Structural Realism & Related Topics

0:00 What do you think is the aim of what is called localized realism, a very strong version of scientific realism? The position aims to defend realism against physical combustion, which of course we know is a burden for scientific realism. It began by Hitcher and Morrow and was very well developed by the series. The aim is to overcome the conclusion of pessimistic induction, which says that since we have, and we actually have, scientific theories in the past that were successful, but from the point of view of our current science. If there are faults, or at least not approximation 2, then that means that success doesn't signify approximation 2, that success can signify also faults, which means that realism will have great trouble. So, most of the realist literature in the past 15 years is attempting to discredit this industry. Euclidian Realism says that even if old theories were not near true, they at least contained some approximately true components. And these approximately true components were behind the success of the old theories.

2:30 There is more to mathematics than that, and we'll find out why. The basic tenets of localized theories is that a scientific reality is not obligated to accept a whole theory or all theories or all components of a theory, and that a theory should be fragmented into its basic structures, and there will be some idle assumptions. Or what each of all three propositions or preposites and some approximately two assumptions and some perhaps deflected assumptions. But as long as there is an approximately two assumptions that pose a theory to succeed, that's okay. And we are entitled only to accept that before the ideas and can ignore the ideas. But the problem now is how to decide which... This element in a theory is a genetic tool because when a theory is confirmed, the prediction which is confirmed, the confirmation extends to every part of the theory. You can say that it confirms this part or that part or this part or that part. It is not easy, practically speaking. However, Salus found an important way to answer that question. He acknowledged the difficulty of determining which part of the theory is true based on the information of the theory. He said we can know from the attitudes of the scientists. The scientists who constructed the theory do know which part was suspect, which part... I have a problem with that because it's difficult to find documented attitudes of scientists all the time because they are not interviewed by CNN and not all of them write their autobiography.

5:00 It's true that Black, who proposed the theory of caloric, which Sirius uses it as an example, had suspicion about the theory of caloric, but he also said it is most probable all theories are suspect. And some, or many physicists, scientists, talked about theories in general, their theories in general, their attitude to their theory in general, but not particular theories. However, saying that the mental attitude, the psychological status of a scientist correlates with the epistemic status of a proposition leads to arguments which Sylvius doesn't provide. There are other problems with this structure also, which is that scientists are human beings like everyone here in this room, and they can have idiosyncratic... There are many opinions, wrong opinions, and some of them have a changeable attitude to their students. For example, Einstein in 1905 presented a table of light quanta, and he was very happy about it, and when all other physicists opposed him, he changed his mind and became ashamed of it. Actually, it threatened his membership in the academy of science. So he withdrew his enthusiasm, but about 20 years later, when Colton's experience was done and proved that light is a photon rather than waves, he accepted it and objected to it. Planck did the same thing after his black body theory was proposed and was found in Solvay's conference that contradicts Maxwell's. He tried to reconcile both theories and to change his original theory, he changed his attitude.

7:30 But then the quantum physics didn't work out. He had trouble with his own ideas. And this was the fighting of his own for the first time. And he was a very great scientist. But when it came to the progesterone, we just, we couldn't accept the progesterone at all. We kept fighting for it at insurgency and PPI. So the attitudes may not correlate with the extremities. And then the idea of idle assumptions. We call assumptions idle when they don't make predictions. However, many assumptions didn't contribute or participate in making predictions for five years, six years, ten years, and later they made predictions. Should we be jumping into conclusions for an assumption that hasn't yet given a chance to make a prediction? I do not exclude it. We should call it potentially productive. Meanwhile, I don't think that making a prediction is the only useful thing for assumptions. Facilitator inferences, others, are very important for the coherence and consistency of the theory, the glue that holds all the assumptions together. To figure them out, maybe no predictions at all can be made. So, the word ideology looks like unemployed and so on. It's a negative term or assumption. That may prove you very useful one day. It's not a big deal. There is no argument against the position, but it shows an implausible point. Sirius uses the example of Laplace's prediction of the speed of sound in air. And I think he knows French, because he used the original writings, and that gave me a hard time to find good translations. But what I found is that Laplace was very enthusiastic about the caloric...

10:00 He varied it with his theory of gases and used the idea of calorific to explain the expansion and movement of gases and expansion by the temperature and so on. So he didn't use in his law which took him. It's about the pressure of air and volume and so on, and specifically the caloric directly, but the idea of specifically the idea of the movement of gas and so on is inconceivable without the concept of caloric and what you know about it. Heat is composed of particles because that was the time when the idea of waves was still not popular in physics and everything that was thought of in terms of particle-material properties. They thought that calorics are repulsive, mutually repulsive tiny particles that move away from each other while matter, like gas, will fuse. All of them attract each other by gravity, and that was the interest of New York University. But when the calorific acid associates or binds to the material molecules, it makes them overcome the gravity and expand and move. And the gravity expands and collapses, we can transmit a wave, a sound wave. And that's bad because it means that false assumptions can lead to predictive success. That's very dangerous for the realists. Other things that show that Laplace believed that calority was a material property called the material fluid. He talked about attraction and repulsion, and at that time attraction and repulsion were understood only in the material interest, not in the house.

12:30 So he was talking about something material, and he talked about heat conservation, at a time when energy conservation was not yet known. Energy conservation was known in the 19th century. If heat is conserved, it must be material, because only material matter is conserved. So I think this example was a great example, a counter-example to the theory. Of course, it could be said that nothing in the law of Laplace refers to physical properties of the caloric, but that's true of any other unobservable particle. When the theory of electron was proposed, It didn't make any prediction based on the properties of the egg. When Fermi proposed this theory of neutrino, the theory succeeded because it predicted the positron spectra of cobalt and sodium and some other sorts, but not because it showed anything related to neutrino. Later, when neutrino became observable, years later than that, The neutrino was used to predict and support other things, and that's something common to everybody, including the neutrino part, so it's not unique to the caloric. Another thing is that every unobservable entity was actually postulated on the basis of inference from the best explanation, because no one sees it. The data of the experiment or observation suggests its existence. So it is not that if we discover the electron, the electron makes a prediction and we believe the electron is true because it made a prediction. It is that we, by thinking, find that the existence of the electron is the best explanation of certain phenomena.

15:00 There are some inferences to the basic explanations or some inferences to the basic explanations are there, but all physics is inferences to the basic explanations, like Newtonian, magnetic field, whatever you think of. And what is more important than that is that Lucaroni says that all theories are not false because they contain an approach metric to a component which caused its success. And that's the interest of the business definition. So we want to apply the interest of the business definition somewhere, we do not apply it everywhere, as long as it is soundly derived and not not in the interest of the business definition, that's okay, but we can't prevent, we can't prohibit using it in certain areas. My last objection, which is the most important, in my opinion, is that why should we need theory support? If we say that the whole theory is not believable and that we should conquer and divide everything and cut it into its parts and take the good ones and throw away the bad ones, why is that? Why don't we just, as scientists, propose assumptions and test them instead of theories and destroy them and take the assumptions? What's the aim of the procedure? And we are, as localized realists, are not committed to accept any theory as a theory. What is the need for a material theory? I don't know if there should be an answer, but there is no argument for that. Meanwhile, we still hold the theory that we don't believe that school of idle assumptions would hold the assumptions we hold, Recommend not accepting good theories, and if theory, you don't want to believe it exactly. That's a point that needs clarification. So, to conclude, localized realism doesn't justify or fits to justify the construction of theories.

17:30 What are the role of theory and theories in science? Only to supply good assumptions? Well, we can make assumptions without the theory. We all agree that pessimistic induction unfortunately is very strong, but we don't need to reflect it at any cost. I mean, denigrating and outplaying things, just escape the effect of pessimistic induction is not very satisfactory. And also, we don't want to give implausible pictures of things that are less credible than they are. Components, full of other parts, unbelievable as new structures and so on. Although localized realism is very clever, interesting, and creative, you will enjoy reading serious books on realism. This is a very new idea. I mean, it has origins in moral and fiction. Very great idea. But it is a small difficulty, equal to the difficulty raised by the scientists. Okay, questions? Yeah, I very much regret the people that you met. I apologize for the disrespect. I'm out of time. Oh, you? Oh, yeah. But don't intimidate me. This was enjoyable. And I cannot respond to everything you said. We can discuss that. But here is a $64 million question that anyone thinking about this issue should face. When we, when we, put it that way, what should we be committed to when we accept the theory? And here there are two and a half answers. The first is Van Frassen, empirical adequacy, observable, blah blah blah.

20:00 The other is the full blown realist answer which says everything. And then there is a half in between position which says, look, let's not be all or nothing, let's try to localize. And I don't think that's an incoherent position. It might face problems, and it does, but I think it's important for the following reason, which I said yesterday in this room actually, that we've got two kinds of evidence for this theory. We've got first order evidence. Okay, which might speak in favor of heat beam material, substance, or whatever, okay? Hard evidence that scientists provide, but then we've got second order evidence by the pessimistic induction or whatever, which says, look, there might be problems, you know, be cautious in what you believe, because some of the things that you believed in the past turned out to be false. And if you accept that, the only way, it seems to me, to get a plausible account of realism is to try to find a way to differentiate your commitment so that you draw a balance between these two kinds of evidence. Now, I don't think you object that the criterion I propose is sufficient. The issue is whether it's necessary. So, idle parts, are they unreal necessarily? Well... That's a whole issue. I mean, I need to think harder about that. But let's not be quick. It's a call for caution. The things that are not so... they don't earn their right to be part of what we should believe in automatically and everywhere. Let's be more cautious about that. Look for more evidence. So that's the idea. Why do we need theories? I don't know. That's a sociological question. I'm a philosopher, and my worry is, given that we've got the theory, do I believe in all of it? The empirical bit of it? Some parts of it? I don't advise scientists to build or not to build theories. That's, they don't matter. Why they do it? They feel comfortable. They get grants. But wouldn't it be weird? The issue for me is what to believe. And I think that's... Would it be more parsimonious for them? Not necessarily, I think because you can actually, you can be lucky, you know, you can hypothesize a few things and they might turn out wrong. And you need the theory to get the connection. But it's one thing, it's something like the logical discovery versus the logical justification here. Okay, so I'm worried about where to put my money. Okay?

22:30 And I disagree with Van Frassen. I take seriously Laudan, who says, don't put it on everything the theory says. And I'm in a difficult position, as many people say, and Warren and teacher and others, how to draw the line. I strongly believe that the line should be drawn. Maybe the criterion is wrong or it's not fully adequate. But I don't think there is hope for realism unless somehow the line is drawn some way. Two points. One, continuing what Stathis was just saying. It could be juristically useful to have theories as a whole, but you can still follow Stathis' advice, or even scientists can follow his advice and say, well, we're going to suspend our judgment when it comes to the whole theory, but we're going to try and see which of these components, which so far seem to be idle, can produce some predictive success. Well, part of the theory would be used instrumentally, heuristically, and the second point is when you talked about idle assumptions I think we're a bit unfair to the people who talk about idle assumptions because they're not really saying if some assumption seems to be idle now it's going to be idle forever. What they are saying is we should suspend our judgment because it's idle now. It could very well be the case, as you said. I wanted to ask you about the penultimate argument you made to the effect that Stathis, on his behalf, might say, well, no one actually used the knowledge of the properties of the relations of caloric, say, to derive the predictions. But then you raised cases like neutrinos and photons where apparently people accepted these things without having that kind of knowledge. I wonder whether that argument seems plausible because we incorporate a kind of idealization into our discussion of these issues. We talk about realism as though it's an all-or-nothing kind of commitment, right?

25:00 Either we believe in these things or we don't believe in these things. But that seems like a, it really does seem like an idealization because really what we have are not just among scientific realists but among scientists. We have graded commitments, right? There are things that people are very certain about, there are things that people really are, you know... And then there's a whole spectrum of attitudes, you know, degrees of belief in between. And it seems to me that when you do have, say, detailed knowledge of the properties and relations of things, and you can use that knowledge to do things, to make predictions, to, you know, manipulate things, that pushes you towards one end of the spectrum. And when you don't have those things, it probably pushes you towards the other end of the spectrum. But the point still goes through, right? Which is the point that... Well, I think that would apply generally, but then you'd have to evaluate, you know, whether this sort of, you know, this detailed knowledge is leading to predictions, etc. That sort of thing you'd have to evaluate on a case-by-case basis, but the general point, right, that we might be skeptical about caloric because it turns out that we weren't relying on knowledge of its properties and relations to do the work of the theory. That general point. But why not, what I'm suggesting is why not take it as a principle once you appreciate that commitment is created. It's not all or nothing. You will not believe in any unobservable way. You take it as a principle. You want first to, if you are supposing a theory about the electron, if you want the properties of the electron to help the theory that introduced electron to predict, you are asking for too much. You have first to convince me that there may be reference by interest to the rest of Spanish, and then further theories to clarify that. Shall we move on to the next speaker? We'll take the five minute break. After the second talk. Would you like to break now? Yeah, totally. Yeah, if you want to build it. I mean, we have only time for four thirty minute talks, right? So, we should probably begin. Yeah, I think let's begin now. Okay, we can take a short break. Five minutes after the second part?

27:30 I think it's five minutes long. Okay. Well, it's a bit of a shame that I came all the way around the world to speak to basically people I know from England. So we could have done this in Leeds or in Bristol or somewhere else. Let's keep it. Let's go for a drink. So, I owe my loyalty to both of those universities because when I submitted this paper I was at the London School of Modern Studies and now I'm at Bristol, so put both of them up there. Let's get down to business. Okay, so what do we mean by structure in this particular version of structural realism that I'm going to be discussing today? Well, informally we mean an excess of relations. More formally, we can define a structure as a non-empty set of objects and a non-empty set of relations defined on both of them. So set theoretically, where relations, we understand relations as extensionally, right? So, and here's the definition of abstract structure, which is, there's something wrong with this because an abstract structure is just an isomorphism class of structures. What it means for two structures to be isomorphic. And all that it means is that you have to have a bijection between the two sets of each of the structures, which means that the two sets must have exactly the same objects, and each individual in each set is uniquely correlated with the individual in the other set. And the second thing is that if there's any relations in the one set, they're going to be in the other set too. So between the corresponding objects. Okay, we've got that down. Very quick example.

30:00 Okay, so we have... There's one structure, there's another structure here, this is the domain of our objects, we have the set of relations, there's two relations in there, and we define the relations like that, and so these two are isomorphic actually, and here I just put the abstract structure which is supposed to symbolize this isomorphism class. So the abstract structure is just saying, it's not even naming the individuals, it's just saying there have to be at least There have to be three individuals exactly and must be relations corresponding to the relations you find there. And now we get to more important stuff. Okay, the upward path to structural realism. Don't mind me, don't come knocking on my door complaining about this thing. You go knocking on Stante's door. Well, he compares this to the downward path. And the downward path is, well, basically the motivation behind these terms is the motivation that structural realists use to promote their structural realism. So downward path structural realists like Worrell and Poincaré have historical considerations, those considerations are primary. So, they look at the history of science, they see whether there's any continuity of structure, and then they say, well, look, I mean, there is continuity out there, there's continuity of structure, we should be structural realists. The upward path to structural realism is basically starting from the foundations, it's a foundational position, and building your knowledge upwards. That's why it's called that way. And Russell is one of the main advocates of this position. There's also Maxwell. Maxwell is kind of an upward-pass structural realist, but Maxwell does something very different with his structural realism. He advocates the Ramsey sentence approach, which is very different from what Rasmussen did. We can talk about that later. But the first very important point is this. What I'm about to tell you is a rational reconstruction of, so the epistemology that I'm going to give you is rational reconstruction.

32:30 It's not that Russell was so stupid to think that scientists actually work in this way. That's how they come up with their theories. Okay, and it's a foundationalism and it's an empiricist foundation. So what's at the foundations is the, are the perceptions or the... The sensory data, the percepts, which are kind of tricky concepts, but so basically in Russell's 1927 book, The Analysis of Matter, percepts are these events that we have in our head, and but you can, we can talk about this later, but I'd rather we just think of them as just perceptions. Not completely unconceptualized, so they're not pure like sense data, but they're the kind of things you have in your head that are as unconceptualized as one can get. And he also has this causal theory of perception, which basically says that, you know, we have these perceptions in our head. And these are, they originate basically from the external world. So we perceive the external world and there's some causation from the external world inwards. And of course, percepts are said to record information about external world objects. So we can, in the end we can infer things about the external world from our perceptions. Now, these three points that I have up here, they're, to a great extent, basically what science would sanction today, right? I mean, scientists do think that we perceive the world, that perceptions are not just, they don't just come out of nowhere. They encode information about the world, and that we can infer certain things about our surroundings given our perceptions. So, what can we infer? Well, Russell says, okay, so we have direct acquaintance with the precepts, and the precepts are not the same thing as the objects, that's quite trivial.

35:00 So we can't really say anything about the objects themselves other than say that they're the causes of our perceptions, right? Are perceptions. Do these relations come out of nowhere? Again, if perceptions encode some kind of information, and when I talk about relations I mean more generally You can think of an n-place property as a, if it's a one-place property, then it's an actual property, normal property as we understand it. If it's a two-place or more, you think about it as a relation. I'm doing the reverse here. So I talk about relations and I include properties too. If the objects and the precepts are different, so the only thing that we can know about the objects is basically that they're the causes of our perceptions, then the only other thing that we can actually infer about these things, other than that they are causes, is that these relations between our precepts don't come out of nowhere. They're basically telling us something about the relations between things out there in the world. Controversial points, we can talk about that later. But notice here now that if you specify the relations without being able to specify anything else about the relata, then that simply amounts to just specifying things up to isomorphism. And of course you need some sort of justification for these things. So Russell has... A few principles that he uses to justify this kind of . And we're going to look at two main ones. They're not the whole story, but they're the biggest part of the story. OK, this is quite interesting, actually. HW stands for Helmholtz-Weil. That's another one of Stathis' terms. Why did he call it like that? Well, because both of them advocated this principle. And in fact, I found out that there's other people throughout history who advocate this principle.

37:30 Descartes in the Meditations, Hume in the Treatise, and more recently Quine. Not referring to other people's works, they independently advocate this principle. And it's a very simple idea actually. The fact that I can perceive two different things allows me to infer that there's two different things out there causing me to have these two different perceptions. And notice that this is, you can think about this in the contrapositive form as well, which says same stimuli, physical objects, imply same persons. So that's actually quite difficult to deny, that principle. It is a fallible principle, but without it, you can do very, very little. Can you imagine if you couldn't really tell the differences, infer differences about the world, about the world you interact with from your perceptions? If you couldn't do that, you couldn't do anything. And the second, this is a much more controversial principle. It's that the relations between the persons mirror or have the same mathematical property as the relations between their non-perceptual causes. Now, I'm not going to try and defend this. I think it's quite difficult to defend. But what I will say is the following thing. If you're going to be a realist, you need to have some correspondence principle. You're going to have some correspondence principle that's going to tell you anything about the world. Okay, here's the whole story.

40:00 Okay, so we have an actual physical system, let's say, here. Should I? Yeah, yeah. Yeah? Okay. Some actual physical system and then if you if you're standing here as an observer you might be caused to see certain to have some certain perceptions of this physical system and then these two are supposed to be isomorphic according to Russell. So you can actually abstract from the concrete perceptual structure to the abstract structure which is very simple you just take you know that Abstract structure is just defined as an isomorphism class, so if you have one structure, then take any other structure which is isomorphic to it, and then you get the isomorphism class. So that's an easy step. And then, knowing that these two are isomorphic, you can say that there's a unique structure down there which is actually causing you to have this perceptual structure. I mean, you don't have it in your mind, this thing, it's just you postulate it after you make an observation. This allows you to say that there's a unique physical structure that's causing you to have this perception whenever you perceive these things. Okay, and here we come to Mouffet's, another one of those, Psyllos' papers, what can I say? But we have to thank, structural realists have to thank Psyllos because any publicity, even bad publicity, is good publicity for us. Why one? Well, I don't have time for more. I actually don't even have time for this one. But here it is. He says, on the one hand, the HW principle, the principle that says different And on the other hand, if you take HW plus the converse of this principle, WH, which Siddharth is going to say you need to do that if you're a structural realist, then it's going to be too strong. Okay, let's see what's going on here.

42:30 Well, this is something that studies have denied. Well, I still think that in your article you are saying this, but we can discuss this. So here's the point I think he makes in his article, one of the points, but this is, it's not the important point actually, so it's not going to be that much of a problem. So he says that HW can only establish embeddings, not isomorphisms, and the structural realist needs isomorphisms. That's the problem. To establish isomorphisms, appeals must be made to the converse of HW. That's definitely something, a claim that you are making. And WH, that's the converse though, is basically this idea that different stimuli, physical objects imply different persons, that's the converse. By embedding, you mean homomorphism? Let's just get the . So without isomorphism, no inferential, this is his final claim, without isomorphism, no inferential knowledge about the structure of the world can be had, so structural realists are in trouble. Okay, so let's see how this thing works first of all. So we have percepts here, the causes here, and then HW allows us to establish injective mappings from this set into this set. And you can see that because this is the definition of an injective mapping. It says that if you have X and Y which are distinct in that set, then their images, the The objects that correspond to those objects in the other set are going to be distinct also. And remember what the HW principle is saying. If different percepts, then different stimuli. So if you have different objects in the one, you have different objects in the other. So it allows you to do an objective mapping. It's just a converse, HW is just a converse. HW also allows you to establish objective mapping, and in this case, it would be from the causes to the persons. And together, if you put the two together, they allow you to establish objective mapping.

45:00 Okay, now, HW seems more problematic than, you know, WH seems more problematic than HW. I'm not too sure about this, but for the time being, let's stick to this idea. In fact, Russell did think that WH was more problematic, so he didn't advocate WH. He only advocated HW. And an example that he raises, and a counter example, is the fact that if you're looking at somebody in the distance who frowns, Then presumably there's some, so we have two scenarios. In the one scenario he frowns, in the other scenario he doesn't frown. Presumably there's different stimuli that are hitting your retina, but you end up with the same percept. Because he's too far away you can't really make out the difference, you have the same percept. So Russell rejects WH, and Psyllos is aware of this, but he rejects this. Okay, now very quickly. Sinus is right about HW not being sufficiently strong. I think we can agree on that. But strong about it being strong enough to establish embeddings. This is a controversial point that we have that he says that he's not really saying that in the paper, but it's not that important. I'm going to go through it very quickly so that we get this idea of what an embedding is and then... And then we move on to the more important points. So an embedding is basically a mapping that is an injective mapping but also preserves any relations. So embeddings are more than just injective mapping because it's not just that this holds, that every object in the one set is going to have a corresponding unique object in the other set. It's also that these relations, any relations you have in the one set, you'll find in the image of the set. So, HW is not sufficient to establish embeddings. To establish embeddings, a field must be made to the MR principle I talked about before, which basically guarantees that your relations are going to be mirrored, so if you have any relations in the one set, they're going to be mirrored in the other set. And notice here that WH is not really necessary. I just used HW and...

47:30 So HW is going to allow you to do this injective mapping and then you appeal to MR and that's going to guarantee that if you have any relations in the one set, they're going to be in the other set. Okay, but more importantly, this is important, I think that Psyllos is wrong in assuming that embeddings are not going to help structural realism. He seems to think that without full-blown isomorphisms, structural realism is not really a viable position. But notice also that embeddings that we can get only by appeal to HW and MR are just isomorphic markings of a particular kind. Well, here's the definition. Structure S1 is embedded in the structure S2 if and only if S1 is isomorphic to the substructure of S2. Much more importantly, notice that Russell made a very qualified claim when he was proposing this theory. He's not saying we can infer the structure of the world. He says he knows that we're limited, and there's many different ways in which we're limited. He says we can infer a great deal about the structure of the world. And in fact, what you should take out of this is that isomorphism is the ideal limit of knowledge. It's not something that we have now, it's not something that we might not have it in the future, but we are talking about the limits of knowledge now. And that's what this is saying. Isomorphism is the most that we can know, according to this piece. There's an additional objection here, which might not fit exactly with this part of the dilemma, but I'm going to look at it here. I think probably for studies, this is one of the most important things, one of the most important objections. That in principle, it should be possible that the physical world has extra structure. Structure not necessarily manifested in the structure of the person. Now that just sounds right. And I agree, it is right. But it's possible under structural realism to have this. That's the point.

50:00 Notice, it's a very simple, simple claim. Structural realism requires the structure of persons mirror the structures of the physical world. It does not require this claim. It's stronger, well, differently, that every physical structure in the world is going to have a corresponding perceptual structure. For some reason, it might just be that some structures, I don't know, they're so minute or for some reason we just don't have access to them through perception. But this is not what we're claiming here. If you're just claiming that, of course it's possible for there to be extra structure in the world. The second part of the dilemma, this is quick, Stathis says structural realism requires both WH and HW, the two together allow for isomorphic mappings, but once you get those then that's too strong for the following reason, he says it should not be a priori false that the structure of the physical world and the structure of the perceptual world can diverge. And again, I think that sounds right too, but... Is it something that structural realism cannot accommodate? Well first of all, I can reiterate that SR does not require WH. As I mentioned before, use HW plus MR and you get whatever you want to get, but even if it was a case that we needed WH, I can give you a good story that's still not a big problem for structural realism. I'm done actually. One minute is perfect. So structural realism, I would say, need not assume that there is no divergence between the structure of persons and the structure of the physical world. This is a different point, by the way, from the point we made before about extra structure. This is a point about mismatch of structure. Not extra structure, but mismatch. Structural release can accept that we have fallible methods to get at the structure.

52:30 In fact, we probably use inference to the best explanation just to posit this concrete, what I called before concrete observation or concrete perceptual structure. And the structure of the external world need not always be correctly reflected in the structure of our perception. In fact, I have a qualified HW principle here. And I could do the same thing with WH. It can lead you to error. These are valuable things, but most of the time, most individuals would, when given the same stimulus, identify the same person. So, in the end, we can say, well, some variance does exist, some mismatch does exist, but it does not undermine the overall reliability of inferential knowledge about the structure of the external world. And as promised, I am finished. I think that Psyllus' Dilemma does not hold. Does this mean that SR is true? Of course not. That's it. We have five minutes for questions. It seems that you have an underlying metaphysical thesis that the human mind and perception are attuned to the external world, which was argued by Charles Kersen in his evolutionary epistemology, but he didn't defend it and I don't know how he can defend it. Well, I wasn't interested to defend that today, and I'm not sure I have a perfect story to say about that. What I do want to do is, given certain assumptions that some of us might think are plausible, what kind of an epistemology can we build? What would you say to the same thing you feel from cold, or heat, or trauma, or chemical? It is the same person caused by various things. But it's not that every... Well, that's precisely the WH principle that I talked about before in the example that I gave with that you can have different stimuli that imply basically the same percept which violates the WH principle.

55:00 So it is a fallible principle. I'm not sure if it's more problematic than the HW principle because you do have at times the HW principle is violated too. On the whole, I think it's pretty sound. I'm going to challenge that you can use set theory to do this. Sorry, can you speak up a little bit? I'm going to challenge that you can use set theory to do this. So, and I'm going to use Berkeley to do that. So, how can you have a perception of... A distinct object and a distinct relation. Don't those always come together as a perception? So Barclay would say, look, I'm going to deny abstract ideas because you always have a particular. And I'm going to say, you're breaking your analysis up into objects and relations, but how can you have a perception that's broken up in that way? Well, again, I'm going to fall back onto the point that I made when I started talking about the Apple path, which is the structural reconstruction. It's not how things actually do happen, but you can think of and do experiments actually whereby you just ask people, are these the same things or are they different, these two things? And they can say, well, they're different things. So there is some kind of, we do get some sort of information about how they relate, namely they're different things. But over and above that, you can have other types of relations that accumulate over time once you observe, for instance, that you first have the, see the thunder, and then you hear the, see the lightning, and then they hear. Can I say a word in defence? But I think I tried to refute the Asselian claim that the structure of the world can be known inferentially, and you answered by saying that part of the structure of the world can be known non-inferentially.

57:30 No, I think Russell does that too. That's why I have this quotation where he says... So one issue is a matter of interpretation. If I'm right and I say that Russell says what I took him to say, namely the structure can be known inferentially, then it seems that the dilemma stands, and what you said is actually true, but ain't there a different thesis for actually proving if anything that's part of the structure of the world can be known? Well, okay. I hate doing interpretations of people, so I agree with you. That's not the important issue. He does say at some point things like we can infer a great deal about the structure of the world. Well, it's very ambiguous. We can infer a great deal. Okay. Fair. Very fair. So if he didn't mean that, really, he should have meant what I'm saying. So I correct Russell. The real issue here is the status of the new ring principle. This is the real problem. How to take it? And I do grasp and try to argue for this using all these assumptions. Whereas you take it to be kind of a postulate. You start with and you use it to be right whatever. And I think that's the issue we need to think about. And also in terms of the actual status of the principle itself. I think you're right about that. In fact, when I tried to look through Russell's text to find out what exactly it says about this principle, sometimes he thinks that this principle just follows out of HW. Other times he uses other things to justify. It's not clear at all how he gets that. I think we do have room for a five-minute break, and we will resume in exactly five minutes. Our third speaker will be Mark Newman from the University of California, San Diego, speaking on Ramsey sentence realism as an answer to the pessimistic... It's an apple, isn't it? I like it. Okay, so that was the title of my talk. This is actually an overhead, so I'll give you the title here.

1:00:00 Feelings. Let me know if you can't see when I start moving down, yeah. Okay, so there's going to be a fair bit of rehearsal here. I'm sure everyone in the room is familiar with the subject, in particular Wall's account from 1989, but I'm going to run through it fairly quickly at the beginning so we're all on the same page before I move into cruising happiness approach to structural realism. So the general scope of the paper will be a quick review of the world, and then I want to look at how cruising happening might be able to respond to the pandemic. There are a number of questions that Wal has concerning the criticisms that have been raised by myself and by Silas and some others on his account, and seeing what would happen if I applied Krusenpacker's account of realism to pessimistic induction, I want to be clear on this right from the start, because they don't actually argue for it themselves. They have a paper called Realism Without Reference. And this is the paper that I'm concerned with. It came out a couple of years ago in a collection. And they don't want to actually delve into the pessimistic induction. They want to show how you can be a realist and what they call a structural realist. Without reference. Now I'd love to choose a problematic for most of us, right, because the notion of reference seems to be important to the science of the realist, but if you look to the history of science, the pessimistic induction codes and the entities that were positive, the theoretical entities, aren't the ones that genuinely existed, or they never existed in the first place, right? So you can get rid of reference altogether. And you can look to the success of your scientific theories to justify your realism without that reference, then you might have some road to answering the pessimistic induction. Ultimately, I think this fails for a number of reasons, but I don't want to cast this in too negative a term, because it would be awfully nice if you could come up with a way of ramifications of theories such that they justify the realism that can answer the pessimistic induction. Okay, so, current state of play is something like, um, no miracles argument, what can we do about it, right? The miracle is something that's worth, and, yeah, it's a bit of a...