Ep. 280: Imre Lakatos on Scientific Progress (Part One)

0:07 Mark: You are listening to the Partially Examined Life, a podcast by some guys who at one point set on doing philosophy for a living, but then thought better of it. Our question for episode 280 is something like, how does science make progress? And we read Imre Lakatos's Falsification and the Methodology of Scientific research programs from 1970. For more information, please visit partiallyexaminelife.com this is Mark Linsenmayer, true by definition.

0:31 Seth: This is Seth Paskin, never in danger of being falsified by a fact. In Austin, Texas.

0:37 Wes: This is Wes Alwan, neither confirming nor denying that I am in Cambridge, Massachusetts.

0:42 Dylan: This is Dylan Casey watching Minerva's owl gliding along the Willamette under dusk in Portland, Oregon.

0:49 Wes: That's fancy.

0:51 Dylan: You guys probably missed it if you didn't read all the way to the end, he says. I hope I've shown all these theories of instant rationality and instant learning fail. The case studies in this section show that rationality works much slower than most people tend to think. And even then, fallibly, Minerva's owl flies at dusk.

1:08 Mark: I did read that. I just didn't read the appendix. By that time, I was kind of done with this. So we haven't done our philosophy of science sequence in a while. I mean, we did like Bacon not terribly long ago, but this is a direct follow up from our Karl Popper and Thomas Kuhn episodes, which are very old did many, many years ago, because he's kind of trying to split the difference between the two. That Karl Popper had this notion of falsifiability, that while you can't actually prove general scientific claims from individual instances, right? Induction doesn't work.

You can definitively falsify them by just, you know, your scientific theory implies a certain observation. You don't make that observation, you make the contrary observation and therefore your theory is false, and you move on to better and better theories. And that's the way it works. Thomas Kuhn had said there's no progressive science in that sense. It is these paradigms, you know, sort of whole worldviews. And if you're an adherent of one paradigm and there's counter evidence, you don't just throw away your paradigm. No, you cling to your paradigm.

But eventually enough counter evidence, enough problems with the paradigm, enough anomalies build up that maybe not you, the believer in that paradigm, but other scientists who are open to other ideas, basically the believers in the old paradigm mostly die off, but there's some sort of revolution, religious conversion, there's some change in fashion that makes the new paradigm finally take over the old one. And Lakatos says neither of those is true.

2:41 Seth: He starts off with actually a bigger project. I think it's not just about, let me examine these two points of view. He's trying to get at the notion of scientific knowledge, period. So he mentions, for example, the positivists, and he's got a lot of ists and isms in here that I kind of glossed over. But, you know, it's really the bigger question of how can we be sure of our scientific knowledge, how can we ground it? You know, it's the same project that the rationalists have had for hundreds and hundreds of years of how can we put together, you know, an epistemological framework where we can say for sure that the scientific knowledge that we have is valid and grounded and that we can trust it.

And of course, the project for a long time was to try and do that from within the system. In other words, you take the theory and it's validated by facts and there's a kind of project like that. And then you get to an understanding that you can't prove from within the system of knowledge that the system of knowledge is consistent and valid. You need something outside the system to be able to do that. And then you have all these strategies about how you can counter that so that you don't have to have a metaphysical theory to feel comfortable with the validity of your scientific knowledge.

And so this goes way, way, way back. It's not just simply a 19th or 20th century problem.

3:59 Wes: This concept of that you're getting at, Seth, of. You know, there's an epistemological component to this, as in what counts as knowledge and in this case specifically, what counts as scientific knowledge that gradually will change. Those two things will get pried apart, or the question of knowledge will get pried apart from the question of what makes a claim scientific, which is not to say that the claim is automatically true or it's a form of knowledge. Right? So Lakatos reviews some of these different approaches to the question of in the beginning it's knowledge, right?

So justificationism, where you say that you can prove things from the senses, that I'm thinking, what is the grounds of scientific knowledge? Well, it's empirical. I sense things and then I develop propositions from experience and then I can in some weird way deduce theories or maybe induce, use induction to get theories out of those things. That of course, doesn't work for reasons we'll discuss. But you might approach this in terms of probabilism where you think of some theories as being more probable based on the empirical evidence, something that also doesn't work.

Popper's radical insight, or at least according to Lakatos, that's what it is, is that you talk about this in terms of what makes a claim scientific, even if it ultimately ends up being false. Right? And that is just the. That you specify what would falsify it. And this is the kind of talk we hear a lot today, right? Which is. It's really interesting. I don't know, just even in popular discourse and political debates, or maybe quasi theoretical debates, you'll hear people say things like, well, that's not even falsifiable, that's nonsense. That's been a very influential idea.

And one of the principal value of this reading is just how meticulously Lakatos shows the weaknesses of Popper's falsificationism, even as he considers himself within a tradition of Popper and developing a more sophisticated variation on falsificationism. But that more naive, what he calls a naive view, just that'll turn out not to work at all. Science just doesn't work like that, as Marx said. It's not just that we have these bold conjectures and then it's Newtonian mechanics one day, and then we find out something doesn't fit with it, and then suddenly, oh, Newtonian mechanics is refuted and now we're on to general relativity.

6:36 Dylan: No, part of that is that so much of falsificationism rests on crucial experiment ideas and that there's sort of a silver bullet solution to showing a theory is wrong. And Lakatos goes through pretty, as you said, meticulously showing how that isn't how it works, and then also providing, apropos of his own account, his own way of talking about science, providing an alternate theory, an alternate account for how science actually works, and that both answers the problem of its dynamic nature, that theories change as well as its entrenched nature, that it takes a long time for things to change, and its messiness, and then all the while maintaining, at least from Lakatosh's account, maintaining its rationality.

I think that Seth's absolutely right. He's fighting a fight that's been around for a long time. He frames it as the demarcation between science and pseudoscience, which he points as having vital social and political relevance and hearkening back to science as being the word for knowledge. And he'll point out that depending upon your authority, at some level, over and over again, different authorities will Render a judgment about what counts as science and what counts as pseudoscience. And he wants to articulate that line of demarcation as something other than the whims of whoever's in power and the whims of fashion.

Which is the problem with Kuhn, according to him, that Kuhn removes the line of demarcation from being anything rational at all.

8:19 Wes: Yeah, Kuhn does something that's very popular today. Right. He tries to give an explanation of scientific change in terms of human psychology and in terms of sociology, which of course the implication of that or Lakatosh's claim about that is that it means that scientific change is essentially irrational. We reduce it to these external explanations when there really is an internal logic to discovery, just like popular thought. There are really rational reasons why we advance from one theory to another. It's not random, obviously, and we should be able to figure that out, how that happens.

8:57 Dylan: And Lakatos is in that way trying to save science itself. And there's a long history of trying to save science as a way of providing us understanding of the world, knowledge about the world, by framing the ambition in the right way and by framing the process and account of the process in a right way.

9:16 Mark: Rational, the progress between different scientific viewpoints. But it is not rational in the way that we're going to get massive intersubjective agreement about it. It's really. It's a rationality that it makes the history of science look rational.

9:30 Dylan: They're like, oh, and in fact, very importantly, part of the process of doing science is constantly re narrating the history of science. And you have both, I'll call it good and bad things. You have to constantly keep that history in mind and reframe it in a way that accounts for it. But you also end up emphasizing or not emphasizing certain things. So part of the activity of science, when you say you get down a wrong path where you'll have brought a whole bunch of different anomalies into focus, you know, you've made your account for it. That's part of the intrinsic activity of science.

And you could frequently point to times where it was went astray and a new rival theory will successfully account for why those anomalies were accounted for wrongly in the past. There's this recapitulation. In fact, the way Lakatos talks about it, it makes it clear why every introductory physics book, every physics book starts with explaining how the whole history of science has led to physics. Now every single one of them does it. Democritus said back in as if it's a single line of thought. But I think Mark's way of saying that that's the way the rationality is working.

You're articulating, you're rationalizing, not in a negative way, but in a way to say how the account works all the way back.

10:54 Wes: What looks irrational in the actual history of scientific development is, for instance, the fact that theories are tenacious. That in fact, lots of anomalies arise, lots of observations are made that simply contradict the theory and quote unquote, falsify it. But of course, no one behaves that way. That's what looks irrational in Popperian terms. You get all this counter evidence and people say, so what? They're not really saying, so what? Sometimes those anomalies have been anticipated and they are according to the heuristic of the research program, and they're put aside temporarily.

Or there's even a kind of plan on how those anomalies might get resolved. Or in the interim, there might be some ad hoc explanations of those anomalies that don't turn out to be right. So in the near term, you might classify the chaos of scientific development as in some sense irrational. But there's a higher rationality to it that Lakatos said involves kind of waiting for a better theory to come along to explain those anomalies before you just simply abandon your current theory. So we don't just abandon a scientific research program based on counter evidence.

We work on a different theory or maybe a new iteration of the current theory. Something that can explain all the content the current theory. So, for instance, general relativity, right, can account for everything that happens in Newtonian mechanics, but it can also account for anomalies that Newtonian mechanics could not account for, like the perihelion of mercury. And it can predict lots of new stuff that ends up getting corroborated. So this concept of predictive power, in a way that's like one of the key phrases in this reading. He doesn't use the phrase predictive power specifically, but it's the kind of thing you'll hear philosophers talk a lot about today.

And I think it's influenced by Lakatos explanatory value or predictive value. That's, in a way, how we assess the scientificness of a theory or the value of a scientific theory. Its capacity to, the way a Lakatos puts it, is to produce new empirical content or to be empirically progressive, but basically to make predictions.

13:16 Seth: Yeah, specifically in his mind, there's just so much to unpack. This was A really rich reading. And unlike Mark, I actually liked the examples towards the end because I thought they were really fun.

13:27 Dylan: They were freaking awesome.

13:29 Seth: They were great. They were great. I think the key theme. So there's two I want to get to, I want to mention now, so I don't forget. The first is he does what I think is a slam dunk refutation of the importance of the empirical aspect of scientific discovery, meaning this idea that we move from facts to theories or that facts somehow prove or disprove theories. He makes a convincing argument that in fact the empirical piece of it is relatively minor compared to the theoretical aspect. And the theoretical work can be completely divorced in some respects and in some examples from that.

But more importantly, that what constitutes an important empirical fact at one point in time may become completely irrelevant later. And so we're constantly in a position of reinterpreting the empirical aspect of it through the theoretical and progressive. The progressive theoretical enterprise. So what I. Dylan, what I got out of that is the irony is Lakatos is in arguing against the irrational aspect of scientific paradigm development or whatever, he actually points out that there is an irrational element to science which is its own retelling of its own story and it's effacing of things that were important at one point in time and highlighting other things.

When it talks about the narrative to create the textbooks that you're talking about.

14:57 Dylan: Why is that irrational?

14:59 Seth: It's irrational because it's literary. It's a story which does not, as he points out, does not accord with what actually happened. You'd never learn the history of science that way.

15:10 Dylan: I think we're interpreting rational differently. Right? I think the rational is making it align, putting it in ratio. I mean, a narrative works just fine that way. I mean, a narrative does the same thing to our world experience. Right? Because it's telling a story doesn't make it irrational. Right. It'd be gibberish if it was that.

15:29 Seth: No, no, no, no, no. I'm thinking of rationality in this sense, which is, in his mind is associated with belief and shared belief. That's what he criticizes Kuhn for, is the idea that you shift paradigms when enough people say, oh, wait, I don't believe the old one anymore, I believe the new one, and that he wants to get away from this notion of belief. All I'm saying is that the narrative of science has normative elements that you want to believe. It's the story that science tells itself about its own history.

15:59 Dylan: Okay, well, maybe we can get into it more when we get down to the hardcore of a research program, the core elements, the hardcore of it and the belt of justification around it and the heuristics and how those work. I mean, I think that he gives a very persuasive account that that is a rational activity. I think it amounts to saying that if you're going to define rationality the way it was defined by the justificationists, then you're doomed to have no such thing as rationality. So part of what's going on here is articulating, when you're doing what Lakatos is doing, that is articulating the line of demarcation between science and pseudoscience. You are defining what rationality is. Other things are inspiration.

16:45 Seth: I'll retract the comment about it being an irrational activity, but I will not retract the part where I think he points out that it's really important to science to be able to tell a story about its own history that has little to do with how the history actually unfolded.

17:00 Mark: Can we just say what the research program, what that term is, since Dylan just sort of brought it up in the middle of a sentence. But it is the key, you know, his solution here. So he's trying to split the difference between Popper and Kuhn, which is kind of a funny way of thinking about it, because he is, in this work, obsessively goes over Popper's argument and different versions of it, and has so much respect for Popper, has very little respect, apparently, for Kuhn. But yet his position kind of, if you sort of were just evaluating it abstractly, is closer to Kuhn's.

It's just that Kuhn has these ideas of paradigms which are sets of theories, sets of practices. So Lakatos just takes that, but renames it Research Program. So he has these elements that somebody mentioned. It's got some certain core theories and then a belt of auxiliary theories. And the auxiliary theories are the ones that typically get changed. Because when you're running a test, you know, if you're testing something about the Newtonian predictions of where the planets are going to be, then, well, you've got the sort of core theories about Newton's laws of gravity, and then you've got the auxiliary theories of exactly how this is supposed to spin out.

And the reason why Popper is wrong is because you could always say when some new thing, you know, this planet is not orbiting the way it looks like it should, well, maybe there's something else going on. There's another planet that's attracting it there's some other force that is pulling it off. So it's these sort of auxiliary hypotheses about what's going on that end up getting changed as opposed to the central. I'm just going to throw away Newtonian mechanics altogether. So we have this change resistant thing that's like Kuhn's paradigm. It's just that I think, as Dylan said, the defining characteristic of a progressive paradigm is it is predicting unexpected new observations come out of this.

And it's not even that, it's addressing all the anomalies that the previous version did. Maybe the anomalies in the previous version carry over into the new research program as well. It's just that a couple really dramatic predictions, you can say this is going somewhere, this is a progressive program. And it's not just that, unlike for Popper, where they're just like Marxism is just unscientific, it's just bad. It's that there's an evolution to these research programs that when they start they're making all these sort of predictions. And if those predictions are confirmed, then it continues to be progressive.

But probably over time, at some point in comparison to some other candidate theories, the things that are being predicted are not coming true. Or rather the alternate research program is giving a more accurate prediction of what's going on. And so that's what's supposed to be rational about. Why you would choose a progressive one over a regressive one at a given time is because the progressive one is predicting more stuff. But that doesn't mean that anybody who's working in the regressive one is completely irrational. By trying to turn that back progressive again to keep working on it, it comes down to some sort of judgment call that each person has to make based on their professional evaluation of the situation.

20:06 Wes: Well, ultimately it's not up to any one person, right? Over time a certain theory or research program will just succeed. The old one will be degenerative. You won't be able to get any novel predictions out of it. It'll become increasingly ad hoc and it'll be obvious to everyone after a while. In the beginning, not necessarily, but I think ultimately, Dylan, you're pointing to the fact that there's something rational, right? To the logic of scientific progress. I think Seth was just saying that. Correct me if I'm wrong, Seth, but you were just pointing out that Lakatos thinks that people often give histories to this stuff which are not as messy as the real history and they're irrational in that sense.

Right? The textbooks falsify the process to look more Popperian, to look more like, oh, we had a theory and then there was a dramatic experiment and the theory was refuted and you had a new theory that I think Seth is right about, if that's what he was saying.

21:03 Seth: So that's what I was saying. But the reason why I think it's important just to point it out and then I'll drop it is that it does a disservice to the description of what actually happens so that people can engage with it more and understand it more and perhaps actually actively understand when a competing progressive theory at what it's trying to do. And the examples we'll go over, we'll talk about that. That was the only point I had.

21:29 Dylan: I think it's a really good point. And in that way the scientific account tells about itself is almost always so clear and well articulated as to completely obliterate the messiness of the process along the way, much less obliterate the status of potential rival theories. It often presents more certainty than there was or it's hard to evaluate the levels of certainty and uncertainty in the state of things in that way. It just makes me feel like, especially for introductory scientific, like a physics textbook or something like that, it reads a lot more like a police procedural.

You know what the end is going to be? They're going to catch the bad guy at the end from the beginning. It's going to be all sort of tied up all the way through and it's going to seem like it's a challenge. But no, actually that's just a challenge that illustrates how awesome your original theory is, that kind of thing, when the details are a lot more interesting and harder to know, especially when you're in the middle of it. And I think his examples at the end, he just does an amazing job in terms of the presentation of the complications. Then in those examples of early quantum mechanics and Newtonian mechanics.

22:43 Mark: Before we start with the text, I guess I want to bring up the issue of what is he actually doing here? And I think what we've been arguing about in terms of is this reconstruction of history, rational or not? The other place that rationality seems to come in is is there some way that scientists at the time, when they're doing something, you know, is there some firm guide that can tell them this is not the research program to going with anymore, you should jump to that progressive one. And I think he thinks that is not the case.

23:12 Dylan: I think he thinks that's not the case. And I think that he thinks that no scientists thinks that's the case. And I think he's right.

23:17 Mark: Right. So that is to say what he's doing here is something about the history of science, about how we tell the story of the history of science to make it seem as rational as possible. And that is the way that he judges what's going to be a good philosophy of science, a good theory of history of science, that if you take Kuhn's theory of the history of science, remember Kuhn himself was a historian of science. Like he was into exactly the same kind of stuff that Lakatos is in here. He's telling these stories, not about the thick of actually doing science himself.

Compare the two pictures, Lakatos's vs Kuhn's. Kuhn's is just describing like just these spasmodic unpredictable changes that it again, it's sort of like the old people die off and the new people. I think that's a lot of what ends up happening. And there does not end up being a picture that we objectively looking at this history can say, oh, I can see that those later people, they were right to reject that earlier thing. And those people were right to reject the earlier thing. Like Lakatos is positing that out of a philosophy of science, what we want is to feel like, yes, there was actually reason and progress and it's not just random bumbling around, but yet knowing that as a historian is not going to be useful to scientists.

And it sounds like you're not disturbed by that, Dylan, because you think that's not what philosophy of science is supposed to do, actually give direction to scientists.

24:41 Dylan: I'm not sure. There's a couple of things that I just completely disagree with you about, but I'm trying to figure out what they are.

24:46 Wes: So I just, I'm a little unclear. Yeah, too on what you're saying, Mark. But the idea is, yes, we can say some scientific theories are better ultimately and that they're in closer accord with reality. He makes that very clear towards the end that that's his position. This is not just sociology. We're not just explaining this in terms of human psychology and social history. What happens with scientific progress and investigation is ultimately tethered to the way that these theories are related to a mind independent reality. So he calls what he's doing rational reconstruction.

So we can discuss what that means at some point. So he's making the story neater than it really is as well. But I think he thinks he's doing that for the sake of just exposing the true logic of discovery that lies underneath some of the more chaotic elements of what's going on. And he does it in a very convincing way. But, yeah, we can look at each stage and say, yeah, Newton was in some sense wrong, strictly speaking, Newtonian mechanics was false, but it's true for most circumstances, let's say.

26:00 Dylan: So you're hitting on just parenthetically. Probably the biggest thing that I felt like he left out, which is why there are numerous theories that we actively work on and leverage and use to figure out things about the world that quote, unquote, strictly speaking, are wrong, and the terms under which we hold those and what that sort of means. He doesn't engage in that. When you're done reading his paper, you feel like you'll have this kind of messy space in which there are multiple theories that are held, but they're all sort of vying to be king of the mountains and that the rational progress would get you to have one king of the mountain.

And that fact is complicated because what do you do with the fact that we use Newtonian mechanics to fly satellites, we don't use general relativity, and we don't use quantum mechanics to do either of those things? And it's also the transition between very, very small things in quantum mechanical level, up to even normal sized things. That transition of the theory is not well understood still. And what it means to talk about the world in the quantum mechanical terms and talk about it in sort of conventional classical mechanical terms. You might be able to talk about the hard core of all of those things in terms of certain kinds of conservation theories and things like that.

But he doesn't go into those parts very much. And maybe we can talk about that a little bit.

27:32 Seth: Yeah, I'd like to transition. Can I read just the last paragraph from the introduction? Because think we're talking about the epistemological aspect of it. But for him, this project is much bigger and much more important. Are we ready to go there? Mark, I see your face.

27:46 Mark: Can I just clarify what this introduction is? That it's actually a separate essay. So the essay that we're mostly focusing on is Falsification and the Methodology of Scientific research programs from 1970. The book that I picked that we're reading it out of is the first full chapter of the Methodology of Scientific Research Programs, Philosophical Papers, Volume 1. And the introduction after the editor's introduction is something called Science and pseudoscience from 1973. So it is in some ways redundant of. But it's, you know, really readable and short. It's like a seven Page summary. Go ahead, Seth.

28:23 Seth: All right, so he sets out saying that the goal here is to try to understand what the criteria for defining scientific knowledge would be so that you can determine what science and what's pseudoscience. And he says the problem of demarcation between science and pseudoscience has grave implications also for the institutionalization of criticism. Copernicus's theory was banned by the Catholic Church in 1616 because it was said to be pseudoscientific. It was taken off the index in 1820 because by that time the Church deemed the facts had proved it and therefore it became scientific.

The Central Committee of the Soviet Communist Party in 1949 declared Mendelian genetics pseudoscientific and had its advocates, like Academician Vavilov, killed in concentration camps. After Vavilov's murder, Mendelian genetics was rehabilitated. But the party's right to decide what is science and what is publishable and what is pseudoscience and punishable was upheld. The new liberal establishment of the west also exercises the right to deny freedom of speech to what it regards as pseudoscience. As we have seen in the case of the debate concerning race and intelligence, all these judgments were inevitably based on some sort of demarcation criterion.

This is why the problem of demarcation between science and pseudoscience is not a pseudo problem of armchair philosophers. It has grave ethical and political implications.

29:43 Wes: I would add to that that this is not really just applicable to science. I think this has very broad applicability to intellectual debates and political debates, even to the concept of free speech. I think in some ways, when people are discussing. Not that I want us to get off on the sidetrack, but just as a general observation before we dig in, when people talk about the question of free speech, they are often talking about what their model of discovery is or what they think it should be, how discourse should be structured and how different theories within discourse ought to be evaluated and how they compete and replace each other. Maybe towards the end, I can explain that a little more if there's time.

30:29 Mark: Yeah, I was thinking of that a lot since we just did the critical race theory, that if you say one of the core tenets to our worldview right now is that everybody is approximately equal, then I'm not going to be interested in your differential measuring people's IQs according to whatever grouping, because whatever you come up, even if you come up with this group says much higher. I'm not going to conclude from that that, well, that is just the genetic basis of that or something. I'm going to say there must be some environmental thing because it won't be the core that's going to shift. It's going to be one of the something about the way you did the measurements. That must be pseudoscience. Just don't even do that. Don't even look into that.

31:10 Wes: So I think a good place to start in section one of the essay that we're concentrating on, he goes ahead and gives a lot of different definitions of justificationism, probabilism, and then Popper's falsificationism, which we talked about a bit, and Kuhn. But it's section two where we really get into the meat of his refutation. I don't know what you want to call it, but his critique of what he calls dogmatic or naturalistic falsification, which he sees as what he calls the weakest brand of justificationism. So I thought maybe getting at those three different, what he calls two, but really turn out to be three different false assumptions would be a good starting place. Although it sounds like Mark, I think you're probably going to want to define justification as him now.

32:00 Mark: Yeah, I mean, just that first step, because I think this is something that came up in somebody's speech very, very early in this episode, was that Popper's first move that Lakatos discussed was to prove that you can never prove a statement based on an experience. Right. He says it that baldly and that sounds just absurd to most people. He says this is a basic piece of logic, but escapes the notice of most people. Do you guys remember where this is?

32:27 Wes: It is right near the beginning. The idea is that the concept of proof, it's about the relationship, a deductive relationship between axioms and conclusions. And what constitutes proof is a logical relation. So to call the relationship between sensory experience and statements proof, in a way, is just a category mistake. And it gets even worse, right, if we do induction. So it's simply not demonstratively the case that if we see enough white swans and we think if we come up with Sabin, all swans are white. Of course there's no deductive power to that generalization and very obviously it can be easily falsified.

So I think some of this has to do with the problem of induction from sensory experience. And then as we get further into his objections, we'll see that there's also the problem of the theory ladenness of experience, which we've discussed a lot on this show, but that basically when you make observations, you're using equipment, even if it's your own eyes, but maybe it's also a telescope. There's kind of an observational theory that goes along with that. So we're never free of a theory.

33:40 Dylan: Mark, you were referring to his boldly saying that no experience can prove a theory.

33:47 Mark: Yeah. This is page 15 and it is in the context of talking about that first form of falsificationism, dogmatic falsificationism. But I think it's supposed to follow from the reason that the justificationist is that the kind that's before Popper. So there's justificationalism, probabilism, and then we get to Popper.

34:07 Wes: So justificationism is the idea to write. Our scientific knowledge is things that we prove via the sense's intellect, which I've called a category mistake. Probabilism, just tries to fix that and say, well it's just some theories are more probably true than others based on the evidence available to us, the empirical evidence available to us. And that's something that Popper was supposed to have rejected and showed was false. And I don't know that story. And that's not a story that Lakatos puts in here.

34:37 Mark: It's too bad that we can't pause for that because that seems really interesting that all theories are equally improbable because like that's something, you know, I felt that was definitely a stage in my even belief in God kind of thing. Like well, you can't prove that there's no God or that there is a God, but you could come up with probabilities. It seems unlikely that there's some sort of being in the sky, you know, just.

35:02 Seth: Right.

35:02 Wes: We did a whole episode on that. Who's the author again?

35:05 Mark: Mackie. But you know, strictly speaking, according to Popper, that's all hand wavy bullshit like that. You can't. For whatever reason that we can't spell out here. Yes, all theories are equally unprovable. All theories are equally improbable, equally probable. Yeah, you can't do the math.

35:23 Dylan: I mean, is it the sketch of that goes back to the provability of fundamental axioms or the givens in a given theory and that those have a status that is just not provable and doesn't land on a scale of probability.

35:39 Wes: That's a good point. Yeah, yeah.

35:41 Dylan: And the fact that they don't land on a scale, probability is directly related to them not being provable because probability just makes it so that you sort of stretched out the probability from being. It's either yes or no into some sort of degree. But they are of that kind. When Wes calls that a category mistake, I think that's exactly right. You've made a claim about the status of those givens in the theory that make them a kind of entity that they're not. They don't have a status of being probable or improbable that you can't number them that way.

36:15 Mark: Did anybody re listen to our Popper episode before this?

36:18 Wes: No. I was thinking about listening to Popper aku.

36:22 Mark: I felt like we probably didn't do justice to the subtlety that Lakatos is pointing out in Popper, that as far as I remember, all we really dealt with Popper is what is here called a dogmatic falsificationist. Right. Science grows by. This is page 13. Grows by repeated overthrow of theories with the help of hard facts. For instance, according to this view, Descartes vortex theory of gravity was refuted and eliminated by the fact that the planets moved in ellipses rather than Cartesian circles.

36:50 Wes: There are a few different elements to this characterization of dogmatic falsificationism. And it's not just that counter evidence can refute a theory, because anyone would say that, you know, that there's a possibility of that. It's just this idea that a proposition is only scientific if you can spell out the way in which it might be refuted. So you specify in experiments in advance how you would refute it. And that's why certain statements and Freudianism or Marxism are just treated as metaphysical for Popper or pseudoscientific because they're not falsifiable.

And as I said, this is an accusation that people frequently level at each other. Today you're speaking nonsense. The statement you're making could. There'd be no way to say whether it's true or not. There'd be no way to falsify it. The other aspect of dogmatic falsificationism is just that there is this sharp line between theory and experiment. So as Lakatos quotes someone man proposes, nature disposes. And that what we're trying to do, or that what happens during scientific change involves this kind of conflict between these two very well defined categories, one theoretical and one observational.

And as we'll see, he gives us three misconceptions operating in this point of view.

38:09 Mark: Well, and one of them. So that quote on bottom page 15, top of 16, the truth value of the observational propositions that supposedly then would falsify something cannot be indubitably decided. No factual proposition can ever be proved from an experiment. This is what's the Thing that I thought in isolation sounds really radical. Propositions can only be derived from other propositions. They cannot be derived from facts. One cannot prove statements from experiences no more than by thumping the table. And that's a quote from Popper.

38:42 Wes: Sorry, you're onto the second assumption, which is.

38:45 Mark: I know. I just thought it was the most dramatic of them, like that we might as well lead with it.

38:49 Wes: Yeah.

38:49 Dylan: Okay.

38:49 Mark: This is one of the basic points of elementary logic, but one which is understood by relatively few people even today. So justificationism, we were thinking, was wrong because induction doesn't work. Right. That if I view swan white, second swan white, third swan white, I can never prove from that all swan's white. And if that's the model that you think that science is supposed to follow, then like way back to Hume, he just showed that that never works. In fact, it doesn't even, as we were saying, make it more probable, strictly speaking, that the next swan will be white. Like you just don't know. Yeah, but this is not just about generalizations. This is about individual facts that I want to say. It's hard for me to. Yeah. Can you explain this a little further? I really.

39:35 Seth: So he uses a convention. I too struggled with this a little bit, but he uses a convention where he talks about observation and then observation in single scare quotes.

39:49 Wes: That's later. And that's complicated. Yeah, it's much later. That's the methodological falsificationism. Yeah, but we can't do that here in the dogmatic section. That's going to be confusing.

39:58 Seth: Okay, well, then I'll hold off on it.

39:59 Dylan: That's the hallmark of dogmatism, is not being ironic about it.

40:02 Wes: Exactly. Yeah. This thing that Mark, that you're pointing to is just. I think it's what we've already said, which is. What he'll say is that no factual proposition can ever be proven from experiment because propositions are. Can only be proved from other propositions. It's a derivation relation. So again, this is just a category mistake about how experience is related to our axioms. But I wanted to back up to the first one, which to me is more important and interesting, which is just that there's no clear difference between observational propositions and theoretical propositions because we've already pointed to this.

But observations imply instruments and the theories related to those instruments. Right. So Galileo's telescope implies an optical theory. So that when he makes his observations and apparently was controversial in his time, you know, someone might object. Well, you're seeing with the telescope isn't Necessarily really there because of x, Y and Z. About the instrument, let's say. So in each case there's a theoretical component to our observational statements. And then when he moves on to the second step, and then each of these is kind of builds on the other, he'll say, well, even if that weren't true, even if there weren't a theoretical component to our supposedly basic statements, we still couldn't call them proved.

41:22 Mark: I guess I read those as those two objections, as saying the same thing, that the reason that you can't derive a factual proposition from an experiment is because the experiment, the observation, is always going to be theory laden. It is always like for you to even say what that thing is that you just experienced. You're coming at that with some sort of theories built in. So this is kind of like that Hegel's. You know, you thought you could have sense certainty, but actually you had to bring in sentences, you had to bring in language, and that involves theories. And those theories could be wrong. So, you know, that's why the observation is not going to actually prove anything. It has to fit into some sort of mental framework that you already have in place.

42:06 Wes: He's granting the first one though, for the sake of argument. So even if there were no such natural demarcation, logic would still destroy these second assumptions. This second one he's already granting for the sake of argument that we could have non theoretical observational propositions. And I think what he wants to do in number two is something different and it's logical as opposed to having to do with a theory of observation.

42:30 Seth: Isn't part of the point that an observation or an experiment, it's simply descriptive. When there's a fact and you put a name to it, you put a proposition to the fact. It's simply like saying swan white. And the jump to all swans are white is not a jump that can be made. It's like the is ought distinction in some respect where you can collect as many facts as you want, but there's no logical way to get from the descriptive element into the theoretical element, which entails the predictive capacity that you're looking for.

43:05 Wes: But it goes even for singular statements. So my iPhone is blue. That proposition is not proved technically by the my experience of my iPhone being blue. Because proof is the wrong word. Proof is illogical. I think it's something very trivial. And that's why it's hard to, in

43:24 Mark: the colloquial sense, we would say prove to me that your iPhone is blue. You show me your iPhone and then I see that is it blue? And it is therefore proven to me. But what if that is overlooking to me is just this again, this theory ladenness of. There are other possible reasons why I could have that experience. I say I'm asking you to prove something and you actually do some sort of magic trick. So you don't actually have a blue iPhone there at all. It is a picture of an iPhone that is blue. You don't even own an iPhone. You know, you fooled me.

You inject me very quickly with something that makes me hallucinate an iPhone. There are all sorts of alternate potential explanations that we make the assumption in an experiment, as he says explicitly later, that all other things being equal, right? So that if perception is reliable, if my color senses are working, if there's not weird lighting going on, then putting this thing in front of me is going to constitute a proof. But that's a lot of ifs.

44:22 Wes: I don't think so. Sorry to keep disagreeing about this, but I don't think it would constitute a proof even if you had what Lakatos called a well confirmed ceteris paribus clause, meaning you'd controlled for all the variables. I still don't think that's proof. But it's hard to arbitrate this because he just gives us so little on what this means.

44:39 Mark: I'm willing to grant the grammatic logical point, but when we're actually thinking about real situations and what would count as proof, as I was just describing, we have to admit, right, that there's a relationship between observations and claims. If we say there's an unbridgeable gap, it's like is and ought, then we could never argue for anything.

45:00 Dylan: But that's what he's saving, right?

45:02 Seth: He says at the very beginning of the essay that the notion of provability, I mean, he alludes to the idea of correspondence like a correspondence theory of truth. And he basically says once you realize that you can't do this, you try to patch it up for the fallibility or these other things. But what you don't understand is you're still working in this correspondence paradigm that's just fundamentally broken. And he's trying to address that. And I think that the brokenness of the paradigm is exactly what you're saying, even if you are struggling to accept it. Mark, which is a mere observation of any sort, cannot act as proof for truth or falsity of a non observational statement.

45:45 Wes: We're getting bogged down on a tiny paragraph on this long reading, but I

45:50 Mark: think it's philosophically important. Right. Wes, you said earlier that what makes Lakatos ultimately see the history of science as rational is because he does think that the subsequent theories, the better theories, are closer to a mind independent reality. That was your words. I didn't see anything like that in Lakatos and wouldn't be surprised if he was completely scared away from that.

46:14 Wes: It's in the Stanford article. Oh, okay, that's a thorny problem. But unfortunately we don't get a lot of data on that from Lakatos himself. So. So I think the third element here in what he calls the three false assumptions of dogmatic falsificationism has to do with what he calls the demarcation criterion.

46:35 Dylan: Yep.

46:36 Wes: And his criticism here, and this is a pretty. To me, I was pretty like, wow, this is pretty radically stated, you know, when I first started this reading. But he says the most admired scientific theories simply fail to forbid any observable state of affairs. And then he gives that, this is on page 16, but he gives that nice account of planetary misbehavior. Right. There's a planet that doesn't behave as we would expect according to Newtonian mechanics. And then what do we do? No, we don't say Newtonianism is overthrown. We could say there's an unknown planet affecting it.

And then if we can't find that planet, we say there's cosmic dust or maybe there's magnetic fields. There's something interfering with the observation itself or it's just ignored. Just no one talks about it until it's time, until a better theory comes along that could account for why that's happening. So that's a really interesting observation, that falsification doesn't work in this naive sense, in this dogmatic sense, because there's no basic statement that's going to disprove a theory.

47:48 Dylan: I think that there's part of this that he. I don't know if it's overstating. I think he's certainly right that theories are resilient in this respect. That if someone says, oh well, look, I have this observation that is inconsistent with what Newton's theory says. The first thing you'll say is, well, it's inconsistent with the way in which we're playing out Newton's theory. But that means we need to understand Newton's theory better or we need to play it out to account for these kinds of circumstances, or we're just not understanding the situation clearly enough.

And he does say that, or things will just get ignored. And I suppose there's a certain amount of that. That's true. But there's also the effect that certain levels of inconsistency will regularly get grabbed a hold of within. I'll use his term later on within a research program that they will be a grand irritant. And the research program will keep coming back to trying to figure out why do I have this inconsistency. And there will be arguments about whether or not you need. I'll call it new science. In order to account for that inconsistency you need to revise the way you understand things or whether you can account for it with what you know.

And he has several examples, you know, the birth of special relativity and the messiness of that and how long Einstein and others had to basically argue that the special relativistic account was the more accurate one. Same thing with early quantum mechanics and the inconsistencies with the black body radiation and the different spectra for the different elements. You could account for those in different ways. And it took a while for that to get sorted out. But those controversies are. I'm just saying that those are anomalies and irritants that, that at times they get held up as something that everybody's worried about.

And it's not that they get ignored. I guess that's my point is that they don't always get ignored, but they may become a feature of the research.

49:49 Wes: Yeah, his examples later on make that clear that they stick around even if they're ignored temporarily or if they're not on the front burner.

49:58 Seth: Right.

49:58 Dylan: There are different classes of them. Sometimes they're ignored outright because no one sees or no one in the right position. This would be like a Kuhnian kind of thing. Right. No one in the right position is being listened to about whether or not it should be paid attention to or it may be the case that they are irritants that aren't slowing the train down. But people have questions about it. And this goes to how well people feel the belt of hypothesis is defending the hardcore of the center of the research program.

50:29 Wes: The upshot of this last point about the demarcation criterion is just that we can't really control for all possible factors. That's why we can't get such a clear cut contradiction between a theory and an observation. And that's where he talks about the ceteris paribus statement, where if you had a well confirmed one, you could say yes, I've definitely controlled for all the variables. So we never get that, we never get complete certainty that we've controlled for all the variables. So we can never be sure that some observation is actually refuted the theory as opposed to just as accounted for by some other factor.

51:07 Mark: And should we say ceteris paribus is all other things being equal, we've gotten the basic theory out, but there's a lot more detail and these stories from the history of science that are very fascinating that we will go through in part two. If you're a PEL supporter, that'll be the next thing in your feed. If you're doing it through Apple, it might not show up till next week, but you'll get it.