Simulating the World & Remodeling Philosophy (Session 2)

Welcome to the second session of Simulating the World and Remodeling Philosophy, Models, Diagrams and Toys. I'm now handing the mic to Reza. Thank you everyone, really great to meet you again and apologies these things happen quite often so if I remember correctly two people offered voluntarily there was no force involved to make a presentations and who were those people I know one of them was present I think

Adam I think yeah yes yep yep oh good so um shall I just dive in um a summary of chapter three of Weisberg which is about the anatomy of models um so here's things I noted. So he offers a concise definition of model to start with. He says models are composed of structure and the scientists interpretation of that structure. He breaks it down into three types of models that we touched on the other week. So concrete models like real physical objects that

are built like the the AECO sized San Francisco Bay model but also things like wind tunnels. He also puts some other interesting things in this category like historical natural experiments, as you might get in geology or in economics or climatology. And he also puts in paradigm organisms, which are very important in biology. So stuff like the fruit fly, which is used as a sort of test bed or experimental test bed, he conceptualizes as a concrete model as well. The second type of model is mathematical structures. So we have like the differential equations for population modeling, quite abstract, but

mathematical models capturing a mathematical relation between different elements. He specifically points out that it's not just dynamical models of traversal through a state space. It could also be expressing something like a graph theory relation. There's a little bit of a discussion about predicates versus other forms. Basically he just wants to leave it as broad as mathematical structures. And the computational structures, he spends a little bit of time on this. He basically concedes from the get-go that computational structures are mathematical structures, but he breaks it out as a separate type because the place that is doing the sort of work of the model is different,

where you have an algorithm in a computational structure that is doing the main representational work, by contrast to a sort of set of equations which states a relation. And that's the distinction he's drawing, which is quite different in practice, I guess. Right. I guess, you know, shelling the race model that we touched on the other week is this key example here. That also reminded me of, I'll share the link, that there's some follow-up work from that sort of modeling. It's one by Lurie and Coat, where not only they look at a model, a quite abstract model of workplace discrimination and racism

where you don't even need a sort of neighborhood effect. Because in shelling, you have this neighborhood effect of if you like your neighbors to look a bit like you, then you can get this sort of emergent racism, even if there's no actual KKK members in that community. Lurie and Coat have an informational model where you can have, if some surface criteria like skin color is a proxy for less observable or unobservable things like actual skill level at the job, then they show how you can get a racist effect without any racist individuals. Right, right.

Result two. Computational models for the world's climate are, I think, a key example under computational models. It's really interesting to think, though, something like the models used for the IPCC report are actually incredibly detailed and have lots and lots of measurement parameterizing the computational model. They're right at another end of representation or detail of representation compared to something abstract like the Schelling model or the Lurie models. Next section is on model descriptions. He just wants to, I think the main point here is about the work that descriptions do in model

definition and use. And they can play an important role in the model itself. So in a concrete model, for example, you will have specification for the concrete model. They drew up full-blown engineering diagrams for this acre-wide San Francisco Bay model, for example. There was an operating manual for all the pumps and stuff like that. And that itself is sort of part of the model, or at least even if you don't count it as part of the model, it's an important part of how a model is used. It carries weight. It reminded me of Thomas Kuhn's point about paradigms. He said, you can get a paradigm shift

simply from a change in representation, right? Okay, change in representation, the obvious one. The point I wanted to make is a change in instrumentation. So a new scientific instrument or changes, like even a new type, like a new version of a scientific instrument can actually trigger a paradigm shift in Kuhn's worldview. That kind of reminded me of this seemingly auxiliary thing actually could carry a lot of weight in the use of models. So that was descriptions. The next section is on construal or interpretation. He breaks down the Construal into four parts. The four parts are the model is intended scope,

an assignment, dynamical fidelity criteria, and representational fidelity criteria. I thought of this as two parts. It's like the mapping and the test you have on the mapping. fidelity criteria is like at the test and the other part is describing what the mapping is. There's a paper by Reinhardt and Rogoff which was a bit notorious called growth in a time of debt where they had a spreadsheet based model and based on that they said hey look in a recession you should cut spending to have a good influence on debt if you do it on growth if you've

got high debt and this was waved around after the financial crisis to say okay we should all be doing austerity um and turned out in this case apart from all the other ideological infrastructure there that there was actually a key error in the spreadsheet and they the fidelity criteria were kind of weak here because they weren't able to pick up the internal consistency of the model or of the mapping against the system that we're modeling. 3.4, last section was on representational capacity. And this is basically like, can any structure be a model? If you're saying like structures are the main thing in models, can any structure be a model? And Weisberg sort of said, he sort of bites the bullet

and then backs away from it at the same time. He sort of said, yes, any structure could be a model. of anything, but not a very good one. And so it uses two main criteria. One of them is representational capacity. So if you have some very simple system, it can't represent much. And then mechanistic adequacy, which is about the quality of the mapping and the ability of the model to execute in a way. So representational capacity seem to invoke ideas of Kolmogorov complexity and other informational measures, to me, they shouldn't be explicitly. JOHN MUELLER- Super. Excellent. That were my notes. That's all I got.

Fantastic. It's fantastic. So I will respond to you once. Jean-Pierre, is it your turn? unmute okay no uh i wasn't scheduled for his okay who's the next person i think it was theo wasn't it theo meredith was responding okay meredith yes okay let's hear meredith sorry I'm on set dementia. Ignore me. You guys go on. Berendif, can you please unmute yourself

and possibly change the location of your camera because you can't see your face? Thanks. Okay. So I just had a few questions. The first was... Mommy! Mommy, can I have some candy? Definitely yes. Okay. Okay, so my first, this is like a reaction more to the text than, Adam, I'm not very good

at responding probably to your individual points, but more to the text. My first was I was interested. I almost finished my food. Okay, you can. Yes, go. Model states are states of values of variable quantities. As a sort of anatomy of the mathematical model. And I was wondering what the relationship between this and sort of Wittgenstein's sort of notion of the state of the world. And if maybe we're thinking about simulated models, you know, is that a, I mean, this is really sort of like off topic,

but, you know, if we're thinking about computer simulations, you know, is that one place where that argument also breaks down? It's sort of like the whole idea of states of the world. I don't know, I thought the, I mean, the whole idea of like variable relationships versus transitional relationships. I thought this blew my mind from last week. My other question was, oh, sort of the notion of populations. This is really from chapter two. So mathematical models are modeling populations, whereas computational models are modeling transitions.

But I do think there's sort of like a breakdown here. One reason is that the modelers intend to target actually behaves probabilistically. So that is a feature of computational systems on page 30. But I feel like when you're dealing with populations, you're not talking about probability necessarily, but you're talking about statistical models. And, you know, if you're doing a computer simulation, you could theoretically model each individual if you're doing some agent-based modeling. So I was interested in sort of like the relationship between populations, statistics versus probability, and mathematical models as a statistical model. Yes, yes.

It's okay, I already had dessert. Oh, that's okay. Mommy says I could have his muffins too. Um... My other, my other, I was also interested in the idea of instantiation and the uninstantiated model. What do you mean by uninstantiated model? So on page 37, an uninstantiated model description is an equation in which values are not assigned to the parameters.

Instantiating a model description means adding in values for the parameters. Right. Yeah. So like is an uninstantiated model a, you know, so I think it's in chapter two, maybe even chapter one, maybe it's in chapter two. Weisberg says, okay, computer simulations are sort of a subset of mathematical simulation or mathematical models abstractly. So is an uninstantiated mathematical model so different? how is that really different from a computer simulation? I mean, I guess you could make it about recursion or sort of like differential equations, but I was interested. I mean, I do think this idea of

instantiation versus uninstantiation, you know, I think it is, I think it's an interesting, I think it's an interesting point. Right. I think I will talk about this today. I generally think that Weisberg's idea of computational model versus mathematical model, as Adam said, really doesn't hold. The way that he describes computational model, really the distinction doesn't hold. But I do think that there is in fact a distinction between a mathematical and computational model And quite contrary to what Weisberg says, I would say mathematical models are in fact subsets of computational models. And computational models are not exclusive to simulations.

But you go forward. Sorry to interrupt you. I agree. And he even says that. He even says that. I mean, I don't even think. I mean, I think he probably could have just because he says that in the beginning of chapter two. I think... Right, mathematical models are abstract... He does... well I agree with that as well. In any case... In the first chapter, I know that he goes on and kind of refines the distinction so people don't come and bite him later on.

But nevertheless, in the first chapter, he actually makes a very, very, I would say, weak distinction in the sense that he says that, well, you know, mathematical models are based on mathematical structures, usually differential equations. whereas computational models are algorithmically implemented in the sense that you would say that it's just a mathematical model for which there is actually an algorithmic program, input, mapping function, and output given certain circumstances. That's it. But I think that this just really doesn't hold up in today's computer science anymore. I mean, I would love, I would love it, you know, I always thought mathematical models were no different from computer simulations.

But I would love it if there really was a difference. And in some cases, like the colorability problem. Guys, guys, you can't bother me anymore. Yes. Cute. Okay, but you know, but I feel like there is something and maybe we're not there yet. So like for example with the colorability, the proof for colorability, like the three, can you have a map with three color, use only three colors to color a map. So this was solved, you can't, like four colors is the minimum,

and this was solved basically by brute force, like there's no mathematical formula that you can use. You know, and like, is that the distinction between a simulation and a mathematical model? But anyway, it's like... Yes, but the thing is that as I mentioned, computational models don't just do simulation. Yes, they do simulation. First of all, I think that this idea of simulation is quite, from an engineering system analysis perspective, is quite limited in this scope. But also, computational models do a lot of other stuff rather than just simulation. They do inaction, they do emulation of a system, so on and so forth. And all of these, what you might call to be exemplars of what you might call to be uninstantiated

model or weakly instantiated model. But we'll get into those things. Well, Rezo, what is an example of like a simulation model? So, okay, three things that comes, and I know that these words are being thrown in a very common sense manner, that's fine, but we are in the business of modeling, we are in the business of philosophy we should use the words accurately. So simulation is essentially what you might call to be replication or reproduction of a set of behaviors without or with minimum regard of the underlying laws or

rules responsible for giving rise to these behaviors like for example a dog whenever in Pavlovian a dog whenever sees a bowl of meat salivates okay now I can I can easily simulate just a behavior I do not need to know that many details of course it would greatly enrich my understanding of the behavior or the modeling of the behavior but nevertheless I don't need to know that much details about the mechanism undergirding the salivation of a dog as soon as he sees a bowl of meat well this is so simili

behaviors are being simulated underlying mechanisms rules or laws responsible for giving rise to these behaviors are being emulated the environment responsible for the evolution of both mechanisms and their corresponding behaviors are being enacted So we have three criteria, simulation, emulation, and reenactment, or enactment, each of which require different computational criteria. And in fact, from a theoretical standpoint, you might say that simulation is the simplest

form of computational modeling, and inaction is the most complex one. Because it looks into the genesis of both rules on the girding behaviors and the symptomatic behaviors that come out of them. The same thing about thermostat. Thermostat has certain kind of behavior, cybernetic behavior. You can simulate it, right? But then, that's easy task. But then if you go a little bit a step further, you see the kind of mechanism responsible for the kind of behavior that thermostat undertakes.

That becomes a little bit more difficult. That's emulation. Whereas, you see thermostat as essentially a cybernetic entity in interaction with its environment and no longer a discrete entity that can be just talked about by itself as if their its behaviors can simply be extracted and simulated that the thermostat requires some sort of what you might call to be a real-time interaction analog thermostat with its environment that's to reenact this environment in which thermostat is a part that requires a fundamentally different computational modeling in fact even the paradigm of computation changes here

from a universal Turing machine discrete input transition state discrete output during which the environment is shut down to the paradigm of computation in which the computation of a machine is always thought and modeled and basically systematized according to its real-time concurrent interaction with the environment of variables. That's interaction I mean as interaction or inaction, sorry, inaction.

Of course there is a massive amount of historical texts about engineering discoveries that go, you know, oscillate between these three paradigms. But I would say that inaction is something far more newer than any of those two simulation and emulation. Simulation and emulation have been modeled quite often particularly after the turn the computational turn but inaction is is is fundamentally complex because essentially you are at the whim of the environment you have you need to have a pattern for concurrent computational

processing so on so forth that makes the job of modeler far more intricate but of course it enriches the model one of the things and then i'll let one of the things that i'm interested in i think from your course description is the idea of philosophical toys or um so you know in what sense is a thought experiment a model or in one sense sense is a is a game like a like a even a children's game a model or they have a play they don't exactly follow rules there is a difference in the philosophical canon

between a game and a play a game requires a minimum acquisition of rules right like the the game of language. But play starts with the idea that you can in fact to a certain extent suspend such rules. Like imagine a good example of it. You know the difference between rule and play is comes to this like when children are given bunch of Lego pieces. the game is more like when you tell a child that make a robot Lego for me and then you hand down the diagram the instruction that comes with the Lego kit

to the child and you know she starts to try as best as she can put these two together according to what she sees but the play is very different thing toying around toying around essentially it is not about creating a target system or a fidelity to a target system like making your robot model in the lego universe but it is about whatever you can make out of these lego pieces as long as the lego pieces fit together she might come with a dragon she might come with death's star so on so forth as long as

the pieces fit together shape wise color wise so on so forth there is an infinite play a suspension of rubies. We will talk about all of these. Artemis, would you be able to elaborate on your question a little bit? Yes. Hello. Because you said that there is this thing that as long as it fits, so it's like it has one rule that you need to follow and not instructions. Yes, yes. It is a rule, nevertheless it's a minimum rule. It's a minimum rule. You see,

is basically what the idea of a structure is you see the idea of a structure by itself is just the minimum rules for it to have some sort of correspondence we do not know yet with a target phenomenon the structure just about minimum rules in fact if it had if it is bogged down with too many rules this is not a good structure structures should always have minimum rules this is the structure of the model the core of the model now you can think about the diagram the instruction that comes with the Lego kit that says that you have to you know in order for you to make a robot Lego robot you have to

follow these steps now this is not simply the structure structure is just says as you said about minimal rule of how to fit these Lego pieces together so as you make any kind of a structure right so the the rule is more like the description of the model plus the fidelity the fidelity criteria the fidelity criteria in the sense that it instructs you to what extent can you put these stuff together these building blocks together and still make some semblance representation

of a robot. Now that's, this is, here we are in the business of rules and no longer toys whose objectives are playing with their structure rather than abiding by a specific representational rules. Meredith, is there anything that you would like to add? You are mute, you are mute, You have to unmute yourself. Okay.

The relationship between model and description. Now, I was rereading Carnap this weekend. And I was... What were you reading? The off-bow. Off-bow, okay. And what is the, you can have a model without a, I would be interested in unpacking the relationship between model descriptions and I guess model constructions.

Model structure or model construction? I guess it's model structure. Here it's being described. So we have right at the beginning we have two, we have like a criterion for what it is to be a model. They all have structure. Interpreted structure, right? All models consist of an interpreted structure. So where does description fit in? Is that the mapping to? Is that? Yes, so essentially you see the interpreted structure is what you might call to be, At the level of interpretive structure models, if it is just interpretive structure, models

are much more like theories, right? The theories, you know, you can toy around with them and they essentially don't need to be targeted on a specific or constructed based on a specific phenomenon. The thing is that the model description is a certain, what Adam was talking about, certain sets of symbolic representations, symbolic representations.

It can be diagram, gestures, mathematical equations, you know, a certain kind of constrained algorithm, so on and so forth, in which the interpreted structure of the model is connected to a certain kind of phenomenon. In that sense, you can say that the model description is about the application range, the application range of the interpreted structure. We will come back to all of these when we talk about Josephus Sneed, which of course

we talked about quite heavily in our previous class. But I will tie it back into this. And this is exactly what you might call to be MPP in the Joseph Schnitt vocabulary, partial potential model that covers the applicatorial range of the model structure. Of course, application is not the same as the interpreted structure, as the kind of possible phenomenon, MP, that you are going to address.

One thing that's not addressed yet, and we're going to probably get there, is analysis of the results set. So we do have how we're judging the criteria, how, you know, the, the, how to judge the fidelity of the model. But, you know, often, as often as the case, you have, you, you, and this is maybe, well, I think we're going to probably address this in other readings.

The, the data, the data that you, that you generate with, with a dynamic model has its own, itself has to be interpreted. I'm thinking in this case of data visualization. And this is probably beyond the scope of now. But sort of how do we think about that second order mapping between sort of the understanding of the data? and I mean you know we're in the case of doing of creating a I don't know I mean that's that's something that I think about a lot I don't even know if that's a real that's a real

distinction are you are you mentioning something like that how can we you know once we have once we have certain kinds of data that are derived from our application of a model, then how can we frame them such that we can interpret them consistently with the model and its structure? Well this is a point where theory comes through. Essentially, sure, you know, models can be applied regardless of the theory infrastructure that undergirds this model. But if you want to go even further such that you can in fact analyze the very configuration,

the very procedures by which you have framed this data derived from your model, then I I don't think that models can actually give you any answer. This is where you are in the business of a meta-model or a meta-theoretical model where you actually start to, you know, think about these. But I wouldn't say, but the thing is that, to be honest with Meredith, being the majority of the models as being used today they don't have such concerns this doesn't mean that such concerns are not important it just simply means that they really do not want to know about this and of course we will talk about this

that what are the dangers of what of not wanting to know about the meta theoretical metamodel assumptions that lead you to frame your data in such and such configurations and not others so you see when you if you are simply working in the framework of a specific model X any kind of that's all that you garner from a physical phenomena or whatever would be framed according to the you know the parameters of the model itself but then that would be just

almost an automatic task in fact this automatic task as so many people have talked about can in fact be algorithmically realized you don't you You don't need a scientist to do this for you. You can even give it to a great computational program to assess the situation. But the thing that is missing here is, as I mentioned, Adam brought it on the Google class with regard to Theo's question, is something like a toy model, a meta-model assessment. you can actually see that okay these are all great with regard to the model there

is a criteria of consistency and coherency here but with regard to how model can be mapped to a broader theoretical structure does it really should it be repaired, should it be left intact? These are not the questions that engineers often bother themselves with, but I think that they are important questions. And that's why there is this kind of tendency in the past decade or so toward this kind toward this kind of meta-inquiry in which the model is no longer taken as discrete but

already as embedded within a theoretical system and that theoretical system is all embedded in a Carnapian constitutional hierarchy in a meta-theoretical sphere and those meta-theoretical spheres should actually give you a better idea how to gauge not only what you have made out of this model but also whether the configuration that you have arrived that is theoretically sound what does it actually cover and what does it it preclude? The latter question is really important because for the most part when a

modeler wants to model a real phenomenon, sure it is about the fidelity to a target system. That's the first question. That's how a modeler makes a model. But then once Once you garner certain data, you want to actually know that once I applied this model, I had this kind of result for configuration. According to my theoretical and theoretical assumptions behind my model, what other kinds are models basically were avoided or were prevented by virtue of me using this a

specific model it's like this when I say blue the assertion of this is blue preclude me to say that this is green this is red this is brown this is not colored right so I want if I really want to get back to the nitty-gritty stuff I want to know why is that arriving at the judgment of this is blue precluded me from such assertions such as this is brown, this is red, and this is not colored. That mechanism is entirely outside of current paradigm of modeling.

It is getting there but it's not yet there, absolutely not there. Maybe some Some semblance of them are in computer science, but in engineering, really there are no such things at this point. So any questions? If not questions, let's have a five-minute break and then we will come back. I'll start reading and then we'll open it to discussion. And thank you so much Meredith and Adam, fantastic, fantastic contributions.

Questions, anyone? Heckling, Joven, Jean-Pierre, all those silent ones, I will eventually come to you. Go on. I just wanted to say I have the presentation too from the reading, so whenever, or I can wait till next time, but yeah, I have one ready. Are you trying to kind of say that you are going to be the next respondent for the next session? No, I was the original one to respond for this session, but yeah, I have one ready, so whenever. Okay, superb. Superb. But I'm always happy to get new respondents and new presenters for the next week.

You can't just jump on it. It's a horse that just waits for you to go on. Any questions, please do. Where's our attorney, by the way? What do you need? Okay, questions. I'm leaving it to others at the moment. Okay. Okay. Okay. Don't worry, Ian. I have a question. Hello? Please go on. Please go on. Don't let me interrupt you. So I'm wondering about making an analogy.

Like, so, for example, the difference that you had outlined between, like, a game and, I think, play would play being the suspension of rules and a game being kind of, like, the almost, I want to say, like, insistence on a kind of minimum of rules. and so immediately there I was thinking of how that relates to like a kind of hyalomorphic understanding of imposing a certain form onto matter versus a more kind of like non-hyalomorphic form of finding form in matter that is like pre-existent and then also just like what writing means like the creation of an object so like can we think of like the writing of a book

as a kind of model and I think that those are definitely interesting questions but I guess what I would be more interested in is like what is the analogy like is the making of an analogy in this precise situation the kind of occupancy of the kind of meta modeling perspective? Is that like theory? Because I feel like one could argue that there is a representational capacity to turn anything into a model. And I think that as Meredith had said, Weisberg emphasizes that, or I can't remember who had said this earlier. but yeah so I just wonder what I guess analogy making is like can we is it like an infinite

regression I suppose the further models or not okay I think these are two questions one let me answer your question by getting back into the history of philosophy to the time of cats you see and the kind of post Kantian philosophy. Obviously, when we are going to talk about an object, we require certain kinds of rules. These rules in Kant are often called categories. It constitutes what we understand as an object.

In fact, to talk about an object without categories would be a leap back into the critical philosophy, right? Into a nominum, into the thing in itself. And of course, Kant says that a thing in itself cannot be known. It can be thought, it can be speculated, but it cannot be known. To say that it can be known, you are in the business of pre-critical philosophy. And the job of the model are about knowledge, giving us some semblance of what is going on with regard to a target system or a specific phenomenon, physical objective phenomenon.

Now insofar as objects requires rules, every object constitution is rule-based. Now this is Kant's orthodox version of transcendental deduction or transcendental philosophy. To conceive of an object you need to have a system of rules. Here, I would say that this is, and of course this is also Wittgenstein, with regard to language. And many people think that Wittgenstein, yes, Wittgenstein was anti-Kantian, but the core of the philosophy of these philosophers, with regard to the presence of rules, is quite

actually similar. Now, Carnap understood this and he launched a criticism against both. And this is not the early Carnap, this is more of a late period Carnap. He launched a criticism against both Wittgenstein and Kantian idea of rule. In the sense that, okay, let's talk by way of an allegory. So when you are handed down a Lego kit with an instruction to make a dragon, then of course

That instruction is a system of rules according to which you create an object called a dragon. Right? But then a problem here arises. If everything that we can make in this Lego universe or everything that we can conceive in the epistemological universe is based on pre-established rules, then how can we ever conceive of new objects right that's actually quite an skeptical question Kant actually addresses this in the

critique of pure reason early on but he dismisses this question as armchair speculation the reason behind that comes to a very very specific weakness in in Kant in the sense that categories are a species of logic for Kant being living in 1700 for him logic is simply what you might call to be a revised version of Aristotelian logic so he doesn't see logic as some sort of free play creating universes of logic but simply as something that always and all the time should abide by the rule of application

that's what he calls a logic as a camp that infinite play of logic for the sake of logic he calls logic as an organon and he calls logic as an organon which is a play an infinite play an illusory business a sophisticated conmanship but the thing is that once Frigge's revolution starts once we are in business of today's logic and once we see what logic actually is as the constitutive edifice for all scientific enterprise we say we we see that even though the rule-based system of logic the

application making a dragon according to the instruction was great gave us so much but it's Also, by giving us so much, it precluded or prevented us from conceiving of other objects, of other kinds of knowledge. This is where world building or infinite place comes to the picture. The more we can play with our rules, so as make new system of rules, the more we are capable of constituting new objects. But that cannot arise until and unless we suspend the established rules of representations and actually deal with rules as toy objects, like Legos.

Just put them together. As long as they fit, that's fine. Then we will look into their application later on. Thank you. Thank you. Any question or should we have our short break, like seven minutes and then we come back? Can I ask you something? I missed the last part regarding the representation capacity of the model. like I was thinking that I mean if we follow the Orthodox count could we say

that every model is not merely representative or yes yeah we opposite thing I mean no no it is not nearly a representative of course when we are in the business of representation nothing is ever perfectly represented that's that's that's that's a given that's a given that's a that is transcendental given since the time of can't to today but the thing that can't tries to say is that we should not play with logic as an infinite toy because that just doesn't have any applicatory rule what we should do is to see representation within a certain, a specific set of rules, categories, and

within which we expand our objectivity. But now, here a problem that arises. That okay, if all rules are required to be subordinated to some representational constraints, and if these representational constraints are also rooted in a certain kinds of logical rules then we are already in a kind of vicious circle in the sense that we are only can talk so much about the kind of stuff in the world after we can represent precisely because we have only so much rules at

our disposal this is the idea of a what you might call to be the Kantian orthodox Kantian model. Whereas a toy model or an infinite play is trying to actually suspend these rules, not terminate them but simply temporally suspend them so as it can play with them and create the range of possibilities of what can be represented, what kind of models can be applied, what kind of logical structures can be made such that we can enrich not only our representational capacities but also our knowledge of the world exactly comes

back to that Lego allegory that I made you know a child who has given a set of instruction comes with the Lego kit while making a robot of course she can make this robot with different blocks and shapes but as long when she has to follow a certain kind of a structure and kind of a minimum structure but then imagine forget about that imagine a child that has given a Lego kit some Lego pieces without any instruction as long as these things fit she can make different rules different structures and of course this is merely a potential

structure a potential model we do not yet know what it may possibly represent but nevertheless in so far as we take for granted the model is what gives us a kind of vision into a real phenomenon as long as we enrich the range of our possible models we may possibly also enrich the phenomenon that we can conceive by those models

any question if no question seven minutes break and then we will come back and talk okay then see you in seven minutes Thank you.

Thank you. Thank you.

Thank you. Thank you.

Thank you. Thank you.

Thank you. Thank you.

Thank you. Thank you.

Jean-Pierre, you have been so silent. as Justin, Mary, Andrea, Alberto, Alan, don't let me to name you. I was actually struggling a bit with my internet connection during that last minute, so I didn't catch much of what was told then, but this was a problem. Yeah, yeah, I had a problem here

with the connection. No, I know, I know, I know. I don't understand, but, well, yeah, actually, I don't know if you already answered something like it, because I think Meredith brought Wittgenstein up, so it seems to me that Wittgenstein's concept of rule, not so much in the Tractatus, but in the philosophical investigations, is much more akin to what you were saying about play. right yes philosophical investigation rather than tractatus yes absolutely yes yes yes definitely yeah yeah because he's always saying about language games and games what what what was that quote um

there are games which rules we make we make along as we play something like this right right and And also this answers to the German word, the German sense of Spiel, which is much broader than game in English or even in Portuguese. So I don't know. My question would be something like, is Carnap's critique to Wittgenstein, does it apply to philosophical investigations as well? That is under research. But so far, I would say, yes, it does.

Yes, it does. Precisely because even in the philosophical investigation, even in the philosophical investigation, not because of the what you might call to be a difference between rule and play but even in the philosophical and logical investigations Wittgenstein still think that language is something about the world for some for language to be this is debatable I think but okay I do think, no, I actually, I can quote certain paragraphs, certain, you know, those phrases

from Wittgenstein in which the correspondence between the language and the world is still upheld as far as the logical investigation and philosophical investigation. Now here a problem arises and you know what that problem is. That's essentially a metaphysical fallacy. How can language be about anything? Language is about nothing really in the world. This is, I would say, if Wittgenstein tries to get rid of this aboutness, aboutness, which is no longer strong. I can understand that. It is not that kind of intentional

stuff in old philosophy of language. Nevertheless, there is a weak correspondence between language and the world, which of course, Sellars also inherits it. Yes. Sellars says that an empirically meaningful sentence, for a sentence to be considered be an empirically meaningful it should be part of a language and is about the world of course he doesn't again me a strong correspondence but simply this kind of weak aboutness but this is where the card out skips it that language is not about anything the language is not about any kind of representation of rule

And precisely to that extent, Estivaldi and Karros liken Carnap's vision of language to this unbound dream of an unbound ocean that can grow infinitely, eat at anything that we take as solid or land. Yeah, I was thinking maybe the critique still applies, but it ought to be weakened a bit, because whereas in the Tractatus there is a very strong rule of correspondence, the meaningful sentences are the possible sentences, the possible sentences are the ones which

respect the rules of enchaînement, we would say in French, of chaining the objects, right? The logical objects. And when the meaningful sentence obtains, it is because it corresponds with the world. This disappears in the philosophical investigations. The philosophical investigations you have like some... the connection from language to world is much weaker, I think, even though you have the stratum of sentences which seems to be about something, but they are actually fixing the rules for the use of other

sentences. So when you say the Parisian meter, which is the model for the meter, has one meter. For him this is meaningless. It looks like a proper sentence, but it actually affirms that which is the property of the model itself. It's not a meaningful sentence about the world. So you don't have any more metaphysical connection between language and world, but you have like a conventional paradigms which we take to be rules in order to do other activities in the world. Right. No, I agree with this. I agree with this.

Yeah, this opens up the gates to have other paradigms, to adopt other rules, even in Wittgenstein himself. This is what I had in mind, actually. right okay everyone what we are talking about and that's actually a good opportunity to me kind of somehow reformulate the difference between play and rule why they are important with regard to the model okay so imagine this that the canonical Kantian idea of rule is essentially about the application of logical

rules to sensible intuitions, to the world, broadly understood, to phenomena, to phenomenal appearances. Now of course this what you might call to be a representational paradigm of rules. Not that rules represent anything in the world but that in the last instance they should be somehow brought in some kind of harmony some kind of correspondence with the stuff with the phenomena in the world

okay this is what we might call the idea of game again in the Kantian sense in the orthodox Kantian sense I'm not talking about games in different kind of context just orthodox Kantian idea of game now play is is a very different child plays the object of a child can any can can every all of you can he can you hear me yes yeah because I have a little bit of noise yes yeah I'm having this nice too this was bothering me yeah before the break yeah yeah apparently went away did it

go away on your end as well yep yeah okay good so so we have this idea of kantian game okay i'm not talking about wisconsin at this point just orthodox can't so this is the idea of game that It's not that the rules of logic or rules of thought cannot be diversified at whim of the said rules, but that they ultimately should be somehow subordinated and applied, applied aka subordinated to some representational ways of how we see and smell and taste and

perceive things in the world okay so this is the idea of the game in the cantina sense whereas the idea of Carnap's play is very different. Carnap, after Aufbau, after the logical structure of the world, he fundamentally gets rid of all the previous stuff and his fidelity to Wittgenstein and can't for that matter. He says, as I was answering to Sean, something akin to this, that okay, so if we can only talk about stuff in the world

in accordance with the rules of this game which is at some point representational then this also means that we can only talk about stuff in the world as long as we have a we have established rules for them but in so far as our established rules are also beholden to representational systems which means that we are in the business of a circularity we can only represent that which corresponds to our established rules of the game and we can only make a rule in so far at ultimately its aim

is is the application to some stuff in the world so basically we are in a in a kind of kantian prison at this point so carna wants to make a jailbreak out of this kantian prison by reinventing the idea of language as the organ or the organon of play toying around with rules rather than just abiding with them. As such, it comes up with this idea that language is not really about the world. In fact, if you take language as logic about the world,

having some sort of in the last instance subordination in terms of their application to representation of the stuff in the world phenomena so on so forth then that only give you a very very narrow idea of what the world is so what you need to do is to infinitely play with the rules of language, minimal rules of language and logic, diversify them, toy around with them, be like a child philosopher, a child philosopher, like that child who doesn't follow the instructions but just put the pieces together as long as they fit together in the broadest possible sense.

and that's how you make different linguistic structures, logical structures. They have the potentiality, once applied to our representational system, they have the potentiality to enrich the very reality that you can represent. This is different. So here you see that it's not that Carnap idea of playing with rules rather than game, as playing with rules puts Carnap against Kant.

No, no, no, no. I think in that sense Carnap takes the Kantian project to a different level rather than diverging from it. shall I asserts the our new session finally Okay. So, I remember that at the end we were talking about concrete models. We talked about mathematical, computational and concrete models. With regard to concrete models, as Adam said it

all you know they can be model exemplars like fruit fly like some some like a rabbit a population of rabbits so you can study their invasion into an ecology how basically they become like pests and vermin and other kinds of stuff. Now, so this is the three categories that Michael Weisberg puts forward, mathematical, computational and concrete models. With regard to the concrete

models I think that one of the weaknesses of Michael Weisberg is that sure a fruit fly drossophila a rabbit and or ari you know what whatever is or is the model of you know those kinds of that rich people usually put in their living room it's that model of the motion of celestial bodies that's called an orary orary there are cephala so on so forth these are all concrete models essentially they are exempt laws of some sort of concrete structure that we use

in order to study the target phenomena relevant to such models. However, I don't think that all concrete models can be bunched up together as concrete models. In that sense, I would oppose and argue against Weisberg precisely because an orari is not like a drosophila, a fruit fly. You see, the kind of inferences that we garner from these concrete exemplars or models are

fundamentally different. An orary is based not on empirical observation of the model, but based on the laws of Newtonian equations. Whereas a Durosophila is about genetic theory within which we have certain kinds of experimental leeway to kind of compare in analogy the genome of the human or other animals with the fruit flies genetic structure. You see here we are in the business of two kinds of inferences.

is based on established laws that cannot be dismissed like Newtonian equations of motions for celestial bodies and the other one is a genetic theory a genetic theory that just not so strong as the Newtonian equation it's just what you might call to be a platform for our experimentations. If we come with a new empirical evidence within this model, we might actually say that, oh shit, the genetic model that we were talking about might not actually work. We might actually supplant that theory with another theory. So essentially,

even though I would say Michael Weisberg idea of concrete model you know is a nice elegant category but nevertheless this the kind of models he brings under uh the umbrella term concrete model are not the same and i don't think that they should be bunched together precisely because they create fundamental theoretical confusions in so far as they are beholden to different kinds of inferences we derive from such models

Any question? If this is not clear? Joven, you have been so silent today. So has Justin, two troublemakers. What is the first rule you're talking about again? The kind of inference? Establish laws? that established laws so you see and what is an orary so you know that it's like that kind of a sphere and then there is a Kimball around it and then you make the handle and these celestial bodies in your mechanical model go around the sphere so on so forth that's an or ever a kind of a model for

celestial motions. Now this model is absolutely based on the established laws of Newtonian equations of motions for any given celestial body. The model starts from an established law. Literally, no matter what you experiment with this model should abide should abide by the newtonian mechanics it cannot transgress and you won't even notice if they somehow transgress it precisely because the model structure is fully entrenched in newtonian laws now with drosophila the fruit fly things are different somehow sure we are also in the business

of the theory of genetic or the general however this theory is not a law it's more like a thesis in which that if you compare a drosophila with you know these kinds of other organisms human at the moon a giraffe a castle on so forth yes you see similarities of patterns in the genome and genetic structure but let's imagine that you come across a new species the very fact that you come across the new different species is because you can actually compare it with

the drosophila you see that cross the pattern and you can in fact put the anomaly that you have just observed back into the model and then make new conclusions whereas for the Newtonian mechanic model namely or area you can't do this precisely because it's just a set kind of stuff a set kind of universe whereas the Drorosophila is what you might call to be a kind of weak universe that can actually accept some anomalies as a new empirical evidences and can be revised according to that and you can actually revise the genetic theory

whereas for an orary no no worry all it gives you is Newtonian laws of motion celestial motion there wouldn't be any anomaly you wouldn't be even able to detects anomaly with it through that model. So you see the inferences between these two models are fundamentally different. Thanks Rizzo. Anyone heckling? Nothing? So disappointed. Okay.

Let us begin. So, coming back to these three categories, Michael Weisberg then asks, you know, why not more categories? Why not different kinds of classes of models? Why just we need these three trivial classes, mathematical, computation, and concrete? He then goes on to talk about a couple of other models showing that even though these models look different from these

three categories, the reason that they come off as different is not because they are different but because the description and structural models are so vague that they appear as different from these three established models, model categories. And if you were to crystallize the description and structure of the model, it would in fact fall under one of these three categories. For example he talks about this experiment that so you know scientists

start to give people geometrical shapes. These geometrical shapes you know you're familiar with them and then they also give you a different kind of geometrical shape. This other geometrical shape is non-superimposable, rotated, symmetrically rotated version of the first one. And then they ask people, you know, are they identical or not? the majority of the people who say identical they do the hard job of

mentally rotating the second version the second geometrical shape and they see that oh once I for example I rotate the second version to such and such way it will actually superimpose on the first one. They notice that the time that it takes for a person to see two topologically or geometrically varied shapes is a function, is a linear function of the angle of rotation of the objects.

But, as Michael Weisberg talks about it, is that, for example, in such experiments, We are not ever getting details about the exact parameters, but more importantly, about the exact mechanisms responsible for doing the job of identification. What does that mean? so when we have when I have this when I have this lighter and this lighter imagine that is just like this this shape and this shape are not super

imposable now in my mentally I have to rotate to this and bring it to this I don't know what make it how I'm going to rotate it but nevertheless say we I rotate it and then once it becomes this then I can identify these as basically identical now the thing is that and what a model is about is not just about the description of this becoming this but how the second image becomes the first or the first becomes the second or we are into we are actually interested in

the kind of mechanisms that allow us to explain why is that the rotation of this shape, of this geometrical shape, mentally led us to see that it is in fact identical to the very shape that we were given in the first place. They usually these kinds of models don't talk about this. As such any kind of category that they come up with, for example Michael Weisberg talks about

verbal models for precisely for this kind of model, rotational geometrical shape, they are not different models than mathematical, computational and concrete. They are in fact a subset of these models. The reason that they appear as distinct models is precisely because they are vague in terms of the mechanism responsible for them and in terms of the description of how the model mapped to a target system. So Michael Weisberg talks about this that you know a purely variable model such as this is in fact a subset of a computational

model. So for example given such and such neural mechanisms and such shapes under such and such circumstances, there will be a function that maps such an input to such an output. You see, the verbal model in this sense is not a different model, a category of or class of models, it's just simply what you might call to be a vague enunciation of a computational model. And by the way, I forgot to answer one of the greatest things that Adam mentioned with

regard to, I know that I talked a little bit about it with Meredith, with regard to the contrast between mathematical and computational models. As I mentioned, in contrast to Weisberg, I do think that mathematical models are in fact subsets of computational models. Now, there are different ways to actually talk about this, how the mathematical model can be a subset of a computational model. Well, one way to go about it is to understand or to somehow accept the theory that computation

all models are in fact naturalized models as opposed to mathematical models. But then in a sense that for example when you know we are geometrically rotating this shape so such that it becomes identical to the shape that we were given. Well of course we can do it by the way of an algorithm, a very simple algorithm. But then this algorithm based on how we think about the pattern of computation, not computational

model, just computation itself, lead us to a new question that, okay, what if every sentient being implements such algorithms? You see, here computation becomes a paradigm of naturalization. This doesn't mean that computation is essentially paradigm of naturalization. All I am trying to say is that now you see based on your meta theoretic assumptions, it's only based on your meta theory

assumptions that you can see computational modeling to be a subset of mathematical modeling or the other way around. And that's why metatheoretic assumptions are fundamental or are of fundamental importance precisely because if you are dismissing them you can bring fundamentally implicit even sometimes dogmatic assumptions to how you treat a model regardless of its application in fact

Imagine you are some pancomputations, a person who thinks that everything that goes into the nature is a form of computation. Of course, for a person like that, everything else would be a subset to computational processes. But you see, to see mathematics as a subset of computation or computation as a subset of mathematics, it is not given in your model.

It is something that is metatheoretical. need to have a meta model in order to investigate the kind of implicit assumptions you have not only to subordinate math to computation or computation to math but also how you go on and talk about your model as mathematical, computational or concrete versus other categories. So in opposition to Weisberg I would say that these categorical concerns are not all given

in the kind of model that we have studied. In fact categorization of models I would say is a kind of business that is that should be mostly done within the realm of the method meta theoretical meta theoretical or meta modeling because within the realm of modeling you're already have taken for granted certain kinds of criteria versus others and only if you go to the realm of meta theoretic or meta modeling then you see that then you have to be you you will be forced to justify what is that i have taken for example mathematics as a subset of computation or

computation as a subset of mathematical so on so forth questions Yes. So I think I got the point about the metamodel stuff for the sort of embedded assumptions that can come with models or classes of models. I guess I just wanted to throw in one thing I liked about the concrete model description, was some of these concrete models seem to have this, this quality of surprise built into them or like

there's some, so the San Francisco Bay model, right? They built it because they didn't, couldn't describe the way that the system behaved using some theoretical mathematical model. So they built this physical model and that was a lot more informationally rich. And it seems like concrete systems can have that quality a bit more readily than some of these at least more simplified Okay, okay, okay. Professional models. Right. I think, I know where you are heading with it.

But Adam, don't you think, and this is a challenge for you, don't you think the reason that we call concrete models like Drosophila, like the San Francisco water log models, so on so forth, the reason that they appear to us rich is precisely because they are not good models in the sense that they are not really mathematically or computationally models as any model should be. In the sense that they are smuggling pieces of empirical data or phenomena into the model such that the model looks always stable like emergent like almost contingent

but sure from a perspective of philosopher this might look like rich but from a perspective of engineer so oh i don't want this garbage junk data get rid of this they don't want this I would say that here we have a kind of a little bit of a confusion between two pictures of model and we are interpreting differently. So just to add to that. Go on Meredith. You go. I guess I would just, I don't totally disagree with what you're saying actually. but I think I think there is some smuggling going on yes yes but I think

that's kind of like it's it's sort of like is it a bug or a feature type question because it's a bug it's a bug I'm not totally convinced because this bug is good because it lets you smuggle in features you didn't realize were there in the system that you're like you're sort of like formal or theoretical model is not rich enough to capture you can see what it lets you represent them even though you can't articulate them yes but but Adam why do you think that computational mathematical models are bug free you see this is a miracle fallacy that we only say that something is bugged in that

sense that you mentioned when a little bit of pieces of the empirical or the real creeps on our model say oh it's so rich but computational model have the same thing you see there are so many paradigms of computation there are so many patterns of computation modeling some of them actually might creep in our own parochial computational modeling that is also bugged but why is that we are are always on the side of the real. Are we empiricists now all of a sudden? Well, yeah, OK. So that's fair too. I mean, you can get emergent qualities in computational models as well, right? That's one of the reasons they're interesting, right? That's the Schelling thing, for example, as emergent sort of property.

I guess if we think of like this is a toolbox, and you're picking something out of the toolbox to use, like am I going to model this computationally? Am I going to model this in a concrete model? When you're choosing to use more of a concrete model, it's sort of like an admission that there's certain, there's probably something in this space that you're not model in a more normal way. Yeah, it's kind of like making it explicit. Right. So you're like, OK, there's things in here that we want to just import wholesale without fully understanding because we think that will give us a better model based on some other, basically a heuristic type thing. The Army Corps of Engineers had like a century of

Aqua engineering experience or Riverine engineering experience and they just had an institutional learning from that that this was going to be a nightmare. So they're like, okay, let's just build it and see what happens. So I guess that's the sort of defense of concrete models. Yes, yes. No, I understand that. In the concrete model, the idea is that to a great extent, not wholeheartedly, not wholesalely, we are we already know that we are in a messy business like we are basically engineers engineers are the ones who get dirty whereas for the mathematical

computation we say that oh okay we are making idealization but in fact I do think that there is no such thing as an idealization in a in a common sense matter. Idealization in engineering already assumes a lot of messy problems. It's just that they are not explicit. Now here, even though I can go along with you having a background in engineering, taking sides with the concrete model, I would say here is a danger. Precisely because I have seen engineers they think that okay so when we are dealing with like a San Francisco Bay model or a Drosophila or any kind of concrete

models we already know the kind of messy problems we are dealing with now I would say that in so far as these models are not fully tractable by mathematical or computational procedures, we just don't know how far we are in the messy business. To the extent that we don't know how far we are in the messy business of representing a real phenomenon, we can actually make more mistakes than we could do, we could commit those mistakes in computational or mathematical procedures, if they were done correctly.

Yeah, I can see that. So I think we have a new typology of models, that this is wrong. That we should instead have discrete models and continuous models, and that these concrete models are really continuous, are just continuous models. almost like you know the you know discrete models or mathematics plus simulations are sort of like the application of calculus. Right. I would say that is that is completely a fantastic distinction between a continuous model and discrete model. A discrete model would be a kind of a very

you know, you have a discrete mathematical system configuration that is only applied to a very, very restricted aspect of a given phenomenon. And then the continuous one is that it takes into account all the kind of messy stuff that goes between these variables, you know, not just the variables but also the interaction of that phenomenon with its Now I would say that even though the contradistinction between a discrete and a continuous model is important, it is not just about mathematical computational models but also concrete models.

In fact I would say, and that comes back again to the meta-theoretical commitments that you had once you started to make the model. You know, imagine that you make a model based on topological continuities rather than arithmetic or differential discreteness. Imagine you make a model not based on a universal classical Turing machine which is a discrete machine, a discrete algorithm, but something of more like the computational model of today's of how we understand computation, namely interaction between the system and its environment in real time.

Then these continuities, continuity versus discreteness debate holds as much in the concrete realm as it is relevant in mathematical and computational modeling. Questions, questions.

Hey Reza. Please go. You had kind of finished your last little segment of discussion talking about the metatheoretic assumptions involved and whether mathematical is a subset or not of computational models. and it actually was sort of another piece in the text Weisberg's text that I was interested in a little bit. He makes a comment about the ontology of mathematics and says he kind of wants to avoid that topic which I imagine is a huge other conversation but I didn't know if in relationship to what you're

proposing as the different meta theoretic assumptions of why computational would be the broader set with mathematics as a part of it if you could unpack or would care to just maybe contextualize the questions in the ontology of mathematics that he is sort of avoiding in that comment right right well basically today's advance is in logic mathematics and theoretical computer science It gives us a new paradigm of the hierarchy of the objects belonging to these domains, namely mathematics, logic and computation.

And the thing about this hierarchy is that it is no longer a kind of top-down hierarchy as we imagined. You see, when Vienna Circle, like Carnap, started to talk about the idea of a structure and then people started to use his idea of a structure as they interpreted the structure for models, they believed that all mathematical objects can be reduced to logical objects. Okay? This was the core of logicist program. But then we see around the turn of I think as late as 1950s onwards that this is not

really the case. Not all mathematical objects can be reduced to logical connections and objects. Essentially, logic program in its classical sense fails. But then later on, we see that there is in fact a very, very deep correspondence between mathematical structures, logical proofs and connections, and computation, namely programs and algorithms. program and algorithms use synonymously here.

This is what you might call to be more or less the new paradigm of any form of system. Computational Triturianism or Computational Trinity. As Robert Harper has put it succinctly, mathematics, logic and computation are like features of one and the same Holy Spirit. Sure, they are distinct, but nevertheless they are also combined, integrated.

In that sense, logic, logical proofs are what you might call to be the primary objects on the surface, on the uppermost. Then you can reduce them to the mathematical structures. And then you can furthermore reduce these mathematical, reduce by reduce I don't mean it in a kind of naive sense but reduction as enrichment in a kind of a process based, you can reduce mathematical structures to computational processes namely algorithms. And in that sense I would say that every single, what you might call to be a formal system,

can be understood as a subset of a computational process or computational processes. To say something like that means that our metatheoretical assumptions with regard to logic and mathematics, sure, can be addressed by way of further metamathematical and metalogical assumptions but ultimately they should be addressed in terms of computational assumptions. What kind of computational paradigm is it responsible to give to such mathematical structures

or logical connectives and behaviors, so on and so forth. And it is in that sense, I would say, that computational modeling is what you might call to be the generalization of all abstract forms of modeling. But of course, metacomputational assumptions all considered. Insofar as even now, there are open questions, open problems with regard to whether computation is just about a discrete state machine, like a classical church-tiering, you know, state

transitioning machine, where you get discrete set of inputs, you go through time-stamped transitions during which you don't, the machine doesn't accept any input from the environment and then you yield an output or you go with something like a little bit weaker, something like an oracle system where there is a semblance of interaction with the environment or you go into something actually very strong in the sense that the machine the computational machine cannot be defined unless and until it is defined in terms

of its interactions namely confrontation of its actions with the actions of the environment so you see these are all the meta-theoretical assumptions that whatever we do in terms of modeling they are just there in the background it's just that we don't talk about them. Yes, a model doesn't need to talk about them but for a scientist who investigates into the protocols of modeling, these things are important. Precisely because if we change the paradigm, things might be fundamentally different.

Anymore? Okay. I'm so sorry. Anyway, our classes started a little bit late today. Unfortunately, I have to go. I have a commitment at 4.35.

Let's just hold here. As I mentioned, no matter, you know, it might be a little bit late, you know, next session I will continue with the stuff about the kind of what I initially mentioned, the difference between who is a modeler and who is not a modeler. between modeling and direct abstract representation, between modeling and theorization. I will talk about this, and then we move forward. If there is something that is left,

I will definitely make a make-up session at the end. But for this session, I think that even though we didn't really go over anything new, we had you know quite address some important points which I think are important so before signing off who is going to not give me a response because we didn't really go over new materials but like a five minutes barrage of questioning who is going to do that two people volunteers Marie are you trying to become a volunteer I

I could volunteer for next session, but I'm a little bit under the weather and overwhelmed, so I'm possibly in taking today, and I'm mostly agreeing with everything, which makes questions a little bit boring, I guess, where I'm inflamed. Don't worry, don't worry. How about, how about Sean? But next time I can present or speak about something, but it would be nice if it was in the context of a text. Sure, sure, sure. No, next session would be just asking questions. Like five minutes of questions. Like the most stuff that... And you have a week to crystallize your questions.

You know, because I do think that asking questions is far more important than asking them. are you a volunteer or not okay Sean is done Ian GP says he will ask him questions volunteering yeah I will ask questions jump here thank you so much any person we need one more volunteer please before I ask I okay I have a presentation from this week okay Michael Mike Mike Mikey

is going to do also the presentation so that's it problem solved but after that session believe me I won't ask you I will force you I will drag you no matter how much you kick and spit Oh, you know, the problem for me is just the time, actually. No, I know. Yes, no, I have to. It's time to formulate something properly. This is not the school. Nevertheless, this doesn't mean that I'm not going to force you. Because it takes one week to formulate it, this is much more appealing for me, because I actually take some time in order to know what I want to ask, actually. Right, right.

right yes superb excellent i'm pleased uh one of the things is that you know um i know that uh first i noticed that's uh so many of you i i really do apologize should have said this before i only have one email i know that i have 300 emails but i only check one of them it's arnagaristani five six at gmail.com if you have questions please send to that one because I don't check the other emails so that's one second thing is that what was going to say yes okay so if you

have a question, you know, you don't, you know, you just abruptly come across that question. Just post it on the Google Classroom. Even if it's not me answering, someone will definitely answer it at some point. But I will definitely check the Google Classroom. Erhaz, is there a text for next week? No, we are still going with these two texts. Okay. Yeah. Reza is the email that you get. The email that... Yes, that is correct. That is correct. Justin actually posted it. So that means that all the students actually already have that email that you always look at.

Yes, yes. They just need to go to the Hangout link and there they will find your email. so that answers Murray's question yes yes absolutely absolutely and Murray sorry I just I was my apologies if I didn't I was traveling for a long time for talks I came back last night I will answer your emails and your email robot has not me so I put it to rest and not taking personally I see that I have to correct me anyway it was it's as always great to talk to you and please you know if I I don't want you to you know I

know that you are spending more or less massive amount of money ah well let's pretend that it's massive amount of money but it's not massive amount of money compared to academia nevertheless money is money i would like you if you are actually have such a commitment to both spend finances and attention for these classes if you ever find something that wants to be rectified please email me if you don't want to email me say it on the class if you're too shy to say it on class email me please voice your opinion justin no I'm not going to speak about my

book like that book I don't want to even talk about it but yes the book launches Patrick could you please turn off the camera mine or yours the whole life thingy I'll stop the broadcast