Hello and welcome to the second session of the object philosophy of science and the 21st century. I'm going to pass the mic to Rezanegor Sany now. Thank you Theo, thanks everyone. So we had a very light introduction last session. Today we start to look at Popper's text that I recommended and also we look at a little bit of Kuhn's debate with Popper, the famous debate that they had, and the clashing of their view of scientific science but also scientific theory, their structure and their dynamics.
Then, if there is time, I will start to slowly move toward the work of one of the greatest philosophers of science in 20th century, Wolfgang Stegmüller, who gives a very thorough analysis of this debate between Popper, Kohn, Fire Rabind and Imre Lakatos. And his own alternative, which is built on, I was just mentioning to Theo, by the work of Sneed, a famous work, and a very, very complex and difficult work called The Structure of Mathematical Physics.
So before starting as usual if any person has question any comments We'd be delighted to hear them I wanted to say something A little bit about the topic we discussed last time about to the science it's a bit outside of How we gonna approach in this course, but I think we should consider how useful it is is to have pseudoscience in a variety of ways so in useful in very strange ways I mean so for example there was a debate a huge debate around the new year in Russia about pseudoscience and about science popular science so where was this famous Russian sociologist who was invited to
with popular science popularizers conference, Neil deGrasse Tyson types as they were called. And he gave a huge lecture about ritualistic practices of these people. And one of the ways with community responded to it is by denying his approach to sociology a scientific status. So by having this line between science pseudoscience some scientists and like people who are near sciences use it to to like dismiss yeah this unpopular opinions to and also so I think it's
very useful for science to have pseudoscience as a way to preserve status quo because you know at the moment we have something like obvious like the moment we have something some ideas that it's obvious when something is science or pseudoscience immediately it becomes a social practice of an exclusion of whatever is not in liking of yes this is in fact a point that that Kohn brought against Popper, that such a, what you might call, to be a strict or restrictive boundary line between science and pseudoscience leads not to conservatism
of pseudoscience, but it actually leads to conservatism of science. It is as if you are trying to really interrogate a creationist to abide by the strictures of, for example, Darwinian theory. And through this course you let the scientists of the book, that how much they in fact abide by non-prejudiced methods. Yes. Yes, I understand this, and I think that yes, the rigidification of the distinction is not
methodologically wise, but I think there should be in fact a criteria of rigorous science of epistemology or epistemology of science from that of what you might call to be a hogwash of basically belief mongering or opinion mongering. Yeah, but then we have since... For most of the people who follow the popularizers of science, for most of those people, being scientific or not, is very close to the distinction between truth or not. And of course, some of the philosophy of science is never going to be scientific, and some of the sociology of science is never going to adhere to the same standards as science itself.
So this not only makes the science more conservative with respect to religious people, it also makes science less reflexive. And yeah, so I think it's, you know, the moment we start to demarcate between scientific and non-scientific view, we always, it's always going to be interpreted in terms of truth and not truth and it's even so like it's a completely, it seems a completely different issue to us. I don't think that essentially in philosophy of science, the distinction between science and non-science, by non-science I mean a very very broad set of knowledge, basically knowledge attaining activities.
And I just don't mean pseudoscience. This distinction is no longer what you might call to be a truth vision of science. This is something that has been fundamentally debunked, that science does not approach an ideal or canonical truth. There are criteria of trueness and falseness within a specific context of a theory, but the broad vision of scientific progress, if there is such a thing, can never be thought in terms of approach or convergence upon truth. Yeah, I'm not even talking about science in relation to truth, I'm talking about
status of science in the society because science is what represents truth for people even if we as philosophers don't like it to be so this is what like this is what people believe no absolutely absolutely they're going to believe so these uh uh distinctions with pseudoscience uh they create like what what is actually being created is exclusion of everything that is not science from the field of truth. Yes, yes, absolutely. But I think that that's why I think philosophy of science can be very much helpful in such debates, precisely because it shows that, first of all, criteria of accuracy, scope, coherence, and truth are fundamentally diverse.
And I'm not talking just about ordinary vision of truth or trueness and falsity. But I'm talking about simply the logical undergirding of such predication as whether something is true or false. These are extremely, extremely diverse. And my opinion, and I don't want to wear the badge of, unfortunately, at this moment, scientistic philosopher, as if I'm into scientistism. But I think that in fact what we really do lack today is not that everyone thinks that science says true. No, in fact I would say that many people still don't trust science.
And there should be some epistemological trust in the scientific method and the broad scope of the scientific enterprise. But yes, those issues, as you mentioned, really, I think particularly in the Western country, has led to ostracization of the scientific enterprise in the sense that insofar as many fields, including philosophy, are being rejected under this pretext of scientific truth. It has done two things simultaneously, and they are, I would say, faces of the same coin. One, you see that there are still such things as creationists. And basically, in fact, in many colleges in the US,
these things are being still taught or even discussed. No, we are not going to discuss creationism. That should be banned and illegalized. But also, so that's one, in the sense that once science rigidify its own criteria and dismiss other disciplines, this austerization leads both to the conservatism of new scientists and epistemological mistrust in the scientific endeavor. Both of them are extremely dangerous, I would say. Yes, I agree, but even if we consider absolutely nonsensical things, things that seem absolutely
nonsensical, we're still, I think, very important. So, for example, there is some anthropological research about how people use anti-vaccination movements as a strategy to defend against biopower. And while, of course, there is no point in denying the effectiveness of vaccination, it's dangerous not to vaccinate, it's still like we don't have a lot of ways to not let like government, you know, decide to stop. Yes, that's what I would say precisely. It's basically that the first one that I said, you see that there are both effects of essentially the same thing. It's a rigidification of science leads ultimately to mistrust in science and hence you see in
insofar as there is not enough debate about what is actually what scientists are doing and the metaphysical assumptions behind their practices, so on and so forth, in the sense that the practice of science becoming increasingly specialized and close to the public sphere, obviously you get such consequences down the line. And I think that that's why really philosophy should open both in the technical sense of philosophy of science, analyzing different assumptions, meta-theoretical, meta-mathematical,
meta-logical, even meta-computational assumptions behind the practice of science and modeling, But also philosophers should all try to advance a popular or a vision of popular science. Unfortunately right now I see that those people who write works of popular science are what you might call to be scientists without philosophical disciplinary thinking. And you see that they simply try to, they dismiss philosophy and by virtue of that, they become sloppy philosophers. They dogmatize vision of popular science.
So you get all these loose metaphors and kind of wishy-washy claims, for example. I mean, you have seen it, for example, in build-up AI, about this and that, but also vaccination, so many of these. Yes? Unfortunately, I think if it was not for sloppy metaphors and far-fetching conclusions, popular science would not be very interesting to many people. Yes, but you see, metaphors are dangerous species. First of all, I don't think that any field can advance itself without the power of metaphors and analogical thinking. But it is one to say that we are using metaphor.
It is another way that we fundamentally use inappropriate or overextended analogies to such an extent that it fundamentally distorts the view of what is going to be said. I think metaphorization is a virtue, but overextension of analogies is a vice. So any more questions, comments, these are all great. And we will get back to some of these discussions. Yes, here. Maybe, you know, I've been reading the paper that you suggested.
And, you know, he criticizes a lot, you know, the idea of making one justify themselves based upon origin. He considers these, like, authoritarian questions, you know. And, you know, a big thing that you, you know, he's definitely, you know, an empiricist in a lot of different ways. and so he always says look at facts or what types of evidences you have according to the tests that you use to falsify claims I'm wondering
how this would apply to things like rules and like you know like like like linguistic rules and things of that sort no no absolutely no I mean first of all I don't think that science by any means uses natural language as its primary linguistic tool it just uses language yeah axiomatization axiomatization of systems This is fun. Yeah, sure, acclimatization and other things where artificial languages play rules. And so I was wondering how this form of the empirical...
It seems like Popper would make a claim that Kant would say that we use the rules of the understanding as the origin of our claims and whatnot. and Popper wouldn't follow that justification based upon origin, but would rather say something along the lines of how these rules line up to the evidence and their application is supported or not by evidence. So I was wondering what you would think about how sort of empirical testing relates against things like the faculty of the understanding, the organon of rules, so to speak.
Yes, yes. I don't think that Popper essentially is against this idea that when he says, first of all, we need to pay precise attention to the fact that in Popper's essay, he always very quite accurate, he says empirical sciences, empirical or empirical exact sciences. Basically, he's interested, he also knows that there is no, he's not an empiricist, first of all, by no means. So the criteria of what you might call to be empirical testability, he would say something
not essentially Kantian, but quite rationalist in that sense. In the sense that he believes essentially not just the data, the empirical data, not just observation, not observational statements, but also the criteria of empirical testability are all theory-related, are all theory-dependent. Hence, the theory as the structure is what frames such information and sets ultimately the criteria of testability. Now, this of course raises the issue that, for example, for a theory that has such a
structure, has such a scope and assignment of a structure, namely it is capable of arranging over a specific sector of reality, how can it find the criteria of its own testability? That's of course, it's one of those lynchpins that is being used by Cohen. And that's why he doesn't like, you know, sort of put a lot of faith in the positive claims of theories, but rather sort of their negative mode in being able to prohibit certain claims that other theories make.
Yes, yes, yes, absolutely. He made that thing quite clearly. The more strictures and restrictions a theory has, the better that theory is. But again, you see, I will start it. This is exactly this whole idea that, you know, okay, if criteria of empirical falsification or testability are also theory-laden, then if we are, for example, working in the context of one specific scientific theory, how can we ever adequately test our own theory within our own theory? It's just almost a Gaudelian Which brings in questions of meta-theoretics and meta-notics.
Yes, and also it opens a new vision as put forward by people like Nicholas Rescher, Grunbaum, Stegmuller, and even Cohen, that the criteria of falsification that Cohen has is not holistic enough in a sense that it does not see that the criteria of testability can in fact be thought in terms of theory dislodgements one theory become the falsification of another theory rather than some empirical data falsify a specific theory can I hop in because I
wrote something on this that I'm sure feel like is relevant for ice so I'm just gonna read what I wrote because it's exactly about it for a theory to be refuted according to qualifiers which are not themselves empirical and are instead some kind of rationally constituted criterion, then isn't this just pitting one baseless theory against another in a more or less arbitrary way? This gets complicated, obviously, because the whole notion that we could have a grounded theory, grounded in an empirical way, is already dispensed within Popper's view. But the problem doesn't stop there. Is there a way in which we can conceive of Popper's theory of conjectures and refutations itself, the critical
attitude, as an irrefutable theory. This can be made more evident in Popper's Darwinian claims that the fittest model survives while the least fit is eliminated. In order to accept this, the reader must accept the frame through which Popper sees the results. Are the grounds for this claim if they're not empirical, entirely conjectural. Another way to formulate this problem is, why is the critical attitude itself justified? It is as if Popper believes that one can contentedly balance upon the beam of criticism without falling off either side into skepticism or dogmatism. The irony of critique is its seeming indifference or uncommitted attitude
to both skepticism and dogmatism, neither of which it takes entirely seriously. Yes, no, I do agree. But I think, you see, when one theory dislodges another theory, as we will look at next session through a similar, it is not that the comparison is always arbitrary. Yes, there is such a thing, we will look at it today, that Cohen calls incoming zero-ability thesis. It does not mean that, so for example, the idea is that if the criteria of falsification should be, for a theory T1, should be given by theory T2, then how can we say that theory
theory T2 and theory T1 can in fact have some sort of comparability, criteria of common zero ability, so that the later theory falsifies the previous theory. I don't think this is essentially an arbitrary quality or arbitrary what you might call criteria, Precisely because there are many criterion of theory comparison. There are not one criteria, there are many multiple criterion of theory comparison, in the sense that the application of one theory to another as a falsifying element wouldn't
be called accidental, arbitrary or dogmatic. And some of these criteria I would say one is, for example, the inferential view of the core statements of one theory having as a later theory, the inferential statements of a later theory have as their premise the conclusions of inferences of the earlier theory. This is one. what is called the micrological view of theory comparison. Another one is the macrological view of theory comparison. What sectors of reality such theories
are being applied to? If the dynamic of such theories cohesively share the same range of application or the same range of phenomenon, that can be another way of falsification, application of a later theory to a first theory. Another one is Ockham's razor, which is very, very tricky precisely because we won't get to the idea of induction versus inductivism versus deductivism. We will see that it is actually quite a messy thing. It cannot be, Ockham's razor cannot be wielded at wind. It's fundamentally a contextual semantic criterion.
So there are many criterions of theory comparison, but nevertheless we need to understand that there is no canonical criterion. And yes, canonical criterion is usually in every field lead to dogmatism. And they esteve up productive skepticism. Okay, so before I start reading and I have highlighted… Question. Oh, absolutely, sure. I just want to know where you got those terms, the micrological and macrological theory, where you got that distinction.
That's actually the first person who put that forward was Sneed in that book that I mentioned. And later on it was adopted by both Stegmuller and Adolf Grunbaum. Okay, okay. I haven't finished with Stegmuller yet, so I'm sure I'll come across it. So, before we look at Popper's essay and some parts I have highlighted, let me just give you a very very brief version of what exactly the issue is in a very kind of a light review of this whole issue.
So very briefly we know that the philosophy of science generally in the most what you might call comprehensive sense, without going into much detail, deals with many issues such as, for example, the relationship of scientific statements to other kinds of statements, the status of scientific claims. For example, should the entities described by science, for example fields, subatomic particles, be understood as real or merely explanatory posits. And another point is that whether science can provide knowledge.
For now we are not going even to talk what we mean by knowledge. We are just going for what you might call to be full understanding of knowledge. So within the scope of this broad vision of philosophy of science, Popper asks a fundamental question, which theories are scientific? For example, take two theories about medicine, Modern medical science and theory of magical medicine. What makes the first one scientific and the second not? One of the...
Prior to poker, many philosophers of science thought that modern theory of medicine, for for example, you know, as an instance of a scientific enterprise versus the magical medicine. The difference between the two is that the first one is confirmed by evidence, whereas the second one is not. But of course, we can make our theory of magical medicine completely consistent with evidence, like conspiracy theories. You can find all sorts of confirmation instances or evidences to support your theory.
For example, I ascribe different maladies to influences of different evil spirits. Because of this, confirmation is perhaps not what you might call to be a criterion for distinction of science from that which is not science. The second view of science and not science was put forward in terms of explanatory power. medicine you can say is better than the magic medicine in so far as it enables us to explain
what is happening in cases of illness. But so does magic. In fact, people are possessed by evil spirits. That would be the explanatory power. So even pseudoscience, even conspiracy theories have, do have explanatory power. So the question then is that why is the explanation offered by modern medicine is better than the one provided by magic? You can say, well, modern medicine is preferable to magical medicine precisely because it allows
us to predict the future in useful ways. But a theory, in fact, does not need to be scientific to do this. For example, I believe that the spirits cause the seasons or sins lead to storms, and I can predict it, retroactively or predictively. So the criteria of prediction in this sense is just too arbitrary. usually go for predictive biases in terms of the purposes that they serve them, their
usefulness. But usefulness is not what you might call to be a non-arbitrary or a solid criterion. So Popper here comes and delivers his work, namely the distinction The distinction between science and non-science or pseudoscience more specifically is falsifiability of science. Popper thinks that in order to be considered a candidate scientific theory, a theory must always be capable of being falsified.
This is to say that there must be possible events that would cause us to abandon it. These are what you might call to be anomalies in Kroenian sense. And these anomalies are not essentially what you might call to be an event in the world, in external reality or in a phenomenon. You might actually say they might be related to the equipments we use, you know, or even
our methods. Of course, Popper at this point does not consider any of these adequate criterion of falsification. usually talks about falsification in his earlier works is really what you want called to be a phenomenal anomaly. This is not so the reason that falsifiability can be used as a criterion for distinguishing science and not and basically set science apart apart from pseudoscience is because we can say, for example, magic or conspiracy theories
cannot be falsified precisely no matter what the evidence they can be made consistent with it. So in this broad, very summary view of falsification of the criterion, there is another important concept and these are called observational consequences. We can say a statement of a scientific theory possess some observational consequences.
For example, we can express these in conditionals. If theory X is true, then Y must occur. This means that if Y doesn't occur, then theory must be in fact false. In conjunction with observational consequences, there is another important criterion, and And that's absence of confirmation or no confirmation. A.M. Popper thinks that a scientific theory can never be proven to be true, confirmed.
It can only ever be proven to be false. Our scientific theories are good ones because they have not yet been proven false despite being subjected to what you might call to be bombardment of observational tests, namely falsification. The more tests we hit our theories with, the fitter that theory is.
Now if scientific theories must be falsifiable according to Popper, then it turns out that magic cannot be a science precisely because it can ever be proven false. It's also what distinguishes conspiracy theories from other theories. Similarly with beliefs in beings that no one could ever detect by any means, there might indeed be such beings. But the theory that claims their experience is not a scientific theory because it is not testable. It does not have what you might call to be observational consequences. I think medicine should be a very poor example for it because...
Yes, yes, because of heuristicness. But the reason that I am putting medicine is precisely I want to show that medicine is still what can be called a science, whereas magical science is not really a science. Even in the range of heuristics, we can launch the different criteria of falsification. And certain methods, certain heuristics, can in fact be tested. Some of them will be ruled out. But some of them, we adopt them precisely because we might say that for now,
they represent a robust range of falsification. Namely, we have used such and such falsification techniques or testability, and it still has hold, so we can't really, at this point, get rid of it. But yes, precisely because I would say that medical is what you might call to be not fundamentally a theoretical enterprise in the regular practice of medicine. Whereas, for example, something like physics, chemistry, biology are all fundamentally, fundamentally theoretical. And medicine is more of a heuristic theoretical
I have kind of a silly question, but I mean given that we're talking about pseudoscience, I feel like it might be worth bringing up. When you're talking about people who are capable of engaging in reasoning behavior, it's possible to be able to leverage these things like falsifiability and what not to what not to be able to persuade somebody uh that that's these sort of magical ideas are false you know but you know there's like a phrase it's like you can't reason a person out of something they haven't been reasoned into to begin with and you know it's like people with these conspiracy theories and magical thoughts they arise from like kind of like deeply like pathological
psychological behavior. So I was like wondering at that point... Whether science can actually... No, I would say that no, it probably won't. It won't be able to convince such people. The question then becomes how does science able to have the type of material traction to be able to abolish these types of psychological maladies? Well, this is where science is not the only thing that is important. There is something called education and philosophy. And education and philosophy are responsible for reformatting of pathologies of individuation, not science. But yes, in fact, education can be implemented scientifically.
And pathologies of psychology and dogmatism can, they are not something that you might call that there is this master key that you can simply banish them to the Pandora's box again. No, it's a pluralistic endeavor and that's why that we need to understand confrontation and challenge to such pathologies not in terms of scientific reasoning or even rationality, but in a broad range of human activities, politics, economical conditions, childbearing, developmental psychology, educational system, so on and so forth.
But nevertheless, that is not a good excuse, however, and I'm sure that that's not what you meant, but so many people actually use this as an excuse to say that, well, science then it is restricted. science develops not because what people think in their head but despite of it science progresses despite human psychology as a way to rephrase like the previous question so it was said that but is not alone believe something without having a reason for it, I would say that there is no useless pseudoscience. Like, all of the pseudomedicine is prompting us to try specific herbs or whatever.
And in this way, pseudoscience is, most of the time, is very practical, even like the practice of it is just getting grants or misappropriating funds. still very practical and it means that if we are to criticize pseudoscience we probably have to compare it to fields like medicine rather than fields like physics because yes of course there is a pseudophysics but it's much less of a problem and yes yes no I agree that medicine but however I would say that the reason that pseudoscience is being practiced is not because of pseudoscience is because the elements of pseudoscience like herbs like for example some plants are being
scientifically investigated so they show that they have such and such properties yes obviously this brings ultimately Selaresian vision here that you might say that pseudoscience is in fact a subclass of Salars called the manifest image of the world. Obviously the way that humans have developed and cultured, acquired ordinary knowledge, is by way of the order of appearances, by way how things seem to be. But of course how things seem to be are not always the case. Nevertheless that is what you might call to be the footfall for the progress of science and scientific image simply
comes as the posterior explanatory enterprise built on top of this manifest image. And the whole point of, for example, this Solarzian idea is that ultimately many of these order of appearances, in fact all of the order of appearances, will be completely taken over by the scientific enterprise. Nevertheless, there are elements of what you might call phenomenological or the ordinary manifest image that are never going to be essentially overtaken by science.
They can be explained but they cannot be abolished, like reason and rationality, like what you might call to be conceptual activities, so on and so forth. But yes, yes, I agree that medicine probably is not the best example, but nevertheless, medicine is in fact something that we can in fact investigate medical theories by, for For example, molecular biology, by chemistry, even by physics. It seems like it could be problematic in a way, because at what point do you decide that
something is unfalsifiable? Because you could just say, well, that's unfalsifiable to anything. Well, the falsifiability is regard to claims or the statements that, for example, such a pseudoscientific theory puts forward. Now Popper thinks that falsification of a non-scientific or pseudoscientific theory, All it requires is that you have, as I will mention, all you require is to get one of the claims that they make. And of course such a claim should have certain kind of consequences and implications.
If whether you falsify its premises or its consequences of just one claim even within the edifice of that theory, then that theory has been falsified. Now, how can we find, well, the whole point is that, yes, that's basically when Cohen criticizes Popper on this, that's the idea that we sometimes we cannot really find proper means of falsification. We do not know what in fact these things are. I mentioned earlier on if particularly if we are within a theory then how can we
find the means of falsification of our own theory because it might be very fact that our theory simply reiterates its own pseudo observational falsification. It's like a picture of Durian Gray. Theory just gives distorted views of how it can be falsified. So I mean it is reasonable to sort of dismiss claims of something being falsifiable if you know it doesn't have a metatheoretical jurisdiction over that theory
and able to demonstrate that certain things are falsifiable yes yes i think um i i think that popper uh makes uh we are just simply introducing progress i'm not by no means defending popper But I think that this is, it can be said, quite a greater step to mapping the criterion of scientific enterprise. But the restriction of the distinction between science and pseudoscience, the criterion of testability or falsification, is neither global nor in fact always practical. So,
sorry, I lost my notes. Another point that Hopper makes in his idea of testability in his paper is the problem of induction. This is something that let's not, you know, spend too much time on it at this point, but
precisely because it's quite, quite complex story that I will make sure I will, all the session, if not two, we will talk about this. What basically the problem of induction claimed that we have no reason to believe in the uniformity of nature over time in space. Therefore, we have no reason to regard science as more than a speculation. Proper thinks on this front that to imagine science could ever prove the uniformity of nature is to misunderstand the nature of science. Science can't prove anything. It merely consists
of falsifiable theories that have not yet been falsified. So Popper thinks that Hume is right that we cannot prove any theory to be absolutely correct, but it is the case that we can absolutely prove that a theory is incorrect by showing that one of its statements or claims has been tested and the test has shown that the statement basically is falsifiable and hence should be ejected, so has the theory itself. So as I mentioned proper things with regard to the problem of induction and many inductors
people who see theory or science simply being an inductive enterprise. And of course, usually an inductivism has some empirical, what you might call to be commitment, obviously the point of the problem of induction targets empirical observation. So Popper thinks that science does not in fact make use of induction. We will talk later on, Popper with regard to science takes a view of deductivism rather than inductivism. Deductivism is what
you might call to be the inferential deductive inferences that are made between a statement of within a theory and a statement of one theory with regard to another theory. And such statements are ultimately theoretical. All data in fact are caught up in such theoretical statements. So regardless of that, Popper thinks that instead of being inductive science is what you might call to be conjectural. We just hang on to the theories that survive all the various slings and arrows, we shot at them, we shoot at them. And the whole point is that
that here a problem arises, we'll discuss later, in the sense that precisely because the view of falsification or testability that Popper offers, you can say within this view if a statement or claim is being falsified, then the entire theory should be ejected. But then if Popper at the same time also thinks science is conjectural, meaning that we have to hang to the theories, then of course these two views are ultimately somehow incommensurable.
How can we at the same time falsify a theory and also say that we can hang on to the theories as best as we can? And this is when Cohen's critiques targets Popperian philosophy in the sense that theories can be in fact falsified, but the idea that we can simply say that science is falsification and or falsifiability and if once we falsify the theory we should discard it, that's just not how science works. scientific theories are being falsified but in so far as they are what you might
call to be there is no better alternative for them to explain a sector of reality or phenomenon under study then we have no choice then to still resort to such theories as Cohen says that it's better than you know leave live in a house with a decrepit ceiling than no ceiling at all it's better to sail inside a broken ship than no ship at all. It seems like, go on, no, no, go on, please.
It seems also that it throws out the possibility of various as-if types of arguments or theories or hypotheticals. And some other things like with Newtonian physics. I mean, Newtonian physics works just fine at the level of macroscopic phenomena. And so with that restriction, it does have material attraction on our reality. Yes, absolutely. And of course, it has falsified, basically, or dislodged Ptolemaic system. yes yeah and and and also you know things like you know there's you know so many different mathematical models of of quantum physics for example um but i mean the the mathematical
structures behind them you know do have their their own right even if they don't actually end up applying to our quantum reality. So you can reject a theory in its theoretical structure as incoherent, but that doesn't mean that the structure of that theory, but if it's a coherent theory, you can't reject the structure of the theory itself. you can just sort of like reject its applicability to various, you know, physical material applications of it. Yes, yes, this is, what you are saying is, I think, more in tandem with Goranbaum's contextuality of theories.
And yes, what you see, I think perhaps we shouldn't in fact talk about Neutron. We should talk about Copernicus for Galilean revolution and the Ptolemaic system. Copernicus doesn't, we will talk about it in next few sessions, Copernican system can be shown in fact not falsifying the Ptolemaic system. It simply rectifies some of the parameters of the Ptolemaic system. And in fact, usually people who are relativists want to say that, or anti-scientists, they
will say that, well, they use exactly this example that Copernicus didn't fundamentally diverge from the Aristotelian framework of the Ptolemy, and that's completely true. But this does not mean that Copernican system is not scientific or is still Aristotelian, precisely because as you say within certain kinds of contexts of its operation or its theoretical structure, it allows us to formulate new questions, make new statements that can be tested in fact. You know, theoretically it's not useless to consider that there could be some universe
with some strange laws of physics where, you know, you do have planets with these strange epicycles that seem in, like, the Ptolemaic type of system. And it doesn't apply to this reality as we know it. But, you know, nonetheless, you know, it doesn't completely throw out the Ptolemaic system. Well, that would be mere speculation and science. Yes, but of course, even in mere speculation, there is a grain of truth to it. And as I mentioned, this was one of the main ideas of Boltzmann, that we can, in fact,
think about other creatures who do not, at the level of observability or macrophenomenal, macroscopic phenomenal observation, observable phenomenon, you can say the arrow of time is in reverse. He says that we should not simply always say that these are just mere speculations because sometimes mere speculation if supported and that's something that Boltzmann doesn't say it's Reichenbach says it on behalf of Boltzmann that when mere speculation is coupled with mathematical sciences in
fact it can be shown that such a reality exists and so as new facts of of experience, like a different arrow of time. So, but let's not go to these details yet, we are still in the preparing phase, we are yet to come to any of such stuff about reality of space and time. But before I go on, maybe we should have a brief break and then come back. Yeah, should we take maybe a seven minute break so people can... Sure, have a bathroom. Okay, thanks.
question as well, which is that if, so like apparently NASA, when they try to like send rockets up to other places, they, like Newtonian mechanics is not, like a certain scale Newtonian mechanics doesn't work, but then they still use Newtonian mechanics to like calculate like certain trajectories because it's easier and more helpful for them. And so I'm thinking about this in regards to a kind of Kwanian mode where theories that we use present problems, but the problems are within some kind of web by, in a simple sense... Coherentism, yes. That's a coherentism view, yes. Okay, so what do you think about that, I guess?
Well, I'm actually myself, I consider myself a coherentist on this front. But you see, I think, well, basically, there is this book, I remember that I have mentioned a couple of times in the past courses, by Nicholas Rescher. It's called Epistemology, and it's one of the most classical works done on this front. You can see that within the, what you might call to be analysis, of traction of scientific theories and reality, we have different kinds of views. One of them is called coherentism.
So what is exactly coherentism? You see, I think first better to make the opposite example. once the axiomatized vision of scientific enterprise that all scientific theories are ultimately axiomatized at their base, which some are, and we will talk about this, what kind of axiomatization. You see, the kind of range of epistemological claims it allows you is that, so you always start with some givens, what you might call to be data or axioms, But these are not sense givens, these are not empirical sense data. They are what you might call to be logical and theoretical posits that axiomatize by
a specific type of an axiomatic system like Hilbertian, Carnapian, and so on and so forth. You can think about such structures always moving hierarchically, bottom up. From the axioms you build more and more claims within which your empirical data is being cut up. In contrast, rather than in opposition, in contrast or in distinction from such view of theories, there is such a thing as coherentism. In the sense that we never start always from the givens. We abductively think about a horizon.
We might call the hypotheticals or data candidates. Data candidates in the sense that as long as such data have not been refuted, we should go for them. The criteria of truthness of theory is not really an important one in the coherent view. It's what you might call the plausibility or coherency of how such truth candidates or data candidates, truth candidates, hold together, namely cohere. And once we set such a horizon for ourselves, all we can do is moving inward rather than
moving upward, we move inward. And we start to see the coherency of such assumptions, like a web, like a web in which all these truth candidates are there. And as we dig it deeper, we see that some of these coherences no longer hold up. We then go step by step piecewise revise some of these. And at some point we might actually introduce new truth candidates precisely because we see that our theory is no longer adequate to support its criteria of its own coherency or explain how is that some of the truth candidates or some of the data were in fact coherent
in the first place. So new elements at some point will be introduced. Some of the webs, it's what you might call to be like a plastic web, like a neural pattern. Some of the networks between your truth claims and statements will be pruned. Some will be discarded and some will be built on. Can I ask about, so it seems like in order not to fall into any form of relativism, this revisionism would be, I'm guessing, objective by means of modeling on your part?
Yes. I, well, Meredith actually brought something in the classroom, Google Classroom, and asked something about, you know, as... What was Meredith exactly what you said? Plural models of objectivity? I think Meredith is not here. But I think something along that line. Yes, this is something that has been put forward recently by some of the philosophers of science, by like people, Objective pluralities, yes. You might think about them as model pluralism of some sort,
in the sense that each of these models give you a certain limited scope of assignments, representational capacities, methods, dynamic capacities, and certain class of equations, which are basically the model description. So very, very briefly, I mean, when we're I'm talking about model, I do not mean it in a common sense model. First of all, we should know that saying, for example, X is the model of phenomenon Y, it's just an ordinary talk of models. Models require a specification. And in fact, all models have a core specificatory characteristics. Every model is comprised of two things primarily, two components, core
components. One is more prior and the other one is what you might call to be derived from the first core. The first core, which is the prior, is called model structure. So model structure is usually an encapsulated theoretical and metatheoretical undergirding of that model. I will unpack what exactly the structure is. And something that is derived primarily from the second core, which is derived from the structure of the model, is called model description. Model descriptions, you can think about it as a bunch of graphs, formulas, and equations
that describe implicitly, not explicitly, implicitly the specifications of the model's structure. There are such equations are always encapsulations, meaning that many details or implicit assumptions whether with regard to criteria of a structure, the criteria of its theory in which it is embedded, or the criteria of method theoretical assumptions behind it are hidden. So these are the first two quotes. Now a structure can, when we unpack it, we see that a structure is ultimately
They hold some important information which are called usually model construals. What are these model construals? These are what you might call to be interpretation or the interpretive import of the model itself. So they are in contrast with the model description. The description is what describes the structure and of course is not equal to the structure. Model construals interpret the range applicability but also criteria of fidelity of the model itself. Now, these construals come in three, again, classes of information.
First one is what you might call to be the model assignment. assignment of what this model basically supposed to do. The second one is models scope. To which sectors of reality should be applied? What phenomenon does it target? What target system does it apply to? Okay. The third one which is a little bit trickier and usually modeler or scientists are not fully understand the important of it and precisely because of that they will distort the data is something called fidelity criteria so fidelity criteria are what you might call to be constraints regarding
the dynamic capacities of uh of the said model uh how the range of a structure can be applied to what range of systems. The second one is representational fidelity or representational constraints. How much does this model represent? How much does it compress the data? how much it is what you might call to be, you know, true to the target system in terms of representational capacities.
And of course we should know that all models, so as all theories, require a certain amount of idealization and simplification, either one or another or both. The third constraint, the third fidelity criterion, is what you might call to be resolution fidelity or resolution constraints. Basically, what is resolution constraints? It's what you might call to be a scale. When we are talking about every phenomenon, when we are talking about every target system or sector of reality, that sector of reality needs to be understood at different scales. Like I mentioned last
session, previous session, about for example think about a metal beam. So you have the macroscopic level, the level of the metal, elasticity, pressure, so on and so forth. Then you have the level of, for example, metal crystals, fundamentally a different scale with different kinds of details and information. And then you have, for example, you might say the nano scale. And of course, there are many ways that these scales can be differentiated, but I'm not going to go to that length. So what is that scale is important precisely because if such informations are simply being
implicitly assumed and never made explicit, you can in fact will end up applying a wrong model to a wrong scale. only you might misapply the model to a target system, for example, if you ignore representational or dynamic constraints, but also you might actually use, for example, some like model couched in terms of differential calculus to a level that requires a fundamentally different mathematical structure for it to be engaged. And so these are why they are important. And yes, so these models, you might say that
the view of coherentism can also be said as model pluralism, in the sense that we have models and these models compress or encapsulate certain kinds of information. And such model pluralism essentially gives us a theoretical arbitrage and plasticity. We do not fundamentally get rid of a theory. We just simply tinker toy with the arrangement and basically the number of models we have used or we are using. so in what sense can we say reality then determines model application so so do we
not in some sense already need some kind of model or some kind of meta criteria yes it actually speaks about reality you see I would say first of all we should know that for philosophers the question of reality is very important, but for scientists it's not that much, or even for a modeler. Of course, such, first of all, even in philosophy we would say that, you know, if you're looking at Quine onwards in philosophy of science, there is no such a thing as being anymore, like these kinds of metaphysical reality that's just out there by itself. You might
say that reality or being is the designation of theories. Object is a designation of a theory. In fact it would be, as Salars would say, that a view of reality as already set in advance that decides how we basically apply such models or divide such models is a variation of the myth of the given in the sense that we have some access to some prior knowledge which you might say to be foundational. And all of our knowledge about the world ultimately bottom out in this foundation which he calls myth of the given. Basically the word myth simply for Selares means an ideological foundationalist fixation.
Now the thing is that yes, this doesn't mean that reality does not exist, but reality is the product of discovery, of the labor of intelligibility that is given by theory. Theory and objects are co-constituted. And in fact, the more we discover, we notice that in fact reality does constrain such models and their applications in the sense that and that's what basically representational fidelity or scale fidelity or all of those construals that I mentioned for any models are attentive to precisely because if for example we notice that at different scale lengths a material
entity or system or physical system has different kinds of behaviors different kinds of regularities then of course it will constrain our model and how we can think about model building so I would say that's such if you know a scientist is more hospitable to the idea of co-constitutivity of theory and object or model and world there are less concerned about whether there is such a thing as metaphysical reality this of course however does not mean that scientists don't have metaphysical assumptions in fact they do have both metaphysical assumption and meta theoretical assumptions and many scientists are
actually fundamentally ignorant of such assumptions I have a question absolutely just continuation of what you just so yes last time you and you just repeated before that scientists they have they are metaphysical assumptions or even like having some biases but then you are saying that question of and then you give examples when I ask what would be the example of the physical bias of a scientist and then people would propose statements like electrons are real so basically statements pertaining to reality of things and then we are saying that scientists, actually the question of reality doesn't matter.
Like you can ask a scientist, he can say yeah they exist or they know they may even have a bar fight about that, but it doesn't really matter in their practice what they think about whether electrons or atoms are real or not. That is true, but I would say that this is when perhaps we should distinguish different classes of metaphysical statements like for example whether electrons are real or not yes I would say that it most probably certainly perhaps does not negatively affect or positively affect for that matter the practice of science
But some do, in fact, some classes of metaphysical statements do, such as the notion of nomological laws of nature, or causality, or causation, or the nature of time, the idea of basically asymmetry such such classes of metaphysical statements do in fact I would say both affect the theory and the practice of science just a brief question I know I'm the one who always ask this type of question where'd you get that three the model the representation and the third I forget
which one the distinction this is I mentioned to you well it's something I came up with a little bit I've worked at this stuff lately but I mentioned that one of the great books recently written on modeling is Michael Weisberg simulation similarity some of some of these specifications but some of these specifications are the ones that I have added, precisely because I think they have been also recognized by other philosophers of science, but not specifically in connection with modeling. Okay, so waiting to read whatever you end up writing about it, I guess.
And also, again, response to this coherentism or model-prolourism as the way that I think how science in fact works or should work is that there is this philosopher of science William Wimsatt there is another concept that is usually associated with such coherentist view or model pluralism or what Meredith called what was it again objective pluralities is the idea of of what Winsat called thickets. The view of science is better be thought as a jungle or a forest, a dense forest, a thicket,
a thicket of models and theories. And of course, these thickets, as I said, these are not just forest properly, it's not really a great metaphor, even though Winsat tries to use the metaphor of forest precisely because forests can be cut down at a specific point. You can create roads, change the course of the river, so on and so forth. But essentially, why I think it's not really a good metaphor is that in a forest, you don't really have explicit criterion for coherency between your constitutive models or your
constitutive theoretical structures. Another thing that I don't want to get into details but something that we will talk about at some point and is that this whole idea of thickets, pluralistic thickets, as the way of understanding the scientific enterprise leads to something that Hentika had already criticized you know in his earlier works it's the idea that these tickets become so much out of control in which you lose it becomes almost impossible to attain any
classificatory or coherentist, non-arbitrary coherentist parameter between how such models hold. The point is to always make, he says that it's better always to make a manageable parking lot rather than grow a weedy forest. Do you think that sort of criteria is possible, like a general criteria of cutting down forests, or do you think like at the limit it's impossible, only like possible across certain domain of theories? I would say it's more, I think such observations should always be context sensitive,
in the sense that, first of all, I do not want to talk about like Salars in terms of convergent realism, the idea that science converges on what he calls a Cauchy limit, in which the entire furniture of the world will be revealed by a complete science, by a convergent progress of science. I don't want to talk about such things precisely because these are just require too much of assumptions about the scientific practice and the nature of the scientific endeavor. But yes, I would say that within, for example, certain kinds of theories within what Cohen's
called distinct disciplinary matrixes or pragmatic projects, such regions can be upheld and can be defended. And in fact, it would be very good for scientists to think about such views of the scientific project. But again, another problem with Thikki that I forgot to mention is that one of the really nasty things about these forests is that they create ultimately metaphysical, what you might call to be, not only metaphysical, but what you might call to be non-existent kind of entities that don't add up to anything. They don't
help us in any regard. In fact, create much more confusion. I mean, a string theory is a good idea of a jungle, you know, a forest. In everyone, a string theorist tries to put something, but it's just worse than a zoo. It's really a forest of a hentikai-ian, you know, world growing out of control. To the point that it becomes almost like any person can say anything without any criteria of coherency, I'm not going to say truth or trueness. I'm going to say criteria of accuracy, criteria of coherency, so on and so forth.
So in order not to succumb to this overextension of Hintikka's jungle and thicket, but at the same time not to merely relegate metaphysics to the practices of sciences. I'm just asking like, and also not like going back to the Selaresian thing that you mentioned earlier, where there's the kind of a unified totality of which we can speak about, you know, metaphysics or being or anything. Then, yeah, I'm just trying to like find out what kind of what alternatives or routes there are? I think I would say that I will go with the Hendekalian vision
of having tickets, but these tickets needs to be thought in terms of manageable lots. Where the cars and lots can be called theories and models. Precisely because Because if we have such a kind of disciplinary matrix, you then, a scientist can actually think about those implicit assumptions that she always have, both at the level of practice and method and at the level of theory. For example, the whole idea and such assumptions can be fundamentally broad, not just metaphysical.
for example, the structure of a certain range of mathematical structures and when you are making a model that has a mathematical model, if you do not know about, for example, the range of such mathematical structure, then obviously your model might not be an optimal model. A good example of this is for example you know these new works are being done in category theory modeling of neural systems. Part of the brain, modular, what Andrea Ersman calls memory evolutive systems. It's in the idea that you know brain has a modular system and
these modular systems at each scale or at each level of the hierarchy, they are constructed from more rudimentary modules at the base. And the way that, for example, some of these new works in theoretical neuroscience are being done, they are using, for example, highly specific categories, theoretical models, like, For example, the concept of core limits, commutative diagrams, such and such. Without the understanding that, and this is one of the criticisms that was given by a number of people within the field of neuroscience, but also mathematicians, that such commutative
diagrams or category theory are fundamentally a structure preserving forms of mathematics. They are part of what you might call to be abelian categories, where a structure is fundamentally thought in terms of a structure preserving rather than, you know, forms that, in which structures are being broken or pruned. So obviously this already, this projection of this kind of mathematical structure to target system, namely the architecture of the brain, creates a fundamentally a limited vision. Of course it enriches our views, but it also, if the modeler is not aware of those mathematical
and mathematical assumptions, can fundamentally restrict the view of how we can find about the the architecture of the brain. Not everything in complex systems are about symmetry preserving or structure preserving. So the key then is to test out different models, I guess? Yes, absolutely. Test different models in connection, in coherence, connection with one another, but also have what you might call
to be metatheoretical view of models, where you can, in fact, investigate the hidden information or the hidden assumptions that each of these models have individually. And of course, you see these are, these coherent views, I wouldn't call it really pluralistic Because pluralism usually gets, you know, pluralism is usually at least in philosophy and even philosophy of science can be a very, very confusing term.
It can sometimes in fact be interpreted as almost a relativistic thesis. It is not relativistic though. But it can be interpreted as such. And that's why I wouldn't call it a pluralistic view of the enterprise of science. I think it would be better to call it really a coherentistic view, in which you have webs, networks of different models. And not only you can see how these models cohere, some of them might in fact be incompatible, but also you can in fact at a different scale, at a meta-theoretical scale, look into each
of these models individually with regard to their implicit assumptions, with regard to a their application to a target system their theoretical commitments their mathematical logical or computational in encapsulation of such theories so going back to I mentioned, so Popper was saying that science doesn't make use of induction, it is conjectural.
And that we just hang on to the theories that survive all the various tests that we throw at them. Now before going to Cohen, Cohen's alternative view of science and scientific progress, we can pose this question, has Popper solved the problem? Which sciences do in fact pass the tests of falsifiability? thought that, for example, Freud's psychoanalysis did not. Grunbaum however shows that in fact psychoanalysis is not a pseudoscience, contrapropor, but it shows that as a science it fails
to substantiate its own goals and if you want to read it's a fantastic work written by Gorenbaum called I think it's a foundations of psychoanalysis right Theo? It might be called critique of psychoanalysis or something like that let me see there is a work called foundations of psychoanalysis by Gorenbaum Yeah, a philosophical critique, foundations of psychoanalysis. Yes, the foundations of philosophical critique, yes. And I don't go, but nevertheless, it is actually a very, very interesting, what you might call, to be critique of poker.
And the really fantastic thing about Grunbaum, we won't talk about his work on psychoanalysis, we talk about his work in epistemology of science and particularly his work on fundamental philosophical problems in modern science of space and time. But nevertheless, the great thing about Grunbaum is that he shows that, first of all, Popper misrepresents Freud, one, two, he shows that in fact psychoanalysis can be falsified, hence it is science, but then he again go against, you know, proponents of psychoanalysis to
show that in fact psychoanalysis falls into its own contradictions within the domain of doing science. It's something called, if you can, I think probably you can find it online as a very kind of encapsulated version of his argument which is called the Talley argument. And it's extremely escaping, you know, logical argument against the method of Freudian psychoanalysis. He then later, in the other chapters of the book, he doesn't really target Lacan or, you know, but he targets Melanie Klein and many other psychoanalysis, not just because of
tally argument but also what you might call to be heuristic biases which psychoanalysis is fundamentally blind to majority of such biases so against this falsificationism or testability view of science One of them, other than Kuhn, there is another philosopher of science which we are not going to read but if you are interested you can research a lot more, Philip Kitcher.
For example, Kitcher takes Newton laws and plus the theory of gravitation. This theory, he argues, predicts that an apple released from a tree will fall and hit the ground. If that doesn't happen, we must reject the theory. This is required by Popper. But it isn't then clear that we would have to. For example, we could just suppose that there were some other factors at work. If that is the case, then Newton's theory is unfalsifiable and hence not scientific. Kitscher then thinks that we could state that Newton's theory contains a series of supplementary assumptions that would cover all eventualities and keep it indeed falsifiable. But anyone could do this for any theory so of course then
this would be just too arbitrary of criterion to be accepted. For example the core claims of Newton's theory are not falsifiable by themselves but only with supplementary claims. Similarly I can claim that magic is not falsifiable by itself but can be made so by addition of supplementary claims that entail observational consequences. For example, if magic is true then grass is green and animals exist etc. So adding arbitrary supplementary arguments. So the moral of this story according to teacher is that once you have to add observational consequences to a theory you can do this for any theory and we are back where we we have started with no distinction between science and the rest.
The response to such quandary is holism, which we will look at it in more details when we are engaging with the work of Wolfgang Stegmuller and his analysis of this debate between all these philosophers of science. So Popper assumed that, for example, individual claims or a small group of claims like Newton's laws have observational consequences. What if actually individual scientific claims do not confront the evidence one by one but in large groups?
This is basically what holism actually is. So holism abundance this idea that we can isolate precisely where a theory has gone wrong. If a theory faces contrary evidence, there is always a choice about which assessments of that theory must be revised. More than one theory can fit the facts, we just tend to make the easiest revisions. So you can think of falsification versus holism, the falsificatory Popperian view versus holism,
and also Quine Duham thesis, which is an instance of holism. You can think of this difference between the two in terms of really the old philosophical battle between atomists and holistics. So imagine that theory has a series of claims, claims made by the theory. So in the Popperian, and each of these are represented by, for example, some core statements. And these core statements are called atomic facts or atomic claims. So according to the atomistic view of falsification, which is a probarian one, for a theory to
be falsified and hence considered to be science, the falsifying evidence only needs to be applied to one atomic fact, claim, or very small group of such claims. And that's exactly when we know what went wrong with this theory, which claim was responsible for this falsification according to tests. Now a holistic view of falsification is not this atomistic view. It's falsifying evidence should be applied to the entire class of claims covered by that
theory. Any claim may be singled out as the wrong one. For example, Newtonian mechanics was falsified by the orbit of Uranus. Scientists had a choice, propose some unseen influence and keep the theory or abandon it. They chose to keep it in defiance of their observations. Why? Because it was simply the most useful theory they had. There are a variety of reasons why some revisions are better than others. Some revisions are testable independently of the theory. Like for example the existence of Neptune, some revisions are more unified than others
in the sense of cohere better, more optimally, more coherent. Some revisions are more useful in getting us to ask or pose new well formulated questions that the other theory could not pose. Allow me to just get a cigarette and I will be back. No problem.
I said you can no longer blame all you're smoking on me Reza. Yes that is true but you were responsible for the trigger. Sorry, let me get my notes. And were you mentioning Philip Ketcher's criticisms of Popper? Is that who you're talking about? Yes. Which work is that from? Well, there are a couple of works. One is The Advancement of Science and another one is Patterns of Scientific Controversies. OK. Any thoughts before I move forward?
Yeah, I was just kind of wondering, this is just kind of lingering here, a question for me. To what extent is falsifiability necessarily like first person? You know, like you always have to, it seems like it's up to you to decide that something should be falsifiable. And then I think it's kind of related to a question I have just about the title of this course. This idea of theory and object as opposed to like subject and object. Yes. Well, theory, you see, that's the whole point. Subject, I think, has psychologistic connotations. Psychologism is a fundamentally serious thing, particularly how insidious it might be actually.
And but the view of theory, I think, can be coherently thought without in fact resorting to the vagueness of the idea of subject. yes, basically you might say that theory, subject, mind, intelligence represent a class of fundamental poles. They are not essentially synonymous but nevertheless they are representing the same kind of class. So that's why I I chose theory and I will talk about why theory, what is exactly a theory, particularly scientific
theory. As for falsification... And also, I don't know that subject is always going to, you know, you can't always reduce it to psychological facts. I think there is an ethical subject too, or a fish subject, it's not just psychologism. Yes, yes. What is psychologism? You see, psychologism is not essentially psychological facts. You see, the idea is that the subject, since the time of Kant, as we saw it, under some criteria of psychologism in the sense that what you might call to be subjective preferences,
options, psychology, even values sometimes in a sinister way are brought and levied against and distort because of that, distort logical structure. You know, I think really one of, I would say that this assault that's launched by Frigge and Carnap on the edifice of psychologism was extremely, extremely important to finally, We can still have the best of, for example, theories of mind and theories of subjects,
but within a broader domain, which is that of theory. So we can in fact see that even though we think that we are part of a theory, but as long as we have a systematic theory, we might be in fact capable of deciding and determining what are of the subjective assumptions are brought to impact or distort objective facts. Objectivity is something that is essentially a product of theoretical thinking.
and of course we didn't hear the call thinking subjects play a role like scientists scientists are subjects but when the reason that it's not called subjects and object precisely because subject as I mentioned has what you might call to be fundamental metaphysical and psychological assumptions That's just what you might call to be what I call these days a metaphysical bloatware. It's a metaphysical bloatware. Yeah, it does.
It seems to me like I just have this feeling of the challenge you're posting here with this class. So I'm interested to see how you develop that and kind of how I'm looking at it. Yes, I will. I will, in fact, when we get to Stegmuller, we start to look at different kinds of theories. And also, I think it probably would be great to finally say what theory actually is. Not just in science, but what theory is. I mean, people use the word theory as if it was some vague notion that you are just thinking, and put that stuff together, and you call it a theory. No, theory is distinguished by its fundamental component, structure. What is a structure? It's something that we are going to think about, and how a structure is ultimately the designation of, in a philosophical sense, being,
or designation of reality, or designation of objectivity. But I think at the same time, I've always thought of virtue as exactly the opposite of metaphysical bloat, where whatever the polar opposite of that would be, would be that tool that you would use to cut through those thickets of theories to arrive at manageable lots of subjecthood. Can you elaborate on this point? I just, I've always thought of virtue as like asceticism, as, you know, arriving at clarity of thought so that you can act in the world in a way that is virtuous, as opposed to this bundle of confusion. I just don't know if that's an accurate depiction of an ethical self or what virtue is.
Yes, I mean basically, but you see, at least in the domain of the science, but also I think philosophy, systematic philosophy, we might say that theory is essentially the vehicle of virtue, of basically get rid of all those metaphysical junk. that we usually ascribe both to the object and to the subject, or which might in fact distort our objectivity, precisely because we are under certain kinds of psychological factors, preferences, choices, so on and so forth. And that's why, I mean, SELARS is fundamentally, I think what you are saying here is fundamentally a SELARSian thesis,
in the sense that those of you who have read earlier essays of Salars, not just the more hardcore ones, you see that he sees the systematicity of theory and particularly science as a tradition of what you might call platonic virtue. It does two things at the same time. It unbinds intelligence from its own limitations, progressively piecemeal, but also it discovers or reveals new sectors of reality.
what you might call to be expanded the scope of the intelligible. And in so far as the scope of the intelligible can change, and by virtue that intelligence and intelligibility are correlated, there is no intelligence that does not require the labor of intelligibility, whether the world of which that intelligence is a part or the intelligence or the intelligibility of why is it we call it an intelligence so the more you expand the scope of the intelligible but through the labor of intelligibility which is the labor of theory the more you can change the
nature of intelligence That's kind of how I'm interpreting this class. It seems to me like that's what you're arguing for. And I just wanted to clarify that. Yes, well, it is essentially the sequel to Plato. Any more before I move forward? So if you don't have any comment, let's pull up that essay by Popper. And I have highlighted, I want you to, basically what I just gave was a summary of that essay. So I want to just give you, like some of, emphasizing some of the parts, and then make a very brief comment
in the idea that I think Popperian epistemology of science or the idea of falsifiability has fundamentally important significances that are not usually actually being talked in philosophy of science. One of them is the concept of information. And I think before even Shannon laid out the concept of information as we know it today, he was prepared to actually laid out the concept of information, which is a fundamental, important
and central concept of science. So on page one, as I mentioned, one of the first things that Pooper wants to do is that showing that confirmation is not really a criterion of the scientific enterprise or what can be called a designation of science. That paragraph, it says, I knew,
of course, the most widely accepted answer to my problem, that science is distinguished from pseudoscience or from metaphysics by its empirical method, which is essentially inductive, proceeding from observation or experiment. But this did not satisfy me. On the contrary, I often formulated my problem as one of distinguishing between genuinely empirical method and non-empirical or even a pseudo-empirical method. That is to say, a method which, although it appeals to observation and experiment, nevertheless does not come up to scientific standards. The latter method may be exemplified by astrology with its stupendous mass of empirical
evidence based on observation on horoscope and on biographies then next page it's third paragraph at the end of it it says once your eyes were dust opened, you saw confirming instances everywhere. The world was full of verifications of the theory. Whatever happened always confirmed it. Thus, its truth appeared manifest, and unbelievers were clearly people who did not want to see the manifest truth, who refused to see it either because it was against their class interest or because of their representations, which were still
unanalyzed and crying aloud for treatment. The most characteristic element in the situation seemed to me incessant the stream of confirmations of observations which verified the theories in question and this point was constantly emphasized by the adherence. And then he goes on to talk about Alfred Adler, Marx and Freud. but you should know that one of the hidden enemies in these early paragraphs is actually not what you might call to be pseudoscience Adler Freud and Marx it is actually positivists particularly logical empiricism so this is one of the
things that you should know that Hooper essentially wants to, it's kind of a sneaky, he wants to, this whole idea of science and pseudoscience is not that much important for him, it is, but really the ways that he wants to set the criteria of science from pseudoscience, what he might call to be, he can extract some conclusions some political conclusion some political tools that can equally be applied to logical empiricism or an empiricism in general
There's a sort of strange part that I wanted to go over or ask your thoughts about where he's talking about inborn expectations. In one second. Let me get this. Page nine. Yes, we haven't. We are going. So keep this so I won't forget it. And then really the most important part of the essay is on page 1, 2, 3, where it gives the list that ultimately
encapsulates its criteria of science. One, it is easy to obtain confirmations or verifications. For nearly every theory, it will look for confirmations. Two, confirmations should count only if they are the result of risky predictions. That is to say, if unenlightened by the theory in question, we should have expected an event which was incompatible with the theory, an event which would have refuted the theory. Now so you see, here problems becomes a little bit entangled, precisely because
How can we expect confirmation as the result of risky predictions? Risky predictions, we can only make such risky predictions within our theory. And to the extent that our theory is the one that ought to be confirmed or falsified, arriving at such confirmations, even via risky predictions, might in fact be the result or consequence of the theoretical distortion of inhabiting within that specific theoretical framework
that supports the risky prediction, and thus the confirmation derived from them. It reminds me of what you said a couple weeks ago about Empedocles jumping into the volcano or something. Yes. Taking that risk. That's a risk. Yes. But this risk, you see, it is risk. But nevertheless, risk, I would say, just like the idea of evidence, it's just too vague to be understood as something that we can talk about it as a criterion of science precisely because of that epistemological frame problem that I mentioned. If your risk, risky predictions,
are within your theory then what are you risking really? What are, why are they called risky predictions? Risky prediction is what you might call to be being outside of your theory to risk it. But of course Popper wants to show that we cannot do that really. Popper essentially, as I mentioned, does not believe in theory dislodgment or scientific revolutions in the Kuhnian sense. And that's basically, as we shall see, Kuhn brings this up. He wants to show that to talk about this fact that talking about a stepping over a stepping or side-stepping
from the theoretical edifice in which we are work and we are making such predictions it is can itself be right for dogmatism and again the same problems of confirmation and evidence So, number three, every good scientific theory is a prohibition. It forbids certain things to happen. The more a theory forbids, the better it is. Essentially, you see, you can now see that why he's a rationalist. It's exactly like the idea that what you might call to be
inferentialism, but set at the level of what you might call to be a scientific theory structure. Once we say, for example, this is red, it means that this is not blue. So making one move deprives you from making certain kinds of moves. And as your chain of inferences grow, if that chain of inference is great, actually it should have quite a broad range of restrictions about the kinds of moves that you can no longer take. Like this chair is brown.
Even in such a simple perceptual judgment, I have already ruled out many, many moves that I could take. But imagine if a theory is a chain of statements or claims and each of them have inferential connections with one another, that of course makes theory to get rid of many, what you might call it, called to be moves, possible moves, hence it restricts you. And within this restriction, within this chain of restricting the limitation of inference is that science can work.
It's basically quite from again, rationalistic perspective, you see that Popper thinks that the constraints of science are not limitations, they are enablements, they are positive constraints, they enable us. to progress the science. Five, every genuine test of a theory, sorry, four, a theory which is not refutable by any conceivable event is non-scientific. Irrefutability is not a virtue of a theory as people often think, but a vice.
Fine. Every genuine test of a theory is an attempt to falsify it or to refute it. Testability is falsifiability. Here a little bit of what you might call to be selfsame paleonism here. But there are degrees of testability. Some theories are more testable, more exposed to reputations than others. They take, as it were, greater risks. And of course, the idea of testability, as we will see in future sessions, I think when
we are talking about Popper's idea of deductivism. The whole idea of testability for Popper comes hand in hand with what he calls the degree of corroboration, the degree of corroboration. He does not say confirmation ever because you see confirmation in the canonical philosophy of science is always what you might call to be an empirical confirmation. And empirical confirmation also, for that matter, has inductive, basically, significance. To the extent that he does not want to identify science as induction, he does not want to talk about testability in terms of degree of confirmation.
That's what Carnot does. And we will see how Carnot does it. Popper in fact goes for something called degree of corroboration which you can say it's identified in terms of logic, probability deductive logic, of how, for example, a false instance, you know, and what you might call to be a test exemplar, a falsificatory exemplar,
does to a theory, does to a theory and basically calculate the degree of how much whether such a theory after surviving the test is being corroborated, essentially deductive logical concept, or falsified. Number six, confirming evidence should not count except when it is the result of a genuine test of the theory, and this means that it can be presented as a serious but unsuccessful attempt to falsify the theory.
Seven, some genuinely testable theories when found to be false are still upheld by their admirers. For example, by introducing ad hoc some auxiliary assumptions, the one that I mentioned at the end of that introduction, or by reinterpreting the theory ad hoc in such a way that it escapes refutation. Like you add more and more, it's like basically patchwork, patching up the system, patching up your theory. But you see, patching always comes with a price in the sense that at some point you might say that you have, if you haven't decided the degree of the falsifiability, the exact degree of falsifiability for the
exact statements of that theory, if you simply patch it up by these auxiliary assumptions, You will never know, in fact, which statements of your theory and to what degree they were either corroborated or falsified. It says again, such a procedure is always possible, but it rescues the theory from refutation only at the price of destroying or at least lowering its scientific status. I later described such a rescuing operation as a conventionalist twist or a conventionalist
strategy. Now, again, so Theo, you want to point out to that part that you wanted to ask a question before I move forward? It's a little further in the essay. Page nine, you said? Yeah, is that all right? Which paragraph? It's the second paragraph at the end. He says, there's no danger here of an infinite regress. Going back to more and more primitive theories and myths, we shall in the end find unconscious inborn expectations. The theory of inborn ideas is absurd, I think, but every organism has inborn reactions or responses.
I still cannot find it for some reason. Let me just, organism, let me just search it. Okay. okay yes yes yes and then the key key thing that i was going to point out was he says in the fourth paragraph he says thus we are born with expectations with knowledge which although not valid a priori is psychologically or genetically a priori for example prior to all observational experience. One of the most important of these expectations is the expectation of finding a regularity. And I was just gonna ask for clarity
about how would you go about developing inborn expectations if not through observation and then the only way that I could see getting around that is by giving some, by saying something like experience itself is conjectural, but then how do we arrive at new conjectures or something like that? Okay, I fundamentally, I think I disagree with Popper and probably this is also, you see there is, I think this is something that And also Putnam tries to put forward in the sense of Putnam, it's the idea that he reframes
such what you might call to be the a priori status of experiential expectations of regularities or invariances as something inbuilt, as something that's just evolutionary given. I do not find resorting to such what you might call to be, you know, inborn reactions and response or inborn expectations, a very convincing argument for how we are actually doing science. precisely because it follows into a vicious circle.
You are using, resorting at the bottom level, some inbuilt mechanisms, inbuilt expectations, which might in fact have a priori status, but then you are using them in order to explain them. So, there are both the premises and the conclusions of how we think about the world. I fundamentally find such evolutionary or what you might call inborn arguments or inbuilt in this argument unconvincing on this fundamental epistemological ground that it is a petit
in a human sense. And I have seen it, it's actually brought up very, very quite often by him, who is actually arch enemy of, in fact, such stuff, by Patna, by Andy Clark today in the sense that he's trying to, in fact, lay out such ideas of expectation, for for example, in the context of predictive processing paradigm. So that's one that I am disagreeing. I would say the alternative would be to understand, basically, let me put it that way.
I would say that such arguments with regard to the source of the scientific enterprise are in fact pseudo-problems. We should in fact not talk about them in terms of philosophy of science. It should be something that should be discovered by the scientific procedures. What I am proposing is more of an abductive view of the conditions of possibility of scientific thinking. Abductive, in the sense that these are merely what you might call to be hypothetical conjectures, which are fundamentally constituted by our logical vocabularies.
By that I mean broadly the logical structuration, math, logic, language, so on and so forth. Within this view, again, I don't think that it's in fact proper for us to speculate it, as if there were such conditions of possibilities inborn or expectations or experiential. We should go, I think, the Persian way, in the sense that we think about this in terms of abductive reasoning or the coherentistic view.
We simply say that these might be plausible truth candidates. We say that they are plausible. Plausible because they can be at the level of the manifest. They can be corroborated, but also more importantly because they can hold together by virtue of some logical or theoretical structure. And only then, once we establish these truth candidates, and there are candidates precisely because they are presumptuous, they are not truth as such, they are presumptuous truth. They are simply candidates. Once we establish this, then the process of abduction starts
from the moment that we have set the horizon like this by virtue of this plausible truth candidate, and then by way of the exact methodologies and in correspondence with evidence, we can hone out, prune some of these assumptions or truth candidates, retain some of them and in fact we might actually end up fundamentally changing them for something else. But I would say that, yes, I just genuinely never understood the merits of such postulates or hypothetical discussions about the origin of science, particularly set within the domain
of experience as inborn, as evolutionary, inbuilt, so on and so forth. So are you taking a more transcendental approach in that sense? Yes, yes, yes, yes, but what you might call to be methodologically abductive. There is this book that I really, really recommend for you, it's Thinking About Knowing by Jay Rosenberg. I saw one comment that was asking me about this Persean view.
The best example that I can come up with is this one. It's fundamentally great work. So let's end up this session. We are still very behind of the schedule, but nevertheless, as those of you who have been part of my classes, we can always take our time and add some free sessions at the end. So there is something actually very interesting in the notes if you have read them. Code 6 is actually a very nice counter-challenge to more of the Wittgensteinian meaning-oriented
view of science or semantic view of science, primarily semantic. He says, Wittgenstein's example of a nonsensical pseudo-proposition is, Socrates is identical. Obviously, Socrates is not identical, its negation must also be nonsense. Thus, the negation of any nonsense will be nonsense, and that of a meaningful statement will be meaningful. But the negation of a testable or falsifiable statement need not be testable. As was pointed out first in my LSCD example page 38F and later by my critics, the confusion
caused by taking testability as a criterion of meaning rather than of demarcation can easily be imagined. So basically, this is another indication that essentially Hooper does not primarily imagine the criterion of falsification or testability as a logical criterion in itself. Another one which I mentioned is something that people don't talk about this that much
is that the degree of falsifiability as proposed by Popper is in fact the first known systematization of the concept of informational content within science. You see, logical positivism in the half of 20th century brought back the older projects of empiricism but you know of course with with a kind of a logisticism undergirding and its ambition was you know to reconstruct scientific knowledge on the basis of direct
observations and logical relations between statements about those observations and we will we'll look into this. This sounds very, you know, parochial, but believe me, once we read Carnap Afbau, Logical Structure of the World, which is one of the most famous works in the tradition of logical empresses, this is actually fundamentally, fundamentally a very, very sophisticated way of thinking. Of course, as I mentioned in the previous session, it's famous that Carnap knew that he became dissatisfied by this project, but nevertheless his dissatisfaction wanted, he
wanted to test this dissatisfaction by taking this sophisticated idea of logical empiricism to its ultimate conclusion, to see that if it fails, then basically this program is no longer tenable under any heading whatsoever, and it failed. The old criticism of Kant on empiricism, again brought back by Quine, within the framework of this positivistic program, induction was invalid and causation could never be established objectively. So in response to this program in 1934, Popper formulates his well-known demarcation criteria
and he positions this explicitly as a solution to Hume's problem of induction. Scientific theories formulated as general laws can never be verified definitely, but they can always be falsified by only one observation, which we have talked about. It's called the atomistic perspective of falsification, right, on the holistic point. This implies that the theory is more scientific, it is richer, and provides more opportunity to be falsified. Now in this regard, in his 1934 paper, Popper says that, thus it can be said that the amount of empirical information conveyed by a theory or its empirical content increases with its
degree of falsifiability. Now this shows that the ambition to measure the amount of empirical information in scientific theory understood as a set of logical statements was already recognized as a philosophical problem more than at least a decade before Shannon formulated his theory of information. Popper was in fact completely aware of this fact that empirical content of the theory is always related to its falsifiability and this is in turn has a relation with the probability of the estates meant in the theory. Theories with more empirical information are less probable.
Hence, he distinguished logical probability from numerical probability, which is basically employed in the theory of gains and chance, and of course in the statistical notion of probability. So and again, in another part of this essay, he says the logical probability of a statement is complementary to its falsifiability. It increases with decreasing degree of falsifiability. The logical probability one corresponds to the degree zero of falsifiability and vice versa. It then continues, it is then possible to interpret numerical probability as applying
to a subsequence picked out from the logical probability relation for which a system of measurement can be defined on the basis of frequency estimates now of course proper you that the reason that I am going to this tangent is to talk about that mysterious footnote at the end with those formulas the idea of the empirical information put forward as probability measurement of the informational content of that theory. Popper is famous that he didn't neither have the cognitive resources nor the enough mathematical
resources to fully flesh out this theory, but nevertheless, yes. So the idea of information theory as formed by Shannon for the first time in a completely systematic way was preceded by Popper's idea interpreting the empirical information and degree of falsifiability in terms of probability. So, we can have a couple of questions and then we can conclude this session and then
next session Kuhn and Stegmuller. where Kuhn particularly debated Popper over basically the enterprise of science. And then we will show through the work of Stegmuller that all such debates depend on what we mean by the structure of scientific theory and its dynamics. And in fact, as the segment shows that all of these debates between Feyerabend, Lakatouche,
him and Kohn were basically talking past each other in so far as their view of a structure of structure and dynamics of scientific theories were essentially conservative. They were only seeing one picture and not the whole. I have a few questions. Sure. So first, have you ever had a thought why, well, this problem of induction and I'm interested in so do I understand
correctly that Hopper says that induction is basically nonsensical when we speak about science and and still he points out that a lot of people still believe that induction plays important role in science and I wonder why a lot of people including Max Born I think why do they stick to induction when frankly not you shouldn't you shouldn't stick to it and yes you need maybe so I was thinking about it and I would think about it but could it be it's related to
to the way maybe how we learn? Yes, yes, absolutely. It's precisely because in fact, that whole point, paragraph that Theo mentioned when he was talking about inborn is absolutely how it shows that the implicit assumption that even though he has rejected, proper has rejected induction, but nevertheless, He resorts to those experiential biases, inbuilt biases. By that I do not mean dogmas, inbuilt biases, what you might call to be in today's predictive paradigm processing, they call it hyperpriorance,
of how we basically see observations in the phenomenal realm, given the kind of, for example, neural structure that, for example, we have, or neural setup we have. Yes, but, and that's why I think that Popper, even though he, as I mentioned, even though he is against this idea of induction, but nevertheless, he brings something, a question, the question of inbornness, which is essentially the question that once is defended, you in fact implicitly commit yourself to induction. And that's, these are good, I think,
inconsistencies in Popperian view. And yes, absolutely, I do agree. And I will, when we are talking about induction, we will show that in fact, you know, Hume, yes, Hume understood a picture of the problem of induction, one of the main things that was wrong, but he really did not understand what the problem of induction in fact was. The only people who really understood it was Carnap and Nelson Goodman. And the thing is that Carnap shows that there is such a thing as inductivism or not induction in the sense that you might
say that I'm talking about regularities of observation according to my inbuilt biases. He says that inductivism is very different from induction and yes, science uses inductivism, but what is inductivism is what you might call to be logic, probability logic of induction in which induction no longer formulated in terms of observational regularities, but in terms of the logical and the probability of logical statements in which our observations are cut up.
But then I will show that Carnap also didn't understand the nature of the problem by way of Nelson Goodman. And even to this day, those people who even refute idea of induction as that's kind of simple observational induction, but they go for the Carnapian version of a probability logic and couch it in terms of what you might call to be formal accounts of induction, like Ray Solomonov or, you know, Andre Kolmogorov, so on and so forth, or Paul Vittany and recently Marcus Hutter. Still, they have done, you know, great at great length showing that,
yes, absolutely, inductivism is a component of science. By that, I don't mean induction. It's basically what you might call to be probability logic that follows the norms of probability logic and not observational regularities or observed regularities. But even then, in the sight of the formal theory of induction, which is essentially an information theoretic computational thesis, is that it can be shown that it has not only contradiction but unwarranted metaphysical assumptions, particularly with regard to the notion of simplicity or elegance. And again comes back to that idea that we were talking about that there are different
classes of metaphysical assumptions. I talked about causation and time or time arrow. Another one is simplicity. Why simplicity should be thought as a universal criterion of arriving at a set of candidate hypothesis from the point of information theory or probability logic? Sorry, sorry, please go on. I'm less familiar with probability logic. Could you just say in a bit more detail how that is fundamentally different from searching out regularities? You see, searching out regularities is what you might call to be, it comes in different
kinds. For example, you might say that searching out regularities you can think about in terms of ordinary observations. All ravens are black. We have seen black ravens, we haven't seen white ravens. So we say all ravens are black. It's essentially such a statement can be constructed by enumeration, enumeration of empirical observations, ordinary empirical observations. Now, probability logic is not really observation-based logic.
It tries to show that if we have within the parameters of probability, and of course there can be many, many different concepts of probability, but usually it's a statistical or frequentist idea of probability, using such a logic, which again is not dependent upon observation, it's what you might call to be degree of confirmation. We can show in that, for example, there are such irregularities, but such a regularity does not, for example, what you might call to be tantamount with all ravens are black. You might say the degree of probability of
such observations, for example, above 0.5, below 1, or above 0. So you create parameters. And also another thing is that the famous formula for degree of confirmation in the probability logic is called C, parenthesis H, comma, E equals to R. So C is what you might call to be confirmation, criteria of confirmation set by not our observations
but by criteria of statistics. So H are basically are, is a, what you might call to be, what is it, is a sentence hypothesis, usually a singular one. E, now many people think, and we will get into this when we are talking about inductive wisdom, many people think that E actually is an empirical evidence, like a piece of observation. No, it is not a piece of observation really. It is a logical reference in a Carnapian sense to an observation or a series or a class of observations.
It's essentially a logical statement rather than observational statement. And R is what you might call to be degree of confirmation, which as I mentioned expands from 0 to 1. and depend on what you might call to you it can give you a degree of whether this observation can in fact be turned into our Ravens or black or not we will talk about this in more details another thing is that regularities are not again observational. They can be in fact, I mean they cannot simply
obtained by observations. They can be obtained, you can have the traces or single atomic observations, but the way that you postulate the regularity is by how they can be compressed computationally which is that's basically the core thesis of Solomon of every data sorry every regularity compresses the data and every every every every every compression is a regularity and every regularity compresses data this is called the duality of compression and regularity, which is fundamentally couched
in terms of information theory and computation. Okay, shall we convene as of my back a little bit, getting again? Sure. Yeah. Thanks, everyone. We went a little over time today, so I'll end it now. Okay, we will start our topics and thank you, everyone, for a fantastic contribution.