 So you have about one hour and then we'll have a discussion. So what I will present is more like a research project. So it's quite broad, it's not very detailed and in fact I'm looking for feedback to know how to make it a bit more precise. So I apologize if it's not a very well worked out presentation. But the general idea, so my project basically is to formalize a pragmatic conception to give an analytic framework for a pragmatic conception of scientific theories. So just to give a bit of background, as you probably know, there have been various ways of understanding what a scientific theory is in philosophy. In the first half of 20th century it was quite usual to think of theories as statements expressed in a theoretical vocabulary of statements speaking about the world. Then in the middle of 20th century people started to say, oh, you know, statements, maybe those are important, but the most important unit in science are models, so we should focus on models and then it became commonplace to consider that scientific theories are families of structures, families of theoretical models. And you have some kind of nice formalism for this kind of view, which is called the semantic view, which was developed by Super, Van Fressen and others. You have a lot of nice formal tools, the model theory, the possible world semantics. The general idea that, so, is that the laws of the theory do not really constitute the theory, they're more like a way of describing this family of structure that is the theory. This family of structures is what is really used by scientists to represent the world. And in this view, there is still some kind of, what we shall call an absolute semantics, so the idea that the interpretation of the theory is not really contextual. Usually you have, first you have the idea that any structure satisfying the laws is a model of the theory. So any, for example, any structure that satisfies the axioms of Newtonian mechanics would be a model of Newtonian mechanics. And you will have this idea that a model is a kind of possible world if a theory is true. So basically, a theory gives you a big set of possible worlds. Okay, and then this idea is a bit disconnected from scientific practice and from what scientists call a model. When scientists use models, they never represent the whole world, they represent a particular system. Well, I will go back to all the differences that we can find, but basically we can say that starting from the 80s maybe with the work of Nancy Katzreich and other people starting to think that maybe theories are nothing without the models, but the models are nothing without the users of the model. And there was a term to practice, the philosophy of science in practice, which is sometimes called a practical term with this idea that to understand what the model means, what's the model represents, you have to know how it's used by its users, you have to take into account informal aspects, contextual aspects, the ideas of this kind already in the work of Thomas Kuhn, notably the focus on models as paradigm and the notion of informal aspect, that building a model is not necessarily like an algorithmic process. You take the theory, you apply it to a case and it's kind of, you have rules to follow, it's more like an informal, it's like Nancy Katzreich, I think, say it's like a nut. Sometimes building a model is a nut. And so these people put forth a certain number of observations about scientific practice. Among them, the idea that models distort theoretical rules, they idealize, they use approximation and so on. Sometimes they even incorporate incompatible theories, so in physics you can have a model of a quantum system that is within a classical environment, an environment that is described in classical physics. And it's not necessarily a problem for scientists to do this kind of, you know, tinkering, yes. So this is an idea put forth by Katzreich, she says the laws of the theory are never applied directly to a concrete system, there are always some steps to apply it. Then you have this idea that models incorporate domain-specific postulates or empirical inputs but also postulate that don't directly derive from the observations. An example of this is the models of superconductions which incorporated some specific postulates that do not strictly derive from the theory of electromagnetism. There's a famous collective book edited by Morgan and Morrison who is dedicated basically to say that models are autonomous entities, they are like mediators between the theory and the applications in the world. Models can caricature the targets, they idealize the target systems and all this is somehow sensitive to the aim of models. So again it's not an algorithmic process and so as I said, model construction is an art and model application, you could also say the same thing, so you have this step of building a theoretical model from the theory and then you have a step of applying a model to an experimental situation, something like that and in experiments too there are informal aspects, how to use the instrument, how to calibrate the instrument and all this is not part of the theory, it's more like a know-how and it's not always strictly formalized. So far so good, we can say that we moved from a syntactic conception, models as a statement, to a semantic conception, sorry theories as a statement, to a semantic conception, theories as families of models and then to a pragmatic conception, theories as models that are part of an ecosystem of users, context, methods, etc. But contrary to the semantic view, there is no analytic framework, there is no standard semantic tools to analyze the content of theories and so my projects basically, the way I presented usually is that I want to make some steps towards building this kind of framework and as I will say later, what my method to do so is to import some tools from the philosophy of language, some from pragmatics in philosophy of language and to transpose all these tools to scientific representation and before to say a bit more about that, I want to give some desiderata for why a possible word semantic isn't fixed, why we cannot really be conservative in this matter and we need something more. So among the problems with the possible word semantics, I take it very strictly as the idea that a model is a possible word if the theory is true, so the theory gives you absolute rules that are applicable everywhere without contextual aspects and so basically the consequence of this view is that any model, any structure that satisfies the laws of the theory is a possible word if the theory is true. And so the problem with this view, this possible word semantics is that in light of all that I've just said, all the observations of pragmatic philosophers, we can conclude that not all models strictly satisfy the laws of the theory because as I just said, sometimes they distort the laws, they caricature the laws, they distort the laws and not all structures satisfying the laws are relevant either because a lot of structures that strictly satisfy the laws of theory are simply irrelevant from the point of view of a practicing scientist. So it's neither necessary nor sufficient to satisfy the laws of the theory to be a relevant scientific model. Then a second problem is that models do not represent complete words, they do not represent the universe as a whole, so a possible word is supposed to represent the universe in all its detail. Every proposition is either true or false in a possible word but that's not true of scientific models, in general they represent a bounded system, a local system and the propositions about what is outside don't matter, so we could say they are neither true or false, they are irrelevant. And even when in cosmology, where models do represent the universe as a whole, well it's a cost-growing representation. The models of cosmology are homogenous, there is this postulate that is often made in cosmology that the universe is homogenous but of course it's not, we can see that there are a lot of homogeneities everywhere. So it's always cost-grained and most of the time it's local and bounded. Then a third problem is that models represent types rather than instances in general, there are exceptions but most of the time models represent a type of phenomena or type of system, it does not represent one particular object of the world. And a fourth issue is that models are often intentional rather than extension. So very often models integrate some, they contain some kind of causal structure and this causal structure is informative not only about what is actual but about what would be the case if we did something else or physics is often made in a state space and usually you have variables and you can say if I did this, if I had measured this for example I would have obtained this and so on. But what do you mean they are intentional or they look intentional? Because it's a desirata, so it's what you want? Oh so this desirata I take it to be what is the case of actual scientific models in the way they are used and the way their content is normally interpreted so maybe we can discuss this but I think that this is the case that in general their content is intentional and it can support counterfactual reasoning for example. You can use a model and say and produce a counterfactual from the world. And last point is that models sometimes I think that sometimes models prescribe rather than merely describe the situation. This is more clear in engineering practice I guess where you can build a model of a bridge and the point of the model is not to describe an actual bridge it's to tell the users what we want to achieve in the world. So there is a prescriptive aspect sometimes. And so I would like a semantics to account for these five features and I think a possible situation semantics can do this but a possible world semantics cannot really do any of this. A small question for clarification so if you say models can prescribe rather than describe it you mean that sometimes they can also prescribe it. It's not their first purpose. It's it sounds like prescribing becomes more important than describing as you formulated as you explained it. You gave some examples sometimes in engineering it can be more prescriptive. Yeah I think we should not assume that they are by default that they describe stuff. Yeah but not prescribe either. Because that might be a model it's almost by different definition something that is what you should what is supposed to be good it's a model behavior it's not. I mean there is a normative. I mean it's debatable and we will enter into details but so a pure case of description would be the model of a remote star. Obviously we cannot act on the remote star then you could say but in order to check whether your model is correct you need to do things here on those two measures stuff. I mean you cannot enter in discussions like that. I don't want to take a strong stance on this because you could say well yes but what you're doing here on this is only true matter so what I'm observing is a real subject matter we'll go to subject matter later. So real subject matter of the model is a real star you do not act on the real star. So I would like to remain I think I think what you wish to say that sometimes they describe sometimes they prescribe they can do both. Okay and then so basically all this the main point of the project is to account for some pragmatic aspects of scientific practice but in a kind of formal way to give some kind of unity to all this work which is done in a bit informal way by philosophers of science. You have a lot of case studies that say oh looks it's interesting but I would like to make some steps towards more of a keen framework to understand all the pragmatics aspect. And luckily I think that there is a precedent in philosophy which is in philosophy of language where almost exactly the same kind of problems occurred like some decennies ago with the philosophy of ordinary language and its opposition to the formal semantics. So just to remind you a bit what's at stake there was all this formal semantics stuff developed by philosophers of language or by Russell, Frager and so on with the more or less explicit objective to put aside the pragmatic features. So you have it's rather clear in Frager that he says yeah ordinary language is complicated you have a lot of contextual stuff but first we develop an idealized view of language purely a contextual formal semantics in terms of correspondence between sentences in the work and then the other is kind of noise that we could it's like a second order of feature or something like that. And then all the philosophers of ordinary language kind of resisted this they wanted to say no context is more important it's more central and they put emphasis on all the contextual aspects of ordinary language such as well just about prescription the performativity of language sometimes we say things you know we give order we make promises and we are acting by speaking speech act theory and so on. The modulation of meaning by context if you say there are many examples one example if you say on my case it's heavy it's not the same as saying my dinner was heavy for the same sense of having one. All this kind of contextual modulations and these people wanted to say that it's central and that you could not first give an absolute theory of meaning and then kind of adapt it to take into account the context but it should be the other way around. You should first account of meaning as used as in terms of the function it plays in context and then you can give a story of kind of more absolute or conventional meaning that's emerged from this. This is for example Grice's theory of meaning who say that meaning is first and foremost an intention to convey beliefs to an audience and then you can account for the absolute meaning the timeless meaning of words in terms of the norms in the linguistic community of the norms regarding how you could normally convey beliefs to an audience. So basically the idea is to take the contextual meaning to be more primitive and then the contextual meaning to be kind of more secondary derivative. Okay so and all these people and eventually all these people were reacting to formal semantics but eventually there were people like Grice or Kaplan or people like this who provided tools for analyzing these aspects in particular indexicality maybe it started with indexicality so the analysis of words like I now this which only I have a reference in context and then it was extended to other stuff and part of it is the notion of situations and the notion and truth and that's then derived to a notion of truth micro semantics or a notion of topics of subject matter and philosophy of language. I will explain a bit more how these in particular are used in philosophy of language because it will be central to my project this notion of topics of subject matter. But for now the idea is well let's let's loop philosophers of language. Let's take all the tools and transpose them to the philosophy of science. And I will be in particular interested in the notion of topics subject matter and why this is so well because it's basically the notion of topics and subject matters are used to account for propositions and additional attitudes so well I will say more about this later but okay but basically so this is my research project for the future that I am trying to make a bit more precise now and part of it was already in my past work. So I started for example in several articles I started taking some the crisis distinction between contextual meaning and conventional meaning I started to apply it to scientific representation, indexicality comes in the same package. And I just want to continue this and also as some of you know in my PhD thesis I talked a lot about situations and as it happened all this stuff about topics and subject matters and there was also talk about situations. So there is a there is a there are affinities between my past work and this stuff so I want now I want to make it more explicit. So I will say a bit more about topics. So white topics from what I've read recently I've read papers and books from Diablo, Fine, and Hopes and I have to read the one from Bertrand that Florian said me and what they have in mine basically is to account for propositional attitudes. So an example of motivation is a case where John said I have chicken wings for dinner and then if you say John said he will have chicken for dinner. This sounds okay because having chicken for dinner is part of having chicken. It's part of what John said. But if you say John said that he will either have chicken wings or chicken breast for dinner. Well no that's not what John said. He said he would have a chicken wing. So what's weird with this example is that in both cases if you take the standard intentional analysis in terms of possible words you will you could be puzzled as to why the first one is correct but not the second one because in each case the second one is a consequence logical consequence of the first one. Having chicken is a consequence of having chicken wings and having either chicken wings or chicken breast is a logical consequence of having chicken wings. And what these people say is that in order to have the right fine bringing for the content of what John said you cannot be satisfied with a pure intentional analysis in terms of possible words. You have to you must bring in something else which is a notion of topic. So according to them the problem is that the second consequence of John had John will have chicken wings changes topics. So it brings in something that is not relevant to the subject matter. Well the first one stays in the topics. That's what they want to say basically. And so they have this theory about topics, linguistic topics and this theory is about a part food relation. Some topic is part of another whether John will have chicken or not is a topic and it's part of the topic of whether John will have chicken wings or chicken breast or something. Okay. And why this is relevant to the philosophy of science because it's about the attitude that users, users of language have to want the sentence that they make. And so the idea is that scientists also have attitudes towards their models. Their models have a purpose. That's exactly what all the pragmatic philosophers say. Models have a purpose. It's important to have the purpose. And so it seems that this notion of topic is really tailor made to address this issue of science too. And then you have some formalizations of topics. Basically one of the first ideas comes from Lewis who say that a topic is a partition of possible worlds. We can take as an example to understand the idea. If you're wondering about how many stars there is in the universe, how many stars there is in the universe, there are several possible answers to this question. There can be zero, one, two, three and so on. And each answer to this question will correspond to a set of possible worlds. The set of possible worlds where there is zero, one, two, three, four stars. And so his idea that if we partition the space of possibilities in terms of how many stars there are, then doing that is exactly presenting the topic of how many stars there are in the universe. So a topic, it can be thought to be equivalent to a question and it can be formalized as a set of answers to this question. So here a partition means it's an exhaustive set of answers. It's either one, two, three stars and so on. But you exhaust the possibilities and every possibility is mutually exclusive. This is what a partition is. Then as far as what I've read so far in the literature, my impression is that Tiablo likes these ideas, but he wants to make it more flexible for ordinary languages. So he says partitions, maybe we should replace them with covers, which are sets of possible worlds that cover a situation of reference, something like that. And this is a way to account for vague answers. So if you want to say give me the number of stars more or less five, you will have a set of possible worlds that overlap. So this is kind of, they have more flexibility, but I'm not very interested. I think it's interesting for accounting for ordinary language, but maybe less so for scientific models, because in general, I think that scientists strive for precision and having exclusive answers. And when they ask questions about them, they want exclusive answers. So I think the partitions would be fine at least at first approach towards giving a semantics for scientific models. Then we have Fine, who wants to get rid of possible worlds. And so he has a completely different approach. He starts from what he calls truth makers, which starts from a real situation from object and properties. And this is a kind of state. And then the state can make true a sentence or make false a sentence. And then he defines topics. And he wants some kind of logical properties for his semantics. So he wants possible states. You can combine the state where something is red and some state where something is not red. You combine the two, it's a new state where something is both red and not red, stuff like that. And again, I don't think it's very relevant to my project, so I'm not very... I mean, it's nice to have this closure maybe from a logical perspective, but I don't really see it. It's very relevant for a project. And then you have Peter Hoax, that I don't know that well what he did, but from what I saw, it's like a bit in between final levels, which could be interesting. But if you have more inputs for me, I would like to take all the inputs that you would have on this literate, but I only started reading it recently. But my idea is to have a kind of leucian picture where you have partitions of possible states for situations. So to take a simple picture first, see where we can go with it. Another aspect of all this literature is that it's focused on philosophical logic and so they want to build subject matters out of the syntax of logical sentences. So they want to say that if A and B, the subject matter of A and B are defined, the subject matter of A and B is the fusion of the subject matter of books. I have nothing against this kind of project, but they're not my focus either because I have the structures of models which are not necessarily expressed in logical sentences. And also a nice thing there are some divergence between philosophers of language on this, but it's not obvious to me that all the subject matter is fully supervised on the sentences. I would like it to maybe to be fixed by a broader context by the aims of modelers on something. So I would want it to be defined independently of the exact sentence that I use or the exact structure of the model that is used. So that's what I say in this part. Yes, that's section 2.3, just repeat what I've just said. So I plan to use partitions of situations. Any questions? Yeah. I did not understand why you ask him for the fine proposition. Why? So what I like about fine is that he doesn't use possible words and I mean I would be ready to use possible words as a kind of tool, like technical tool, I might, but I don't think that they are very relevant in the end. So that's one part I like, but what I don't like is the impossible state. But I thought that was the strength of the proposition of fine. It's the impossible state is to have that. At least for philosophers of language. But I don't see how it would value it. I mean models in science never describe impossible states as some of the possibilities. You have states basically I can't wait for it here. Anyway, so I started thinking in terms of possible words because it's easier, but I'm open to other propositions. But I started thinking in terms of situations as sets of possible words. So for example, a situation where my computer is on a table, in front of me is a set of possible words, where my computer is in front of me on a table. But if we do that, we should not think of possible words as metaphysical entities, such as some kind of logical possibilities. And in the end I just want to use them as kind of tools to describe a system. So I don't think it matters a lot. But the idea is a situation, you can think of it as a set of possible words, or as maybe a set of information about a word, something like that. And then just doing that allows you to have some kind of relations between situations, and notably the fine-braining and cold-braining situations, of situations. If I say that my computer is grey, I'm adding more information into my picture, so I'm restricting the sets of possible words. It's tinier. This is fine-braining. So fine-braining is adding more information. It can be adding temporal information that my computer will be before the ground in two minutes. It can be adding external information that has nothing to do with my computer, that has also a microphone there. Or it can be making more specific the properties of my computer, that it's a kind of, I don't know, antracite grey or something. And cold-braining is exactly the contrary. And then I think it makes sense to describe experimental context in science as finite partitions of the possibility space for a given cause-grain situation. So I will take examples from physics. If people have no better about other fields, I'm open to remarks. It doesn't apply to biology or stuff like that. I'm really to hear that. But the idea that I have from physics is roughly that you experimenters will typically try to measure specific properties, and they expect, a priori, they can expect a range of values, because this is a range that the apparatus is able to detect. And this is, roughly speaking, this is a partition of the possibility space for the system that they are measuring. And you could easily expand this picture for a dynamical situation then you have a set of possible histories, which will be a succession of measurements and outcomes, something like that. So roughly the idea that you have a situation, which is a set of possible words, and this is your experimental situation, and this defines your experimental situations. So if you don't have this, your model simply does not apply. It's kind of the relevant conditions for your experiment is given by a situation of reference. And then you have the topic is completed by a partition of possibility space. This is the question you ask, you want to ask to your system. You want to know whether the electron will go up or down or something like that. And this set space of possibility is a priori, it's a question of relevance again, but which of this possibility will actually occur is not a priori. It depends on the external world, on the object that you are measuring. And then you could say that roughly speaking an applied model will wait, will give some probability whether or not nothing wait maybe to this partition, to the cells of this partition of possibilities. If we say if you do that you will obtain that, if you do that you will obtain something like that. So it gives weight to possibilities. It can give weight to conceivable histories. And in classical mechanics at least, usually the model will exclude some possibilities. Say this cannot happen, it's impossible. If you have a pendulum the model will say if the starting position is there it will behave like this. If we have this history and only one basically will be selected as possible if you give the initial conditions, but you don't have to put in the initial conditions. Then you still have a set of possibilities, but usually it will exclude some possibilities. If you start there you cannot have a... Or just to take a simple example, if you drop the object it won't go down towards you. It's a small question for clarification. So if it's about extruding possibilities or something like that meaning less probable or whatever, why is the weighing on the cells of the partition and not on the possible worlds? Because the model does not discriminate between the possible worlds in the same set. Because the model is only about your system and how it will behave. So if you apply a model to a pendulum the model says nothing about what time it is. It doesn't say the sun is up in the sky or down because your model is about a pendulum. So it gives the same weight to all the possible worlds where the pendulum behaves in a certain way. But it won't discriminate between the possible worlds within a cell. Which would contain much more information, but everything that is the case as well in the universe. But that's the case at the emulator. And then the last aspect of my account, this is stuff that is already more or less worked out this part. In my book for example I present this kind of view more or less. The last aspect is for the instance type that I was mentioning earlier. I think we should consider that an applied model gives weight to the possible world of an experiment or cells of a possible sea space. But a theoretical model is more general. It's a general model that could be applied to any system of the same kind. And so it should be modeled. The semantics of a theoretical model should be something along the line of indexicability. It should be a function from context to applied model. The context is exactly what I've presented. It's a partition of possibility space for a situation. And the theoretical model is a function from this to... So it's basically a function from a partition of possibility space to a weighting of this partition. For any partition give the weighting. If the partition is relevant. Because the model does not apply to anything. And so yes, I said all this is more or less worked out. Recently I published an article that applies these ideas to quantum mechanics. I don't know if I should go into the details now. Because it depends on the time. Ah no, I don't have time. But basically I think it fits pretty well with the structure of models in physics when they are based on state space. At least you can consider that the partition of possibility space is a partition of the state space of the system. And usually when you apply a model to an experimental situation. That's what you need to do. You need to have some kind of partition of the state space. You have to partition it to set that correspond to the possible outcomes of your experiment. So I think it works pretty well at least for the model of the state space. And I applied it to quantum mechanics. And it corresponds quite exactly to the consistent history approach to quantum mechanics. Which is nice, but it gives a different interpretation. Another aspect that is important is that quantum mechanics will put constraints on the structure of the context. So that's an interesting aspect. There is a consistency condition. For example, you know that in quantum mechanics you cannot measure velocity and position with an infinite precision at the same time for a system. This corresponds to incompatible contexts of applications that are not consistent together. So the theory looks like a kind of theory that if you apply this picture in terms of context you arrive at that quantum theory is roughly a theory that gives you consistency conditions about the context. And in this article I defend that you should reify contexts. So this is something that is part of my project. That is... You are a cryptorist. Assuming you are a cryptorist. I always knew. I will assume being a cryptorist and say, yes, contexts are real. They are not in the head. And that's a kind of... I think that that makes a difference between this project and what philosophers of language are saying, because they would be tempted to say that a topic, subject matter, are something like in the head of the law of controls. We can discuss that as well. If you think it's crazy, I don't know. But the reason is first quantum mechanics. Because if you have this consistency condition that is not a priority, you could say, well, that means that there is something to say. I have a kind of proto-argument for this. If you have this... Like the set of possible... You have a possible world where you have an infinite precision and an infinite velocity for a system. This is part of the possible world semantics. You make every probability very precise. But this possible world will be part of two possible situations. But in some sense it's impossible. There's no... Quantum mechanics would exclude thinking of this as consistent situations. But it's still part of two situations that are acceptable. So this is an argument to say that situations are more primitive than possible worlds. And then that we should have some... We should reify the situation and the context. I'm not interested in this argument. You can have a look at my paper. Consistent histories through pragmatic lenses, which was published this year. And the second reason that I like to reify situations and context is the prescriptive aspect. I think if we want to take it seriously, we say that models can prescribe. And if you want to say that the context are not in the head, but they correspond to an experimental situation that is implemented. It is actively implemented by experimenters. And this gives another reason to reify context. Some kind of stuff in the world that is actively implemented. And then what we have is roughly the idea... The physical picture is that in the world there are what I call six situations. These situations with a partitioning of probability space. And with weights given to them. Dispositions associated with the cells of the partition. There are six situations in the world. There are contexts. But the theories of physics or of science in general do not tell us which situations in the world exist or not. Only direct experiments tells us that there is a situation of this kind with this context. The theory only tells us that if you have this situation and this partitioning of possibility space, then you can attribute these dispositions to this situation. So this is the idea that a theoretical model is a function from context to weighting of the context. We do not know which context exists in the world. It's something that only experiments tells us. We don't know it from the theory. But the theory can tell us from any context that we are interested in and that we know exists or we suppose exists. We can attribute this position to it. This is the general picture. And if we accept this crypto-realist picture, then we address all the decisions that I mentioned earlier. So we have an autonomy of models. I say it of course, but I don't know if it's of course. I don't know why I wrote that. Models are autonomous from theories. Well, I didn't really talk about theories. So maybe it's not clear why, but there's no reason to assume that they cannot be autonomous because they apply to bounded situations. Models do not represent the whole universe. They apply to contexts which are situations with partitions of possibility space. Models represent types and not instances. I just explained that theoretical models are functions from context to weighting of the context. Models are intentional. They give weights to possibilities for a situation. And models can prescribe if you take that context and are known by having some kind of intentional state towards them, which is an epistemic limitation that we only know by intentionally, by being intentionally related to a context we know that it exists. I don't mean that contexts do not exist outside of our intentionality. I just mean that it's an limitation that we have. Then yes, models can somehow prescribe the sense that they have relevant conditions that are distinct from their accuracy conditions and the relevant conditions can be prescriptive. You should implement this situation and this will be the case. And then, so I didn't go into the formal stuff that is part of what I wanted to discuss this week, but time was fast. Thank you very much. Maybe I can... I proposed a take five minutes break and then let's have a discussion. Good. So let's start again with a discussion. Thanks a lot for the talk. Okay, we will have a discussion now. Questions. I have a question. I wonder about how do you feel what would be your answer with this philosophy? At some time, especially at the very beginning of the presentation, I had the feeling it was quite formal. So of course you describe all models and the positioning of pragmatism and the position of all that, but eventually if you say it's incompatible model, what does it mean when you study at physical and counter-objects that position and speed are incompatible models? That's okay. What's your interpretation as to what? I guess the answer will be related to... your crypto-realism. What's the sign? It's going to stay that way. So if you apply what I described to quantum mechanics, you get something like the consistent histories approach and graphists who introduced this interpretation talked about framework, which is, roughly, a partition of possibilities for a dynamical situation. So a set of histories for the system. And there is a condition that they must be consistent. The histories must be consistent. You cannot have a story with a particle as a precise position This is a way to avoid all the problems you have for the interpretation of quantum mechanics, the way it makes things... It's almost like a classical model in the end if you respect this consistency conditions. So I guess your question is how I interpret this condition? Yeah, I mean, if it's the classical model, if it's coconut interpretation, there is no interpretation. So graphists, when he introduced the consistent histories and all the ones who had similar views, they were saying it's a modeling choice. So according to them, basically what graphist says, ah, measurement is not a physical term. We should avoid talking about measurements, measurement outcomes. It's not proper term for physical theory because we are released. And then he says, for this reason, I will take the choice of the framework to be just a pragmatic choice from the model. But then this creates a lot of difficulties and everybody criticizes the consistent histories specifically for this reason. Because if it's arbitrary, if you have an arbitrary choice at the start of building your model, then your model is not realism after all. They cannot say, ah, I'm realist, so I won't talk about measurement, your model is completely dependent on an arbitrary choice at the start. And so it was criticized for this reason and also because you cannot account for the classicality of the model. You have a lot of criticism of consistent histories which say you cannot explain why the world, the macroscopic world is classical or almost classical because saying it's classical depends on the choice of a framework that is classical, but it's an arbitrary choice. And you have other frameworks that work just as well that are not classical at all. Even if you have a model of a full universe, because some are still working on this idea of a model of the universe, even if it's classical up to now, your framework can be completely non-classical in the future. So whatever constraint you put in from the past, you still have... it still depends on the choice from the modeler that the world is classical. A lot of people say now it doesn't make sense, we know that the world is classical, that it will continue to be classical in the future, it won't have quantum effects everywhere. So the consistent histories approach is bad. And my interpretation is different and that's what basically the topic of my paper is to say that you can apply this interpretation but the framework should not be interpreted as a modeling choice. You should go more pragmatist, more neo-continagian and say that the framework corresponds to real situations. That's part of the reason to refile the context. That you should say that you should refile this context and say it's not a modeling choice, it depends on the empirical inputs from the situation to which you want to apply the model. And then the problem of classicality, well, I have other responses to say why we should expect the world to continue to be classical but basically what I'm saying is that it's not a job of quantum theory to explain classicality. Alright, if I understand the way, it means that you are not free to choose the model and so according to you, what's a good model for quantum physics, for instance? If I understand you, you are not free to choose any model. Yes. So what would be the right model for this model? It's a matter of coordination between your theory and the experimental situation to which you apply the situation. So you have some rules that come from experimental practice, mostly, that tells you that if you have measurement apparatus this way, this way, this is this kind of situation and so you should apply this kind of model. In quantum physics, what kind of answer will you suggest? Or maybe you don't want to be committed. Answer to what? For instance, about the measurement, about those quantum particles that exist or something like this. Sorry? I mean, is it true that measurement exists? Is it true that quantum particles exist in some way? That kind of answer. If you choose a model, you answer this question. Yes. Many of you don't want to commit. I say, yeah, basically if you want to go realist but I don't think I'm really a realist but it's more about that I'm a bit tired of defending non-realisms or if people want to. So I just talk as a realist. But basically, if you want to be a realist you can be realist about applied models. That's my main approach. You can be a realist about applied models. You have an applied model with a framework. You have an applied model of quantum mechanics with a framework. You can be realist about the content of the model. There's a way that you have these possible stories. They are consistent. You have these properties of systems and you can be realist about that. I wouldn't put the metaphysical weight of realism but that's another topic. For instance, the question is is it possible to have backward position that can be a good question in quantum physics? But if I understand you you are more providing a way to think than a specific answer in the... Yeah, I don't think we... I think you will never get backward position because that would mean that you could intervene on the future and change the past that won't happen. I have two follow-ups already to your questions. It's just a follow-up. So it's two follow-ups. After that I will end about that. You could have inconsistent history, backward causation, if the fact in the past has never been measured directly at all. But if you have in a way associated to a state in the past where the value is... You can... Maybe, but I don't see any good reason to put this past event into your framework, into your possibilities because it's in the past, it hasn't been measured. Why do you need that? If you want to model the quantum eraser experiment, for example. But that's because you want to know something about the past event that has not been measured. Exactly. But then you are applying a framework that is not strictly speaking. You introduce in your framework a distinction between two possibilities that are not directly measured and so you could wonder if this framework corresponds to a real context because I want to verify the context of applications. And since it hasn't been measured you could say, well, no, it's not a real frame. Well, it's metaphysics. But you have constraint on it. You have constraint on it because why you talk about the result is because there's some efficiency to this non-measure qubit, for example. But maybe there's not one way in conscious of history to build it. Yes. Because I would tend to think that roughly speaking what framework is or what the context is this partition of possibilities space it corresponds to the information, it's an intuitive way of understanding what it is it corresponds to the information that is made available by the system to the outside world. Okay. And then this event in the past you could discuss is this information made available? It's not available yet. Yeah. But, yeah. I have an other follow-up but so your argument that quantum mechanics force you to reify integration will not apply to classical mechanics so not as strongly so my question is that what is specific in quantum mechanics give you good arguments to reify? Is it the bizarre way in the Hilbert space or is it the fact that quantum mechanics has a very strange way to talk about experiments or at least quantum mechanics contrary to classical mechanics or experiments implicitly what part is doing the work? Yeah, I would say that it's all the weirdness of theories that in general that forces you to think this way and to think that actually it was already the case in classical physics but it's just that information is so much available so practical information at that it's not easy to detect the fact that it's important to reify the framework but, yeah but the main argument is that is that you have to choose a fine-graining and in classical physics the fine-graining you can choose the absolute fine-graining for everything infinitely precise position and velocities and a continuous space and there's no problem in doing that but in quantum mechanics you have to make a choice you have to say when you refine your framework you start from a most grand framework I measure this code, these macroscopic properties and you want to refine it as if you are postulating about what is more precise what exists more precisely in the world if you're making this kind of postulates at some point you have to choose how you will refine your framework and you cannot do both all refining because they end up incompatible in classical physics you have no problem with this kind so that's the reason that I need to answer the question Questions? Good. I have two questions one which is graphical, right? I'm sorry, in terms of exactly what's the role of the initial possible worlds too because generally quickly you get rid of them and you just take a refining situation and you get rid of the possible world so why do you start by taking a possible world just to do the best way possible? Because I think that this point we can visualize the motion of fine-graining cause-graining like it's a restriction of possible worlds and so on so you have a nice way of picturing the relation between situations but then if I could just retain the structure of situations and throw in the which means possible worlds I would be happy with this Yeah, because I mean I suppose that if you are a semantic you have theories that have kept the possible worlds as a thing that's structured all the situations are structured by the possible worlds the theory the theory and so we see that as a question apparently going back to what he said earlier about the inconsistency of internal consistency of model you gave a nice symbol of semi-classical with the semi-classical models of direct learning systems so why do you just get the don't-go best way for your for a lot of drastically very good inconsistency models that like a quantum system in a classical environment because I think is that if you apply the framework in the end because the series that's because I didn't talk about series presentation but the series basically picture will be very pragmatic that they are instrumental for building models that work and that's there are some leeway in the way they are applied so it doesn't only matter if you take them as tools to build a model in the end what you have is a kind of structure of causal structure of events you will end up with something like that like a cheat no it's just not being very risked about series so it doesn't matter you don't have to be risked about to see that you have a constant hypothesis in your model model works as a real tool but they are inconsistent if they are taken to be both everything in the universe but if you there is no inconsistency in describing a quantum system in a classical environment if you could say if you are a crazy metaphysicist you would say my system is the only system in the world that is quantum ontology and all the rest of the universe has a classical ontology and I don't see any inconsistency you just put in arbitrary considerations of the domain of application of your series you say the external world obeys to these laws everything except my system and my system obeys to quantum laws and as a quantum ontology but only my system and everywhere else everything is classical if you think it's obviously I don't think it makes sense to claim that seriously I don't but I don't need so the end product of your process does not rest on formal inconsistency yet if you are instrumentalist about series it's kind of stuff that leads to frigate laws to be more futuristic because they have a lot of models that you know are non-possible and stuff that still works in practice but then again the main point is being realist about applied models and so if your theoretical model is inconsistent it doesn't matter you take it as a function from context to applied model is just a waiting of possibilities and I don't see any inconsistency in this waiting of possibilities and so I don't need the inconsistent states and stuff like that but yeah I see how you resist that there's in the model there should be no contradiction or something that will make the model explode or okay but my impression is that your position force you to adapt to really separate quantum from classical or semi classical to be separated from quantum to have to represent them as different modelization a priori that they could exist independently because they strictly apply a certain context and they cut the context blah blah blah and so in that form you lose something about the heuristic that I don't know when one of the more spectacular of course is that you use, you say that particles are a quantum but the electromagnetic field is classical and you have the Ahranov-Bohm effect and you have you discover a lot of stuff and you can modernize a lot of phenomena new phenomena and after that you say I learned something about quantum quantum world new stuff because it's not classical and the effect is not cannot exist in classical but the Ahranov-Bohm effect is modernized using non-quantum photons and my impression is that you would say okay in that context it's applied blah blah blah but of course it's not quantum quantum mechanics is another context because the experiments are different and the auto-consistency is different so my impression is that your formalism force you to separate these case much more than you would like heristically, pragmatically or am I mistaken? I don't know I'm not sure I understand what the problem is but I can say one thing I'm not sure it will answer the question but one nice thing about the consistency studies framework is that you have a nice transition between quantum and classical case so the idea that some frameworks are classical which means that they are arbitrarily close from a classical framework and if you apply this kind of framework the resulting model is exactly what you get with a classical model so in the context where classical mechanics apply if you apply a framework where you are interested in the position of a particle you will get something that looks like events that follow the classical trajectory if you apply a classical framework so there is a smooth transition between the series and maybe it doesn't answer completely the kind of worry you have but I think maybe if I understand better I think that would be maybe a pass to argue that in the end in the end everything is kind of effective applied model is kind of effective and you get in many contexts you get the same effects from classical and quantum theories and in some contexts no quantum models are very different predictions but then you should apply quantum models I don't know it does not answer but I will if Peter allow me are there other questions just go ahead so it does not answer but it's logical your answer is logical near framework because your framework is driven by experiments it's experimental context that fix the reference and example that are really important in the history of physics like discovery of the atom of bone effect that it's discussion about theoretical theory and you don't do you don't do an approximation to the classical you just take part of the classical electromagnetism classical take part of quantum mechanics pure quantum mechanics do something bizarre together fits it in the world and experiments win a Nobel Prize you know and at the end they say we learn something the atom of bone effect exists in the world as a quantum effect not as a classical one because it's excluded from the classical but of course we don't know because we don't have a let's imagine that impossible or almost impossible to do it in pure quantum field theory we're just we can't visit it's works because we have this are you making kind of non-miracle arguments that it works so the series has been I just want to see that these theoretical context could they be implied that there's some indexality there too but in the theoretical side I don't know because to me it looks like more a question of epistemology that we have this theoretical heuristics and they work and you build models and they make predictions that you are not expecting so the theory works as a good guide to build new models that were never applied before and that eventually correspond to the experimental to what you get with experiments so to me it sounds more like some kind of epistemological question how is it that theories work so well or something like that so your question so you would say how the fact that we confirm the existence there are a number of problems why does it, why people believes they learn something about the quantum pure quantum and that would be a good question yes but then the question answer would be epistemological and this my thesis are more semantics and I don't I'm not sure there is really a clash between my semantics and the kind of story maybe maybe this semantics is too your point would be my semantics my way of interpreting models and theories to deflationary and so you cannot explain this why theories are such good guides or something like that and then yes it's a fair question and I should come up with epistemological answers but to make my semantics compatible with sorry enough and I'll give myself the words so I have quite a lot of questions there's all things around here to me at the same time I feel quite promising seems quite promising what you're proposing and I've had that feeling already since the PhD but there's some things I'm confused about and maybe this is all because of course I'm coming from the positive language point of view and maybe putting too much of that context into what you make of it so you like switch between possible context, possible situation possible fixed situation and there's like the thing that is possible is sometimes context, sometimes situation I'm a bit confused why at least in Kaplan's picture Kaplan's picture context is quite different from possible circumstances or possible situations or something like that and it is quite important different so context is all about the performance itself when something is set or written who is it written by, who is written to what are the intentions of the people involved what is the specific locust time and place of it happening and so on what are the kind of topics they are discussing and all these matters that are about the specific act itself, the specific sentence not sentence but the performance of the sentence while then the thing you call possible or at least in Kaplan's picture where the modality comes in is in a completely different level where you're going to evaluate what has been said so you first I can do that independently so there can be something that Newton has said long time ago he said something I believe that blah blah blah and of course we have to know the I and the context of what all these words and the amount and so on to know the kind of proposition that he tried to utter and as soon as we have a proposition by knowing enough about the context we can evaluate the proposition in different possible circumstances or impossible circumstances we can say well maybe the person who uttered it was true but we can imagine we can find ourselves that it's not true we can imagine a world where it would be false we can verify the situations for evaluation so I'm a bit wondering where your situations lie are they on the evaluative sites or more in the the performance of the experiment and do you want to make the distinction between the two or do you want to be like is this for your specific purpose not an interesting distinction to make context of evaluation and context of yes it's a good question yeah maybe I'm not familiar enough with with these distinctions between context of evaluation and context of maybe a good way to call it but it doesn't require much for familiarity my question is not meant to be deep just about do you care at all about utterances or something are you interested in the experiment things that are actually happening with people involved and so on and research questions and these things and do you think we can maybe evaluate science try to still see possible ways to evaluate what comes out of the experiment without like already after having fixed a context so can you make this distinction my idea would be that there is I think what I'm talking about is more like the equivalent of a context of utterance so there is a like when I say relevance conditions where people are speaking they're speaking about a specific situation with some stuff that is there that must be there if the model has meaning at all if it is to apply to something it must be there and then you can come up with a story about how to evaluate the model but it's like a different story that I didn't know about but you could say that if the model says okay so the context is that you have this situation and you have three possibilities and the models say the first one is impossible it must be one or two or three then you can evaluate you say oh imagine the first one happens then the model is is a bad model imagine two or three of them and the model is still adequate to have this story then maybe to evaluate the empirical adequacy of the model but what I'm trying to formalize is more the context of utterance I guess okay just want to follow up but you would hope that in science there's more transparency between you know utterance and the reception at least you it's not like the example I gave Newton said I blah blah blah it's in the past I don't know if there's a Newton if there's every someone I have the person receiving this proposition has to do all this evaluation of contextual information I believe that you would hope that there's a lot of work at the beginning of the thing the utterance context let's say production of stuff that would be almost transparently interpreted by people around hopefully it's probably false but I mean it's a bit, part of the protest is idealistic and I don't want to because I don't think that we can claim to provide a framework that would be describing exactly everything that happens but part of it is admittedly an idealization of what happens and but yeah another question I'd have is unless again please interrupt I don't want to monopolize the rest of the discussion another question I would have is about these weights and weighing I really don't see the intuition there accept if it's some kind of a weight just have a smaller situation you call it fine graining but I don't think that's actually a very good word because adding information is not adding grade necessarily, grade you know it's because the word is used a lot in this literature and then specifically the kind of thing you're interested in and having models that can distinguish more things it's just like the lens has become more detailed that's a fine grain while in your case if it's just if the situation becomes smaller it's not that you have all of a sudden more detail it's just that you have shrink your possibilities it's more information but it's another kind of so I don't think the term is very practically choosing if I can I think I agree but maybe for it's more adequate for when it comes to partitions well yeah, then it might be useful partition can be more fine grained or coarse grained but that's not to have you have to put it in here I believe you said you spoke about fine graining and coarse graining and the size of the situation what sets a possible world but there's this idea that if you describe the condition of application of your model without describing the outcome it's kind of a blurred view of the situation you say there's a pendulum I'm not seeing what it's doing but there's a pendulum and it will have some that's it and then fine graining could be how this pendulum is acting this way precisely so that's the idea maybe that's just getting more information about something and it's not necessarily going into more detail at least in this if you refer to that literature then maybe talking about purely how detailed the partition can go that could be a fine graining in the more traditional sense of the word but that's just a frankly suggestion but I was actually wondering about the weighing so of course I can see that sometimes it's useful to or you can apply a model and then your situation gets even smaller because it's not the general thing but more specific but a weighing suggests that you can also have like other probabilities or something and then I just don't could you give a simple example not in physics if possible that could capture the idea why application would be a form of weighing I don't see the link between applying and weighing no so the weighing these are the predictions of your applied model but an applied model was if I don't make a mistake it's a situation plus so for example I should find an example it's not from physics what you said it's applying models weighing the cells of the partition of context of application so you have you understand space but it's not in physics it's in the atmosphere you have the temperature and the pressure of the atmosphere okay then I understand the pressure and temperature that amount of physics I can handle well your model says the atmosphere is such and such you have such and such constraint that comes from the sun and everything the earth etc which means that if your system is there at 0 it should do something like that I invent something trajectory in space space should be so the temperature will rise and then it will stay stagnant and the pressure will rise but then you plan to experiment on your atmosphere but you will just measure with the final precision so I mean you are a pilot and precision in range so basically you have a kind of partitioning of your space space so this is your experimental situation in this case it's just an observation situation I use experimental situation in general way just observing is experimenting and what your model would say in order to plan your model you will project the theoretical model which is the continuous trajectory this is a theoretical trajectory I'm not saying that the atmosphere really follows this continuous trajectory this is what the theory says about this kind of system and if you apply it you will see that basically you have to project the continuous trajectory onto your partition which means that basically this will be your applied model will say these cells have probability you could index in time but let's simplify these cells can happen there are possibilities and this one cannot happen one path this one can happen this one for example cannot happen your system will never be found in this cell and maybe if the trajectory spends more time here it will give a higher probability way to this cell and if it's just there's a tiny perception it will allow probability here for example that could be the idea it's an intuitive idea and so what your applying your model does is putting weights here it's almost a 1 or 0 but it could be a probability in some cases to the cells of your context and saying this is a permitted state this is an impossible state a forbidden state forbidden by the theory conditional to t equal 0 to initial condition but then yes but if you don't want to specify what the history is ok so the the possible outcomes you had in the beginning before the weighing mean nothing at all that just whatever is it's how the variables work or something the world does not have no impact on you mean to continue trajectory or the cells the theory of the weighing the picture before the cells or the content of the cells just the the the plane itself is without restriction the modality comes in at the weighing parts the theory only has influence I mean are the no if you go to mechanics you want you want to say this black stuff it should be consistent so then that's the specificity of quantum theory that even the grid has the theory that some grids are consistent but you shouldn't apply this grid so if you wanted to go infinitely precise in both if you wanted to apply an infinitely precise grid just take the full state space on the continuum that would be forbidden by the theory for example but that's a very specific situation in general but you want to have the idea you have behind these state spaces or possible worlds they're not possible in any kind of deep sense it's about purely theoretical constraints doesn't say anything about the laws of nature or whatever it's just what you conceive of as possible as mathematically conceivable kind of but with the caveat from mechanics yes you could take it like that but the point is that I would like to verify this in the end so I would like to say we know that this grid exists because we know we are capable of measuring it because we will make this measure and we're confident that we will be able to do it so we are confident that the grid is somehow real and we can postulate that final grid also because information goes from the atmosphere that if we had more capacities we could refine the grid to a point but this kind of influence that final grid exists out there in principle well the idea that from quantum mechanics we should be we should think that maybe it's not true that it's a postulate it's not necessarily the case that infinitely precise grid on the system is actual so I want to refine the grid as well I don't really see what that means to refine it I mean where does it live qua object I mean it's and you know my whole problem maybe is related to the fact that I see these topics or these partitions you talk about as questions as we talked about one day as Lu is also interested in them and I think this is really interesting for exactly to capture the sort of experimental context so you have a context in which you have certain information right among a group of researchers and they ask a question to the world and so it is indeed very important to understand correctly what goes on in such experiments if we can have a theory of questions in there but the question is what we do right but there could be restrictions on the questions we can ask that I find very natural and that comes from the world but it's still something else then so the partition is the question but there is maybe some partitions only that can be possible questions due to the restrictions in which we live but then if I wanted to do some it's because in the end I'm going a bit towards making a small step towards metaphysics and I wanted to do some kind of metaphysical job as well and that's what I think as far as I understand that's what you criticize so I can what I can do is just give an intuitive picture that there's a question and we ask and then we could say we restrict our view of reality to the question we actually ask to the world we cannot say that only that this exists so we cannot say that when we look at the world it doesn't exist because we're not looking we're not asking questions to the world by looking at things it's a bit too radical to relativistic to think that and so let's make a step towards metaphysics which is I think remains common-sensical and we have a common sense and assume that things still exist when we're not looking and so this implies somehow that there are questions asked by nature to itself, by part of nature those are part of nature that line in the interaction between the between stuff and that we do not necessarily know much about what these questions are but we could kind of postulate that there is a state that if there's a state of the matter whether we will measure this but there is a state of matter about whether it's in this self or this self so in principle we could refine and our analytical model saying what can inform us about this situation that we're not directly this question that we're not asking directly but after me it still informs inform us about it and the fact that there is an answer means that somehow the environment of the atmosphere asks this question to the atmosphere there's information that goes from the atmosphere to the outside world or something like that yeah I understand it's a bit it's not very precise way of talking but there's a there's a person a fine man who was talking about it I am but the question is do we need it we already have time to talk about it it talks about questions as to how do we but do we need it that's a good question because I mean the kind of thing itself yeah if you make this distinction between the partition which is maybe up to us to just be interested in a certain partition at some point but there might be restrictions on possible partitions yes and that might be due to the world or to our restrictions as people but the question doesn't have to happen in the independent existence or something like that for it to be a possible question we don't have to say that because of quantum mechanics you want to say at least that if this question exists then another question cannot exist in the world you want to say something like that yeah so that's a restriction on what possible questions can be asked which is fair enough there's nothing contradictory for me against that but saying that the question is already out there I don't know that seems like turning around but it's just when you say information transfer between cognitive stuff what do you mean it's saying that there are questions out there it's just metaphorical way of speaking but I think that yeah we don't need it strictly speaking and the only reason to accept that is just the kind of commonsense view that I was expressing that we believe that there are still some stuff out there when we are not looking and that is kind of structural I mean you cannot be some kind of canthian or something and say no the question are just the questions that we ask and when we are not looking at the world the universe is kind of blown with no structure I don't know I mean that's you can say that I can say that I mean if you assume that there is a fact of the matter even if no one knows that there is a fact of the matter of what lies and I know in a dark room here at the university there is an object on the table no one is looking at it but you assume that the object is still there and it's kind of a structural way of speaking it's not like it's kind of a global etc so we if you have this minimalist commonsense realism I would say maybe then it's tantamount to saying that there is an answer to the question there is an actual answer to the question is the object there on the table yeah sure I don't deny that but that doesn't mean I mean we are the question so the question is supervene on the fact that there is an answer no that's weird that's weird the reason that's weird to say that the question supervene and has some common ontological if you don't accept that you need to accept that there is an infinitely fine grain question that determines what there is in the world I want to say that whether there is an object no this fine grain is a question that you won't have any answer because it's a bad question that's all doesn't mean the question exists supervene it exists in principle in the sense that there is an answer to it in the sense that there is information about this stuff that goes from the situation in the the room where the world is looking there is a fact of the matter which means that there is some kind of information about it that goes from this situation to the outside world somehow and this can be codified or formalized by a kind of partition of possibilities and if you're not doing that I don't know how you can express that but it will commit you to things like that there is a fact of the matter if you don't say about what there is a fact of the matter how can you say there is a fact of the matter but it will commit you to buy things like the electron and the hydrogen atoms is not at the same location as it's the core because if it was if he was asking if it was asking a question about its position it would be it would have to force a variability on the speed and therefore it would just get away from the center and that's exactly like Feynman is talking there was a two measurement between the electrons and the proton and the hydrogen atoms like they were cognitive agents asking questions each other so it's metaphorical they're not cognitive agents I'm not saying that everything is a cognitive agent but if it's a metaphorical why do you not buy the very sensible so again let's take an electron then if there's a fact of the matter whether one of the electron here at this table has we know that it cannot have a very precise velocity but there is a fact imagine there is a fact of the matter as a matter of fact it has a very precise position and not a very precise velocity then you must say that there is some kind of but first we cannot measure that does it mean it has not you know according to bone maybe you would say as I said you can say the refining stops there there is no fact of the matter about the electron because we cannot measure it but then you have much more radically idealist that I would like to be I would like to follow common sense and say that this fact of matters exists somehow but I don't want to go as far as saying that they are ultimately precise and described by possible words so I must say that there are situations out there ok thank you very much