¶1Emily Adlam, welcome to the podcast. >> Great to be here. >> So, why do we need to save science from quantum mechanics? I was under the impression that quantum mechanics is part of science. So, what what gives?
¶2>> Uh yeah, so the worry is that um in order to uh sort of make sense of science and to have confidence that science and scientific theories actually match up with the world in a in a sort of realistic way, uh we need to have a complete account of the way in which uh the scientific theories map onto the world and in particular how the uh the evidence for the theory gets connected up with the content of the theory. Um and the problem is that uh because we don't have a complete account of measurement in quantum mechanics, something really important seems to be missing from the story that we have to tell about how the theory as we understand it uh maps onto the world in sort of the terms of the evidence that has led us to arrive at it. Um so, in particular, many of the well-known approaches to the measurement problem, the sort of proposed solutions, uh seem to have quite big gaps in them in terms of making sense of how the epistemology is actually supposed to work. So, the many-worlds interpretation, many of the observer-relative interpretation, cubism, like all of these certainly have a have big problems with the story that they have to tell right now. So, there's kind of a worry that at this point we don't really know how to uh interpret the theory in a way that actually uh consistently accounts for how the evidence is supposed to marry up with the theory itself.
¶3Um and you know, the reason this this takes science beyond just quantum mechanics because um uh quantum mechanics uh is is sort of fundamental and lies underneath all of our other theories. So, insofar as we don't understand what's going on in a quantum measurement, there's a worry that that sort of it potentially leads out beyond just the science of quantum mechanics and affects our theory of observation in other regimes, too. >> Okay, so there is this reference to the measure problem, the measurement problem, and then this idea that we need our theory of quantum mechanics to recover the empirical confirmation that quantum mechanics went through, the experimental results that we've relied upon in order to confirm quantum mechanics. So, can you explain a bit about this this problem of recovering the experimental confirmation for quantum mechanics? I mean, I I know from the broader context what you're talking about, but it's not immediately clear.
¶4>> Yeah, I mean, it's it's a precondition for for sort of any reasonable scientific theory or interpretation of a scientific theory that [snorts] it should be able to tell a sensible story about the nature of the evidence that we've used to arrive at it. So, you know, as we go about trying to empirically confirm a scientific theory, we're making sort of certain assumptions about evidence and our relationship with evidence, where, you know, assuming that we remember the past accurately, we're assuming that the sort of relative frequencies that we see in experiments map on roughly to the probabilities uh predicted by the theory. We're assuming that um when we communicate with other observers, we are able to communicate information about what we've observed to them reliably. So, all of these things are part of the sort of uh presuppositions that go into the empirical confirmation of any any theory. Um And so, the this sort of essential point is that the theory must then, when we arrive at the theory, the theory's got to be consistent with itself.
¶5It's got to tell a story which like maintains the idea that the evidence, as we have used it, actually does provide support for the theory. And so, what I worry about a number of important interpretations of quantum mechanics is that they seem to sort of undermine that story. They seem [snorts] to uh imply that the nature of the evidence and our relationship with the evidence is very different from what we would originally have assumed in ways that seem to seem to undermine the what they the sort of support it offers for the theory in the first place. And obviously, we can't accept that. You know, you can't can't accept an interpretation of a scientific theory which says, "Oh, well, it was a mistake to arrive at the theory in the first place." That would be very straightforwardly self-undermining.
¶6So, those problems really do have to be fixed, otherwise we just couldn't take these interpretations seriously. >> Right. So, it's a problem of self-undermining, or at least undermining the justification for the very theory that this this is supposed to be an interpretation of. That would be a bad situation. >> Yeah.
¶7Yeah. You It wouldn't make sense to believe in an interpretation of quantum mechanics that says we shouldn't have believed quantum mechanics in the first place. That's what would be very straightforwardly self-inconsistent. >> Right. So, it's kind of like the the interpretation could be true, but it undercuts any reason to think that it's true.
¶8Is that sort of the issue or the the worry? >> Yeah. If If, for example, you think that uh the many worlds interpretation undermines the evidence for quantum mechanics, that doesn't show that the many worlds interpretation is false necessarily, but it does show that it would wouldn't make sense to believe it on the basis of quantum mechanics. Uh the quantum mechanics is supposed to be the reason why we would believe that. Uh so, if it if the theory tells us that's wrong, it just doesn't make sense to believe any of it.
¶9>> Right. Okay. Okay, good. So, let's let's get concrete to to fill out the idea here. So, let's look at the many worlds interpretation.
¶10We'll start with just painting the picture. What What is the many worlds interpretation of quantum mechanics? >> Yeah. So, the many worlds interpretation uh solves the measurement problem or tries to solve the measurement problem by suggesting that anytime we make a quantum measurement, in fact, all of the outcomes occur. So, it's sort of one of the central problems of the measurement problem is that uh we don't have a very good story for how we go from a superposition of possible outcomes before the measurement to a single outcome after the measurement.
¶11The many worlds interpretation resolves that by saying, "Well, we don't do that. We just keep all of the outcomes. We branch off into a bunch of different worlds and the one copy of the observer sees each one of the outcomes in each world." >> Okay. Okay. And how I I asked this question to David Wallace, but but how literally do we understand this talk of like different worlds?
¶12Naively, I suppose, when I first heard with interpretation, I thought that like space-time, the universe literally splits apart into two parallel realities, but that they're completely spatiotemporally disconnected. And then I have gathered that no, that's wrong. They're in some sense all part of the same space-time, but then of course the question is like where are these alternative histories? Like we only experience one. So, what is this broader story about the the metaphysics of it?
¶13>> Yeah, but I mean the metaphysics is quite disputed, I would say. Um so, people do have different views on this. Uh I think probably the most most common view would say that uh they're all in some sense within the same world because they are sort of causally connected to each other. Um you know, they came from the same event and in principle they could still interact with each other. It's very unlikely.
¶14It doesn't Once things get sort of uh big enough and distinct enough, it never actually happens, but in principle these worlds could still interfere in various ways. So, that that indicates to us that they're still in some sense part of the same world. >> Right. >> the question of of space-time is interesting. I think particularly in the context of a theory of quantum gravity, you might expect that space-time itself is going to branch.
¶15So, it's not necessarily the case that they're all in the same space-time, but again I think the metaphysics there is is a bit complicated and nuanced. >> Right. Right. I mean that's fair. Okay, so I guess first though we should back up and just make sure we're up to speed on what the measurement problem is.
¶16What what is that in broad outline? >> So, the measurement problem uh is basically well, a bit People have different views of exactly what the nature the the central problem is, but it's to do with making sense of the formalism of quantum mechanics uh in a way that sensibly incorporates the process of measurement. Uh so, the the the basic worry is that quantum mechanics seems to have sort of two separate parts. It has the sort of uh ordinary part of the theory which involves systems undergoing continuous linear evolution. It's all very mathematically straightforward and it's nice and simple.
¶17Uh and then also there seems to be different process that takes place during measurement in which we go from a sort of superposition of different elements of the wave function to just a single out outcome which is witnessed by the observer. So the the question is how to incorporate those two different frameworks together. When does this process of collapse happen? Does it Does it literally happen or is it some kind of Is it Is it some kind of misreading of the formalism to think something literally happens? So all the different interpretations have different stories about how to put those things together.
¶18>> Right. Okay. So one way of putting it is we have these superpositions of distinct states, but in our experience of the world, we only experience one definite state or outcome. And the question is how do we have both of these things? That's sort of it it.
¶19>> Yeah. Um one way or another we have to go from a quantum state which looks like a sort of mixture of all of the different outcomes of the measurement to in the end each observer only sees one outcome. So you've got to have some story to account for how that happens. >> Okay. Good.
¶20So then many worlds so defined, what's the problem about recovering empirical confirmation for quantum mechanics? >> Yeah. So the central problem for many worlds interpretation I think is the probability problem. So the worry here is that main point of many worlds is that when you do a measurement all of the outcomes occur. So what probability then might you assign to those outcomes?
¶21Well, so perhaps naively you might think if you just assign probability one to all of the outcomes because the theory says every one of them definitely occurs every time. But of course quantum mechanics does not predict probability one for all of the outcomes. It assigns various sort of non-trivial probabilities between zero and one. Um and and that's important because nearly all of the evidence for the theory comes from looking at like long-term relative frequencies observed in laboratories and matching them up with the probabilities predicted with quantum mechanic by quantum mechanics and saying, "Oh, look, they match, so the theory is right." Um In the context of the many worlds story it seems a bit hard to make sense of that procedure of looking at relative frequencies and matching them up with probabilities because um if all of the outcomes occur every time then it seems like as an observer in that world you have would have no particular reason to expect that you're going to see relative frequencies matching the probabilities predicted by the theory. It seems hard to understand how those probabilities would mean anything at all.
¶22Um so and if if we can't make sense of that story then it seems like a very large chunk of the evidence for quantum mechanics doesn't really make sense anymore because this process of matching relative frequencies up with probabilities seems like it wouldn't be a sensible thing to do in that kind of setup. >> Mhm. One naive response to the problem would be to say, "Well, you know, if quantum mechanics asserts that this outcome has a 2/3 probability, probability of 2/3, and this other outcome has a probability of 1/3, then that means that in 2/3 of the branches uh the branching post-branching branches after this experimental setup, 2/3 of them have this result and only 1/3 have that result. Why is that like branch counting not right?" >> Well, for a start, that isn't the correct count of branches. Well, the first problem is that it's difficult to give any any precise count of branches because in the modern Everett interpretation the branches are kind of approximate emergent structures, so it's not clear that you can actually count them at all.
¶23>> Mhm. >> Um moreover, if you there's if you do try to count them, the sort of most obvious and naive way of doing that would just have one branch for every outcome. So, if you just count the branches it looks like all of the outcomes are always going to have the same probability, whereas of course in most circumstances quantum mechanics does not assign the same probability to all of the outcomes. So, in particular it there doesn't seem to be any straightforward sense in which it's true in the case that you described that 2/3 of the branches have one have one outcome and 1/3 have the other outcome. The the natural thing would be to look at them and say one has one of the outcomes, and the other one has the other outcome, and so you would seem to get 50/50 probabilities in that case.
¶24>> Mhm. Mhm. Okay. Yeah. I think nobody, as far as I'm aware, nobody seems to think that branch counting of any kind is is going to be the solution here.
¶25Is that basically right? There's consensus about that? >> Uh I think Simon Saunders has a more sophisticated version of branch counting. Uh I I mean, I'm not sure if that's still his current view, but he has definitely proposed that relatively recently. So, that's a That's That's a different approach where he tries to get get show a way to recover the sort of the correct number of branches.
¶26>> Right. Okay. And I am not familiar with this approach, so if you don't know enough about it, then that's where this will end. But like the decision-theoretic approach to nailing down the probabilities, what what does What does that mean? What is that getting at?
¶27>> Okay, yes. No, I have many things to say about this. Um the decision-theoretic approach uh is based on a sort of generalization of classical decision theory. So, one of the sort of keystone results of classical decision theory is a representation theorem uh which shows that uh if we assume that rational agents uh have preferences which obey certain sort of natural principles, then we can show that they will make decisions um as if they are maximizing utility with respect to some utility function and and probability measure. So, you can show basically that that agents with sort of rational preference orderings are going to act as if they are assigning probabilities to outcomes and then trying to make decisions which have the highest probability of getting a good outcome.
¶28Um So, the uh quantum mechanical version of this proves a similar representation theorem, but then goes one step further and also tries to show that rational agents in an Everettian universe will be maximizing utility with respect to the probability function that matches the predictions of quantum mechanics, the Born Born predictions. So, the thought is uh a rational agent in the many worlds many worlds picture will end up acting as if the Born rule probabilities are probabilities. So in some sense, if you particularly if you're willing to adopt some kind of like behaviorist view of credence and probability, then in some sense it just those just are the probabilities and that's all there is is to be said. >> Right. Okay.
¶29And it sounds good. What's wrong with that proposal? >> Yeah, I mean there are a couple of problems. Um Uh One is that uh obviously in order to arrive at this result, they have to make various assumptions about what a rational agent would be obliged to do in the many worlds interpretation. Um and I think many people have questioned whether the axioms that they lay down are really uh rationally obligatory.
¶30One of the worries is that uh the axioms require that the agents should have preferences only at the level of individual branches. In particular, they're not allowed to have preferences about like the structure of the universe as a whole or the number of branches of the of the the form of the branching. Um And I think that that type of axiom seems fine if you're concerned only about how an agent should make choices regarding their own personal future, in which case just the sort of individual branches matter. Um but if you're trying to do like empirical confirmation of a theory and learn what the world is like, it seems that facts about the structure of the world as a whole and about the branching overall should be very relevant to the kind of you'd be making. So it's not obvious that it's reasonable to ignore that.
¶31Um another worry is that um what this at best what this uh this result proves is that it's rational to uh when you're sort of looking looking forward to the future making decisions uh to act as if the branches have certain probabilities. But in the problem of confirmation, we're trying to do the opposite thing. We're trying to look back at the past and say um given that a sequence of outcomes has been observed in the past, I should believe various things about the structure of the world as whole. Um, and I don't really know if there's any There doesn't seem to be very any clear reason why you you uh these uh proofs based on sort of forward-facing probabilities could give you any right to believe anything at all about what's happened in the past or what what the branches were in the past. Um, and there are also worries uh uh this is quite sort of practical proof.
¶32It's about what an agent should do practically at the level of practical rationality. It's not uh unless you take a sort of quite strongly pragmatic um uh and behaviorist attitude to belief, it's not clear that these facts about what it is rational to do practically tell you anything about what you should believe. Um, uh and uh relatedly uh at best these proofs seem to show show that it's rational to act as if the branches have certain probabilities. It's not clear that that means we should think they really do have those probabilities or that we can infer anything uh about uh non-pragmatic matters um from the claim that they have these probabilities. So, yeah, there's a bunch of >> Right.
¶33>> uh sort of more detailed problems with the specific assumptions made in the proof plus also higher-level principled worries about whether this is the right kind of approach to recover robust probabilities of the kind that we could use in empirical confirmation. >> Okay, so that's decision-theoretic approaches to grounding the relevant probabilities. Is there a third type of solution that is uniquely focused on so-called self-locating evidence? Where does that whole Pandora's Box fit in here? >> Yeah, um so uh in so far as there is possible to assign any sort of credences over the branches in the many-worlds story, they are probably going to be self-locating credences.
¶34Uh They're going to be credences not about what the world is like, but about when where or when you are in the world. Um so, uh in order to arrive at uh sort of non-trivial constraints on what uh those what those credences should be you have to hold the view that there are at least sometimes you know uniquely rationally able to assign those kinds of self-locating probabilities and so that's going to be some kind of solution to the probability problem which is specifically based on some kind of thesis about self-locating credences and how we should assign them. So people have made various attempts to do that. One common approach is based on using symmetries and arguing that these self-locating credences should be constrained in various ways by the relevant symmetries of the problem and then trying to pick out symmetries which are going to have the function of making you arrive at the right quantum probabilities. >> Okay, walk us through the distinction here between a self-locating and a non-self-locating credence or hypothesis just so listeners can get a sense of what we're talking about.
¶35>> Yeah, so a non-self-locating credence is a credence about what the world as a whole is like. It's a credence that you could specify from a sort of third-person point of view. Like when I flip this coin does it land on heads or tails? There's a well-defined third-person answer to that question which is a fact about the world itself. Either it's heads or it's tails.
¶36Self-locating credences are credences about when or where you are in the world. So they don't they don't describe They don't they're not assigned to any fact about the world as a whole which could be specified from a third-person point of view. They're assigned to your personal location. They can only make sense of a first-person point of view. So for example, if if we clone you and the original you you get killed I'm sorry and the two clones going to like make two clones and one of them is placed in a room with a red door, the other is placed in a room with a blue door.
¶37When you wake up you haven't seen what color the door is yet. You might assign credences or probabilities to red and blue >> [snorts] >> and those credences are self-locating credences because they're not about what the world as a whole is like. You know there's a red room and a blue room definitely. And the only question is which one are you currently in? So it's self-locating.
¶38>> Right. Like where where is your location broadly understood? >> Yeah. Mhm. >> Okay.
¶39So what is the the distinction you draw is quite subtle between pure self-locating credences and superficial self-locating credences. And I I correct me if I'm wrong, but I think your point is that there are certainly are rational constraints on superficial self-locating credences, but not on pure self-locating credences. And then the point is that uh the credences involved or appealed to in the Everettian case are going to be pure self-locating credences. Dot dot dot. Okay.
¶40Okay. So what what is this distinction? >> Yeah. Um so uh the question basically is about whether the locations you could be in belong to the same possible world or different possible worlds. So in the case I just described with the two clones with different color doors, both of those locations are in the same world by construction.
¶41Both of those clones exist within the same world. And there are also sort of more superficial cases where the locations are in different worlds. So that's that's more common and in terms of the kinds of self-location uncertainty we really encounter. So for example, if you wake up in the morning and you don't yet know what time this is, uh you might assign credences over the various times you could be. So that that's a self-locating uncertainty.
¶42You're not certain what time you're located at. And but of course, in every possible world there's exactly one time at which you actually wake up. So those different times that you could be located in right now are [snorts] in different possible worlds. That's only a superficial case. And the main difference here is that the superficial cases can always be rewritten as sort of a third-person uh question.
¶43And so we can just treat them like ordinary non-self-locating probabilities. So in the waking up case, we can just treat this as a non-self-locating question about you know, here is a certain person. At what time do they do they wake up on a certain day? And so it doesn't have to be you. You don't have to be involved to meaningfully assign probabilities.
¶44And so those those kinds of credences can just come from the same place that other types of uh probabilities might come from. We can just get them from empirical inquiries, from the usual science scientific theories. You know, in the waking up case there's going to be various biological facts about you and how tired you are and what your environment is like and we can use those to assign probabilities in the ordinary sort of way. And that's not going to work in the sort of at case where you're looking at two locations in the same possible world such as this this case with the clones with the two doors because in that case there's no like process which drops you into one of these places rather than the other and so you can't go and do empirical inquiries about the process about the nature of the process that chooses which one you are and to decide what the probability should be because there is no such process. So, wherever the the credences come from in that case, it's going to have to be very different from where they would come from in an ordinary non-self-locating case or in a superficially self-locating case.
¶45>> Right. Okay. Okay. So, in the this is helpful for me also in terms of writing up my paper, responding to some of your material on self-locating credences. Because it's helping me understand um when it comes to superficial self-locating credences, maybe one way to summarize is that like there clearly are in terms of the before I open my I've woken up but before I open my eyes and thinking what time is it?
¶46There clearly are like some rational constraints on what credence I should have in a particular hypothesis about what time it is. So, for instance, if I know from the past that I almost always wake up 5 minutes before my alarm at say 6:55, I should have a pretty high credence in 6:55 and a correspondingly low credence uh that it's 10:05 p.m. It's only been 5 minutes after since I fell asleep. That seems right. Like right, there is a rational constraint upon what credences I'm like rationally allowed to have.
¶47>> Yeah. >> Okay. And so then the claim is that in these more esoteric cases, which these are really esoteric. Assuming I'm a clone and I don't I know what the world is like, but I don't know say one variation on this is imagine there's a thousand rooms, a thousand clones. Uh 999 are behind red doors and one is behind a blue door.
¶48And I've woken up so to speak, but I haven't yet looked at the door and the question is what credence should I have that might that I'm behind the blue door? That is a pure self-locating credence. Um and a lot of people are naturally going to say you should have a very low credence. You should have about what would that be? 0.1% if I did my math there right?
¶49Um because there's only one blue door, but there's 999 red doors. So so clearly Emily, there are rational constraints here. Um but you don't agree and why is that? >> Yeah, um so I think that the intuition we have that there are rational constraints is coming from like a mentally making analogy to more familiar non-self-locating cases. So in a comparable non-self-locating case, um you know, if you're picking one out of a set of a thousand, uh there's a natural assumption here that the picking process is uh is in some sense uniform across all of the options and so they all have an equal equal probability and so in that case of course it's very unlikely to pick the one that's much more likely to pick the 999 and because they all have equal probability.
¶50The problem in the self-locating case is that there's there isn't any picking process. There's no physical uh process which chooses between which which room you're in. Uh and so there there would be no there would be no sensible way to say either that the picking process um chooses with equal probability over all of them or that it does not choose and that it favors one more than the others. There's just there's no fact about that at all. Um and so this [clears throat] kind of what's what's going on in the sort of intuition that says well, red is much more likely is that some kind of sort of principle of indifference is being invoked which assumes that they're all equally likely and so Yeah, 999 is much more likely, but there's no I think there's there's there's no rational justification for using a principle of indifference if there's no process choosing in the first place which could be uniform or not uniform over all the options.
¶51>> Right. Okay. Okay, good. So, let's connect some of these dots. That's helpful for me in my more technical explorations, but for the audience what is the idea about the um self-locating approach to grounding the relevant probabilities in an Everettian context?
¶52You said that the idea is we can appeal to certain symmetries such that I gather the idea is that, you know, prior to opening our eyes, so to speak, to observe the outcome of the experimental result, whether that's you know, particle spin up or spin down, um we should have a credence that we will say see spin up that is in accordance with what the Born rule would specify. Uh that that is what should constrain our self-locating credence. We And it's self-locating because it's we're taking as given that the Everettian branching multiverse exists, and we're trying to determine where in that branching structure we are. And we should do that in accordance with the Born rule. But then that really starts to resemble the case of like clones, where we know which world is actual, uh but we're trying to determine the credence we ought to have in seeing a certain result or in being in a certain place.
¶53And if you're right about pure self-locating credences, then there just are no rational constraints, including the Born rule, on our pure self-locating credences in the case Have I Have I captured it? >> Yeah. >> Okay. >> mean the the worry about the Everettian case is the Everettian case is clearly a appeal self-locating case. There is no physical process that chooses one outcome.
¶54All of the outcomes really occur. So, in that sense, it's pure self-locating. And so, just as in the clone case, because there's no process that chooses the outcome, you can't go and do empirical inquiries and learn about the process that chooses the outcome and use that to decide what credences you should you should assign. this. no no meaningful question to ask there.
¶55Um so in particular, yeah, one of the popular approaches to trying to constrain these self-locating credences and argue that they should match the Born rule involves looking at the experiment and picking out certain symmetries of it and using and arguing, well, your self-locating credences should match the symmetries of the experiment and so that narrows us down to the Born rule. Um and this this sounds at first convincing because it's very common in non-self-locating cases to do something similar to look at symmetries and use them to constrain the credences that we should have and use that as a way to decide how to use the principle of indifference, basically. Um the reason I think this doesn't work is because in a non-self-locating case, what matters is the symmetries of the process that you are using to select the outcome. So, for example, if I'm throwing darts at a dartboard, um uh and I want to know what's the sort of expected distribution of darts, it's no good to me to just look at the dartboard and say it's a circle and so the darts will probably have a rotational symmetry because the shape of the dartboard is irrelevant to that question. What matters is how am I throwing the darts?
¶56Am I throwing the darts in a symmetric way? Is this process selecting the outcome symmetric? Um so in the Everettian case, the analog is that we have a dartboard, but we don't have any process of throwing darts to to ask questions about. >> Right. >> The these symmetry proofs are doing the analog of looking at the dartboard and saying, oh, it's a circle, so I should expect a rotationally symmetric distribution.
¶57Uh they're sort of they're getting they're confusing the symmetries kind of of the outcome space with the symmetries of the process. >> I see. >> And in this case, there is no process, there's no question you can ask, and so there's no reason to expect any particular thing. >> Right. Okay, good.
¶58Good. Let me ask you a tangential question. Is there an interpretation of quantum mechanics, and it might be a a version of Everettian mechanics, wherein there is really only one conscious subject throughout the branching structure, if you just imagine your branches. And your consciousness, that consciousness that is you, is like placed into one particular branch. And so that the other people, the other versions of you on other branches, they're like philosophical zombies.
¶59I may be making this up, but I thought there was a bit interpretation along these lines. Now, if you went that route, maybe you could say there is some sort of process that literally does place you into one or the other. Or have I made all this up? >> Uh so there is an approach called the many minds approach, which >> Okay. >> rather having one conscious version of you, there's a whole bunch of conscious versions of you, which are there's a count of them, and they're kind of splitting off into different branches.
¶60Um I'm I'm not sure that anyone has seriously suggested the analog with just one mind >> Okay. >> choosing an outcome, partly because I think the the existence of all these philosophical zombies is is very unappealing for various reasons to do with the philosophy of mind. Um So uh no no, I don't think anyone is seriously proposing that, but we're we're someone to propose that. Um >> Maybe I could propose that. >> Yeah.
¶61Uh were someone to really seriously propose that, that you know, that would I think probably solve the probability problem. Um Uh yeah, it would have a whole bunch of other problems instead. >> [laughter] >> Um in particular, well, one of the main problems is that the branches are not sort of sufficiently well distinguished. They are only sort of emergent and approximate, and so I not clear that sure that they sort of exist in a robust enough way to to choose one like that. Um but yes, uh the probability problem would go away, other problems would come instead.
¶62>> Okay. Okay, good good. Uh this is this is fun. I'm I'm learning a lot. Um what would be another Well, let me let me ask you a meta question about uh the complaint that these different like quote-unquote interpretations of quantum mechanics are really better called different theories of quantum mechanics.
¶63Do you agree with that idea? I think Tim Maudlin made that point quite forcefully to me. >> Yeah, I mean I think I I mean it's to some degree it's just sort of semantic point, but I would be inclined to use the word interpretation only for the ones that don't want to make any change to the formalism. Um, so, at least some versions of the many worlds interpretation, for example, that's true. They don't want to change anything, they just want to tell you how to interpret it.
¶64I think it's fair to call that an interpretation. Uh, if you look at like the pilot wave theory, for example, or the spontaneous collapse theory, uh, those are concretely changing the mathematics of the theory in various ways. Uh, not in ways that we can currently detect, but in ways that we could possibly detect in the future. So, because they're both mathematically different and in principle empirically different, I do think it's reasonable to call those different theories, yeah. >> Okay.
¶65I like that. I'm going to run with that forever. Um, so, we have the many worlds interpretation facing the probability problem. And we we may return to it to talk about other problems. But, what are what are some other interpretations of or theories of quantum mechanics that you think also face this this issue?
¶66Or is it just the many worlds? >> No, there's a bunch of others which have similar issues. Um, the most obvious, uh, comparable problem that occurs in various, uh, what you might call observer relative interpretations. So, what the observer relative interpretations want to say, basically, is that both quantum states and quantum measurement outcomes have to be relativized to observers. Um, so, when one observer makes a measurement and sees an outcome, uh, the outcome that they see is not necessarily the same as the outcome as other observers would understand it.
¶67Everybody kind of quantum description that produces outcomes suited to them. Um, so, this faces not the same problem as the Everett, but a sort of comparably sort of epistemic problem in that if you if you take that picture to its logical conclusion, what it means is that I see one set of outcomes describing the world and you see a different set of outcomes describing the world. And most of these views don't want you allow want to allow you to have any connections between the world as I see it and the world as you see it, so we just end up in like completely disconnected realities. Uh I can't act I can't really tell you anything about what's going on in my reality. Uh you can't tell me anything what's going on about your reality.
¶68Uh well, when we talk to each other, it will seem to me as if you're talking to me and telling me things that make sense and are consistent with the world as I see it, but there's no real connection between that and what's happening for you from your own point of view. Um So, this this I think looks self-undermining in a sort of different way because obviously the empirical confirmation of quantum mechanics relies on many people working together and >> Right. >> gaining experimental results, which we've all combined and arrived at this theory. Um but this kind of story seems to undermine all of those kind of social processes involved in science, and means you know, at at the at best I could arrive at the conclusion that quantum mechanics is a good description of my observations. I could never have any reason to think that it's a sort of universal theory, which is true for everyone, because I've got no idea what's going on with anyone else.
¶69>> Mhm. Yeah, that that doesn't strike me as a good theory at all if you want to preserve I don't know, our stock of common sense beliefs that we are inhabit the same reality. But I guess it's not necess- I mean, maybe it's one step further away from common sense than is Everett. But if you've kind of swallowed the pill on this branching world structure every single moment as it were, reality is branching into >> Yes. >> entirely separate domains, then I suppose the observer relative idea wouldn't be that much more radical.
¶70It strikes me as a many worlds interpretation of a kind. >> Yeah, I mean, I think I think for me it's important to say I I am very happy with an interpretation of quantum mechanics being weird and radical. I think that's fine. I think probably the right interpretation will be weird and radical. Uh but I think it needs to be weird and radical in a way that is self-consistent.
¶71And in particular, needs to make sense of the epistemology of the theory. So, yeah, I I make in my work a distinction between uh weirdness that is uh benign and uh weirdness that is less benign. So, the benign weirdness is like strange, counterintuitive, but it doesn't undermine the sort of status of the theories empirically confirmed, whereas the other kind of weirdness is the kind of thing that's going on in Everett and these observer relative interpretations, which really seems to lead to a sort of severe kind of self-inconsistency. >> Okay. Okay, very good.
¶72So, there is the probability problem. That's one respect in which we need to save science from quantum mechanics. Are there other independent problems that you see facing these various theories and interpretations of quantum mechanics? >> Um So, I mean, I I find it helpful to divide the interpretations into two classes. So, the the unitary only ones are the ones that the ones that I would call interpretations, the ones that don't make changes to the formalism.
¶73And yeah, I think that the formalism as it is doesn't really have the resources to tell a straightforward story in which observers see the see unique outcomes and see the same outcomes coming across everyone. And so, basically all of the unit unitary into only interpretations face this kind of problem where there's some kind of weirdness to do with the evidence and it's very difficult to tell a sensible story that way. The other class of the the non-unitary only approaches, the ones which do alter the the formalism in some way. So, usually the way in which they alter the formalism is is to solve these problems and to ensure that we get, you know, single well-defined outcomes for measurements in ways that everyone everybody can share. >> [clears throat] >> So, that means that that these those approaches don't tend to face these kind of like epistemic problems about the evidence.
¶74Um but because they alter the formalism, it's harder to make them work technically. And in particular, if you're going to change the formalism, you're going to have to be able to change it in a way that works not just for non-relativistic quantum mechanics, but also in the long run it's got to work for quantum field theory and maybe quantum gravity. And of of all of these sort of non-unitary approaches that change the formalism in some way, none of them right now works for all of quantum field theory. Um and I think there are sort of some strong structural reasons to think that probably none of the existing ones are they ever going to work for all of quantum field theory. So, we're in stuck in this kind of dilemma where like the unitary only approaches work technically but have these really severe epistemic problems.
¶75And then the non-unitary approaches don't work technically um but don't have the epistemic problems. And so, either way there's sort of big big difficulties. >> So, what do we do? I mean, surely we think that quantum mechanics is true in some sense. And we've got to recover that.
¶76And we probably can't conclude actually this whole paradigm was a waste of time. And I don't think that's what you're suggesting. So, what's what is in your mind the way forward here aside from despair? >> Yes. Uh well, I mean, one of my main conclusions is that I think none of the sort of mainstream well-developed approaches is right um because they they they all fall sort of under this dilemma.
¶77Um it seems to me very unlikely that any unitary only approach can work. I think in the end they all kind of collapse into some version of Everett. And they all face this probability problem or some kind of analog of it. Uh and I don't think that problem can be solved. Uh so, that that tells me that I do think something has to be added to the formalism.
¶78Um I think the problem with the existing approaches is that they've they've tried to add things to the formalism in ways that are very sort of classical. They've just tried to add some like essentially classical underlying ontology to go turn quantum mechanics back into a secretly classical theory. Uh and it's that attempt to make to have a sort of underlying classical story which really gets them in in trouble when we move it to QFT. Um so, it seems to me that the right answer is that we do have to add something to the formalism but we shouldn't be trying be so keen on so, um we shouldn't be trying to make it more classical or to have an underlying classical story. We need to be a bit more expansive about what kinds of things we can add which will solve the epistemic problems, but won't necessarily be just turning the world back into the previous classical world.
¶79>> Okay. Um well, let me ask, do you do you have a rough sketch of what that kind of proposal will look like? Cuz I agree, that's probably like the desiderata. That's what we want. Yeah.
¶80But but is there anything on offer today that gets even close to that? >> Yeah, well, I what one route that I find um interesting are the the relational approaches. So, I've I've talked about the observer relative interpretations. Uh the problem with the observer relative interpretations is that they don't want to allow us to have any relationships between the different relative perspectives, and that means we can't talk to each other or share anything. Um so, one sort of obvious alteration to make to fix that problem is to add um relationships between different observers.
¶81Just to put in some kind of mechanism that in certain appropriate kinds of interactions, we can come into agreement. Uh so, Carlo Rovelli and I did propose an alteration to relational quantum mechanics a couple of years ago, which which does that, basically. Um And I think that that that's still I think a definitely an ongoing research program. There are questions about how to formulate that precisely. It take It leads you away very much away from the original version of relational quantum mechanics and towards something that looks looks sort of conceptually and formally a bit different.
¶82Um but I think that's an interesting route because um it it's making an alteration to the theory to solve the epistemic problems, but it's not just adding like a bunch of classical particles underneath or something. It's So, it's pointing towards a different a quite a different way of altering the theory. And I think that that that seems to be the right kind of approach to think about different types of alterations you can make rather than just adding a classical ontology underneath. >> Right. Okay.
¶83Is this theory deterministic or is it indeterministic? That's one way I've seen the different theories grouped together is according to whether or not they are deterministic. Um is that the right question to ask in this case? >> So, I the re- I think it's a complicated question because um uh the the the definition of determinism is a bit difficult. Um, so the traditional definition of determinism is based on a time evolution picture where you start at the beginning and you evolve forwards.
¶84And so, uh, the question is is the state at the beginning of time adequate to predict everything that happens later in history. Um, I think the right way to think about both this version of RQM and various other interpretations of quantum mechanics is to move away from that time evolution story towards a more sort of all at once story where the laws of nature apply to the whole universe and just constrain the whole thing. Um, and once you do that, it's much less clear that the traditional definition of determinism is really the right way to define determinism. Um, because if the laws aren't starting at the beginning and evolving forwards, it's just much less interesting to ask does the initial state determine everything else. Um, so, you know, I think probably this type of theory would not be deterministic according to the original traditional definition, but I also think that's not the right definition to use in this context.
¶85>> Okay, and maybe if you can spell that out for me again as to why it's it's because there's a more global perspective rather than thinking of at first the initial condition and then evolution or or how would you how would you phrase that? >> Uh, yeah. So, um, the sort of traditional picture of the way the laws of nature work is that they take in an initial state and just evolve it forwards and produce the rest of history like, you know, like a computer taking in input and producing some outputs. Um, and that's that's what the way the laws of sort of Newtonian mechanics work. Um, what the way the laws of the Schrödinger equation sort of looks like it works that way.
¶86But certainly once you get into the more relativistic context or even into um, uh, a more detailed analysis of what's going on in quantum mechanics, uh, I think that picture becomes increasingly implausible. And I think there's there's enough evidence in modern physics to tell us that picture is just definitely not right. Um, the way a more realistic way of thinking about the laws of nature is to think about them constraining the whole of history together kind of in the the way that like the rules of a game of Sudoku constrain the whole grid together rather than starting at one side and moving toward the other side. Um Yeah. And so how that relates to determinism is that if the laws are not starting at the beginning and moving toward the other side, then there's just no no reason really to ask the question, does the initial state suffice to predict all the rest of history?
¶87There's no reason why it should if if your laws are not functioning in that kind of forward time evolution way. >> Okay. Does that understanding of the laws of nature entail a certain approach to one's theory of time? Like to me it sounds most naturally attached to an eternalist ontology of time where past, present, and future are all equally real. And in fact, past, present, and future are just observer relative to where you are in the block.
¶88Or or is this compatible with a more robust kind of presentist ontology? I guess my concern would be like, how do you constrain the past and future if those things don't exist as they don't on a presentist metaphysic? It's a pretty obscure question out of left field, but I'm curious. >> Yeah, no, I think I think it's definitely not compatible with a presentist ontology. Um not least but you know, presentism at least in its natural formulation seems like it you have to have a sort of standard of simultaneity so you can identify where the two moments belong to the same present.
¶89So presentism is already in very big trouble as soon as you hit relativity. Um and I I I think probably not in a not rescueable. Um The all-at-once picture certainly looks more eternalist. Um you know, one thing I would say is that the traditional eternalist story um is a sort of four-dimensional block that's three dimensions of space and one dimension of time. Um So in the long run, I think it's very plausible that space-time is is is emergent and that time sort of emerges as a sort of kind of relational local description.
¶90So I think probably that in I suspect that in the long run the story is not not a a of four-dimensional block but rather that is kind of timeless in some sense until you till you go inside it and look from the right perspective. Um so, it's not quite the traditional sort of block universe picture of eternalism because this is lacking that kind of temp There's sort of no time at all in some sense from the external point of view. But yeah, certainly it's not the case that like it's certainly not going to be a present just for you. It certainly is going to have in some sense all of the times existing together. >> Right.
¶91Okay, good. Good. Let me let me step back for a second. I probably I should have asked this at the very beginning. When uh a lot of people anyway think about quantum mechanics, they think about a specific interpretation, the like so-called Copenhagen interpretation, where you have like an actual collapse of the wave function that's supposed to occur on measurement.
¶92So, so vaguely the idea is like at first, before you observe the outcome of the experiment, you do have this like superposition of all of the different possible outcomes, but then once you observe it, it seemingly collapses to one outcome. And that is really all that exists. There's no other branches, there's no other outcomes, there's just that. I think that interpretation has fallen on hard times. Why is that?
¶93And why is that still most people's very rough-and-ready understanding of quantum mechanics? >> Yeah, well, I mean, I think a lot of people have that understanding because that is kind of how we use the theory in practice. We >> Okay. >> we unitarily evolve up until the point at which we want our experimentalist to make a measurement, and then we then we collapse the wave function when the measurement occurs. Um the reason that's not a satisfactory theory is because I mean, in order for that to be a scientific theory at all, you're going to have to say something precise about which which specifically which types of interactions count as measurements, uh and you know, when exactly when and where and how does the collapse occur.
¶94And you know, the standard theory does not allow us to do that. Um so, obviously there are we we know in practice that there are certain things that experimentalists do in laboratories that count as measurements, but we have no clear sense of outside of that class what else counts as a measurement or what the line is between something that is and isn't a measurement. You know, John Bell has this sort of sort of sarcastic quote about it where he says, you know, did the wave function collapse the first time you know, a bacterium was was again consciousness or did it have to wait for an observer with a PhD? Like and so there's a you know, there's a he's being sarcastic, but there's a a real Yeah, there's a real problem here that you have if you're going to have this story where okay, the wave function collapses when an observer makes a measurement, you're going to have to say what do you mean by an observer or how do you draw the line, what kinds of interactions are measurements which can what kinds are not, exactly how does this work? So, you just have to be a bit more precise basically.
¶95>> Right, right. All right, that's fair. Um what are the demographics on this on the interpretations of quantum mechanics? What do most physicists think? What do most philosophers of physics think?
¶96It's my understanding that at least a plurality of maybe philosophers of physics think that the Everettian approach is correct that or at least that's up at the very top. Um is that right? And and like how do how do the views fall? How does that demographic pie chart look? >> [clears throat] >> Yes, well I mean I think within physics a lot of a lot of physicists just don't have an opinion or a strong opinion.
¶97Um you know, this is not a problem that you have to resolve for most of the problems that physicists are concerned with on everyday basis. So, many many physicists I think take the view that it doesn't matter. Um I think that we should just sort of take a pragmatic approach and not worry about what's kind of really going on in the measurement. Um in so far as physicists do have positions, I think quite a lot of physicists are drawn to sort of either the original Copenhagen approach or some various kinds of modern variants on it like Cubism. Um and certainly there are some physicists who who do believe the many worlds interpretation and that's cuz that is if nothing else quite a useful model to think about the theory, so physicists will use it for that.
¶98Um within philosophy, oh I yeah, I mean I definitely the many worlds interpretation is popular. I don't know if it's a majority. Uh I would say probably the many worlds approach is more popular than anything else. Um but I would guess that less than 50% of philosophers overall would would advocate it and the rest would be split between various different approaches. >> Mhm.
¶99Mhm. Okay, fair enough. Well, okay, one of the last things I want to ask you is going to be a total curveball, but it'll it'll prove to be I think a very uh useful clip to take from this video. >> Right, sure. >> Which it rises more in philosophy of religion.
¶100>> Yeah. >> Uh there's been a paper published recently from a friend of mine, Matthew Edelstein and Amoss Wollin. I guess that's two people, so they're friends of mine. Um on the so-called anthropic argument for God. >> Yeah.
¶101>> And I ask you this because it builds upon machinery with which you are very acquainted, so this isn't going to be um completely uh unfair. But the idea is that we've established that the so-called self-indicating self-indication assumption is true. That's the right approach to understanding the evidential significance of statements like I exist. And the rough and ready idea of that principle is that my existence is more likely according to hypotheses which posit large numbers of observers compared to those which posit only a few. So, if uh the the the intuition pump that Matthew will appeal to here is he'll say, "Okay, imagine that a fair coin is flipped.
¶102If it comes up heads, a billion people are created. If it comes up tails, only two people are created." Now, your data here is that you exist. And so then the question is, should you think the the the coin came up heads or tails? And he'll say, "Well, obviously you should think it came up I've already lost track of heads, whichever one corresponded to the billion people." And you think, "Okay, I mean, there's something to that." Um and then he says, "Based upon this principle, we get essentially, like, infinite confirmation for the logical logically strongest hypothesis here, which is that all possible people exist." Okay, so that's the first half of the argument for grand old that machinery. And then the second half is that the existence of all possible people is much more likely if God exists than if God does not exist.
¶103And so, having established this huge number of people, we can see this as strong evidence for theism over at least certain forms of atheism because the idea is God would not be all that unlikely to create lots and lots of people. It's a good thing to do. Okay, so there there you go. It's a it's a fire hose approach to the anthropic argument for God. Um I am assuming you're going to take aim probably at this first step and say the self-indication assumption that's just not as bananas.
¶104That's not correct. >> I think I would take aim at all of the zips. Uh yeah. Uh self-indication is a a a a claim about how we should assign our self-indicating credences. It's a specific constraint that people want to put on self-indicating credences.
¶105It's therefore subject to the same worries I have about all other constraints that people try to put on self-indicating credences. I don't think there are any rational constraints. Uh and so I don't think you're obliged to to believe that your existence is more likely conditional on the billion people than on the two people. Um you know, I think that I think that that that particular uh estimation of the principle is clearly a mistake because that um sort of presupposes that you can identify yourself first and then ask which of these problems do you think is more likely, and I'm not sure that there's a meaningful way to make that identification prior to sort of having all the other data in any case. >> Um okay, wait.
¶106Say say more there. Are you saying that even if you accept the principle, it's been misapplied? Or are you saying, "No, no, no, the the principle itself has this inherent uh fault because you can't identify who you are. >> Um, so So, I think Yeah, I I first off don't accept the principle. I I think in that particular application there's a problem where that that's assumed that there's a uh, you can first identify a person and then ask, uh, which of these cases is more likely?
¶107Whereas, I'm not sure that there's a sensible way to sort of make that reference to a person first before asking about the circumstance in which they exist. >> Okay, and could could you say more there? I just know Matthew's going to be in the background being like, "What are you talking about?" So, why can't we or why do we need to identify I guess I guess what I'm thinking of is like, you know, you're told this setup. Maybe God tells you this setup. Um, and you've just woken up.
¶108>> Yeah. >> W- Why do we need to identify anything about me for me to reflect upon the evidential significance of the fact that I exist? And then I'm asking like, "Okay, which of these two hypotheses renders that fact more likely?" Like, where have I made a misstep in identifying me in that inference? >> Uh, well, I'm I'm not sure that the the the this term me has a good reference. >> Okay.
¶109>> A a a way to establish reference which makes sense prior to specifying the world in which you exist. >> Okay. Okay. >> Uh, so this this argument seems to rely on first being able to identify a a sort of specific person who exists in a cross-world way and then asking which of these is more likely and I'm not sure that that there's any robust way of establishing reference to the person prior to saying which world you're talking about. >> Right.
¶110Okay, and then there's just the more general problem with any principle that posits rational constraints on pure self-locating credences is it's just not clear where these constraints are supposed to come from. Um, >> Right. More than that, there there is no way they could possibly come from. They could not possibly be any such constraints. >> Yeah.
¶111Well, you know, it's funny. I I think I disagree with that, but I'm still working through my thoughts on it. One one question I have for you is whether or not you think that all constraints on credences have to come from the underlying like physical chances or objective chances or whether or not you think there are are other rational constraints, it's just that none of those apply to pure self-locating credences. I'm also asking that as the paper writer and I'm it will help to make sure that our statements are accurate. >> Yeah.
¶112Yeah, I mean certainly I don't think constraints on credences come from objective chances because I I think it's very possible that there are no objective chances. Nonetheless, there are certainly constraints on credences. Um Uh I I think that constraints on credences can come from a variety of sources, but in general are empirical in nature. You need to be able to go and out and do some empirical inquiries that are relevant to the question that you are asking. Um the problem with self-locating with pure self-locating cases is that there just are not any empirical inquiries which could possibly be relevant to the question that you are asking.
¶113And so there's no way to go looking for any evidence that's relevant to those credences. >> Hm. Okay, so now I'm confused. Um and I guess we won't have to get too far in the weeds on this cuz this is more just a conversation you and I could have after the podcast. Um but I took it from your writings that uh physical probability did play a large role.
¶114Um And so I guess I wonder when you say that objective chances don't exist on your view. Like what do you mean by that? Certainly you agree that like there are certain frequencies for example of outcomes and in that sense the objective chances exist. >> Uh so that I don't I think objective chances definitely don't exist. I think they they may very well not exist and it makes no difference to the argument.
¶115Um uh I'm using the word objective chance in the strong sense that, you know, are there are there probabilities in the laws of nature themselves that sort of irreducible objective chances? I think for sure you don't need irreducible objective chances in that strong sense to have sensible constraints on credences. Um I do think that constraints on credences need need to come from empirical physical facts of some kind uh such as, you know, going and looking at past relative frequencies. So, you know, and I was talking about the symmetry case. Uh in the symmetry case what's relevant is to go and do empirical inquiries and learn about the process that chooses the outcome and what its symmetries are and what relative frequencies produced in the past and so on.
¶116Um and the problem in the the South Coast case is just that there's no there's no comparable process that you could ever do any inquiries about. So, like where what possible inquiries could ever be relevant to the question that you're asking? >> Okay, good. Yeah. Okay, good.
¶117I think I think where you and I disagree will be on whether or not underlying physical considerations are necessary conditions for constraints on credences or whether they're just sufficient. Um but I think that's more a conversation you and I can have. I I want to ask you then cuz this seems like uh you would say no, but what do you make of the multiverse explanation or style of explanation for uh cosmological fine-tuning? In so far as that involves self-locating evidence, would you say that it's not successful or is there a way to spell out the multiverse as a as an explanation for fine-tuning such that it avoids this concern or would you say no, even that speculation is ill-grounded? >> Yeah, I mean I think there are some styles of multiverse arguments which can plausibly work provided that you you do you you can do the argument without having to assign any particular set of credences over the multiverse.
¶118Like plausibly there's a style of argument which says um I it's it's very unlikely that given given one universe that there would be any life in it. Whereas, if you have lots of universes, at least one of them is going to have life in it. Um and that explains why we exist. To make that argument, you don't have to like assign any particular probability over the different universes in the multiverse. It doesn't matter what the credence distribution is.
¶119What matters is that as long as there's enough of them, at least one of them is going to have life in it. Um so, I think that that style of argument can potentially work. Um It becomes more difficult if you're trying to do things like explain the value of various constants because in order to do that, it does seem like you have to assign credences or probabilities over the individual universes of the multiverse and then say, "Oh, look, this particular uh this particular um value of the constant is more likely than the others." And so, that's that kind of argument does seem to get to trouble. >> Mhm. Interesting.
¶120Okay. I mean I'm thinking out loud. I wonder in that first style of argument, though, like what ultimately would ground our claim about a certain probability of life existing in a single universe versus the probability of life existing in say, a collection of a billion universes. >> Mhm. >> May- maybe that could come from the underlying physical considerations.
¶121I guess I just I'm not sure how to think about that. >> Yeah, I mean, well, I think to to make this argument coherently, you've got to have some like relatively concrete model in mind of like how the universes are being generated and how the >> how the the constants are sort of coming about. Um but nothing about that story involves any self-locating. You just have a bunch of universes being produced with various constants in them, and you have presumably some mechanism which assigns values to those constants. And so, then you can write down some sort of well-defined, non-self-locating you know, what is the probability that a given universe produced this way has life in it.
¶122>> Okay. >> And you know, and and and I think you can see that, you know, for for for most ways of um formulating that kind of process, it's presumably going to be the case that the more universes you have, the more likely it is that eventually one of them has got life in it. >> Right. >> And none of that involves self-location. >> Okay.
¶123Now, fair enough. Um okay, one of the last questions. Are you familiar at all with John Norton and his work on probability and the material theory of induction? Seems to me like the two of you would get along pretty well. I think you've you say similar things.
¶124>> of his. Yeah. >> Okay. Okay. Are you broadly in agreement with that approach to inductive reasoning, say?
¶125>> Um I'm very unsympathetic to it. I mean, I think induction is a difficult problem and >> Yeah. >> you know, I'm wary of any attempt to like fully and completely solve the problem of induction. I'm I'm not sure that there's ever going to be a completely satisfactory resolution to it. But yeah, I think that the direction is appealing.
¶126>> Okay. Well, that's where I'll leave it on that one. I had a nice and fiery conversation with him that I published a few days ago. So, if you're intrigued >> I'm a listener, so yeah. >> Yeah.
¶127Um okay, last question. I like to end on this one. The name of this podcast is the In Certitude podcast. So, uh of your philosophical commitments, um which of them are you most uncertain of? >> Uh I'm pretty uncertain about the hard problem of consciousness.
¶128>> Mhm. >> Uh You know, I'm as as a physicist, I'm very uh drawn to a broadly physicalist uh approach to the problem of mind, but I don't have a clear clear picture at this point of like how you could probably give a physical account of consciousness. So, I think there's there's something big there that we we haven't understood and that we need more clarity on. >> Okay. Do you see any connections between the hard problem of consciousness and resolving some of these issues in quantum mechanics?
¶129Or is that just a kind of fallaciously lump together two mysterious things and say that because they're both mysterious, they must have a common explanation? >> Yeah, I mean, it it would be nice if they could be explained together. That would be very satisfying. I don't I mean, I don't don't see an obvious route. I think you know, plausibly there's some I do think that because I'm interested in relational approaches to quantum mechanics, I think that perspectives in some general sense play an important part in that story.
¶130So, maybe there are connections there to be be drawn out. Um but yeah, I I I don't see a sort of very straightforward approach approach there yet. >> Emily Adlam, thank you very much for taking the time to come on the podcast today and have this fascinating conversation. I really appreciate it. This was super interesting stuff.
¶131>> [music]