Simulating the World & Remodeling Philosophy (Session 1)

Hello everyone and thank you for joining Reza Negri-Stani's course Simulating the World and Remodeling Philosophy. This is going to be a 12 session course, is that correct Reza? Yes, 12 sessions. And I think I'm just going to read through the course description here before we start. Long ignored by philosophers of science and epistemologists, only recently the concept of models has come to the forefront of philosophical scrutiny. Nevertheless, since the dawn of discovery and invention, engineers, designers, and craftsmen have in one way or another treated models as the most treasured tools in their armature of equipments and methods. This 12-session seminar engages with models

broadly understood. We will study a wide range of models from diagrams to scientific models and children's toys, only to arrive at their true significance in simulating our world and refabricating it. The first six sessions are centered on the study of models, old and new, in order to grasp what actually counts as a model and learn its range of applications. In the second half of the seminar, we'll investigate a model-centered vision of discovery and invention, where philosophy, art, design, and science not only are integrated, but also help us to rethink and build a different world. The ideal audience for this seminar are people interested in history, philosophy, science, engineering, and design.

All right, I'm gonna pass the mic to Rezanegar Sanyo, and thanks everybody. Thank you very much, Theo, and thank you everyone for participating in this class. So, I mean, as I was just saying in the other channel, that I know many of you, but I also don't know some new participants, which is always good. It diversifies the class. So To that extent, it would be great, even those of you whom I know already, go one by one, introduce yourself,

tell me a little bit about yourself, what you are working currently, and why is that you are taking this course. And then after that, we shall begin in earnest our own course. Adam, would you like to start? You're the first on my list here. Okay, sure. So I have a computer science and finance background. I, in terms of research for some interesting software studies and models and complexity, which is something that Reza has also talked about a lot. It's pretty fundamental to all of that and seems like a pretty rich

area to begin to. That's why I was keen to be honest. Thank you so much, Ada. and the next person I have is Alan if you're ready to go hey my name is Alan I live in Mexico City I am an architect and I'm a certificate student here in the new center this is my second class I also I'm also doing a master in critical teaching what the reason this course seems really interesting to me because well architecture is really the whole architectural practices always in in relation to models, like typical models, digital models, and also plans.

And so it's something that it was never really a topic of discussion in architecture. It was, I mean, models were mainly seen as a function of representation. So I'm really interested in the way that models can be taught to reinvent architectural practice. And I'm also working right now on the subject of smart cities. So I'm actually really interested in finding ways of thinking how the paradigm of smart cities implies like a way of modeling the data and algorithms.

I think this course can help me think through those subjects. Sure, thank you so much. Andrea, if you'd like to go. Yes, hi. Hello everyone. So I'm Andrea and currently I'm in London. I'm studying cultural studies at Goldsmiths. Smith but in February I graduated philosophy in Italy and I would say that my current research was which started from my dissertation basically involves the question concerning technology in in some Edgarian sense.

And I'm trying to intersect this with some researches of Stigler, Simondon, Yukui. And meanwhile, I'm currently studying more deeply all the new movements of philosophy, like from speculative realism to non-philosophy in order to kind of build a sort of philosophical critique to artificial intelligence which is i mean where is the the the the field where i'm most interested currently and uh and i will try to connect this uh the critique sort of critique of machine learning on uh out to kind of on how

this this particular technology produce some cultural knowledge in order to yeah to to to intersect it with uh with my current uh institutional studies and so yeah i i am i are really um um really uh looking forward to this course because i i also would love to enhance my knowledge on the epistemological consequences of models and of other ingredients of science and see how them then can provide a new perspective of the world. Excellent, thank you so much. Very fantastic, excellent.

Artemis, would you like to go? Hello, my name is Artemis. I'm also an architect as Alan and I come from Greece. It's been almost a month and a half that I just arrived from New York where I was working and experiencing the metropolis as an architect. I'm trying now to find a job and move to France and I'm interested in philosophy because also in our studies in a really experimental university of Thessaly, this is in Greece, it was always promoting to have it as a base, to reading. Of course, I don't have a big background in it, but as much as I could, I was always wondering and reading.

and that is the reason that I needed a stimulation in that field and that's why I applied in this program and this particular course for the same reason as Alan is really interesting for me I mean we're always around models either digital or physical models and we were always wondering if they are only kind of process in our design stage or they have sometimes some kind of autonomy so yeah I'm really excited to starting this thank you very much thank you thank you you bet that's I'm here to in give you my

enthusiasm. Yeah, Ian Foxworth from Washington, D.C. I'm an attorney. You know, there's no great story here. There's no, like, I'm not coming with this sort of, this philosophical project that I'm trying to finish. You know, I was a philosophy undergrad while working as an attorney, I picked up a master's in philosophy and I'm just sort of, after all that and while also working, it's more that I'm still looking for a place to start. Then I have some particular project and I've heard great things about Reza, of course. Most intriguing.

Thank you so much. Really, actually, it's fantastic that we all have such a very, very diverse, you know, participant and audience. I've never had an attorney in my life, but hopefully you won't be the first, you won't be the last one. Well, I would congratulate you on that. Thank you so much. I totally appreciate it. J.P., would you like to go next? JP, sorry. Sorry, it's me, JP. Okay, all right. I didn't hear you clearly. So, well, I have a background in philosophy and music, working on a project about form

and scale. So I think towards the problem of scale, the theme of this course, it seemed to me fitting to see if it goes somewhere I couldn't go, I mean, in my individual studies. And besides that, I'm very appreciative of Reza's work, so that's why I'm here. Well, that's it, I think. Ah, I live in Rio de Janeiro, Brazil, so that's it. Fantastic, Jean-Pierre. Thank you so much. Jovan, would you like to go? Yeah, hi. I work on non-centered philosophy, philosophy of science, interested in philosophy

of science in an analytic and continental tradition, and I like Rosa's work. So, hi. Super, thank you. Thank you, Johan. And Justin. hopefully we have sound hey what's going on um i just i just got in late actually there was something weird with the links i was in like another side chat room and there was actually another student still over there so the haunted yes i'll do that one thank you for letting me know um it was lenka or something like that um i could probably sign into that other one but i have to sign out of this to go back over there no no it's okay we need to connect the wounded

and break them back to camp pretty interesting um i don't have much to say i guess i'm just kind of a renegade um lost thinker just picking through things and also inspired by resident glad to be here and learning thank you so much justin thank you very much And Mari, would you like to introduce yourself? Okay, I'm Mari Bostashevsky. I guess I'm trans everything in the finest sense of this world because I'm now in Lausanne but I'm in no way Swiss.

And I don't fit here at all and I don't fit in any disciplines straight forward either. I worked for many years as a visual artist on trying to parse conflict economies and the way in the identitarian fictions fit into that and I kind of use photography as an excuse to have philosophical performance where I would enter places like State Department or Foxconn factory or the port of Jebel Ali. I've been to many fascinating places and have very direct conversations and also expose myself and then turn this into an artwork. So I don't do that anymore. I'm working on two, working on it's a big work. I'm looking at two projects. One is about transmematic opportunities for environmental survival and seeing ourselves more as both

animals and machines. And the other is close to what Andrea was explaining also about machine intelligence and algorithmic philosophy. I think the reason for me to join this class is triple. First of all, I'm hopelessly depressed and I think everything is completely meaningless on personal and planetary way and it feels like a good place to be in that state without being told that everything will work out if you send another email or if you work just a little bit harder and I think second is art is always for me a form of a model already and third more importantly I like the way Reset thinks and I think it's rare to find sort of no bullshit philosophy and way of being and intuitively I just feel

like it's a it's a good place for me to be in and that's not nothing. Thank you so much Murray very very great really absolutely thank you so much Mikey hey I'm Mikey from Portland Oregon this is a Sanka can't do that in the university but yeah so I kind of came to like philosophy like when I was a teenage anarchist back in the day and discovered like lots of work from there and it kind of evolved. And it's mostly been like a personal like passion of mine in the background. My formal training is actually in culinary arts where I worked in like high end fine dining

for about 10 years. And now I am a coffee educator. So I just teach people about coffee and senses and how to be a barista. So recently I started, or about a year ago, I started this project called Aleatory Books, focusing on critical philosophy and theory. And it's just kind of a way for me to branch outside of my own little inner world and connect with more people and which led me to being at the new center so I'm a certificate student here and I've been following Reza's work for a long time and this class in particular is really interesting just because the ideas of models and creating new ways to frame thought and you know to communicate them in between people uh especially in interdisciplinary work and also

um to form new ideas i think there's a lot of potential and i'm very interested to uh explore those thank you thank you monique yeah and sean hello uh my name is sean matharu i uh let's see i'm not sure what the question was so i'm just going to introduce myself. I'm a PhD student at University of California, Riverside. I study French, Francophone, English, and Arabic philosophy and fiction, although the philosophy part has been something that I've had to kind of bracket. I mean, comparative literature departments are definitely amenable to the study of philosophy, but I feel like it's something I've had to kind

of engineer on the side. And it's gotten to a point where I really want to do a project. I'm writing my dissertation right now, and I'm super interested in Reza's particular approach to it. Yeah, so I'm excited to be here. I study, let's see, fiction. I'm interested particularly in fiction where consciousness and landscape are kind of interwoven, so a very like very influenced by Ballard, weird fiction, and surrealism, modernism, a bunch of things. And I'm interested in kind of understanding how the study of fiction might produce models that we can apprehend outside of phenomenological subject positions. So particularly, I'm interested

in the way that environmental and racial violence are kind of coded in literature, and I'm interested in trying to present that objectively. And so I'm interested in the study of models, and I'm excited to be here, and it's nice to meet everyone. Absolutely. Thank you so much. Fantastic background. Really, really thank you. And Lenka, we're just doing introductions. I know you just came in. Would you like to introduce yourself? Can you hear us at all? Okay, I think your mic isn't working, but I'm just gonna...

Is it working now? Yeah, I think so. Oh wait, no, that's Mark. Okay, great. No. Oh, it is you. I don't know what is it. You're just a little delayed by 20 milliseconds. Wait, so you can't hear me or can you now? We can hear you. We can hear you loud and clear. Okay, great. I just, yeah, it's confusing. I just missed my flight and I'm kind of, yeah. But I, yes, I did my bachelor studies in Prague and it was in Department of Aesthetics. And then my MA was in London

and I wrote on DELAS and Adorno. And now I want to apply for PhD and I want to continue to work on DALES and on pluralization of time in DALES that I would like to somehow suspend the dessert synthesis and decentralize it. And yeah, I know I tend to think in terms of models, but I didn't reflect on it properly. so that's why it feels like something I should do.

Thank you, thank you very much. Really appreciate it. And my name is Theodore, just in case people haven't met me before. I'm going to be assisting moderating the course, so feel free to contact me if you have any technical problems. and I've been studying with Reza for a little bit now and I'm interested in how we ground our model making practices. Yes, thank you Theo, thank you very much. Okay, thank you everyone, really really glad to have you here today. Okay, let's not waste any more time, unpleasant reads, and just get into our own discussion.

So the first thing is that how does this course evolve over, you know, 12 sessions? Well, CEO already gave a kind of abstract of the entire course. But in addition to that, I would say that today we are not going to go into any kind of technical details. It's just way too early to do that. We won't go into very technical details, but hopefully in future sessions. Today we are just simply going to talk about models, but not just any kind of model.

We are actually going to talk about scientific models, models as such. You should understand that, and I'm sure that you already know, that the concept of model is usually used in a very common sense, loose sense. such that model when we are using the word model we can mean a lot of different things some of them might actually be contradicting other connotations of the word model for example we just think of models as these approximations of reality out there sometimes we think that models are forms

of representation sometimes we think that models are simplifications of the items in the world nevertheless there is something in common among all such loose talks about model models are supposed to give us some insight we do not yet know what that insight is about the world of which we are part and the very distinction between a scientific model and the loose talk of a model is that such criteria as how a model gives us

some insight about the items on the world by which methods it can gives us some insight about the world are made explicit they are tractable now they can be replicated it can be investigated further they can be recalibrated in fact so as to yield more refined model whereas in the loop stock of the model We cannot do this kinds of stuff systematically. This doesn't mean that we are not going to talk about loose models. Like for example graphs, diagrams, gestures, so on and so forth.

But in order for us to actually grasp the significance of models, particularly in the realm of sciences engineering so on and so forth we first need to begin with a very strict non-lavish parsimonious description of models instead of beginning with this kind of forest where we have different concepts of models we begin with a territory in which the forest has been cut back the metaphysical forest has been cut back and we have tractable assessable criteria of epistemological assessment

peculiar to science So, this is how we are going to start and of course we will expand, we will make examples. I don't think that we should make any technical example at this point. Technical examples, I would say that if we bring them up at this point, they might actually create more confusion. Nevertheless, please do ask questions if something that I say might strike you as vague. Aside from that, feel free to question any moment that you want.

I just don't want you to listen to me as if basically everything is fine. Anything, any word, even a word that appears vague to you, interrupt me, ask me a question. I really don't mind about that. Well you can ask questions either on the sidebar and hopefully Theo will monitor it or you can just simply unmute your microphone and ask me a question anytime you want and for those of you who are new we'll usually have a break at some point like 30 minutes from now at the middle of our class and then we come back it's like

usually a five minutes break so we can have a cigarette of course I'm having my cigarette here now, but nevertheless having some water, so on and so forth. And then we come back. Usually I teach and go overboard with courses. This time I actually want to stay within the time frame so you don't get tired. Because unfortunately I understand that these kinds of materials can actually bore even the most staunch, enthusiastic students of philosophy. We don't want you to get bored and that's really important for you to ask questions.

Heckle me whenever you want. So I think the best way to start our discussion of the scientific models and what actually models are starting from Michael Weisberg, a new book, Simulation and Similarity. It is a splendidly lucid and rigorous work. It is what you might call to be the distillation of many, many studies that has been done in the history of science with regard to models. but it also actually gives something more precisely because from a historical standpoint in the philosophy of science and epistemology, models, not all models, first of all not all

models were considered to be models, but also what used to be called model was deemed as some sort of ancillary or auxiliary theme for philosophy of science it was not what you might call to be a central point of philosophy of science the more scientists and engineers have looked into what models are scientific models are the more they have realized that models do in fact enjoy a central point in their study real phenomena and also the epistemological methods of modern

sciences in that regard you might say that the serious study of models began from something like early 70s by the work of Supes the semantic model or the semantic view of models but even then their work were fundamentally rudimentary that the true genuine a study of models began in tandem with the study of complex systems something around I would say meet 80s and then it got refined

over time and now in 21st century finally we have the first you know systematic works about what models are what the procedure of modeling entails so on and so forth so let me read a little bit of Michael Weisberg And do you have the PDF of it? Yes, okay great. So let's just go a little bit, I'm going to read this because to be honest with you, I couldn't come up with a better, more lucid exemplification for what can be considered as a scientist.

model so page one two aquatic puzzles San Francisco drinking water travel travels 167 miles from Yosemite National Parks Hetch Hetchy reservoir how do you pronounce it is a polgas polgas wire temple in San Francisco the temple marks the end of the water journey with the biblical promise I give waters in the wilderness and rivers in the desert to give drink to my people. These words were inscribed to remind the residents of the Bay Area of the fragility of their water supply. This fragility weighed heavily

on the minds of many San Francisco's, but on no one more than John Rebber, an amateur musical theater producer whose business cards admonished, without water there is no life. cut a story short rubber began to think about a very very ambitious project to create two dams such that these two dams or regulate the water flow and in that

sense San Francisco has what you might call to be an infinite reservoir of water for the years to come. Now everyone around that time backed him up, politicians, scientists, many many engineers so on and so forth. However he had also an opponent and this opponent brought up a very, you know, what you might call to be a strong debate against Reber, reminding him or basically telling him that this is all good on paper but if you actually

implement such a plan, it fundamentally leads to a disaster for aquatic life to the point that it creates a natural disaster. People told him that you are vague, you are reactionary, you are just trying to undermine a great mind, you know, everyone is, you know, who are you to, you know, make such a statement. Nevertheless, he went on with the back of a few supporters to create one of the most majestic models in the history of engineering, San Francisco Delta Bay System model.

thing is that they started to measure of course you know we are when we are talking about measurement we are in the business of approximation already the tidal waves the kind of torrential flows in the streams of water the salinity level of water in San Francisco Bay area. You know the elevations, the topographic elevations of the what you might call to be the bed, basically everything had some sort of data

about it gathered. So they started to create this kind of massive, like imagine like a big warehouse. It's not just like these kind of small architectural planes that we make or what engineers make these kind of a small toy things. No, it was a massive, it was a massive model, the size of a whole compound where they started to simulate the torrential flow by way of these water pumps they had already created uh at various scales relevant to the

data they had the you know the elevation of the bed the salinity of water so on so forth and then they started to look at the behavior of the system over this time over a period of time unfortunately rubber died before the data came in the data showed that according such approximates given such a model which tries to be infidelity in in completely you know fiddle representationally

dynamically the target system at hand namely Bay Area Delta system if these two dams are being built it would lead to an absolute disaster now you see here a lesson to be learned sometimes it is actually wrong to deal with the real system and come up with a solution from what we take to be a real system sometimes we have to make a model and only through their study of model indirectly we can shed some

fundamentals insights about the real phenomenon that we took for granted and because of we were taking it for granted we came up with you know all sorts of whimsical solutions this is a very you know famous engineering example which all of course Mike Michael Weisberg also cites it question before I move to the second example anything anyone please if you don't ask me questions I will force you to ask me questions

and you don't want me to force you go on go on go on Maria I have a question but I'm afraid it's too early, if haven't read the book, I kept thinking about Flint and I think in terms of targets it's a very fascinating case because it seems like a problem of water and you can build models around scientific and engineering issues which is what is continuing to happen in Flint but then we need a very different kind of model to actually understand why it didn't work. Yes. And my question is, I guess, because I want to have a question here is, is it at all possible to have this analogy here or it's completely inappropriate and...

No, absolutely. Yes, absolutely. Absolutely. Is this a working analogy because I keep coming back to that case. Yes, absolutely. But I would say that the case of Flint, I mean, and my apologies, I mean, I know just basically the the what you might call to be the pseudo facts that have been circling on internet in the news media and some papers i don't know any details about it whereas i do actually in fact know the details about the san francisco bay uh delta system uh modeling the thing i would say with regard to flint we just don't need uh something like what you might call to be a model of a natural phenomenon but also we need something like a model of the social distribution

system and you know of course I will I will come back to this the third model that I'm going to talk about which is of course it's a very classical model Thomas Schelling which of you know Thomas Shalin by the way. Just from this book. Yes and do you know who he was? He was an advisor, security advisor to many US presidents. He's essentially the man who coined the word MAD, mutually assured destruction. He was the ultimate game theoretic theorist who came up with a game theoretical framework upon which

the the competition or rivalry in the Cold War can be sufficiently understood and of course he he actually became famous because of developing a model called the segregation model it's essentially a distribution model of population that leads to segregation regardless of any belief so on so forth it's just what you might call to be a game theoretic population model so I would say that a Flint case requires bunch of models integrated precisely because it is not just what you might call to be a natural disaster it is actually the consequences of a series of phenomena at different

scales social economical so so on so forth so if you don't have any question Let me just get some of my stuff here, notes. So another classical example is Vito Volterra.

He was a senator in Italy and also a very, very famous mathematician and a statistician. The thing is that after World War I, fishermen in the Adriatic Sea, they noticed that they are not getting a good yield anymore for some strange reason. Nothing drastic has happened. know there was no pesticide there was no oil leak so on so forth it's just that

they fish area yield was drastically suboptimal so they started to approach a person who then began to go to these fisher markets you know keeping the tally of the kind of products they are getting every day from any you know boats markets on so forth and he also actually realized that it is not just that they believe this is the case it is actually the case in fact like you know stingrays sharks these big fishes are actually come up more on the market in

the in the nets of the fishermen then you know cod you know salmon so on and so forth like literally these small fishes they don't have them anymore this person went to his I think father-in-law if I'm not mistaken who was Vito Volterra Vito Volterra also did the same kind of exercise he began to roam around fish there is the boats and get keep the tally and then unlike the first modeler he didn't actually create what you might call to be a concrete model in

order to solve the problem but he started to use pencil and paper turning the statistical evidences that he had garnered from his study of the fisheries into some simple equations calculus in fact not any kind of equations but you know differential calculus equations one of these equations stands for one population of prey and one of and the other equation stands for a population

of predators is what you might call to be locked look through a you know Walter Walter's equations of prey and predatory population once he developed these equations he noticed that given the differential calculus implications of these equations in any society where you have at least one population of prey and one population of predator adverse effects if they are mild like

for example just let think of oil leak for example in Adriatic Sea or for example the abundance of pesticides or biocides if they are mild they actually lead they are more in favor of predators and have adverse effect on the population of price but if there is an intense level of biocides, pesticides or any kind of adverse factor within this dynamic it

would be in the favor of prey. So then you can easily guess what his solution was to this problem. In order for you to get more fishes, cods, whatever you think rather than just these big predators, you have to actually increase the intensity of fishing. Many people then actually noticed it was completely right. The reason that these fishermen couldn't anymore get at an optimal level the fish that they could get before the World War was because World War I had already forced them to quit

fishery. Essentially fishery during in that region had become minimized throughout the years of World War I and hence the predator's population had increased unproportionately And the only way that they could balance this is by intense fishing to create a more balanced population dynamics such that they can have more fish. Number two. Any question at this point?

Everyone is so shy, always on the first session. I just don't like this. Jean-Pierre, you are laughing. It's your turn to say something. Actually, what is going on? Is this sound okay? Yeah, okay. Not really, actually. I was just thinking that perhaps my questions would be too soon as well, because one thing is to build a model which, in some sense,

replicates the physical conditions of a phenomenon at another scale, which is the first example. And another thing is to mathematize the conditions of the phenomenon itself, which it seems to me to be the second example. So I'm sure you will get to this eventually. I'm not sure if I'm if it's, I mean, an appropriate question for now. No, that is good. That is absolutely good. The thing is that I will actually talk about this that not all models are mathematical and This is of course is the topic of our today's session the classification of models scientific models

We have concrete models at the very least concrete models mathematical models and computational models concrete model example was the Bay Area now the mathematical model is is Volterra's you know prey dynamics and the computational model after the break the cigarette break as if I hadn't have enough cigarettes would be Schelling segregation model and of course yes each of these models are beholden to the metatheoretical and

theoretical implicit commitments or the framework in which they have been made. But this is not just exclusive to the idea of modeling. The same hidden assumptions about theory and metatheoretical assumptions plague as much the science of modeling as they do the method of theorization or direct representation. Which of course I will also make example of it by way of Darwin and Mendeleev's table of elements.

Now, shall we have a five-minute break and then come back? Sure, that sounds good. Excellent. See you in five minutes. Thank you.

Thank you. Thank you.

Thank you. Thank you.

Any questions before we begin? Are you going to talk about the fiction modeling stuff, Rozo? Yes, but not this session. Basically, precisely because as I mentioned, the word modeling is being used in a very quite loose sense, it is important for us to start to actually look at what science of modeling is today, what it is about. And then surely we can branch to fictional models, you know, diagrams, graphs, so on and so forth. Yes, yes, and we will do that, yes. And of

course, a very special class of models, which I would say that only in the last decade or so has been brought up and it's fundamental for the study of not real phenomena but the study of models themselves which are called toy models and we will also look at those what toy models are small toy models and big toy models fundamentally different kinds.

Jan, do you have any feedback, input, anything? at this point no okay don't feel free though heckle me whenever I say something is stupid oh no no so so far so good I will save up my heckles for you know the big things yes

you See what about you no questions yet. I'm just listening Artemis anything Well, I have many questions, but it's actually around architecture models. Like, we were always struggling on abstraction of how we represent the existing and how we

represent what we propose. And we were always fighting all those kitschy representations of the real. Yes. why you put the trees you don't need it but it's totally different I mean the example you gave with scientific experimentation when you need actually some data that they need to be almost the same let's say as a real so yeah I'm just think that we are going to talk about you see I would say to a great extent architectural

models you know when I was younger many many many many years ago I used to work in an architecture firm and the thing is that with architectural models they are not exactly scientific models there are models nevertheless in the sense that they try to abstract away from a real phenomenon okay but not all forms of abstraction of reality or real phenomenon or approximations of a real phenomenon can be called a model for something to be called a model should have first

an interpreted structure now this interpreted structure is a fundamentally a theoretical business as we will talk about and at the very least at the very least within the theoretical framework upon which this structure has been installed there should be also a model description usually this model in script description is laid out in terms of simple equations or algorithms whether it's mathematical or computational or or an ancillary model a smaller model which captures both the theory and the

framework of data under study. Now this latter one is not called either mathematical model for a computational model. This is usually peculiar and exclusive to what is called in science a concrete model. Like Drosophila, the fruit fly, is in fact a model organism. scientist but also a lamprey do any of you know what a lamprey is the delicacy usually Roman emperors used to get rid of their rivals by shoving them into a

pool of lampreys so they can basically they're worse than sharks they just like there are these like jets, massive amounts of teeth that they can just simply make dig through your flesh. But the thing is that with lampreys they are considered to from a biological perspective on evolutionary biological perspective they are considered to be representing one of the oldest instance of neurological development like the rosophila that gives us information across the

board about the genome the genetic regulation of many many species across the board lampers also give us information general information about neural relation or simply the evolution of the nervous system across the board of all organisms now these these kinds of models cannot are not really mathematical or computational nevertheless they are models precisely because they shed light in about you know the kind of phenomena that we are studying the genome the nervous system so on and so forth and the thing is that the description of such

models as you see comes down to a different a smaller model rather than some form of equation or an algorithm Shall we start to you? Please do. Okay. So, the third example.

And let me just read line by line Michael Weisberg about Schelling's model of segregation. Page 13, computational model, Schelling's segregation model. Thomas Schelling's Micro Motives and Micro Behavior, 1978, contains a very simple model that shows how racial segregation can occur even when no racism is present.

The model is agent-based, meaning that each individual is explicitly represented, usually by a dot. You can in fact go and from a statistical perspective think of it as a coin having two sides or dice. In his particular implementation of the model, dimes and nickels represented two types of individuals A and B and the squares on a chessboard represented a spatial location. In this model, each individual prefers that at least 30% of its neighbors be of the same

type. 30% of its neighbor should be of its, either if it is A, 30% should be A, if it is B, at the very least 30% should be also B. This is what you might call to be satisfaction criteria. If it is not, which is from a game theoretic perspective, this is indexed by a function called a utility function, reward and punishment. If the criteria is satisfied, A remains where it is. If it is not, it moves away. It punishes the neighborhood.

Each individual prefers that at least 30% of its neighbors be of the same type, so the A's want at least 30% of their neighbors to be A's, and likewise for the B's. Schelling neighborhoods were defined as a standard Moore's neighborhood, a set of nine adjacent grid elements. An agent standing on some grid element E can have anywhere from zero to eight neighbors in the adjoining elements. Although Schelling didn't explicitly provide a utility function, utility function broadly can be understood in a game-theoretic way

as what you might call to be reward and punishment function. The preference described above is usually interpreted to mean that each agent is indifferent between having 30 to 100 percent of its neighbors be alike. What will be dissatisfied with fewer than 30% like. Because of the constraints of the grid's geometry, in case of a full neighborhood, the preference boils down to wanting to have at least three of one eight

neighbors to be alike and to equally prefer three to eight. The dynamics of Schelling's model involve agents sequentially choosing to remain in place or move to a new location when it is an agent's turn to make a decision based on the utility function and the utility function is that if it is less than 30 percent I'm going to move out of this thing regardless of whether I'm a racist or not this is a game theoretically we are not talking about ideological views here. If this condition is met, the agent is satisfied and remains where it is. If it is not, the agent then moves to the nearest empty location. This sequence of decisions

continue until all of the agents are happy where they are and do not try to move. Now, the dynamics of the model unfold as a cascade and are depicted in the figure on the page 14. Agents that were originally satisfied can become dissatisfied as soon as a neighbor leaves or a new one moves in, and this leads to a new agent becoming dissatisfied. A small patch of dissatisfaction will result in widespread movement and ultimately segregation. Though there are a few grid configurations that are integrated where every agent is happy, these These are quite rare and nearly impossible to resolve from agent movement.

The equilibrium state of the model is segregation. Is segregation. Thus Schelling's major result is that small preferences for similarity can lead to massive segregation. This result is quite robust across many changes to the model, including different utility functions, different rules for updating, different neighborhood sizes, and different spatial configurations. In fact, it is extremely hard to avoid segregation when agents have some preference for like neighborhoods. Now the thing however, so first we have the Bay Area Delta System model, concrete model,

then we have Volterra's prey and predator equations which was a mathematical model. The segregation model differs from the previous two. In so far as this model and the prediction of equilibrium state, namely segregation, can only be done for any target system by way of an algorithm, an algorithm, a computational program. Hence, we see that we are in the business of three main classes of models, concrete,

mathematical and computation. Questions? Yeah, I have a question about the computational model. So, in this situation, is the algorithm – could it be understood as a kind of mediator between the model itself? So in this case, the, I suppose, tendency towards segregation or rather equilibrium as segregation would be mediated by an algorithm.

So I'm kind of wondering what is on the other end of that. So would it be in this situation, the particular persons, for example, that are so are they like affected by the algorithm or is it the algorithm that is mediating between the individuals and the equilibrium understood as segregation, perhaps as model? You see, the thing about neither mathematical model nor the computational model, there is what you might call to be no, also even in fact concrete model. And this is the whole point of models. Models are indirect representations in the sense that basically according to some metatheoretical assumptions,

we already have presupposed even though implicitly that model is in a way isomorphic to the real phenomenon according to some theoretical assumptions now so we absolutely don't even talk about the real population okay that's just out of question we just talked about the behavior of elements of the model itself. For example, here in the segregation we are simply are interested in the computational, recursive computational elements which undergird the behavior of agents A's and B's, dimes and nickels, over this grid

platform to see that given for example such and such constraints how under what time frame under what kind of behaviors the state of equilibrium is reached namely segregation given these doesn't so initial and boundary conditions now of course this doesn't mean that the model is in the air is a kind of an idealist kind of you know stuff but the modeler first wanted to actually look at the behavior of the model itself before even submitting it to reality at the end of the day however all of these models should have their isomorphie or one-to-

oneness with reality be investigated by extending the model into the reality into a real phenomena and see how it actually what kind of behaviors that are going on in the real phenomena it distorts which ones it with you know a relative accuracy represents and which ones it ignores this is what actually a model is anything other than that would count as a direct abstract representation so essentially when we are trying to

to theorize about the world from a scientific perspective, we have a direct approach and an indirect approach. Models are always an indirect approach, precisely because they are mediated encounters with reality. Surely, the very fact that we create a model requires some sort of traction on the real phenomenon according to our theoretical commitments. But once we make the model, once we have the model, then we are no longer looking at reality.

We are analyzing the model itself. In fact, I mean from a philosophical perspective and maybe I am mistaken but this is a good challenge for you, those of you who are so into Kant and this kinds of stuff, I know some of you, you can think about that whether the statements that we derive from models when we are analyzing the model itself regardless of the reality are analytical or synthetic please do mari

I don't want to sound stupid and I don't know if it's beside the point or maybe too soon, but for me this model fell apart because it seems the premise of likeness is impossible to establish because I mean what about synchronicity of the asynchronous? Someone thinks I'm alike but I'm nothing like alike and then who's like the whole notion of likeness and I don't mean just in terms of ideology but I mean where is the science and fiction and fiction here because we can't really if if to idealize the way we could say why but if we have to zoom in again to face reality or very quickly realize that look how there's no one alike or everyone is alike and very premise of

the model is falling apart for me there right right no this is absolutely a genuine question and something that we will talk about at length. The principles which undergird any kind of scientific modeling are based on two fundamental criteria, simulation and similarity. Which means that sure, the thing is that once we analyze the model we shouldn't actually look at reality. But the application, you see there is two different things. We are dealing with two different phases and these two different phases should be included in the in the in the science of modeling one is what you might call to be the construction and analysis of the model

the construction of an analysis so phase one construction it requires some sort of isomorphy with a target system namely a real phenomenon Then, the analysis of the model. At this point, we just really do not want to look at the real phenomenon because it actually gives us too much complexity, too much unnecessary information. Now, a stage three. A stage three is what you might call to be the application to the target system, the real phenomena at this point is what you might call to be the optimization occurs sure we have made some really fictional you know ideal model that

doesn't have any roots apparently doesn't have any roots in reality but the thing is that as long as we have some sort of what you might call to be a semblance of isomorphic correspondence, we would be able to not only enrich our model by new details garnered from the data we have got from the real environment, but also revised portions of our model. So it is not just that model is just like this free floating thing. comes in stages each stage requires its own procedures

so i i kind of thought um maybe just some clarification i'm trying to like frame this uh and i'm mostly thinking about kind of uh analyzing um what where when you start deal the model you're actually looking at the model and then looking at the concepts that went into going to build the model um so it's just like stepping like backwards particularly i'm thinking about the predictive policing algorithms um in the sense where it's you know they take data from uh you know different neighborhoods where it's most likely that a crime will occur and they send out basically higher patrols or setting up, you know, police always kind of like stay in that block. And, you know, we'll give numbers of arrests that are expected there. And

in like practical application, it becomes, you know, it kind of validates itself because they can reinforce that information through kind of enacting it in the sense that the algorithm creates this like feedback loop of itself but it's all based on those original base concepts that went into it that were kind of based on a flawed system to begin with um so yeah i think it's like there's taking those steps back like you know those movements where it's like even when you can apply it you find um it working in environment but the flaws beyond the model okay if I am please please please

Mikey tell me I'm wrong at or if I'm just representing your algorithm essentially you are saying that if I'm not mistaken that once we make a model this kind of idealized approximation so we have some sort of algorithm or mathematical equations and these mathematical equations once taken as way of how we understand the real phenomena precisely because they are already idealized they can create distortions down the line so such that the dynamic interaction between the model and the real phenomena within this framework

create more distorted data you get a stop from an idealized model you transport them back into the real world and then those data are already beginning distorted even more and then back and forth and back and forth and just simply you are creating an environment that wouldn't be actually shouldn't be called by any means a scientific epistemic procedure it is as if like a political manhandling or gerrymandering of the model. Yeah, it's always like curious where, and you see this a lot, obviously in this kind of climate where people use these models

or use these statistics to build arguments and where the actual, yeah the actual like kind of science is completely lost because they have data it's not even in politics I mean you can see a lot of this kind of stuff in evolutionary biology yeah now the thing is that well this is something that we are going to address under the question of model pluralism that surely every singular model is going to distort. In fact, all models distort. But the thing is that we just cannot go on and try to predict a certain kind of behavior, describe or explain

certain kind of behavior, certain kind of phenomena without models. This is necessary. This is necessary. Now, there should also be supplemented with what you might call to be strategies of model distortion mitigation where we mitigate the distorting effect or effects of models to do that of course we can refine our models calibrate them ever more but that has its own limits there is another way to go on about it by supplementing models with ancillary models models that actually regulate

and monitor how the our previous model coordinates with the real phenomenon this is what i would call a big toy model and we will get back to this stuff go on yeah There we go. Now, I was just thinking, I mean, that if you have models, you know, algorithms mixed with other algorithms, you know, managed by further algorithms, at what point is there a difference between like scientific theory and modeling? Does that make any sense?

Yes, okay. Scientific theories and modelings are not mutually exclusive. Science models, but not all scientific theories rely on models. You see, here there is a fundamental distinction between what you might call theorization about the world and modeling the world. Theorization about the world ultimately only give you generalizations, not particularities. you have to zoom in on a specific details of a very particular phenomenon precisely

because that phenomenon might actually have different scales some of those scales are not covered by your scientific theory which you use to represent it and hence you have to kind of a little bit of optimization here course any distortion and and it's not just about modeling theories also distort reality you see scientific theories also distort can can potentially distort but the thing is that what is important here is some sort of license or guarantee a kind of regulatory procedure that can actually track this

kind of back-and-forth coordination between this theory and the real phenomena or model for a specific phenomenon without that we are not in the business of science at all of course science can go we both know it I mean And look at today's contemporary neoliberal science in America. The scientists are just like, they are the worst theologians of the medieval ages. This is why that these things require epistemic investigation. Attorneys who actually look into the epistemological procedures involved.

these procedures. Well we cannot just simply say that we cannot trust them on some you know kind of psychological assumption. We absolutely need to know what is exactly the fundamental such procedures in order for in fact to know how we can formulate the right kinds of questions. Would you mind if you are clarifying exactly what the distinction is between the algorithmic computational model and the mathematical model?

Well, you see, at this point I would say that this distinction is no longer a hard distinction. It's actually a soft distinction and will eventually evaporate. thing is that from a classical perspective usually the mathematical model is what you might call to be non-recursive from a computational sample whereas algorithmic ones are recursive there they can be arithmetically encoded as simple as that of course this you see all of these stuff we will talk about this all of these all of these classifications are predicated on the

fact that we have an abundance of method theoretical hidden assumptions here is mathematics difference from computer science can geometry be said to give us elements or properties of abstract entities which cannot be encoded by arithmetic so on so forth you see this is the moment that we understand that even these such classifications are prone to negotiation or renegotiation precisely

because they are predicated, they are undergirded by a certain kinds of assumptions in the domain of mathematics, computational theory, so on and so forth. Another example is Carnap, Carnap logical structure of the world. So, his models are essentially logical. Precisely, why, I mean, not the late Karnapp, I'm talking about the Karnapp who wrote the logical structure of Wohl, of Bauer.

Now here, the method theoretical assumption that Karnapp later actually admit that was behind his initial assumption was that All mathematical statement can be reduced to formal logic of its time. But this is unwarranted. This is unwarranted. In the background of, I don't remember who asked the question or made the statement about the sort of predictive policing model and the way that moves from like descriptive to prescriptive activity or like starts to distort the data but it actually in an interesting way

and when you mentioned Carnap it like it points like the way I'm reading how you're you're sort of building this class and relating it to models is very much like to me parallel to the move from early to late Carnap. And there's a lot of deep assumptions in there about reality. And I guess like the nature of reality as like stabilized or the relationship of like thought to reality. And like the way that to some degree models are this like linguistic tool by which thought sort of prescribes and changes reality. So it's not necessarily to assume like we can always reduce bias to nothing but actually it becomes refining what biases we want to amplify in the system like maybe there is no pure descriptive models like they always have this prescriptive

side that that builds to where yes yes no really fantastic fantastic fantastic uh uh insight justin i i was gonna ask you if you would talk about like the assumptions in there about reality and maybe that's too soon or whatever but... It's probably too soon but let me just give you a very... you actually said something that I have completely forgot. So we say that okay models distort reality but doesn't this already assume that you have some sort of hidden assumptions about what reality is? is. Selar's myth of the given rears its head here. You see, when we are in the business

of science, we really don't take for granted what reality is. Reality is not a tale-tale heart that tells it tells what it is to us by way of some sort of private access it is a matter of engagement of a procedural and tractable engagement with phenomenal appearances such that we can refine this engagement over time. We can select what kind of biases we want to choose, what biases we want to mitigate, so on and so forth. But yes, absolutely. I would say that

and to be honest with you, I just don't want to talk about this. I mean, you have heard me, but for our new friends, is that this is really the kernel of what I would call Orthodox Marxism today. Their enmity toward the scientific method. And not all Marxists, I'm a Marxist. I'm a proud, you know, I want to have the torch of scientific rationality. But the thing is that, okay, so my Marxist friends, you say that these models can fundamentally give us these bad distortions of reality.

Sure they can, of course, but the whole point is that the very fact here is that we have a better grasp of what is exactly being distorted and what it distorts it precisely because such procedures are tractable, are more or less transparent semantically. Whereas if you say that as if we had some sort of real traction under real forces, you're in the business of mysticism, not science or philosophy. You already have some hidden

assumptions in your closet about what reality is and then you are essentially in the business of dogmatism. Can I, so the statements that we derive from models in the sense they would neither be analytic nor synthetic, right? Well, that was your challenge, that's not my challenge, that was your challenge that I wanted to answer it. Go on, Joven. Yeah, I mean, I mean if it's analytic or synthetic then there's an immediate relation into the real, at least an idealized relationship towards which we can,

I don't know, think from there. And so if it's neither analytic nor synthetic, and I don't want to go to the whole Koinian thing, but yeah, I don't know. Why don't you go? You should go for every one of us. no i just like the the you have been reading some wimsett lately and i like a lot of what wimsett is saying but um i just still that i just i still think that the the the quenyan movement is always dependent on a kind of

there's still something relative about it that is not satisfying and I think maybe models is a kind of satisfactory like answer to it or something oh I don't know in what sense okay not for me but for our friends what you what would be the relative component of the coin position so as far as I understand the coin in position takes let's call it the web of I don't know what he calls it the web of belief or something where each where every time we change it's the transcendental project in a sense where we transform what is given and so what is given changes it's changes the way in

which we have a mental or cognitive relation to reality. And so our cognitive understanding changes as well when that happens. But I think it's still a problem insofar as Khoianians still take... There's still a speculative project where it's still somewhat analytic, I think, that you can still speculate on the procedures of the futuro, et cetera, et cetera. Right, right. I see. And then there's also the La Ruelleian way of doing away with the analytic synthetic distinction.

But I'm not going to go there. But yeah. Yes. To be honest with you, I really don't have any kind of settled idea about such matters. However, today, tomorrow morning, I might say that what I said to you was complete and utter bullshit. it. Nevertheless, today I can say that this really, this whole idea of analyticity versus synthetic-ness, you know, the Kwanian form of procedures or Carnapian form of procedures

to arrive at phenomena are not diktats about broad, broad diktats about how we should approach about the world. But as a matter of fact, they are necessary and scale sensitive ways of implementing our methods. At some scale, things can be synthetic. At some other scale, things can be analytic. You see, models, when we are simply looking at their structure, their correspondence with

their descriptions, namely equations, and so on and so forth, they can actually be understood their own right as analytic procedures, analytic statements. Once we try to somehow coordinate this or apply it to a real target system, then we are in the business of synthetic. What is really important here, I would say that for us to in fact to talk coherently about such procedures as whether they are synthetic or analytic, we should be also capable of distinguishing the different scales, the different contexts in which we both study

the real phenomenon and the model. of you have been silent I will come after you if you don't talk Alan well I was just trying to process what you were talking about the discussion of where to treat models between and I'm just trying to process that so I know

I'm not a Kantian or anything but it's an interesting discussion. Yes, yes. No, I mean, to be honest with you, I taught Kant, what like, Theo, how, how, what was it? Like how many months ago, like, so I started to, I had this idea that I'm this staunch Kantian. but then when I started to teach Kant then I noticed that oh my goodness I am actually anti-Kant you never know and this is the beauty of philosophy you start with the presupposition but you end up in what you always afraid of to be

thought takes you to places that you never expect in your entire life this is what I say it's the freedom of thought the freedom of thought trumps sorry to use that word trumps over psychological convictions in every instance and this idea is actually sublimated in the scientific method science as Ray Brasier said develops not because of what you think is psychologically speaking what should be done but despite of it

People thought that, you know, according to all this great canon of Aquinas, Bible, so on and so forth, the great court of Catholicism, absolutely earth should be flat. because it's just according to all of these connections that we have, if Earth is not flat, then we are not good species. We are not children of God. Well, well, well, well. Science not only showed that Earth is neither flat, but also it showed that we are not children of God. There is no such a thing as an intelligence design. And the whole point is that the scientific rationality, even though it has been appropriated

by the neoliberalist system, still it can be reappropriated back by the philosophical system, by the method of thinking rigorously. As such, it is still offering us the best way of shedding our human biases. Regardless, let's get back to our modeling. parochial as it might be. So, we talked about Schelling segregation and Schelling

segregation, have you noticed that the kind of metatheoretical assumptions that we were talking about are actually more prevalent in Schelling's computational modeling of segregation than the other previous two methods that I mentioned, the concrete model and the mathematical Volterra's model. The thing is that in Schelling's proposition with regard to the idea of preference, the choice of 30 percent satisfaction that if it is under 30 percent i'm going to leave i'm going to get out of this stupid neighborhood that is a utility function

of course on what grounds we have such a percentage on what grounds we couldn't have any like a higher or lower percentage. You see precisely because the utility function is dictated by the theoretical assumptions which are already presumed and taken for granted within a game theoretic framework in which Schelling upgrades. But then, so if this is a hidden assumption that should be further investigated then you can even go further and say as Brandon says, Robert

Brandon says, why do we need to actually have a purely game theoretical model here or again, theoretic theory here as basically the platform of modernization. All I am trying to say is that the interpreted structure for any given real phenomenon has hidden assumptions and these hidden assumptions can be categorized to two kinds of assumptions the explicit ones and the implicit ones the explicit ones are the ones that are given by

the theory under which the model behaves and should behave the implicit ones are actually quite scary which are usually called the metatheoretical assumptions and the metatheoretical assumptions are the ones which we take for granted precisely because we are already inhabiting within a specific theoretical framework and not others you know in fact you can say the same thing about volterra's equations why differential calculus why not some other kind of field of mathematics as undergirding for our equations of prey and predation.

Questions here? So, I mentioned that... My apologies. So now that we have a little bit of information about models, we should also talk about, and we also introduced models as what you might call to be indirect abstractions. In the sense that when we are dealing with models, once we deal with the models, at that

actually talk about as if we were talking about reality. We are already assuming two things. One, that this model has some sort of partial, at the very least partial isomorphic, some sort of correspondence with a real phenomenon. Precisely because how we construct our model either according to the measurements like the concrete model of the Bay Area or some sort of equations we have some already, you know, we have observed that they can actually reveal some patterns in the universe according to some sort of selection of mathematical structures or for that matter computational procedures.

Now models, if they can be called indirect ways of dealing with the world, in order for us to understand the indirect, we should also propose a direct form of abstract representation, so that we don't confuse models with other kinds of theoretical procedures which are usually used in the history of science but cannot be called modeling. One of the greatest examples of this abstract direct representation, which is not a modeling

per se, is the story of Mendeleev's construction of the periodic system, which of course also has a very humble beginning. When assigned to teach courses on inorganic chemistry at the University of St. Petersburg, Vendeliev found that there was no good inorganic chemistry textbook available at this time. Inorganic texts lacked an organized and coherent structure from which to characterize the known elements and inorganic reactions. In order to deepen his and his students' understanding of the elements, Mendeliev essentially wanted

to develop a classification system that could explicate or elucidate their underlying properties. This of course would have allowed him for a more systematic understanding of the properties of each element, the reactions each element could participate in and trends underlying these properties and such reactions. Mendelius faced, to do these kinds of tasks, he actually confronted or faced a daunting theoretical challenge. Samples of pure elements have many, many chemically important properties, any of which might form the basis of a classification system.

You know, to the point that, you know, when, of course, here, you know, this is another thing that I want to say. So people say that, well, all models are idealization, idealizations are just distortions. No, no, no, no, idealizations are pragmatically useful. If there in fact was no method of idealization or simplification, we couldn't even talk anything about the world. And now here we are actually not in the realm of modeling, we are actually in the realm of what you might call to be theoretical representation, like such as Mandelius' table of elements. So given the complexities of reactions and properties of elements that he had,

he couldn't really say anything about how to actually talk about this stuff. Because, you know, any kind of talk that might come up can be actually arbitrary. You see simplification mitigates arbitrariness and that's really important. So as I said, Mendelia faced a daunting theoretical challenge. Samples of the pure elements had many chemically important properties, any of which might form the basis of a classification system. Now one might sort elements by color with various you know other kinds of properties bunched together with it. In

the end however Mandeliev decided to focus his attention on finding trends in properties of valency, isomorphism and most importantly atomic weight, abstracting the way from all of the other properties so you see this we are still very very similar to this stuff that we were talking about modeling abstracting away abstracting the way this is kind of almost like the AI problem a frame problem you know a robot cannot navigate the environment if you cannot abstract away from some details are unnecessary for its navigation.

So atomic weight is a familiar concept but Balancy and what 19th century chemists called isomorphism may not be. Elements 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. So Mendeleev's first step was to organize the elements by atomic weight. This gave him a one

dimensional ordering of the elements which served as an initial organizational device, but did not reveal any information about the elements underlying structure or their unity. Focusing next on valency and isomorphism, Mendeleev tried to find other dimensions along which to organize the elements. In modern terms, 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, Mendelius is said to have put the names and properties of elements on cards and played chemical solitaire on long train journeys until he found satisfactory

ordering of the known elements. Mendelius then announced his ordering of elements according to their weight and properties. ordering which later became known as periodic table of the elements organized the elements in in the order of atomic weight and then in columns or groups in virtue of their chemical properties when the elements were properly ordered mendeleev argued one could see the periodic dependence of elemental properties under atomic weight this principle which mendeleev called the periodic law is one of the bedrock principles which organizes the entire science of chemistry it is still recognized one of the most basic patterns among all

chemical phenomena now the thing is that Mendelius theoretical achievements are sometimes overlooked because of the suspicion that the periodic table is merely a classification. It's just like you might say the Library of Congress. It makes certain trends explicit but it has been argued that the table does not actually explain anything. The Library of Congress did a service to humanity by developing a relatively rational tractable system for organizing books in our libraries. But surely we would not want to treat this

as a theoretical achievement of scientific weight. Similarly, it has been argued that Mendeleev articulated an important classification system but not a theory as such. For example, You know, Shapir claimed that what Mendeleev discovered was an ordered domain and that orderings of domains are themselves sensitive of several different sorts of lines of further research but not themselves can be considered as theories. I think, and of course Weisenberg also says this in his book, that this is actually quite

a mistaken view for many reasons. The first reason involves the remarkable predictions that Mendeleev was able to make on the basis of his periodic system. For example, in 1869 he noted that there were gaps in his table for three elements. In 1869 also on the basis of this gap about tantamount coinciding with other information about chemical trends and coded already on the table. Sorry.

He hypothesized the existence of what he called eca-aluminum or eca-silicon or eca-boron. The properties of these novel elements are listed basically in the earliest tables that he draws where he basically predicts that such elements, these lost elements, these missing elements, can in fact be discovered quite adequately based, at the very least, on their atomic weight, a specific gravity, and atomic volume or density. Mendeleev's predictions might look like trivial exercises making inferences about missing

books on the shelf if we were going to use the library of Babel or library of Congress or filling empty slots. This underestimates however the significance of his achievement. Ndendeliev had no empirical knowledge that there were, in fact, any empty slots to be filled, an element yet to be discovered. His task was thus 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 analyzing the theoretical structure he had created. 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 a theoretical, not merely classificatory achievement. However, here, and we probably talk about this at more length next session, where we try to separate the direct abstract representation from an indirect mediated abstraction, namely monolated. But nevertheless, we are in both cases in the realm of theoretical structure.

In the procedure of science, we not always use models. we use things like Mendeleev's ways of abstraction but also in the procedure of science particularly a specialized sciences we not always use direct abstract representation sometimes we choose according to a certain criteria which shall be discussed at length we go for modeling namely mediated representations questions here

Earlier in the seminar, people mentioned architectural models. And it seems to me at some maybe hand wavy level, an architectural model is a bit like a taxonomy. This sort of like it presents the structure like the periodic table of this is how how certain things are in relation to one another, but it's not necessarily as complete as a model with moving parts. You can see the whole system interacting. I'm not sure if that's a useful connection. I would say that it's actually also different

from the abstract model in the way that we talked about Mendelius, but another example would be Darwin's theory. is a study of fossils the thing is that does a simple architectural model I'm not saying that it doesn't is the question for our architect friends a simple I'm not talking about some sort of lavish computational modeling a simple architectural model as it is usually defined in the field of architecture Can it actually predict that there might be some additional elements that can be added

later on into the system? Essentially, you see what Mendeleev does, sure, it is a classificatory system, but it's not just a simple classificatory system. It's a classificatory system with predictive balance. It can actually predict that there are such and such elements are missing, and as long as, according to these categories, such and such properties are being satisfied, we can say that we have those elements on our table. It doesn't explain anything, unlike models. Models do explain. Models do explain. Yeah. Can I add something? Can you hear me?

Absolutely. Okay. Yeah, I think, as you said, it's totally different because actually the reason, I mean, I was wondering, I also wrote something that each model proposes a rule, like a new mechanism. and actually it starts from the reason of its creation, right? I mean, you make a model because you want to arrive in something. In architectural models, it's a totally different story, I agree, but you have two steps there, not two steps, two levels, let's say, of why you're making it. And the one is that you first represent the existing part.

I mean, you have to somehow represent and show how the existing space is and then you make an addition to that space with a totally new fiction. I mean, the building or I don't know the scale you're referring to with its model. which is the totally new thing you are proposing. And then there are so many things about it. I mean, I'm talking about the physical models now in the states that you, yeah. Like you represent, you have to make a decision of how you represent the existing. You need to make many abstractions because you don't need all this data.

You don't need the trees around it, except if all the project is about the trees. You need to make the rules, the decisions that how you represent it and why you put that data in it. So it is also an experimentation though, because when you're proposing things and you are using the tool of modeling, you are doing experimentation of how you can create the best space in that existing space. it's analyzing its special with the T data and you need to find a solution and you try this and that it is an experiment it's not that it's of course it's arbitrary because you don't have the the total bad results let's say with science of course we can you can have with with bad

architecture but I mean it's not like you have to find a truth as in science sure May I ask you a question, Artemis? So do you think that, I mean, you know, of course, as you said, you know, contemporary architectural models can, of course, be supplemented with, you know, computational models as we have seen, so on and so forth, algorithms and stuff. But in a classical, like for example modernist architecture, you think that what counts as an architectural model can actually predict anything about the target system? Or is it just a representation? Yeah, that's the point. If the reason of making this model is your proposition, you're going

to experiment on how the form you are proposing is going to, approximately always, act in the existing space. But if you're making a physical or a computational, let's say, model of a space, again, it is all about the reason. It adds some new data when you are focusing, as you said when you're zooming in you you augment information about reality yeah some certain data in order to find something or maybe you can find it accidentally but if you don't experiment by augmenting and abstracting other data you cannot understand it I see I see thank you so much

anyone else by the way unfortunately to we have five more minutes I have to go a little bit early today I have another talk but as I mentioned any kind of early closure we will compensate it at the end so don't worry about it but can we have one more question I had one announcement just a logistical announcement for next week i think um we have two presenters i think of the readings not next week next week we don't have any class deal oh sorry yeah yeah not next week but the next session yes next session yes um two presenters and then uh and who are who are the presenters

i guess we can uh if people want to volunteer to present on the reading I'm here or I will come after you. Yeah. Fine. So two presenters and two respondents. You'll have a little bit more time to do the readings and stuff. Can we just pick the presenters randomly? Any person, any person. I mean, don't worry. I'm not a dictator teacher. Any person would be fine. And I would actually, we can exchange email and, okay, Mikey, what is the next person? Okay. Great. So two presenters and then the two respondents.

And by the way, our reading is still the same. Mary Hess, Michael Weisberg. We are not going to get any to... yet any any question any course last question do they represent the whole book or with the book feels to us and we expand on this in depth because it's which are you know big books yes don't scare people now you are scaring up no No, no, no. Is it possible to just like zoom in and talk about fiction or Galilean? You're making a model of it actually, right? You're zooming in and you're choosing what you're going to simulate.

Okay, can we, we can, how about this, how about this? Okay, let's not get too overexcited. how about we actually read the third chapter of Michael Weisberg the anatomy of models it is I would say like 20 pages I mean it's doable somehow I would say I don't know well anyway don't worry about it we are not academics Sean yes Sean everything is in the Google Drive if you go to the class or

Google classroom if you click on the folder of the Google Drive you should be able to see them. Okay. I have a just kind of a quick question looking at the more the implicit versus explicit and then use of models, particularly in like interdisciplinary work, like if a certain model with the understanding that it has it's built for like a specific reason content um is there much application in using that model in a different um in a different discipline or different use um without kind of like the implicit explicit content of it or

intentions of it getting in the way yes well well the thing is that uh there is no rule about it but But in the canon of engineering at least and complex science, this usually is not advised. Unless we have actually looked carefully into the theoretical range of the said model. Yes, sure, if the theoretical range goes over different contexts of application, we can try it. in fact if we try it if the range the theoretical range doesn't cover any other context that would lead certainly absolutely to distorted data in fact it is it is it is forbidden in the realm of

model Cool And please you know Theo is going to of course upload the sidebar and you know that the video to the Google classroom any of you who have a question Please feel free to ask those questions on the Google Classroom and either me or someone else will reply to you. Just don't be shy. We are just trying to go over some certain topics.

It's not just new to you, it's also new to me. Okay, the respondent Justin is just going to kind of, maybe if you just had a couple questions to ask about the text or if you have questions about what Mikey and Adam present. It's just going to start the conversation I think. Yes, okay let's not get way too over ambitious here. academia where we are the you know there is a whip and we have to do certain kinds of stuff no no it's a heuristic experiment here let's Adam and Mikey do their you know responding and then we will see how people interact with the

you know those people who are you know creating a response and then after that we can come up with some sort of a more standard procedure but until then I I don't think that we should get it too involved with details. That sounds good. I think Joven also said he would have some questions. Sure, that would be fantastic. That would be fantastic. Yes, yes, absolutely. All right, so we're at our time and I'll go ahead and end the broadcast now, if that's okay. Absolutely.