Sean Carroll's Mindscape: Science, Society, Philosophy, Culture, Arts, and Ideas - 215 | Barry Loewer on Physics, Counterfactuals, and the Macroworld

Episode Date: October 24, 2022

The founders of statistical mechanics in the 19th century faced an uphill battle to convince their fellow physicists that the laws of thermodynamics could be derived from the random motions of microscopic atoms. This insight turns out to be even more important than they realized: the emergence of patterns characterizing our macroscopic world relies crucially on the increase of entropy over time. Barry Loewer has (in collaboration with David Albert) been developing a theory of the Mentaculus — the probability map of the world — that connects microscopic physics to time, causation, and other familiar features of our experience. Support Mindscape on Patreon. Barry Loewer received his Ph.D. in philosophy from Stanford University. He is currently distinguished professor of philosophy at Rutgers University. His research focuses on the foundations of physics and the metaphysics of laws and chance. Web page PhilPeople profile Google Scholar publications Wikipedia Amazon author page

Hello, everyone. Welcome to the Mindscape Podcast. I'm your host, Sean Carroll. Some of you may know that one of the bonuses of my new job here at Johns Hopkins is that I got to choose my own title. I mean, my title is Homewood Professor, but then Homewood Professor of What? And I got to choose what that is, knowing that I would both be involved in the physics department and the philosophy department, I thought it would be fun to call myself a professor of natural philosophy. Those of you who know, realized that back in the day, before we had had separated out something called science and something called physics from philosophy, people like Isaac Newton or Galileo would have been considered to be philosophers. Newton famously wrote his Principia Mathematica, principles of mathematics, but it's actually mathematical principles of natural philosophy that is the full title of the book, and that would have been their labels. And I like that idea because I think that philosophy and physics, there's a certain kind of philosophy and a certain kind of physics that really, really, really, really, really, you know,

overlap that are almost indistinguishable from each other, asking the biggest questions about what is the world, what is it made of, where did it come from, why does it exist, those kinds of things that really intersect with more down-to-earth physics questions like, how does quantum mechanics work, what is fine-tuning in cosmology, things like that. And there's a set of people, it's certainly not just me, I didn't invent anything here, but there's a set of people who actively work in exactly these kinds of questions. We've had several of them on the podcast already. Genanne and Ismail, David Wallace, David Albert, and we have another one today, Barry Lower, who is friends with all these people. You know, we all hang out, et cetera. There's more

of them that I'm going to be bringing on as time goes on. What I like about Barry in particular is that he is absolutely well versed in modern physics and how it thinks about quantum mechanics, relativity, what have you. But he came up as a philosopher, right, as opposed to, you know, David Albert or David Wallace, who actually got their PhDs in physics. before realizing that the kinds of physics they were doing were heavily philosophical, Barry was there as a philosopher from the start and became interested in physics. And what we're going to talk about is truly philosophical questions, like what is a law, right? What is a law of physics?

What does it mean to talk about other possible worlds? David Lewis was the famous philosopher who really put the idea of possible worlds front and center in the philosophical imagination. How does causality work in the real world? Barry, along with David Albert, has been championing an idea called the Mentaculous, the idea of a probability map of the universe with the name stolen from a Cohen Brothers movie. So we'll talk about that also. How do you go from physics to the Mentaculous, the probability map of everything that can happen in the world? And from there, start talking about human-sized things.

The special sciences, as they are called, is that. their free will? Can we talk about human beings in a scientific way? All those kinds of things. It's a great, great example of an intersection of real science, real philosophy, and Barry is a really good person at explaining some of these esoteric ideas. So let's go. Barry Lower, welcome to the Mindscape podcast. Hi, Sean. So this is something where we'll cover a lot of ground that I think mindscape listeners have been exposed to before, whether it's the arrow of time or causality or things like that. But I thought we could start at the widest possible view just to really give an impression of how dramatic what we're talking about here is. You know, on the one hand,

you want to talk about, you want to reach the special sciences, as they're called, right? Why there is a discussion of things like psychology or sociology or political science or economics. But on the other hand, we're starting with statistical mechanics, you know, the stuff that was invented in the 1800s by Boltzmann and Maxwell and those people. Like, am I wrong to think this is quite a dramatic and long journey and we should be impressed that we can try to make these connections? Not wrong at all. I think it's incredibly astonishing. And one of the amazing discoveries that statistical mechanics is at the root of all the special sciences and causation and the era of time and much else.

And is this something where, you know, the philosopher on the street who is interested in these areas basically buys into that statement you just made, or are you fighting an uphill battle as a plucky minority? Well, my street doesn't have many philosophers on it. I live downtown Manhattan. But I think it's a new idea. I mean, the way it's gotten somewhat into the community is via a boath. by David Albert that was published about 20-something years ago on your book on, from a turn to here. I think these are the main sources for me. I studied statistical mechanics in college,

but I didn't have an inkling that it was going to be so important philosophically, not an inkling at all. In fact, the only way it came up in philosophy was over the issue of reduction and discussions of how thermodynamics gets reduced to statistical mechanics. So it was supposed to be the paradigm of reducibility. Right. But it's an enormous importance. From my view, it's more important than anything else in physics to philosophy. Good.

I'm not going to disagree with that. So then in that case, let's get into it. And I thought that the entry point, given other things we've talked about in the podcast, might just be the very idea of counterfactual talking, right? I mean, it seems that we can't in our everyday lives get through the day without talking about how things could have been different. And as a philosopher, I think that you're going to tell us, this brings in a lot of baggage and assumptions. I mean, what does it mean when we talk about if this had not happened, then something else would have been the case? Well, I think the sort of the origin of counterfactual thinking is actually somewhat, is what you might think of as subjunctive thinking about how we're going to act.

So we're interested in what would happen if I were to do this or to do that. And then that gets extended to think about what would have happened, were I to have done this or that. And then that's got extended to what would have happened, were this to have occurred or that to have occurred. And it turns out that this way of talking has been enormously useful as a way of talking about what actually happens. What really is is just what actually happens. and subjunctive or counterfactual thinking is just a complicated way we've developed, human beings have developed, in order to talk about what actually happens. Okay, maybe just say more directly about that.

Like, why do I need to talk about things that didn't happen in order to talk about what actually happened? Is it just more clarifying for me, or is it kind of crucial? That's a good question. I think I would answer my thing in two ways. one is that what we're really doing when introducing certain notions into scientific language, like counterfactuals or probability, is we're providing a way of simplifying to talk about complicated actual events. That's one part.

But of course, also it's very natural for us to just imagine how things go one way or the other. So imagination is going to be very central in them and important. and how these two fit together is a good question that I have to admit I haven't thought about it. So I'm glad you asked the question. All right, good. We like to spawn new research programs here on the podcast. But, okay, but clearly one closely related concept here is that of causality, right? Cause and effect.

I recently did a podcast with Judea Pearl, who is a big name in causal reasoning these days. And, I mean, maybe... I heard it. It was great. Oh, good, good, good. maybe put into your words how you think about the relationship between causality, which I bet seems to most people in the street as something that is here in our world and its relationship to counterfactual reasoning. Like in some sense, it needs to be a comparison between the actual and the non-actual. Well, of course, we all somehow acquire a notion of causation.

It may be it's even in some way innate that we have some sort of concept like that. and we mostly connect it to things that we feel with the motions of our bodies. But we extend it beyond that to think about what were to happen, where we'd move our bodies one way or another. And there's a close connection then between causation and counterfactuals. I'm not saying that you can literally say what it is for one event to cause another just in terms of counterfactuals. There have been attempts, as you know, by David Lewis and others to just characterize causation,

in terms of counterfactuals, other attempts to characterize causation in terms of probabilistic relations. I think these are going to be closely connected to each other. I don't think anybody has really ever worked it out just how they're closely connected to it. Another research program that I'm interested in. In the background of this, I always have in mind Bertrand Russell's famous remark, let's see, how does it go, that the concept of causation is like the British monarchy? I guess the British monarchy has just celebrated what, it's 70. fifth birthday of its current incarnation, Russell thought that like the British monarchy, it's as

useless as the British monarchy and should be dispensed with. Well, the British monarchy hasn't been dispensed with, and despite Russell's remark, causation hasn't been dispensed with. I don't think it can be. On the other hand, I'm not sure that anything exactly corresponding to our notion of causation will be found in reality, just something that does this work of causation. So we can understand why it is that we can talk about causation, even if we don't find literally some causation. I think a lot of our ordinary concepts are like that, and something that philosophers need to pick up on free will is another example.

We'll certainly get to free will near the end. But maybe this is actually a point worth highlighting because I think that a lot of people have in mind that they use. again, people who are not experts in these fields, think about cause and effect as just law-like behavior, right? And you're trying to draw a distinction between fundamental physics where there is law-like behavior but not of the cause-and-effect variety. That's what Russell had in mind. And this higher-level emergent world in which we live where you need to talk in terms of causes and effects. Yes, I mean, that's right. And the way I like to think about it is something

I learned when I was, I guess, even an undergraduate. You probably know this, but it's a very famous philosophical essay by Wilfred Sellers. I forget what it's called even right now, but in which he introduces the terms manifest image and scientific image. And I think that one of the main jobs of philosophers is to show how to fit together. Our manifest image, how the world appears to us in our ordinary way of thinking about it. It's a sort of biological and psychological endowment developed by culture with what science has discovered that the world is like by applying scientific methodology and discovering laws and so on.

All right. Sounds like a good thing to be able to do. So you mentioned David Lewis. Famously, he thought a lot about possible worlds. So this is the typical or the good philosopher move when you have an idea of like counterfactual reasoning and try to think through what is really going on. in some sense, when you say, if this had not happened, this would have happened, you're invoking the concept anyway of a whole other possible world. And Lewis took those concepts

really, really seriously. I think Lewis is often misunderstood. Okay. I don't want to say that you're one who's misunderstanding him, but let's see how this goes. Maybe it's me who misunderstands him, but I spent a lot of my life thinking about David Lewis. I knew him, but I wasn't a student of his. And for me, if I can say it's a preamble, I was at a party of philosophers. Those are the only kind of parties that will invite me a few weeks ago in which there were a bunch of philosophers talking. And I heard him using the talking, one of the people that was talking about someone they called the philosopher. And at first, I thought they were talking about Aristotle. But no, they were talking about David Lewis. Pretty good. All right. And right.

For contemporary philosophers of my ilk, David Lewis is the greatest philosopher of the 20th century, maybe of all time in many ways, although one could argue about what one thinks philosophy is in that. I don't want to get into an argument about that. But in any case, Lewis was really interested in actuality, what happens in the actual world. But what he thought is that in order to talk about the actual world, in a good way, in a simple way, in a convenient way, we need to introduce the notion of possible world. But then he realized that if we're going to really take this talk seriously, he followed his teacher, Quine, often think of Quine and Lewis as being very different, but in fact, they're very close to each other. Quine was Lewis's dissertation

advisor. He followed Quine in thinking that he, if he's going to talk about the counterfactuals using possible world semantics to make sense of counterfactuals, even though ultimately what makes a kind of factual true or false is just what's going on in the actual world, we could come back to how that's supposed to work, maybe. He's going to have to take the notion of possible worlds seriously. So he asked the question, what are possible worlds? And he came out with a particular view about it. Many philosophers saw his view about that, and they thought, well, Lewis, he's a great guy,

he's really smart, but he must be out of his blank, sweeping mind. And they greeted him with what he calls an incredulous stare. I have a very different view. I think when you really understand what he thinks possible worlds are, and you think about the rest of his philosophy, although I don't buy it. It's not for that reason. It makes a lot of sense, given the rest of what's going on, a lot of which many other people would accept.

So I think you can make good sense of it. But this may be a discussion that's more for the in-house philosopher than the one you want to get into. Well, maybe, but let's just sort of summarize, I think, what is going on for everybody else. The typical view of Lewis is that he really thought all the other possible worlds were just as real as ours. And you're saying that's not quite what he thought. He did think that. Okay. We understand real correctly.

But he also thought what they really are are just geometries with space time. points, which are decorated with what he called perfectly natural properties. When you really understand that, and you think you really need this notion of perfectly natural property to make sense of everything else in science, as Lewis did, I don't, but Lewis did. Then this idea of possible worlds gets reduced into the notion of possible geometry, which is almost a mathematical notion, because you do need something like space time points, And he's got to take them in some senses real, but they're not real in the sense that you can get from this, anything, any point in this world to any point in any other world. There are no distance relations between them.

Right. Okay. But they're real enough. And then they're decorated with real, perfectly natural properties, just as real as the perfectly natural properties that decorate the actual world. Is there anything special about the actual world? Yeah, of course. It's really special. That's what we are.

Good. And we're special. I'll take that as an answer. But okay, so let's home in. That's Lewis's answer. That's fine. I honestly do accept that. And Lewis had this idea that not only, well, I mean, maybe we're glossing over this for people who haven't listened to the Jadaa Pearl podcast.

When we try to understand causality in the language of possible worlds. So we're saying that something like. like X causes Y means in the possible worlds where X didn't happen, Y didn't happen. And then all, there's a lot of work trying to clean that up, yeah? Yes. That's the semantics. But what makes the statement true or false, the statement that the throwing of the brick caused the breaking of the window, what makes that true or false, are just events that happen

in the actual world. Sure. Events, which can be described without bringing in the notion of causation or law, at all. So for Lewis, what there is fundamentally is just space time points and perfectly natural properties instantiated at various space time points. Think about it as kind of like a big field. The points with space time, with perfectly natural properties instantiated. Given that, every other thing about the world is made true or false, including whether or not the brick broke the window, because the window to break, or whether or not you or I are talking right now,

whether or not my cat is purring or anything else. That's an astonishing claim. That is astonishing. That should be greeted with an incredulous stare until you see all the work that Lewis did to make it plausible. Wait, sorry, say what the claim is that is astonishing? The claim is, here's the philosopher's way of putting it.

All the truths about the world supervene on the distribution of perfectly natural properties throughout all those space and time. Okay. That is say, the distribution of perfectly natural properties that all space and time make true or false every claim about the world. Good. And the philosophical fun comes in because someone says the reason why the window broke is because the brick got thrown through it. And someone else says, oh, that means if the brick hadn't been thrown, the window wouldn't have broken.

And then someone else says, oh, but there was another brick that was being thrown at the window also. So the window would have broken anyway, even in that other possible world. and you have to be clever about saying which worlds you're comparing, right? Yes, that's a problem with causation in particular. That has to do with preemption or there's another related overdetermination issue. I think that's one of the problems that Lewis struggled with. And I think one of your other friend of mine who gets to the podcast with, I don't remember whether you talked about it's Ned Hall.

It's written quite a bit about that also. And my own view, I should say, is I haven't really, grappled with that enough, but lest I look, I didn't think it was fully worked out. Okay. But it brings up, I mean, Lewis, in trying to tackle problems like that, and there's other problems, maybe you can tell us what some of them are, but he introduced this notion of similar worlds, right? Like, how do you compare one world to another? And I still don't know what that's supposed to mean. How do I know whether a different world is similar to mine? Is there some distance measure that tells me how different worlds are? Great question, which is why I think,

while Lewis is a count of, so the notion of counterfact, can I back up a little bit and talk about counterfactuals? Please. The idea of talking about counterfactuals among philosophers, well, it goes way back. David Yume characterized causation in terms of counterfactuals. But there wasn't a really focused discussion of it until Nelson Goodman in the late 40th, And Goodman said this. He gave his example was, if I had struck the match, it would have lit. And Goodman says, well, what does that mean? Well, that means if here's the situation, the match wasn't struck, but now think of a situation of the match is struck. And otherwise, the situation was the same as it actually was. And the laws, does that entail that the match strikes? But what does it?

doesn't mean the situation was exactly it was. It couldn't be exactly it was because the match wasn't struck. So what he meant was the circumstances, the conditions had to be the same. Well, what does that mean? He called those conditions that were the same. He called them cotenable with the matches being struck. But not every condition that actually holds is cotenable. So for example, you might have someone who would only strike a match if the match is wet. In that case, the matches being dry would not be cotenable. in which case it wouldn't lie. So now I have to say, what does cotentable mean?

What David Lewis did was to get rid of the notion of contendability and exchange it for the notion of similarity. And then thereby produce a much more general account. I think this was all a big mistake. I don't know if you want to follow up, but why it's a mistake. I think everything can be done in terms of probabilities, and all those probabilities come from where we started, from statistical mechanics.

Yeah, no, that is where I want to go, but, you know, it is fun to sort of sift through the mistakes along the way, because, I mean, I'm not going to be cagey about this. I can remember my whole bleeping career to do it, right? I'm completely on board with where we're going to end up, so I want to cherish the mistakes along the way. I mean, when we do compare worlds to each other, one way or another, you know, our world is thought of it by physicists as described by some initial

conditions and then some laws of physics that tell you where you go from those initial conditions. So when you say something like the match wasn't struck, I could imagine either that I change the initial conditions in such a way that the world looks really, really similar except the match doesn't get stricken, or that just the laws of physics were temporarily violated in that particular striking of the match. And so they didn't follow from the initial conditions. Is there some preference for one of these over the other? So Lewis did the second. That was a big mistake in my view.

My preference is keep the laws the same under any counterfactual condition that's consistent with the laws. Not necessarily consistent with those being the laws, but consistent with the content of the laws. And that's the view about counterfactuals that I've developed. Right. Okay, good. So we're, it certainly, again, it makes sense to me as a physicist to, to, change the initial conditions a little bit, that doesn't change very much. It doesn't do violence to our system, but changing the laws of physics, even temporarily,

just seems bad to me. It's bad for many reasons. Here's one reason it's bad. If it were the case, as Lewis thinks, that if I had struck the match when I didn't, would require there to be a violation of the laws, fundamental laws of physics, before the striking was supposed to take place, then it should also be. be true that if I had struck the match, I should have won the Nobel Prize for physics. Why? I would have falsified the laws of quantum mechanics. Surely it deserves a Nobel Prize.

Yes, good. And the good news, just to anticipate, to foreshadow a little bit, is you don't have to imagine that. I mean, you can have this conversation without ever violating the laws of physics, right? Yes. Good. So what is a law of physics? Well, of course, that's the question that mainly fuels what I'm interested in. I'm writing a whole book about trying to figure out what that is. The way I like to put it, I hope you don't mind my repeating a little slogan I like, is that while it's the job of physicists to tell us what laws there are, it's the job of philosophers to tell us what laws are.

I don't know who deserves bigger salaries, but physicists have made a lot more progress than philosophers have on their question. There are a lot of different views about what laws are. They go way back to when the notion of law was introduced in the 17th century. While there's some notion like that going way back to the Greeks, it really wasn't clear that this is what is going on in physics, which should be searched for as mathematical laws that describe how things move around until sometime in the 17th century. In fact, the philosopher who enunciated this is appropriate. program is one of the main people that all philosophy undergraduates read is Descartes, but in a book that isn't read very much called Limon's.

And in that book, he literally says something like, it's the job of us guys, by which he meant both philosophers and physicists, since they didn't make a distinction in those days. I see now, given your new position that we're coming back to this situation now, which I think is good, that this is what. we should be done. We should be finding what laws are. And he had an idea about what laws are. Can I say what he thought laws were? Please, yeah. Okay. What he thought laws were is something that might surprise many current physicists. You tell me whether you'd go to your physics colleagues and tell them what they card. And Newton, for that matter, thought laws were. Newton literally says,

or one of his henchmen, says that what laws are are gods with volition. I knew you were going to mention John somehow, yeah. Theological origin to the notion of law when it was introduced in the 17th century. Of course, that's been falling apart, another project. Not for me, but for a historian of science, so if there's any historian of science who's listening to your podcasts, maybe one of them can take this up because nobody has done this. There is no book-length study of the history of the concept of fundamental laws of physics or the special sciences, if there are laws of the special sciences. One thing has happened is the notion

of the theological part has kind of fallen apart away, in which case the question arises. Well, if laws aren't God's volitions, what are they? Okay. Well, so what's the answer? Okay. So I think there, the number of ideas that people have given it answers. Some, I'm going to back up a teeny bit. One idea is just to find a replacement for God. Something that does God's work. What was God doing? He was pushing things around. So some philosophers introduced something into the world which push things around.

One philosopher that was David Armstrong, who put this question really clearly on the table of philosophers about 1979 or 80 in a book he wrote called What is a Law of Nature? He thought that laws were extra things, facts outside of the ordinary facts, which necessitate the facts that happen. The way he did it was very, very incongruous with the way physicists actually talk about laws. I won't get into details about, wow. A philosopher who's done it in a way along the same lines, but in a way much better because he's a very sophisticated philosopher of physics is my friend Tim Maudlin. So Tim has the idea that what laws are are elements of the fundamental reality, call them ontological if you don't mind that. Anyway, they're in reality, and they do something. They produce or govern by taking the state of the world at a time and making this state at the next time or nearby times.

That's Mordland's idea. It's an interesting and influential idea. My idea is not like that at all. I want to get rid of God. Not because I'm a non-believer. That's not the reason. I just don't think we need God to do this work. Right. So what that means in practice is that this idea of sort of guiding or pushing was the false step that people made. Yeah, I should very quickly mention. I doubt that Tim will listen to this, but if he does, I should make it. that whenever I bring up the theological arch and I want to do it making sure he's not in the room because he wants to kill me. But I do think it's a reasonable point to bring up because many of the intuitions that we still have about laws, and including the language that we use to talk about laws like governing reflect this. And I think it prevents us from accepting what I think is a much better view, which I can now go on and describe to you. That much better view derives, unsurprisingly, from the philosopher, David Lewis.

Okay. What Lewis said, and he took this idea from John Stuart Mill and from Frank Ramsey, who was an important mathematician, philosopher in the first part of the 20th century, he said, look, what physicists are interested is in describing all the facts that will hold throughout all its base and time. we want to have a way of describing them all, which is it somehow makes them accessible to us. What accessible means there requires much more discussion than I'll give it right now. I can give it. But it has to be simple.

Tell us a lot about them. It has to be simple. Lewis called this the best systemization of the fundamental facts. As I said earlier, for Lewis, the fundamental facts were just the distribution of perfectly natural. properties throughout all of space and time. So what you want is an axiom system which will best balance informativeness, giving as much information as one can about all those facts, in as simple way as can. And maybe it has satisfied other features too that make them accessible and important for giving scientific explanations and so on. And something you mentioned

earlier is very, fits into this very well, because dividing the facts up into initial conditions and general dynamics, the way initial conditions evolve over time, it turns out to be a way of doing that for a world like ours. It didn't have to be. This is a contingent feature, nature of our world, but it seems to be so for our world. And so, I mean, let's give our listeners some jargon to hold on to. Is this the humian versus anti-humian distinction? Yes, the God view, remember, scare quotes around God, so I don't cause problems for myself. It's the anti-humian view. Lewis is the Yumian view. My own view is a development, I'm not going to get into it as you ask me, but it's kind of like Lewis's, but does away with a lot of Lewis's

baggage and makes it much more fit in with contemporary physics. But that may be more than we want to get into for right now. When I talked to Ned Hall on the podcast, I brought up the fact that there is a name associated with the Humeian view, namely David Hume, but no name associated with the anti-Humian view. So maybe we should just call it the God view. I think that would be perfectly acceptable. No.

Well, call it that or the Mordland view, but don't call it both. All right. That sounds about it. Another couple of people, I should get this in, who hold the view, who I assume you know, on the anti-Umian part, who just gave an interesting talk and written a paper about it, is my former student Eddie Chenin and my colleague, Shelley Goldstein. So, Eddie, you probably know from, he's made a giant splash in the philosophy of physics community. And Shelley is, of course, famous for being the main, I don't know what's to say, promulgator of Bohmian mechanics. Yeah. All right.

There are people out there who are anti-humian, the gods I view. But the humians, among which I count myself, they're saying that the laws aren't governing anything. They're just descriptions, right? Is that too simplistic? Well, they're not just descriptions. You know, that's like calling, I don't know, the queen, just another woman. It's not like that.

They're very special descriptions. They're descriptions which are both combined. They come from a systemization that combines simplicity and informativeness and other features important to explanation in an optimal way. Right. Fair enough. They're very, very special descriptions of what happens. But at some level... Maybe there's no other junk in the world other than the ordinary facts. Right. There's the facts and that is what makes up the world. The whole fact, nothing but the facts, right? Nothing but the facts. But, So there's two questions that even though I'm on this side, I really do struggle with. One is, if what we think of as the laws are sort of ex post facto, there's the world in all

the facts, and then we sort of abstract the laws from them, why are there laws? Why is the world so regular and predictable? I mean, we can easily imagine any set of things happening in the world, and ex post facto look for simple descriptions thereof. but we have really special things going on, right? That's a great, great question. If I thought it was going to come up, I should have. I would have had it ready a quote from Einstein,

who asks himself that question. And he just, I don't remember the quote exactly. It's a pretty well-known quote, so one could find it, in which he says something to the effect of, it's awesome. But it doesn't lead him to believe that the world has been made this way by God, And it certainly won't help to posit some extra, what, force that makes it like that? Because that just raises the question again. Where does that force come from?

So, okay. So what he does say is the only thing you could do is be in awe. And I do think that's a good state of mind to be in a lot of the time for a philosopher or a physicist, I guess. Okay. So, I mean, I'm willing to sort of tentatively accept that. But I do think it's a little bit of a point for the opposite. side, right? I mean, if you think... How? Okay. Tell me no.

How is the opposite? What is the opposite side? Their philosophers have written papers. Some of them are quite good. Uh-huh. Arguing that. And I have a section of this book I'm writing trying to rebut them. So I'm not saying that it's totally closed matter. I get the pull of it, but I just don't see how positing something else, as long as it's literally a scientific posit, that's supposed to make things happen. gets you anyplace because the same kind of question arises about whatever it is that you posited. Okay, no, I think that that's perfectly fair. And it is, like it or not, it is very similar and analogous to discussions about God, where they have a feature of the world, and a theist

will say, I can explain this feature of the world by invoking God, but then you haven't explained, you know, where God came from. And the laws of physics are in an analogous situation here. If you count them as something extra, right? Yeah. So Einstein's point, I think, in this quote was some people satisfy their religious impulse by just saying God and going to church. Others satisfy it by just being in awe. And Einstein was preferring the second of Rose. Good.

Okay. I go with that. All right. So, yeah. I mean, just like why is there something rather than nothing, there's some features of reality for which it's, there's no deeper explanation. It's something that is. You know what Sydney Morgan Bessor said about that question? I do, but tell the, tell our listeners.

Okay. What Sydney Morgan Bessor said is that, well, even. there was nothing, you would still complain about it. It's very apt. So, okay, if I accept that there's the other slightly more down-to-earth question for the Humian view, which is how do we think about counterfactuals? You know, if I think about the laws of physics as summarizing what happens in the world, the facts of the world, then how can I extend them to other worlds, where those other

worlds have their other facts and therefore other laws in some sense? Well, that's a really good question. And I try to deal with that, too. I would say that there's a deeper question that has bothered philosophers, has written a bit about, is explanation. Because if laws are just descriptions, how can they explain? But, of course, this is closely connected to your question. Let me deal with your question for a minute.

Remember what we were saying about counterfactuals before? On Lewis's view, counterfactuals are made true or false just by the actual facts. that's it okay turns out that the way the laws are is they invite this distinction between initial conditions and dynamical principles

which say how initial conditions evolve this fits very much in with the way we're thinking about counterfactuals and so what we do in kind of factuals we say well under these alternative conditions how will the dynamical laws say that things evolve And the reason we do that goes back to a point also made earlier in the discussion.

And that is counterfactuals introduced for us to say about what would happen were we, were I, to make various decisions. One thing is I know that when I make various decisions, all that I affect is some little part of my brain or mind. I don't affect what's going on at the same time around me. Maybe I don't affect anything that I know about in the past. Maybe we'll come back to that. I can influence what happens in the future. If we understand influence properly, I don't influence the laws.

So I keep the laws fixed when I vouch to evaluate counterfactuals. I think the Umi-an account can provide an explanation about why you keep the laws fixed. Because the U-Mean account, by emphasizing informativeness and simplicity, leads in a world like arts that is a division between initial conditions and dynamical principles. We know that the most we can do is to make a little influence on certain parts of the initial conditions at a moment. And keeping everything else fixed, we then say, what would happen next? And to come back to our original way we got into this discussion, it turns out that statistical mechanics is the key to answering all that.

So let me just rephrase it to see if I got it correctly. Because again, I think I'm on the same wavelength. Even though the humian doesn't think of the laws as extra ontological, entities that are governing anything, we can still figure out what the laws are. And then it's perfectly sensible to ask, what would a counterfactual possible world look like with the same laws but slightly different conditions? Yes. I mean, you brought up a slightly different issue, and that is figuring out. There is a question about the epistemology of laws, how we find out about laws. But we somehow have good guesses as to what the laws are. We can, could talk about

that. Once we do that, well, we, of course, most people don't know in much physics, so they don't go back to the fundamental physical laws to figure out counterfactuals. They just come up with generalizations about their world, which ultimately

are made true or false by the fundamental laws of physics and the conditions in our world. And they use those generalizations to evaluate counterfactual and figure out what would happen were they to throw a brick at a certain kind of window or or jump off a building or whatever. Good.

So it does hang together, I think, in a decent way. But you brought up the issue of explanation, which is one that I've not thought about at all, really, other than in trivial ways. So what should – I want to move on to other things, but what should people have in mind about what needs to be explained about explanation? Okay.

So let me play the devil's advocate. Right. Although I don't know why anybody would want to play the devil. than what he's supposedly responsible for. So here's the objection. The objection was made very, very clearly by Demodle I mentioned before. Another philosopher named Mark Lang makes his objection very clearly and so on, but many other people have too. And that is to say this, look, if, as you Yumi and say, laws are just descriptions and the descriptions of the ordinary facts,

and the ordinary facts make the laws obtain, how can the laws explain anything? Aren't you just running in a circle? Yeah. That's the objection. It definitely has a pull. But think about it for a minute, or I thought about it for more than a minute. When you think about it, what's going on there is that I think what they're doing is making use of their anti-Umian, what you call God view about laws, as producing events. If you think about laws like that, then of course you're screwed because you can't have both the fundamental facts that are producing the laws.

and the laws producing the ordinary facts. But the Umean doesn't think about the laws as producing facts. In fact, the main Umiian reason for being anti-Umian is that when asked, how do the laws produce facts? The anti-Umian says, beats me. I'd no idea. What are these things outside of the world doing to make for fact? They're just employing our ordinary notions of causation and using as though they understand

and applying there. We don't understand it. I think. Okay. So I think the right way to think about laws is not as explaining by producing. They explain mostly by unifying. What the laws do is they unify. Okay. What the best systemization does is it finds the patterns in the world which fit together in various ways in terms of concepts which human beings have used to develop because we're human beings. There's certain concepts that we have. because of our biological nature, psychological and cultural nature, so that certain similarities are very important to us. Obviously, it's important. I don't know who the people of different genders are or whatever you might think, where the food is and so on.

So we make certain similarities. But of course, there's nothing from physics point of view that makes anything more similar than another, as far as physics is concerned, just the initial conditions. Okay. So good. So I don't want to dive in too far into that, but that does make sense to me. But you brought up once again this issue of causation. And now I think that we have enough on the ground to get into where Lewis might not have gotten it right about causation because, you know, one of the fundamental features of causation that we all know is that causes precede effects. They come before the effects in time. Or depending on what kind of causes you're talking about, maybe they're simultaneous or. something like that. But we don't say that the brick was thrown because the window was going to break, right? That's not a formation we would ever do. So how can statistical mechanics? I know that

I know the answer to this question, but you can say it in your own words. How can statistical mechanics help us understand why it causes precede effects? I will answer that. But first, I should say, that Lewis thought that he had an answer to that question. He thought his answer to that question was, it's because of the nature of our world that it's that much smaller violations of law would be required to get a divergence from the actual situation prior to an event than to afterwards to get a convergence to the actual world. And he had an inkling that the reason for that had something to do with statistical mechanics. He just didn't know how.

He does this in a paper of his called Times Arrow and Counterfactual Dependence. For my money, it's one of the great papers of 20th century philosophy, even though it's full of, for my money, errors. I only know that there are errors because I've thought a lot about that paper. I think they're errors. Okay. All right. What statistical mechanics does, and like you said, I mean, I learned this in part from talking to David Albert. I have to admit that this is before I read your book, but I also learned it from teaching your book to students and so on.

that the fundamental dynamical laws of the world are temporally symmetrical in an important way. You've got to be careful if you want to please physicists about this, but the important way is that any macroscopic sequence of events that are compatible to laws that can happen in one temporal direction can happen in the opposite temporal direction. That's true for all the physics we've ever had anything to do with. Why do we never see ice cubes form? spontaneously out of water, but we only see, I keeps melting. Why do we only see people growing old, smoke dispersing, and so on and so on and so on? And the answer to that, as I, of course, you know,

and which is I think most beautifully, if you don't mind, my complimenting you or your turn. Do not mind. I mean about this is expressed in the best way I know in your book. That's why I teach from it. it comes from statistical mechanics. It's because Boltzman originally discovered that since material objects, macroscopic objects are made up out of many, many, many, many, many parts with many, many, many degrees of freedom, very high number, 10 to the 20-something degrees of freedom, more than that, really.

And they're moving around in some way that he wants to describe as random. We can come back to what that means. just a second. Oh, we have time for that, because that's for somebody the coup de graf of this whole thing for me, is it turns out

that if you're in a situation in which the ice cube is sitting there in warm water, it's very, very likely that those random motions will lead to that ice cube being more melted in the future. Okay?

And that's why the ice cube melts. unfortunately for Boltzmann, as pointed out by, I think, his teacher, Lutschmidt, it turns out that if you use the same reasoning, it's also more likely that the ice cube was smaller in the past. And so that it did form spontaneously out of the water. How are you going to deal with that? Here's where the idea came in, which I think has only become sort of accepted. I'm not sure you can tell me whether this is a consensus. I was just being asked by a student recently whether it's a consensus now. But I know the idea was sort of suggested by Boltzman in a funny way.

Later, Eddington suggests it. Feynman presents it. David Albert, given a name. He calls it the past hypothesis. You not only give a name to it, you try to find an explanation for why it holds. And it's the claim that you look, you can answer this question once and for all about the whole universe

by saying that the whole universe very, at the time we think of around the Big Bang, at that time the so-called entropy of the universe was very, very, very tiny. That means that in a certain sense, the degrees of, there are many fewer, in a certain sense of fewer, ways in which the constituents of the universe were arranged moving around than they could have been, which makes it, given the random motion, much likely that they will evolve in such a way so that they get into ways of higher entropy. And this explains why the so-called second law of thermodynamics, that the entropy of any system in the universe isolated system and the entropy of the universe increases over time. It explains it, but only by modifying it slightly by adding very, very likely will increase over time. Yeah, I'm not sure whether that's a consensus or not. I think that 10 or 12 years ago, when I started talking about this, or 20 years ago, even, I'm getting older, it was certainly not a consensus, but not because people didn't believe it. It's because they hadn't thought about it in the physics community, right? They just hadn't thought very clearly about these questions. And so I did get a lot of pushback saying, no, you just don't understand statistical mechanics at all. I think today I would get less pushback, but people would

aren't going to say they change their minds, they're just going to sort of have absorbed that little bit of wisdom as obviously true. So I'm puzzled. Feynman definitely has this view. Yeah. Feynman was an important guy. Why didn't it spread from Feynman to others? He must have told others about it. No, I mean, it's right there in the Feynman lectures.

It's there in the character of physical law. And no one paid attention. Yeah, it's really bizarre. It's really bizarre. Feynman is very explicit about hating philosophers. Yep. There's a lot of self-hate in that. I always say, like, you can't not do philosophy of physics.

You can only do it badly. But he is actually pretty good at it, except for his explicit statements about it. So, yeah, so Feynman, like you said, Eddington, even Boltzman suggested it as one of the possible things, that we begin in a state of low entropy. We call that the past hypothesis. Penrose has a lot of other people. Penrose, right, exactly. And then just one little clarifying technical point here. when we have this discussion at all, are we thinking in the back of our minds classically,

like deterministically, like there's some atoms bumping into each other, or do things importantly change if we take quantum mechanics seriously? So I would say the best answer I could give is, well, I understand quantum mechanics pretty well. I've studied it, and I've written about the philosophy of quantum mechanics. I don't fully understand enough about what quantum field theory. two will look like once it gets combined with gravity. And people have made various proposals about that. So I can't say that things won't change.

But so far as I can understand things now, I'd be interested, very interested in what you and others think about this, that it will make no difference. Well, I think it's interesting because I also suspect it will make no difference, except that when you're thinking classically, you have to do more work, I think, because you have to imagine that there's a deterministic map from the early conditions to today, and there's a lot of information contained in what's going on today.

I don't think it has to be deterministic. The fundamental laws could be indeterministic, even classically. I mean, they're not, but they could have been. And you'd get the same kind of arguments going. No, that's true. I guess all I'm saying is that quantum mechanics, because we'll get into this later a little bit also, but there is some randomness, at least seen by observers. So you can imagine starting with a very simple, essentially zero information condition and creating all of the wonderful specificity of our current world just through randomness happening along the way.

So it's your podcast, not mine, but you just brought up the issue that for me is the most important thing for me. And that is what is randomness? Okay. What is probability? Okay. Could I say something about that? Please. No, it's really your podcast.

I'm just the host. So this is what drove me into philosophy. I was taking some undergraduate classes, three, actually, one in physics, one in statistics, one in history, and the notions of probability and causation came up in these classes all the time. And I was asking the teachers, what are you talking about? What is probability anyway? And they said, don't ask me, go to the philosophy. I went to the philosophies and then said, you go figure that out, go to graduate school.

So that's what I'm interested in. What is probability? And this view about laws that I told you that David Lewis has, the reason it really stuck with me is that I saw it had the key to the answer of what probability is. If I can describe that, you'll see how it brings together a lot of what we talked about before. Remember how I said before that count of factuals in Lewis's view are just a way of talking about the actual facts. Laws are too. They're just ways of talking about the actual facts. So is probability. Probability isn't like a fundamental feature in the world as an extra thing. What it is is a way that we've introduced in physics and science to provide good descriptions and simple way of a lot of information. So if you have, let's say,

coin tosses, many tosses, heads, tails, tails, heads, heads, tails, heads, heads, heads, heads, heads, heads, heads, tails, tails, heads. And someone says, tell me about that. You might say, well, first it was a head, then a tail, then a tail, then a head, then a head, then they say, I'm getting bored. Tell me about it in a simple way. And you say, well, there are a lot of heads and a lot of tails. You say, no, I want more information.

You might, for certain sequences, be able to say, well, there was heads and tails as the outcomes of processes, each with, probability a half and the tosses were independent. The processes were independent of each other. And that will give you a certain, a lot of information about that sequence in a very simple, succinct way. I think this is what probability is. So Lewis's view about laws and as view about chance come together in a beautiful, beautiful way for me. And in fact, for him, laws and probably has always come together with laws. And if I can say another thing that I find astonishing

in the history of the subject is that the concept of law was invented in a certain sense by Descartes and his friends in the middle of the 17th century in Paris. I'm going to be in Paris this summer. I'm going to want to go to the spot if I can. And the notion of probability was also invented in Paris

about the same time. Ian Hacking Sites of the year is, I think, 1954 or something like that by Pascal, who was solving a problem that some friend of his gave him about how to settle a gambling game that was interrupted by the king's soldiers. And he invented how to think about probability. And they were invented about the same time. And they were thought of as applying to very, very different things. Laws to the motions of the planets and pendulum and cannonballs, probability to dice. games and then eventually to how to set insurance policies for ships going to the new world and

things like that and planting agriculture planting and they were thought to apply to different domains although it was thought that god had something to do with both of them god was what's pushing the matter around and then when people ask why is it that frequency stabilize you actually find a few people. It's not so widely known, but actually saying it's God that makes for the stable, they didn't know about independence and mathematics. They didn't know the central limit here. Okay. All right. And it wasn't until Boltzman and Maxwell and later Einstein and the beginnings of quantum mechanics brought probability and laws together and unified them. And I think this is like one of the most smacks you in the face interesting things about the history of science. Have you thought about that

before? It's a question right for you because I did a little bit of reading when I was writing from Eternity to Here, you know, of those papers around Boltzmann and Zermelo and Lozschmidt and so forth. But I'm not up enough on the history to really understand the context of what they had previously thought before they were faced with statistical mechanics, right? So I don't have an informed opinion about where they were going. I know where they got to. I'm not enough on the history other. I have an opinion. I wouldn't say it's informed. I gave you my opinion. I would really like to talk to somebody who's a really good historian. I have talked to some historians of science. And basically I get as well, this complicated, that complicated thing. There's much more written

about the history of probability than about the history of laws. There are good things written about that. But I think these have not been written about together. And like I said, if there's someone listening to your podcasts who takes me up on this, I think it's a great thing. And they could write in the introduction to their book that they heard about this, doing this from me. Love that. Okay. I'll be long, yeah, I'm going dead by then, but I'll still be happy. Not if they hurry. They can move quickly. Look, if, so what do you want to say is that probability is kind of, sort of an idea that we use to describe the actual word. It's a sensible idea even if things only happen once, right?

Right. Okay, good. That's right. And okay, good. Because if it's rolling unifying, what goes on in the actual world, right? So this is going to, once again, a slight diversion, but that's okay, because it's really interesting. Can we think of the credences that we have in different scientific theories? Like, let's say, you know, the axions, the dark matter, or are we?

weekly interacting massive particles to dark matter. Is that a matter of a probability distribution over possible worlds, one in which the axioms are dark matter and one of which it's weakly interacting massive particles? Is that the same thing? So in the probability literature, it's usually a separation between two ways of thinking about probability. One is probability just in terms of degrees of credence or degrees of rational credence. Yeah. Another is probability is some feature of the world. Earlier, we were talking about it as a feature of the world. These are not unrelated to each other,

because when the world has a certain feature, that the probability of the coin landing heads is a half, then this needs a little qualification, but basically that tells you what rational credence, what rational degree of beliefs you should have in the next flip of the coin. Now, I don't think there are objective probabilities like that for things like scientific theories like theories about dark matter. So I do think this notion of credence about theories is applying the notion of credence,

rational credence, beyond where objective physical probability applies. I think that's fine. I do think it's really a question about whether you can talk about rationality in that domain and where it comes from. And of course, there's this whole literature within philosophy of science. There are good ideas about how you should change your credences. That's called Bayesian updating. But where the initial credences come from that you change, that's a big open question.

And there are a lot of philosophers thought about that, too. Something that logic tells you where they get them from. So John Maynard Keens, the famous economist, had a view about that. Rudolph Carnap, the famous philosopher, had a view about that. Those are not my views, but those are views. And I shouldn't say anything more about it because I'm not sure what to say. Well, I'm not sure what to say either, but I do have a soft spot in my heart for a potential unified account, which says there's no objective probabilities about anything. All there are are credences for different things happening.

And it's just that in some cases, we're sort of more guided toward what the credences are, like flipping coins, than dark matter candidates. So I know you have that view from your views about Everettian quantum mechanics, but this is a totally different discussion. Maybe we'll get together. I know I'll visit Johns Hopkins next year. So that would be lovely. Janana's going to organize something. We'll do that. Well, especially because we're getting to the punchline here. We have the laws of physics on the table. We have the past hypothesis on the table. These are two of the main ingredients in what you and David Albert have called the mentaculous. So explain that a little bit to our readers, our listeners. Okay, since it's very likely that most of the people who listen to this, they've never even run across the term antagonist. I should explain where it comes from. And everybody understands why I'm using this term here. So there's a film by the Cone Brothers. Probably many more people know about that than about, David,

my use of this term, called a serious man. And in this movie, it's about a physicist who goes to a midlife crisis. I assume you've seen it, Sean, and maybe you've experienced the crisis. I don't know about that. I'm dealing with it. Maybe you're not at that point yet. I'm not sure.

But in any rate, there's a minor character in the movie who's scribbling in a book. It's clear that this minor character is out of his mind. And somebody asks this minor character, what are you doing? And he's saying he's composing the mentaculous. And then he's asked, what is that? And he says, it's a probability. map of the universe. Well, that's what statistical mechanics when the past hypothesis added to it

gives you. I want to make that clear to everybody in just a second. When you see that, you're given probabilities of every single sequence of nomologically possible events, events that fit in with the laws that can happen in the universe. Okay. It's a probability map of the universe. Now, of course, David and I were perfectly aware that we were engaged in a little bit of self-mockery when we use that term. The reason I want to make that clear is we wrote another paper about Everettian quantum mechanics in order to make fun of it, which we call the many minds the kind of quantum mechanics, and there are people who wrote papers taking it seriously. Because we didn't say at the beginning, this is in order to show what's wrong with many, whatever

You're quantum mechanics. And a lot of those problems that we thought were wrong have been dealt with. Subsequently, when we wrote the paper, it was written a long time ago. But I want to make that clear to whoever is listening to this. But you do take them spectacular seriously. I mean, your idea is that this is how we should think, given our higher level knowledge of the world, about the probability distribution of everything that ever happens. It's a consequence of... That's right.

Let's just make it clear to the listeners, why that's so. what statistical mechanics does is that signs of probability density, or let me call it, distribution, over all of the possible initial conditions of the universe. Now, of course, there's only one actual initial condition, but of all the possible ones, that probability distribution is the simplest one possible. It's what's called a La Beg, as a kind of uniform distribution. That will entail, together with the dynamics, the probabilities of everything. all a macroscopic description of the world is is a collection of microscopic descriptions. If I say my cat is in the room, I'm saying what the microscopic state of the world is. It's one of those states which gets my cat into the room.

And that will have a certain probability. So this probability distribution yields a conditional probability distribution, given such and such conditions of the world, macroscopic conditions of the world, these are the probabilities of other microscopic descriptions of the world. This is deep, deep and interesting because these probabilities should tell you what your degrees of belief,

what your conditional degrees of belief should be about everything. If you only match them to what physics gives you. Of course, we don't know. know what physics gives us. But what we do do is we learn about the world in such a way that we do come up with, I think, matches that to a large extent. And we can explain why it does in terms of explaining how we come up with them and what they are. So I have a little tiny footnote. I mean, maybe you said this, but I missed it, because you seem to say that there was a uniform distribution over all the possible initial conditions. But you mean subject to this past

hypothesis that the universe started with really, really low interest. Sorry, very important footnote. Without that, the whole thing falls apart. Okay, good. Okay. Another thing about the past hypothesis, although David called it the past hypothesis, that makes it sound as though he's assuming a distinction between past and future. But a point that's often neglected, even by David, I think even by you, but to correct me if I'm not right, is that no, no, no. What this does is explain what the difference between past and future is. It's not assuming a distinction between past and future. It's rather by using this to explain why it is that we can only influence events in an interesting way in one temporal direction but not in the other,

why we can have records in one temporal direction and not the other, why smoke disperses in one temporal direction, not the other, and so in the other, by explaining why all of times arrows are oriented in the same direction with their tails. at the past hypothesis, explaining why that is the past. Yeah, no, I can promise you. I never get that wrong. I completely agree with what you just said. I might sometimes misspeak, but in my mind, I know exactly what you mean. I'm not sure you really emphasize.

I know when I taught from that. Yeah. Maybe the students didn't get it right away. I really had to emphasize. I'm sloppy about that. Because students objected. Good. Okay.

They said, well, aren't you assuming the distinction in past and future? No, they're right. Is the past hypothesis a law in the, the same way that Einstein's equations is a law? Other really good question. If you take David Lewis's view about laws, the answer would be yes, because what it does is by adding it to the system of the world, it's adding a great deal of information about the world in a very, very simple way.

I think that argues in favor of Lewis's view about laws. But isn't it so this is an area where I don't have a firm opinion about. this. If I take the definition of Humianism at face value, then I get why we would be tempted to put the past hypothesis, which is a statement about initial conditions, and the dynamical laws like Einstein's equation on an equal footing. They're both very, very simplified, helpful, compact descriptions of what happens in the world. But they're different also. I mean, the past hypothesis is it only one time, and the dynamical laws keep being true from moment to moment, So she'll really throw them in the same bucket as laws of nature?

You just mentioned Einstein. One of the lessons of Einstein is what's so important about time anyway. Now, I know that's an exaggeration. But what Einstein did is he stuck time and space together. And I know in theories of quantum gravity, there are views in which what's fundamental doesn't even involve space and time. They are emergent features. I think they are emergent depending on whatever it is that occupies, whatever is, whatever occupies the fundamental arena.

Its distribution is what determines the dimensionality and the fact that the dimensions can be divided up in a particular way and that they divide them in such a way that there are certain best ways of characterizing all of that, of systematizing all of that. So another related question then is, why is the past hypothesis true? Why is there low entropy at the Big Bang?

And I know that's not what we're here, to explain, but maybe you can give me your feelings about, is that even a good question? Is there a future answer? There is a philosophical important point to make about that. But this is not a question to be answered by philosophers in the way you asked it. Like I said at the beginning, it's the job of physicist to tell us about that. However, somebody might make the mistake. I know of some philosophers who have to think that because I've given a good reason to think that Yumians should consider the past hypothesis to be able. law that I've argued by that that it's a fundamental law. I haven't. It might not be a fundamental law. That might be an explanation. It might be that a better theory of the universe explains why it is that the past hypothesis is true. And I know that from, you know, teaching, reading your book that you've come up with an account of that. I'm not sure anybody believes that account. Does anybody believe it? I think it's simultaneously the best account out there and nobody believes it.

Which is fair. I mean, but I do think it's a good idea to try to come up with an account. Yeah. If you can't, maybe the last thing to say, yeah, we'll accept it as a fundamental law. I'm not sure. Okay. That's for physicists to do. Yeah, no, that's for physicists to do.

Fair enough division of labor there. But let's get to the payoff here. So what we're going to do, I guess, is use the mentaculous picture to do better than Lewis did at coming up with an explanation for things like, causality and stuff like that, the way that we use possible world semantics. Yes. Now, I wish all of this, in my view, was a completed project. It's not quite, but I can give you an inkling about how it's supposed to go. So let's go back to counterfactuals.

Instead of thinking about similarity that way, let's just think about the metaculous probability distribution. Let's think about the counterfactual if the match has struck what would have happened. What we're doing there is we're asking to look at a time when it's somewhat likely that the match is struck and then saying if it's at that time, given the macroscopic state of the universe at that time, it's likely, there's a certain reasonable likely that the match is struck. What are the probabilities of various things happening? Well, given the macroscopic state of the universe at that time, that will entail things, in accordance with the metaculous, about what it will be going on at later times, including whether there's still oxygen around, whether the match is still dry or not, and so on. And so it will entail how likely it is that the match will light or not.

That's where the truth conditions of kind of factuals are going to come from. I think that this is really, for me, personally, now it's just something that. I've just finished. So it hasn't yet been subject to the incredible, you know, I don't know, scrutiny and destructiveness that happens when you publish a philosophy paper. I've been there. But I think this is the best idea that I've ever come up with personally. I think that this idea works to answer all of the questions that bothered me when I first think about counterfactuals is an undergraduate. Good. So again, let me do what I like to do, which is to rephrase it in my own words to see if I've gotten it. So Lewis had to, you know, weasel a little bit about miracles and

violating laws of physics to talk about worlds in which things were slightly different. What you are able to do with this statistical mechanical picture is to compare the macroscopic situation to the microscopic situation and say, we don't need to change the laws of physics. We just need to find a specific microscopic situation in our macroscopic situation where the specifics would have gone differently, right? Where you don't know. exactly what every atom is doing, but all you need to do is change that micro-state to get, in this particular case, a different macro outcome. Yes, except it's not a unique micro-state. Sure. Because there'll be a lot of microsays. We're doing about probability distributions.

Of course, as you know, if we just talk about microscopic states compatible with a macroscopic state, there'll be some microscopic states in which all sorts of weird things will happen afterwards. You know, I might turn into something that looks like a frog. Yeah. in the next moment if the molecules are moving around in a certain conspiratorial way. Okay. But this is just immensely unlikely, I'm happy to say. But it's very likely that what will happen in the situation we described is that the match with light in that case. Now, what I have done is gotten rid of miracles by just limiting counterfactuals to talking about macroscopic states.

and that gives rise to a very deep, important question, which isn't dealt with enough by people who think about statistical mechanics of what exactly is a macroscopic state? Yeah, what is it? Do you know the answer? It's a really good question. I've been thinking about that. I don't know.

I know what to say if I stick to thermodynamics, and I know some more to say about it. But here's what I think about it as a philosopher at the beginning. I don't know the real good answer to that question yet. I'd like to. But what I think of is that, well, if we think about all the possible microscopic states compatible with the world that's around me, all the microscopic states can be divided up into groups in all sorts of different ways, weirdo ways of oil,

so that one microscopic state which has me sitting in the room is in a collection with a microscopic state that has Joe Biden drinking wine, who knows what, okay, all sorts of different things. We don't do them in a lot of ways, only certain ways. We do it in just certain ways. Why some ways rather than other? As I was suggesting earlier, I think the answer to that ultimately is going to have to refer back to us and our biology and psychology. But how to fill that in?

I don't know. The other person I've talked to a lot, and maybe, I don't know if you've done a podcast with them or not, but I know he's written. about this is Carlo, Carlo Revelli. He was my second ever podcast guest. I don't know if this came up with him or not. We talked about quantum gravity. He talked about somewhat.

And I know he's thinking about it too. And he's smarter than me. So he may come up with something better than I could come up with. But that's, you know, I think to me, yeah, this is a crucially important question also. And it's the question of emergence, right? You know, how do we come up with these higher scale descriptions that work so well, despite the fact that they're not microscopically complete. I am not actually going to lean on biology or psychology.

I think there are physics-based explanations for this, but I don't have the complete one to present to the world. I think that's a good problem. I thought that from my satisfaction, we might end on my general philosophical view, which will be about that. But if you wanted to talk about something else first, like free will or anything, we could do that,

and then I'll come back to that. Yeah, no, good. Let's come back to that. We're homing in on the end here, because I think that we've covered a lot of ground and some really good stuff. But I just first wanted to give you a chance. And maybe this is in the same kind of domain, but give a chance to talk about emergence and the special sciences, right? This is a feature of the world that we know and that the metaculous should have something to say about. I mean, why are we able to talk about economic, or psychology, rather than just saying everything is physics, I guess. And I guess it was Jerry Fodor who first put things in those words. Yeah, so, you know, Fodor was my colleague for a while. And I can't say the most fun of my philosophical life was going from traveling from New York. We both live to Rutgers with Fodor and listening to him on the bus. The reason I say listening to him is that I could, despite what you may think of me,

from our conversation. I can hardly ever get a word in edgewise. I believe it, yeah. So Fodor asked this question, why is there anything but physics? His answer was, he doesn't think he's going to find out about answer to that until the very day before he finds out why there's anything at all,

the question that came up earlier in our discussion. Sadly, Jerry is no longer with us, so I don't think he ever found out. Or maybe wherever he is now. He knows the answer to that. I don't know. But in any case, I do think you're right that if everything I said before, the metaculous should say a lot about the special sciences. In fact, if, in fact, all of these probabilities of everything, including all of the truths of the special sciences, are already specified by the metaculus, then whatever probabilistic laws there are in the special sciences must already be in some sense, I say derivable, but I don't really mean that anybody is ever going to be.

position to derive them because they're way, way, way, way too complicated. But in some sense, they sit upon on top of the metaculus. How could they do that? Well, it must be, because they're not adding any probabilities, they must be telling us what we can conditionalizing on. So what the special sciences are doing is they're finding features of the world such that when we conditionalize on them, we get regularities

of other features of the world that interest us. I think, again, developing that idea has not been done by anybody and there are other philosophers I know of, including people who are students in mind who disagree with me about this a lot. But I do think it's a very interesting thing to follow up on at any rate.

I do think that in special sciences, it's more than just finding those probabilities because special scientists like to also qualify by introducing terms like Keteris, paribus, whatever that might mean into their special science regularities. I'm not sure looked at like this. Special science is what I'm saying,

don't add any laws to the laws of physics in a certain sense. And they don't add any probabilities to the probabilities of physics from statistical mechanics. What they do is they find them and characterize them in new and interesting ways. Good, yeah. I'm sure that I could, I bet there's occasional sociologists or economic

listening who would not going to like that way of characterizing what they do for a living. But I think it's exactly right. They should because it's pretty hard to do and they should get paid a lot if they can do it. I don't think there's anything bad, taking anything bad away from them doing it. But if they want to be physicists, they should go be physicists, be sociologists. And I mean, you've mentioned it a couple times, so let's get it on the table here, the free will question. You know, we've had on the podcast both people like Robert Sapolsky, neuroscientist, who says there's no such thing as free will, you're just doing what your neurons tell you and they're programmed by genetics. And also Jan Ismail, who says, oh, you're just misunderstanding free will. It's completely compatible with the laws of physics. I can guess what the answer is, but where do you and the metaculous come down on questions like this? Well, as you know, Jan Ann and I are big pals and I'm totally with her about this. I don't think my neurons have ever told me anything.

I've never even heard my neurons speak about anything. My neurons, though, what's going on when I'm telling anybody else anything? What I do think is that there are a certain conception of free will that we somehow have, and which particularly philosophers have sort of put out there, which says that whatever free will is, it's got to separate us from everything else that goes on in the universe, from physics. I don't think there is anything like that. So I think, like I said much earlier in the podcast, one of the things that we can do as philosophers

is take our ordinary notions and sort of reconcile how we ordinarily think about things with what science tells us about things in part by saying what the ordinary ways are not quite right. But we can come into the same ballpark. We can see why they apply to the extent they do by finding something in nearby. Now, I don't think Jan and I are completely on the same page about how they do that. I think statistical mechanics and the metaculans is the key to showing how they do that. I think it does it because if you go back, it's very easy to say, given back what we're saying about counterfactuals, what we're interested in free will is how we can influence. Why is it that

we can influence the future, but we can't don't have any influence over the fast? What is it to influence? It's for it to be the case that small differences in what's going on in our brains can make a difference in one temporal direction, but not the other. Now, you might ask the question, and the metaculous answers that, because the statistical mechanics it says has as a consequence that small differences in our brain

will yield the same probabilities for anything in the past that we care about it while making for big probabilities for certain things in the future, like, you know, what you eat for lunch or not. Okay, but it won't make any difference to the probabilities of what you ate for lunch. yesterday, given the rest of the macroscopic past now. The reason, of course, is because the rest

of the macroscopic past contains records of what you ate for lunch yesterday. Okay. There's still a question of what is going on. How can we influence what differences in what's going on in our neurons? I think the answer is that that's just what influence is. There's nothing more to it. That's what we are. We are such that our neurons are such that they make a difference to what goes on in one temporal direction, but not the other. If you think of the laws of physics as something like an iron grip that grips the states that forces your neurons to fire one way rather than the other, well, good luck to you. But that's not how I think about laws of physics. Laws of physics just summarize the whole history of the world.

And I think when you think about it like that, you can think about it being the case that all it is to have a free choice is for, given the macroscopic state of the world right now, for there to be various microscopic states that are compatible with that microscopic state. And that's what it is to be able to make different choices. And it can make a difference in one temporal direction, but not the other. Yeah. And so the way that I sometimes put it, because the anti-free will people will try to phrase it as the question that they, They would admit you had free will if you could have acted differently. And then they'll say the laws of physics don't allow for you to act differently. But I think the response to that, correct me if I'm wrong, is, well, conditional on what?

I mean, conditional on my macroscopic information of the universe, I could have acted differently because there are a lot of microstates that would have led to different behavior. I didn't go further. Conditional on the whole macroscopic state of the universe right now, I could have acted differently. not conditional on the exact microscopic state of the world. But needless to say, nobody knows what that is and couldn't know what that is. Okay, good, good. We're on the same page. Okay.

Two final questions. One is I'm giving you the opportunity to say what you just wanted to say about your philosophy of physics and how it, I guess, deals with macro states. Okay. So my other hero in the history of philosophy, contemporary philosophy, besides for David Lewis, is Hillary Putnam. Now, Putnam definitely did go off the reservation some of the time, but he was a very deep and interesting philosopher. And I did study with him a bit, not in graduate school, but later on with him. I think he was a really, really interesting thinker. And he has a phrase that I really, really like.

I'm not sure that he meant the same thing I meant by it, but here's his phrase. He says, somebody wanted to know, is he a realist about physics or not? And the opposition there was between philosophers who say, there's the physics and it accounts for everything. And there's the mind. And somehow the world is the product of the mind. That's the anti. Putnam said, here's his phrase, the mind in the world make up the mind in the world. If you understand that properly, I really like that.

Here's the why I like it. I think that my whole conception of laws is a way that. that it's in a certain way human-centric in the sense that it depends on the fact that human beings want to get a simple and highly informative account of the world. That's other things. Maybe aliens aren't like that. David Lewis was worried about this feature of his worldview. He gave it a name.

What did he call it? He gave it a nice name. He called it. Oh, shit. Sorry. That's okay. Who did he get it? People can look that up.

I can't remember it. But anyway, he had a nice name for it. He thought it was a big, big problem for his view. You would love the name because it's an Australian bit of slang. Anyway, his worry was that his worry was that his view would be too subjectivist, too much mind-dependent. But I don't think it is. I think it's because I think we want the notion of. science of laws to be mind dependent in a sense it's not to say that any facts that the laws say

our laws is mind dependent but it's that they are laws is mind dependent i do think we want that because the concept of laws is mind dependent it's something that we've concocted in order to account for the facts so it's not subjectivist or realist in this bad way okay good so I think that's my view there and it's my view about what's macroscopic, what counts as a macroscopic state too. It's dependent on how we encounter the world and our psychology ultimately and fitting us all together in one giant package is something that I don't know how to do. The view of laws I have, though unsurprisingly, it's called the package deal, account of laws and chances and fundamental properties. Because it turns out,

the way I differ with Lewis is I think also what counts as fundamental is also going to be part of this whole package. Okay. And we don't get into that, but we've talked long enough. Yeah. Well, okay, except I do have one final question as a bonus dessert treat for the readers. Explain, I keep saying readers. They're listeners, obviously. Why don't you like the many worlds interpretation of quantum mechanics?

You're so correct about all statistical mechanics. And, you know, what is this quantum mechanics issue that we can't quite agree on? So I think there are two big issues by it. I think a good place to read about the problems. I don't know whether you've written in response to it is Tim Maudlin's book on philosophy, physics, the quantum mechanics discussion. But one of the problems has been much, much discussed, is going to be a whole conference about it. My guess is you're one of the people going to the conference in Israel in October, are you? I'm busy with a new job, so I'm not going.

I was invited. I can't make it. Sorry. I'm probably not going to go either, though I was going to go in a town. Act in many worlds account there. But I'm just too busy with dealing with doctors and things like that. But anyway, but it would be a lot about that because one of your fellow travelers, Lev Weidman, has organized the conference and he's a big fan.

And I think that at the end of the day, you can't make use of the way I understand probability. It's driven you, rightly so, to take probabilities out of the world and putting them entirely as indexicals, applying to indexicles. They say probabilities. I think that gives rise to a lot of questions, which really deserves a whole other podcast, how to think about indexicality and so on. But I do think you were right to take the problem very seriously

and be driven in that direction. That's one of the problems. The other problem is sometimes I think it's not such a big problem, sometimes I think it is. And that is I do want to ultimately understand how it is that whatever physics says is fundamental, gives rise to the macroscopic, to the special sciences, to my cat, and so on. And when I'm told that the stuff of the world is what, what do you actually think the stuff

of the universe is? It must be something like, what, a field? It's a vector in Hilbert space. It can't just be a vector. Vectors in somethings are mathematics. It's not a vector. No, the world is sui generis. It's the world, but it is mathematically represented as a vector in Hilbert space.

It's some stuff. Okay. Can we say something about that stuff? Well, if it's represented by a vector, and the vector is an element in the Hilbert space, but what is the vector? The vector itself, does it occupy an arena or not? No.

I mean, all the different vectors occupy Hilbert space that I get. Right. But the vector itself, it has parts. Does that vector? Well, you can divide it up in different ways. You can factorize Hilbert space, and then there's a operational question about how best to do that, which helps you understand.

where there are emergent descriptions of things. Okay, but I want to understand that emergence a bit better. I understand how it is that Adams banging around can make for smoke dispersing. I want to understand better how the fact that this vector, which represents this sui generis thing of the world, I don't dislike that. I like that point of talking about it. whatever it is, it has an arena. Part of the world is an arena, some sort of space time.

I'm not sure what its dimensionality is or anything like that, but I assume it has one. Okay. And that somehow, whatever's going on in it is going to give rise to the smoke dispersing. I want to understand that as well as I understand how it is that molecules banging around gives rise to smoke dispersing. I'm not giving you an argument saying it can't be done. I'm just saying I haven't really fully understood quite how it's done. And maybe it's just my lack of imagination. No, I think this is a wonderful research program for the future that I, myself and I'm engaged in. I've written one paper about it. You'll come to Baltimore. I'll come to New York. We'll talk about it.

Okay. Okay. That sounds good to me. And this was great. I think that this is a really great way of thinking about the probability structure of the world and the importance of it. And one of the great things about it is how many questions there are still to be answered. There's still a lot of work to be done by young natural philosophers out there. Yeah. One may add thing, which didn't get emphasized earlier, is that the sort of Louisiana-Umian view about probabilities and laws and the meticulous way of thinking about fundamental probabilities fit together hand and glove because if you think about the fundamental laws is deterministic, then it becomes very puzzling as to what in the world probabilities are.

But Lewis's account makes perfectly good sense of those probabilities. There are other ways to do it too, but that's one way to do it. Who says there is no progress in philosophy? Yeah, I agree. There is something like progress. Certainly there's personal progress. I can see in my own life there's definitely a lot of personal progress. I don't know that someone looking back 200 years from now will say that.

I think they will. I predict that they will. All right, Barry Lower. Thanks very much for being on the Mindscape podcast. Thank you, Sean. Just before you sign off, I can't tell you just how much you help me figure out what I think by asking questions. Good. That's, you know, I, that's why I'm doing this.

It's a great little.