to your model so on so forth for abstract models model descriptions usually take the form of equations but graphs and other kinds of representations like diagrams can also be serve as model descriptions for example in volterra's predatory prey model the two differential equations you know one about the prey and one about the predator

population are in fact the model descriptions. In the model of molecular structure, the model description takes the form of a computer, for example, to a graph. In the original semantic view of modeling advanced by Suppi's, scientific models are equated with logician models and and are said to satisfy a set of axiomatic statements, which he also calls model descriptions. In another account, for example, by Geyer, G-I-E-R-E, model descriptions are taken to define models.

Both of these accounts of the relationship between a model and its description are in fact too strict because model descriptions often lack the precision to pick out a single model and this vagueness or partial specification is actually a benefit in the earliest stages of theorizing. Rather than characterize the relationship between models and their descriptions as one definition let us characterize the relationship between the description and the model as one of a specification this highlights the fact that the relationship can be weaker than the finishing

for satisfaction but that models are picked out by their descriptions now here are some philosophical accounts collapse the distinction between models and their descriptions. For example, Orzak and Sober treat differential equations as models. Now, for many reasons this is a mistake. One of the most important insights behind the semantic view and other attempts to reconstruct theories as sets of models is that a theory should not depend on a particular linguistic formulation, formal linguistic formulation. More importantly, for understanding the practice of modeling,

a modeler often conceives of a model in a vague way, writes down some equations to describe the model she thought she had in mind, studies the model actually specified by the equations, and determines whether or not they pick out the right model. Situations can arise where the modeler's imagination picks out some sets of models, and her model description picks out a different sets of models, necessitating and demanding a refinement either to her imagination or to her model description. Modelers often use models in order to learn about real-world phenomenon.

In these cases, the model must be similar to a real-world phenomenon, the idea of similarity. What counts as similarity? We are going to talk about it. For example, Quine has pointed out similarity is a vague notion and we therefore should not be content with such a simple formulation of the model-world relationship. One of the most active and contentious areas of the structure of theory's literature concerns the question of how to give a more precise and detailed formulation of the model-world relation. For example, some philosophers argue that the appropriate relationship between models

and the world is one of a structural similarity. On this view, models are imaginary structures that would be concrete if real. Their similarity to real-world phenomena lies in some parts of the imaginary structure literally having similar properties to parts of the real-world phenomena. Other theories, for example, such as Supus or Van Fressen, conceive of similarity more abstractly, describing it as a relationship between mathematical properties of the model and of the real-world phenomenon

described mathematically. A third view, related to the second, holds that models have partial isomorphism, one-to-one mapping to their intended target phenomenon via a series of models that are ultimately tied back to data, empirical data. This means that some substructure of the model, for example the relations between its properties, stand in a one-to-one correspondence to properties of another model which can in in fact be understood as the very model or the end framing of the empirical data. None of these views has clearly emerged

as dominant in the history of modeling and the scientific uh you know literature about theory construction. Each of them has its own critics and supporters. So Essentially, for us to move forward, namely to explicating the distinction between modeling and abstract direct representation, we can just get rid of some of these extra details of such debates in the history of scientific literature, philosophy of science and modeling.

Naturally, a complete and final account of modeling will require this issue to be settled, but for us, precisely because we are trying to just simply get into the, an introduction to what counts as a model and why they are important and why we are constructing them, we don't need to talk about these issues at length. Some accounts of models that treat their relationship to the world as determinable simply by knowing the structure of the model and the real world phenomenon being represented. For the purpose of understanding the practice of modeling, this view, however, is too restrictive.

do not have a single automatically determinable relationship to the world. As also, you know, we talked earlier in the session that it's just simply there are fully through and through our mediated axis. Different modelers employing the same model may intend different parts of it to correspond with different parts of a real-world phenomenon. Some modelers may require the model to faithfully represent the causal structure of the relevant phenomenon, as well as make quantitatively accurate predictions.

Others may only require that the model make accurate predictions. Still others may only require predictions in qualitative agreement with the properties of the real world phenomenon. For example, Volterra's model is a good example of these more nuanced properties of the modeler's intention about the real world relationship. Volterra believed that his model captured the essential causal relationships that gave rise to the unusual fishery data following the First World War.

modern ecologists think of Volterra's model as a minimal model, a template for building models of greater complexity, namely with much more details about the dynamics of a population of a prey and a population of a predator in a specific ecology. Thus, if a modern ecologist deploys Volterra's model to the study of a real ecosystem, she does so with a much lower standard of fidelity. Because it's not just an exact representation, because there are so many other complex details

that are going on in this dynamic of not only between the real and the model, but also between those properties of the phenomenon that have been picked up by the model. Now, her use of model is only intended to give a first approximation to the most important dynamics of the systems, abstracting the weight from other kinds of details. Now these important constraints which the modeler includes in according to her intention

with regard to what aspects of the model should pick up what kind of sectors of a real phenomenon are what actually Michael Weisberg called construals. Now construals are what you might call to be interpretations of how the structure of the model is mapped to a phenomenon in the real world, which sectors of it which aspects of it which properties to what extent at which scale so on and

so forth questions Well, not sure if my mic is functioning. It's functioning. Okay, okay. Well, it seems to me many of these definitions use a primitive concept of similarity. Yeah.

and even when they try to define it in terms of other things you still get similarity underneath them so I'm thinking if you if it's possible to define what actually similarity yes I will definitely do it in the next few sessions not not the next session but in it but but for now you can think about similarity as essentially the idea of isomorphism. Yeah, yeah, yeah, this is this is like an intuitive starting notion of similarity. I get that. This is this is fine. As a map to territory kind of stuff. Yeah, yeah. And of course here the idea as Carnap and Otto Nouroth and also some other people have talked about

And I don't think that essentially, we will talk about this, but I think that essentially the concept of similarity should be explicated. It is a vague concept and as such it should be refined. But what counts as the refinement of the concept of similarity? I don't think that Michael Weisberg talks about this coherently. He actually talks about it very informatively, we'll go over it. But the thing is that the concept of similarity is essentially an S-scale sensitive concept. Yes, yes, for sure. In the sense that, like imagine an allegory of a map and territory.

So when we talk about the territory, first of all, the idea of a territory is quite vague. So has the concept of the map. are talking about territory there is no flat universe this territory has multiple multiple levels phenomenon are stacked up in a very kind of convoluted complex way top-down bottom-up mid level so on and so forth and so each of these have different scales and of course the map cannot just

be applied to the entire whole territory as if all these different scales on the territory have have been flattened. In order for map, the concept of map, to make sense, it should be a scale sensitive. When we are dealing with one territory, like for example, this yard out there, the map should be capable of being sensitive to different contexts of what counts as territory at which level. At which level? This is, come back to that engineering example that I have always been mentioning, you know, a beam, a metal beam.

the concept of hardness for metal beam varies across different scale lengths from elasticity of the solid metal beam to the nanometric and crystallographic structure of the metal. So these are different scales of the metal beam. Obviously the concept of hardness should be malleable enough, part of the pun, to be sensitive to this move from one scale of the metal beam, elasticity on the macroscopic level, to elasticity or crystallographic structure or nanometric

or atomic scale of the metal structure. essentially this notion of similarity what you might call to be Jean-Pierre is like a higher order concept and higher order concepts are just philosophical capris they don't have any use in real sciences precisely because science only use explicated concepts a scale sensitive concepts and so as engineering. And yes, that becomes a fundamentally different issue in modeling, which I don't think that Weisberg talks about. Actually, when you read Weisberg's book, there is no mention of the scales, there is no mention of approximation techniques for you to actually

map. For example, you see in engineering, like imagine that you want to map the structure of a metal beam that's your territory at the level of macroscopic observe eye observable elasticity or hardness of the metal beam well the thing is that when an engineer works with a metal beam that's all he gets but we know that the metal beam also has a crystallographic structure nanometric and atomic level structure which fundamentally put new constraints on how we can formulate the concept of

elasticity or hardness there are too many details obviously in order for us for an engineer to use and correctly use this metal beam for example in a bridge where there is always tension you shouldn't just know anything you shouldn't just know about the the macroscopic hardness of the metal beam but you should know some should have some knowledge also of the atomic level or the crystal crystallographic structure of metal imagine you are using a metal which is extremely fragile and it under certain kind of tension it might break

you know so you need to have these kinds of multi-scale knowledge of this metal beam and hence you should create maps for these but the thing is that in real world engineering this requires massive amounts of research massive amounts of basically fiddling with the stuff that engineers don't want to get involved in if that was the case we wouldn't have even making a fucking bridge until now but the thing is that engineers have actually a novel solution for this it's It's called a normalization technique. What is a normalization technique? It is essentially a kind of approximation that maps certain important features of a lower level, more detailed features,

like atomic level of the metal beam to the macroscopic level. and as long as you have implemented this approximation technique even if you look at this metal beam with naked eye with no microscope and no fancy equipment or mathematical equations as long as you have this approximation techniques you know that your elasticity your hardness of the metal is constrained by such and such factors at lower scales of the metal beam. And hence, you know that if I use this metal inside this bridge, as long as these tensions

within the range, nothing will happen to the bridge. this approximation technique is really important we'll talk about it's just essentially a kind of a different kind of modeling okay so we're um seven minutes to the end of the class is that do you want to select people for the presentations next week? Who is going to talk? We always start with democracy. If no one is going to talk,

I will unfortunately have to resort to fascism. I can. What would we be presenting on? So, the idea of construals. Okay. Construals. Michael Weisberg, you can easily locate it in the index where he's talking about construals. So we are talking about construals, these kinds of interpretations about how certain aspects of the structure of the models correspond with certain sectors of a real phenomenon. Yeah. It's page 39 of the book, actually. I got it open here. Yes. Page 39. So about construals. Who is going to talk? I'll do it. Okay, superb. one more one more

don't let me to resort to cruel means Mari you are going to I can't talk next week I have two exhibitions and millions of deadlines and it's really overwhelming but i would happily take the new touch okay yan how about you untenable um is it i mean i'll do it if we need it but i would really be happy getting past this week uh just again for the same reason it's just super busy works okay joven I can't do this week, I'm sorry. Next week or something.

Let's say that this week is when I'm actually giving you some mercy. From that point onward, there wouldn't be any mercy. So, Theo, how about you? I'm going to pass my baton to Mikey, if that's all right. it looks like mikey just gave a talk today you can't just put it again on mikey i i'm not able to next week but okay how about sapide alberto artemis please someone okay andrea okay thank you so much

you see there is a reason that such a thing as academia where a students can just say no to their teacher here we have too much liberal democracy okay don't worry okay that's good let's go oh by the way I'm ready for the punishment I don't know about that decision I made well and that you might give him a day here. Okay, go ahead. No, don't worry. So basically, so, Andrea, the thing is that also...

Barrett, other than you might actually be interested, Michael Weisberg's book has, there is a fantastic review of it on NeuroDham review of books. So you can also read, and there is a section actually on construals and kind of, you know, criticize parts of it and, you know, endorse some other parts of it. You can also tie this into your response. Okay. Okay. can I also like elaborate some questions if I come come up absolutely absolutely absolutely

okay because I had I mean some question that I'm trying to prepare yes yes they're not No, absolutely. They're not structured yet properly. So yeah, I'll probably prepare also some questions. Excellent, excellent. That would be actually even more encouraged. Yes, definitely. OK. Closing remarks? There is no closing remarks. All right, then I'm going to terminate the broadcast.

Thank you.