00:00:00.000 Welcome to the topcast, episode 3 of my Breakdown, my Readings and Reflections on, the
00:00:32.040 This is the book explaining the new physics of constructor theory.
00:00:35.840 The new theory of constructor theory, it's really much broader than merely physics.
00:00:40.160 By David Deutsch, Kiara Mileto, and their collaborators.
00:00:45.120 In the last two episodes, we've looked at Chapter 1.
00:00:49.400 And in that chapter, in the last episode, we looked at how dynamical laws and initial
00:00:54.760 conditions actually work within physics in order to be able to provide you with a prediction
00:00:59.920 of what's going to happen as some kind of physical system evolves over time.
00:01:07.160 Today, we're moving on to Chapter 2, and Kiara is going to explain the limitations
00:01:12.680 of that approach, as well as the benefits of that approach and why that approach has worked
00:01:19.640 So without much more preamble, I'm going to get straight into the reading.
00:01:23.680 I'm going to do a fair bit, probably about 50-50 reading and commentary today.
00:01:34.760 Chapter 2 is called Beyond Laws of Motion, Question Mark, and Kiara writes as an introduction
00:01:43.680 Where I explain the logic of the traditional conception of physics, which uses exclusively
00:01:49.480 explanations by dynamical laws and supplementary conditions, why it cannot capture
00:01:54.840 counterfactuals such as information, work, and heat, or knowledge, and why physics
00:02:00.040 needs to resort to a radically different approach based on counterfactuals, statements about
00:02:05.760 what is possible or impossible, to incorporate those entities in an exact and fundamental way.
00:02:12.920 Now immediately, I'm skipping a few pages as I just suggested.
00:02:18.840 What Kiara does in these first few pages is describe people of the past who are somewhat
00:02:23.600 primitive to us, I suppose, and how they struggled to survive and encountered problems.
00:02:29.880 But at least they had one thing, better than other animals, they possessed knowledge.
00:02:34.920 So people of the past, early explorers, in particular, that's the example that she uses,
00:02:39.640 had knowledge that enabled them to make predictions, predictions of the weather, of the conditions
00:02:43.960 of the seas, and in particular of the winds to enable them to navigate the ocean somewhat
00:02:48.920 better than people who didn't have such knowledge.
00:02:51.800 And so, as Kiara goes on to write, knowledge allowed them to make predictions about favorable
00:02:59.320 winds and currents, about where they might encounter rocks or dangerous shallows, about
00:03:04.600 how long their journey would last, such predictions tamed some of their doubts and fears,
00:03:10.120 and eased them through perils and uncertainties.
00:03:13.320 As for those early explorers, predictions are still the most sought after, output of science,
00:03:21.280 They will be one of the focuses of this chapter.
00:03:24.080 I shall explain the logic of the traditional way of making predictions in physics, show
00:03:28.920 its limitations, and suggest how counterfactuals can remedy some of those limitations.
00:03:34.760 A prediction, in physics as well as in other fields, is a conjecture about some piece of information
00:03:40.360 that is not known prior to the prediction, like any conjecture, a prediction could be false,
00:03:46.120 as one might discover by checking whether the prediction is or is not met in reality.
00:03:51.080 False does not imply useless, pausing their just my reflection on this.
00:03:55.800 Remember that Kiara, like David, like myself, perperience, so all knowledge is conjectural,
00:04:04.040 or any claim that we make is conjectural, anything that come out of your mouth or out of your pen
00:04:08.760 or out of your computer is going to be conjectural, and that includes predictions.
00:04:14.840 So even if you have a robust scientific theory, a good theory or physics, for example,
00:04:20.680 that allows you to make predictions. The predictions themselves, well, they're still conjectures as well,
00:04:25.800 because they're being derived from a conjectured explanation.
00:04:30.040 So they can't be anything more than conjectures. They're not more certain.
00:04:34.040 They're not going to give you more reliable knowledge or anything like that.
00:04:38.360 What they're going to do is tell you what's going to happen on the assumption that this particular
00:04:47.880 So the prediction works on the assumption that the underlying explanation works.
00:04:52.920 And even then, that you haven't made an error somewhere other in moving from your explanation
00:04:58.200 to your prediction. And this brings me to a slight distinction here that Kiara doesn't make
00:05:05.160 in the science of canon Kant, but I've made elsewhere following the beginning of infinity,
00:05:10.440 for example, that there are species of guesses about the future.
00:05:15.560 There are predictions which are something like, in my words, derivations, logical derivations,
00:05:23.960 given a good, typically scientific theory. And in opposition to this, there are prophecies.
00:05:31.240 Our prophecies come in various kinds. They might come completely uncoupled from any consideration
00:05:38.120 of good experimental knowledge. And in particular, they can come uncoupled from knowledge
00:05:45.080 about knowledge. A prophet is, after all, someone who guesses about the future. Either at random
00:05:51.400 or in some cases, and this is the more technical sense that David Deutsch seems to use the word,
00:05:57.160 prophecy is where the prophet purports to explain what will happen in the future,
00:06:02.280 while ignoring the effects that people might have. Namely, what knowledge they might create.
00:06:08.280 Scientists, unfortunately, can be prone to this, which means politicians will be as well,
00:06:15.400 because they tend to take scientists very seriously, even when they're being prophetic.
00:06:20.360 And we'll come to some examples of this. But long-term predictions about the behavior of
00:06:27.640 civilizations or what might happen to Earth or even the region around Earth, given the existence
00:06:33.080 of people who are bringing in to existence knowledge, which can create technology, which can change
00:06:39.320 the outcome of whatever your prediction is going to be, is always something that needs to be kept
00:06:43.480 in mind. And something that politicians, for example, the people whose hands are on the levers of
00:06:48.840 economic power to a large extent, can end up making decisions based upon the prophecies of scientists
00:06:57.000 who assume that no such technology and no such creativity can ever affect their derivations from,
00:07:04.680 they're otherwise usually good scientific theories. And of course, that's in the best case.
00:07:10.120 That's in the case where the scientist is really using a good scientific theory,
00:07:14.840 rather than a hypothesis which might indeed not contain a good explanation at the heart of it.
00:07:21.480 Okay, let's continue on this line that Kiara has about prediction, and she writes,
00:07:27.960 an example of a false but far from useless prediction in maritime history is that made by Christopher
00:07:34.280 Columbus in the 15th century, he predicted that by traveling westwards from the coasts of Europe,
00:07:40.840 one could reach the east, the Indies, his specific prediction was, as we know, erroneous,
00:07:46.600 ought to be precise, incomplete. He had not guessed that another continent was in the way. In fact,
00:07:51.880 Columbus's ocean exploration is how renaissance Europe discovered the Americas. Still, his prediction
00:07:58.120 was powerful, useful, and contained some truth. Had he been able to continue traveling westwards
00:08:04.680 beyond America, or thousands more miles south, round Cape Horn, he would have reached India,
00:08:10.840 pausing their smart reflection. This is the point I've made elsewhere before as well. I should
00:08:15.800 stop saying that. But I tend to repeat myself in various podcasts. I should add here.
00:08:23.240 I should add to this idea of a false but far from useless prediction. Of course, lots of things
00:08:29.320 in physics, like for example, any prediction made from Newton's theory, any prediction made from
00:08:35.080 Newton's theory, must strictly speaking turn out false because the theory itself is false.
00:08:41.320 But that doesn't make it useless and it doesn't make it a bad explanation either. It's a false
00:08:46.040 explanation. But it's very, very, very useful and it contains some truth as that phrase there that
00:08:52.600 Kiara has used comes into its own, containing some truth. Not at a quantity of truth that we can
00:08:57.960 measure or anything like that. And in fact, not only are the predictions from Newton's theory not
00:09:06.200 useless. They can sometimes be more useful, in a sense, depending upon your problem situation,
00:09:14.440 than the more true theory, the more accurate theory, the best explanation, which is Einstein's
00:09:20.200 relativity in this particular case. And it's all to do with precision and perhaps our efficiency.
00:09:26.760 So if you want to do something like, let's say, determine the cause of a car accident because you're
00:09:31.320 a crash scene investigator working with the police and you need to retrodict what happened,
00:09:37.640 what previously happened. You arrive on the scene and it's just a mangled wreckage of two cars
00:09:43.080 that have crashed into each other. And now you're trying to figure out, based upon the amount of
00:09:47.400 damage done and the tire marks and so on, the evidence that you've got there before you,
00:09:51.720 to determine whether or not, if one or both of the drivers were breaking laws in some way like
00:09:57.960 breaking the speed limit, then strictly speaking, if you really wanted to get the highest precision
00:10:04.440 most accurate, most correct answer, you should be using Einstein's special relativity.
00:10:10.440 But no one's going to do that. No one's actually going to do that. Because, although your
00:10:15.560 answer would be closer to the actual truth of the matter, presuming that all your assumptions are
00:10:19.800 correct as well, it's going to be swamped by the errors in your assumptions anyway.
00:10:25.000 Even if you had perfect assumptions about what was going on, and you were able to find out the
00:10:29.240 velocity of these cars prior to the crash to figure out if either of the parties had been
00:10:33.640 breaking the law, Newtonian physics is going to do just as well in our court of law to try and
00:10:40.600 make your case. Because the difference between the two answers that you get for the velocities
00:10:47.160 of the cars will be found in like the fifth decimal place or something, a fraction of a percentage
00:10:52.840 difference between them. It only becomes significant, special relative, the effects of special
00:10:57.400 relativity, only becomes significant once the velocity gets really, really high. So this is why,
00:11:04.120 in fact, the false theory can sometimes be preferable because more people will, I understand it,
00:11:10.040 be able to apply it more efficiently and quickly, especially if time is a factor.
00:11:14.440 And C is going to be well within the error bars, we might say, of whatever the assumptions might
00:11:21.960 be in the first place. So the differences between special relativity and Newtonian physics are going
00:11:29.320 to be completely swamped by all the other elasticity in the calculation, namely what you assume
00:11:37.320 to begin with and what the evidence enables you to measure as a matter of fact. Okay, back to the
00:11:42.520 book and Chiara writes. An notorious case of a useless prediction appears in the legend of the,
00:11:48.200 and I'm going to butcher this completely, cumian sibil, the priestess who resided in the Apollonian
00:11:54.680 oracle at cumay, an ancient Greek colony where Naples is today. The story goes that a pilgrim
00:12:01.800 came to her asking for a prediction about whether he would return safely from an imminent war.
00:12:07.560 This was the sibils reply, Ibis, Radibis, non-mariress, in bellow, the cryptic sentence contains a
00:12:15.560 prediction, which is what the pilgrim was hoping for. But unfortunately for the pilgrim,
00:12:20.840 it is hopelessly vague. According to where one pauses, I, after Radibis or after non,
00:12:27.960 that statement can have two completely different meanings. One is, you will go, you will not come
00:12:33.080 back, you will die at war, the other is, you will go, you will come back, you will not die at war.
00:12:39.320 Apparently the statement was uttered only once, and with a flat tone of voice. So it was
00:12:44.200 impossible to tell which one of the two meanings it had. Okay, pausing that as my reflection,
00:12:49.000 it's of course quite typical for an ancient prophet to be vague. Modern prophets of course,
00:12:55.080 are similarly vague at times. All the claims of looming disasters of various kinds,
00:13:00.280 economic, environmental, civilizational, AI, apocalypse. It's always coming, but we're never
00:13:06.360 really told precisely when or what means, or exactly what the nature of the disaster will be.
00:13:13.240 It's all prophecy. Perhaps one step up from sibils there, but not much. While sibils is kind
00:13:19.240 of vacuous, because it's a contradiction, there's two competing predictions, both of which rule
00:13:26.200 the other one out. At least our modern day prophets are sometimes making an actual claim of a thing
00:13:33.160 that is purportedly going to happen. For example, the AI will kill us and take over the world.
00:13:39.560 Of course, many do not quite do this. Many make sibil-like claims. That being AI might kill us
00:13:48.360 and take over the world, or global warming might melt all the polarized caps before we choose
00:13:53.240 to do anything about it. But of course, anything might be the case. Anything not
00:13:58.440 precluded by the laws of physics, and all that might be the case. Back to the book.
00:14:05.400 Chiara Wright. What is the difference between Columbus's predictions and the sibils?
00:14:10.120 The former is powerful and worthy, even if false. The latter useless. But why exactly?
00:14:15.880 The answer shall not be found by examining the contents of the predictions themselves. We have to go
00:14:20.360 deeper. The difference lies in what the predictions rely on. It lies in the underlying explanations.
00:14:26.120 The prophecy of the kumian sibil did not rely on an offer or explanation of why the pilgrim
00:14:31.480 would or would not come back from war. Without any further explanation, it is impossible to tell
00:14:36.520 which of the two opposite meanings that statement has. Columbus's prediction instead relied on a good
00:14:42.520 explanation that the earth was round, pausing there just again. To this I would add that even if
00:14:51.880 there is a good explanation on offer about the future, it sometimes needs to take into account
00:14:57.400 other good explanations, which if it ignores them will get the prediction wrong. The one I
00:15:02.760 mentioned often is the prediction quite scientific in a sense that the sun will eventually run
00:15:08.440 out of hydrogen fuel expanded to a red giant and boil all the oceans of the earth as it does so.
00:15:14.040 That seems like a reasonable prediction for 5 billion years hence, given what we know,
00:15:19.800 which is quite robust, good explanations about stars like the sun. The problem is, it ignores what
00:15:25.720 people might do, it ignores other good explanations, it purports in other words to know a future that
00:15:31.480 creativity may have an impact on. It presumes that creativity cannot possibly have an impact upon
00:15:37.960 that scenario. And by the way, supposedly, the oceans will boil well before 5 billion years hence
00:15:44.520 due to increasing temperature of the sun anyway if we don't do anything about it. That, of course,
00:15:49.960 is always the point on this planet. For as long as we're here, any prediction is always,
00:15:55.640 unless we don't do anything about it. So, any prediction in science is going to come true,
00:16:02.520 unless we do something about it. But we will often do something about it. By the way,
00:16:07.880 you can simply Google when will the oceans boil to find a debate over the timeframe. It's about
00:16:12.760 a billion years or so apparently, back to the book, Chiarites. The quality of a prediction depends
00:16:18.920 ultimately on the underlying explanation. This point is so important that we need to spend a little
00:16:23.400 time reflecting on it. It is just like what happens on a long hike. When you reach a spot with a
00:16:29.240 wonderful view, it is good to pause and take a little rest while contemplating the beauty of the
00:16:33.400 landscape from that particular place. Our gaze now moves far away from the gloomy land of oracles
00:16:39.880 and comes to rest on a boundless, shimmering prairie, a field where the connection between good
00:16:45.320 explanations and powerful predictions is clear and immediate. It is the field of physics.
00:16:51.800 Predictions in physics are powerful. They supersede religious and mythological predictions and also
00:16:58.520 those made by rules of thumb. Rules such as, in order to grow good carrots in your garden,
00:17:03.720 you need to sow carrot seeds in February. Often, laws of physics are so general that they make
00:17:09.560 claims about the universe as a whole. I'm not going to read the next section. It's an interesting
00:17:14.840 story. It's the story of the discovery of Neptune and the discovery of Neptune was basically about
00:17:20.920 the fact that as Uranus went around the sun or Uranus as some people say. As it went around the sun,
00:17:27.320 it's orb deviated from what was predicted from Newtonian physics. This led astronomers to be able
00:17:34.760 to predict where Neptune should be found and training their telescopes towards that region of space
00:17:41.240 indeed they found it. This was so successful of course that it led scientists later on to
00:17:48.280 presume that a similar effect was happening with Mercury and so they presume that there was this
00:17:54.920 hidden planet that was perturbing the orbit of Mercury. As it turned out, no such planet
00:17:59.160 was found or ever could be found or indeed exists because the solution there was not another planet
00:18:04.920 perturbing the orbit of Mercury from what was predicted behind Newtonian physics but rather
00:18:10.920 it's because Newtonian physics turns out to be false and you need general relativity to get the orbit
00:18:17.720 right. Okay, back to the book in Chiarites. In physics and in science in general, both explanations
00:18:24.920 and predictions must satisfy strict criteria. In particular, explanations must generate predictions
00:18:31.400 that are testable. End quote. Okay, I haven't read much there but I feel that this needs lingering
00:18:38.520 over. She's just written there. In particular, explanations must generate predictions that are testable.
00:18:45.960 Now after years of doing this to me of course and to anyone who listens to me at all,
00:18:53.480 you'll be bored of hearing something like that. It's quite clearly the case.
00:18:59.640 This is in science, okay? In science, you need explanations that generate testable predictions.
00:19:09.240 I'm astonished now that it still needs to be said and this thesis still needs to be
00:19:14.520 defended. Not because we want to say that science is the king of all subjects and has nothing
00:19:20.440 else that's important because we have testable theories and so we are the superior subject.
00:19:25.720 No, nothing like that. It's simply we need testable theories so that we have a measure by which
00:19:31.320 we can actually talk sensibly about how to determine what is really going on in the physical world
00:19:36.680 and which of our competing theories is going to be the correct one or the more correct one
00:19:43.240 and which one has been refuted. If we've got no experimental way of refuting two good
00:19:49.480 explanations, then how do we choose between them? How do we take action scientifically?
00:19:55.560 Now morality on the other hand is not testable and in many, many cases you certainly shouldn't
00:20:01.080 go testing moral claims. It would be a great mistake. For example, we shouldn't be going out
00:20:07.240 and testing what kind of torture actually causes more suffering. On the one hand that's a scientific
00:20:13.560 question and on the other it's a moral question but we shouldn't do it and no one should do this.
00:20:17.240 There's all sorts of moral things that we shouldn't try to test. The same is true of certain
00:20:23.720 theories in economics, certain theories in politics, a claim like you know Marxism is true.
00:20:31.800 It's not actually testable but to the extent that it's already been tried, it's failed everywhere.
00:20:37.320 So there's no need to keep on trying it because there's a moral claim at the center of it.
00:20:43.480 There are claims about people that we know are false, claims about society that we know are false.
00:20:49.320 There is science, there is non-science and there is pseudoscience.
00:20:54.520 Morality is non-science but that doesn't make it unimportant. It is crucially important.
00:20:59.400 Marxism on the other hand, along with astrology, is a pseudoscience. Both of them can be dismissed
00:21:05.400 as false because they're bad explanations and to the extent that anyone has ever taken them seriously
00:21:11.160 as explanations, they have suffered. In the case of astrology, pursuing a false system for directing
00:21:17.400 one's life in a particular way when better ideas are actually on offer and in the case of Marxism,
00:21:22.840 well, the death of hundreds of millions of people. So there's a spectrum of bad effects that
00:21:27.960 these pseudoscientific ideas can have on people. But anyways, some people, well as I've said before,
00:21:34.120 let's face it by some people actually mean some theoretical physicists, get their noses quite
00:21:40.040 out of joint on this point. They want to argue that testability is itself some kind of antiquated
00:21:46.200 notion that we are attacking their favorite ideas, those favorite ideas of theirs arising in a
00:21:52.680 sense out of physics because we say they're not testable. I've seen people get upset that claiming
00:21:58.920 universes with other laws of physics are not properly part of science because they're not
00:22:03.480 testable because we cannot make an observation of them or do a crucial experimental test of them.
00:22:08.760 But there's no reason to be upset about that. Metaphysics is absolutely fine. And what wasn't testable
00:22:15.320 yesterday might turn out to be testable tomorrow if we can figure out how. I think this almost needs
00:22:20.840 to be written on the walls of any academy of the future. Let no one enter here who is ignorant
00:22:26.520 of areas outside of science. There are scientists who think that being scientific is a virtue in
00:22:33.240 a way that being artistic or being philosophical is not and so on. But mathematics, for example,
00:22:38.680 is not science. Constructing mathematical structures that describe alternative physical laws of a
00:22:44.680 universe that is not our own is just that constructing mathematical structure. It's not strictly
00:22:51.080 doing science. It's doing theoretical physics of a kind, I guess. But this would just
00:22:57.240 place that kind of theoretical physics within the realm of pure mathematics, which is absolutely fine.
00:23:03.480 As I've said before, pure mathematicians, Hardy was one. Have been proud of the fact that
00:23:09.320 their mathematics was utterly disconnected from the real physical world in a sense. It was about
00:23:14.920 purely abstract things, at least that's what they thought. In Hardy's case, he actually said it,
00:23:20.200 and I'll quote from his book, a mathematician's apology. He said, quote, I have never done anything
00:23:25.400 useful. No discovery of mine has made or is likely to make directly or indirectly for good or ill
00:23:32.600 the least difference to the amenity of the world end quote. And so he was he was proud of the fact
00:23:39.160 he was kind of like almost an abstract artist. He was just doing stuff with his mind disconnected
00:23:46.760 from the physical world to a large extent. And he thought that nothing about his mathematics could
00:23:51.800 possibly make a difference. But I'm just going to read from the Wikipedia article on Hardy,
00:23:57.720 actually. And the Wikipedia article says, aside from formulating the Hardy Weinberg principle
00:24:04.280 in population genetics, Hardy's famous work on integer partitions with his collaborator,
00:24:10.600 Romana Jan, known as the Hardy Romana Jan Asymptotic Formula, has been widely applied in physics
00:24:18.200 to find quantum partition functions of atomic nuclei first used by Niels Bohr to derive
00:24:24.280 thermodynamic functions of noninteracting Bose Einstein systems. Though Hardy wanted his maths to
00:24:31.960 be pure and devoid of any application, much of his work has found applications in other branches
00:24:39.000 of science end quote. And so even if you're doing pure mathematics, so you think you are,
00:24:44.360 it could be applied later on. If you think you're doing metaphysics, it could be applied later on.
00:24:47.560 It's no seem to be doing these things because they might come within the purview of science.
00:24:53.960 And so actually Hardy's work is now a part of science. So too for people working on
00:24:58.760 megaverse theories or the claims about alternative laws of physics and alternative universes,
00:25:04.040 it's also the good kind of like string theory at the moment. None of it seems to be testable
00:25:09.640 and so therefore strictly within part of science, so to speak. But the lesson of Hardy's pure
00:25:14.840 mathematics should resonate. His work was important regardless of whether or not it would eventually
00:25:19.960 be applied, that it was brought into science in a sense eventually, just makes it now doubly important.
00:25:26.680 Okay, back to the book. Chiarites. Perhaps you have noticed that being testable is itself a
00:25:32.280 counterfactual property pertaining to what can be done with the prediction. There are indeed
00:25:37.080 counterfactuals at the heart of most fundamental scientific theories and of the process of scientific
00:25:42.680 discovery, specifically testable means that it must be possible to set up an experiment to disprove
00:25:48.520 the prediction if it is false. I, if it does not match what it's actually observed in reality.
00:25:54.200 Okay, so end quote. That's that's perfect. Chiarite then goes on in the book to give some simple
00:26:01.240 examples of what it means to be testable in science. I think people listening to this probably know
00:26:05.400 what it means. So we'll skip over those and she also goes over some untestable ones,
00:26:09.880 like for example, that the universe might be supported on the back of a dog, or many dogs,
00:26:14.600 or infinite dogs, which is a variation of that whole joke about its turtles all the way down.
00:26:19.480 Okay, these metaphysical theories that can't be tested. Okay, also, as far as we know, can't be
00:26:25.480 tested. Of course, if we could observe this dog on which the universe was supported, then it would
00:26:30.120 be part of the universe, part of the observable universe, and therefore it would be testable.
00:26:35.160 But, presuming that these things are outside the universe and therefore by definition,
00:26:40.040 outside of what we can observe means they're untestable. Okay, so I'm skipping quite a substantial
00:26:46.440 bit and going to the part where Chiarite's quote, why is testability of predictions so
00:26:53.320 central to the progress of physics and science in general? The reason is that it provides a
00:26:58.760 particularly efficient way to find mistakes in the explanations and correct them. I want to open
00:27:06.120 a digression to illustrate how predictions, explanations, and testing are all intertwined within the
00:27:13.240 method that allows science to make tentative progress. To this end, let me stir the cloudy water
00:27:19.720 in the pond of history and bring up the spirit of the thinker who pioneered the scientific
00:27:23.640 method as we know it. Galileo Galilei. Galileo's experiments to test these predictions are striking
00:27:29.560 for their beauty and simplicity. He intended to test these theories predictions about the
00:27:33.000 motion of systems against those that Aristotle had proposed in antiquity and which had been considered
00:27:38.600 the authority ever since. Galileo's predictions were about the motion of a hard, smooth,
00:27:44.760 bronze sphere left to roll inside inclined smooth groove without friction. In particular,
00:27:50.360 he predicted that spheres of different sizes or masses would undergo the same motion down the
00:27:55.160 groove, the same speed in particular. This prediction was in sharp conflict with Aristotle's theory
00:28:01.560 which predicted that spheres of different masses would roll down with different speeds.
00:28:05.880 On the face of it, Aristotle's idea seems intuitively true, which makes Galileo's prediction
00:28:11.480 all the more interesting. Okay, pausing there and just skipping a little bit as well,
00:28:15.640 and Cara goes on to explain how. Well, this same idea and the one that's usually taught in science
00:28:23.960 class is about, well, if you've got a small mass, one kilogram mass, one kilogram sphere of metal,
00:28:32.840 and you've got a large mass, a five kilogram sphere of metal, and you drop them simultaneously,
00:28:37.480 which one hits the ground first? And the intuitive idea, even amongst some adults that I've done this
00:28:42.840 with, who haven't learned physics for whatever reason or have never done the experiments,
00:28:47.880 to think that the heavy one must hit the ground first. And this kind, it makes intuitive sense,
00:28:52.840 the reason it makes intuitive sense is because, of course, people think of feathers and leaves
00:28:57.080 fluttering to the ground, and so they think that it's due to the mass. Now, you have to try and
00:29:02.680 eliminate the air resistance and always do the experiment properly, but once you do control the
00:29:07.560 experiment as well as you can, then indeed you find that the two masses hit the ground at the same
00:29:12.440 time. Brian Cox does a wonderful version of this with actual feathers and bowling balls in a
00:29:18.360 huge vacuum chamber, or put the image up on the screen. It's a YouTube. And you can see, in fact,
00:29:24.440 that the feathers do fall to the ground at the same time as the boulders. And I think there
00:29:29.640 was an astronaut, of course, as well. I've talked about this before, who dropped the feather and the
00:29:34.920 hammer on the moon at the same time, and I hit the ground at the same time as well. But you don't
00:29:39.800 even need, well, you do need to do the experiment to disprove the different theories, but there is
00:29:46.120 a thought experiment that you can do as well on the assumption that the rest of physics is unchanged.
00:29:52.760 And Kiara talks about the thought experiment, and so let me read the thought experiment,
00:29:58.200 quote, from Kiara. Galileo reasoned like this, if one joins a smaller sphere to a larger sphere,
00:30:05.640 via a string, and then drops them both from a height. What happens, according to Aristotle,
00:30:12.360 is that the smaller sphere has a small velocity. In this thought experiment, the smaller sphere
00:30:17.320 should lag behind. If the two spheres can fall down for long enough, the smaller sphere would
00:30:23.960 slow down the larger one by pulling on the string. So the combined system made of two spheres
00:30:30.440 would go down at a speed that is slower than that of the largest sphere by itself. But here is
00:30:36.040 the glitch. This contradicts Aristotle's idea that that systems with larger masses should have
00:30:41.640 larger velocities. After all, the system made of the two connected spheres has a larger mass than
00:30:46.920 the largest sphere by itself. If Aristotle's idea were true, it should be faster, not slower,
00:30:51.960 as we concluded. Therefore, by this thought experiment, Galileo was led to conclude that Aristotle's
00:30:57.640 idea was false, and that spheres of different masses should undergo the same motion when falling
00:31:02.680 freely. He then conjectured with another leap of creativity, but they should display the same
00:31:07.240 behavior when sliding down the groove. Paul's idea just might reflect that. My version of this,
00:31:12.440 of this thought experiment, if you like, is if you have, let's say five people who like to go
00:31:20.760 skydiving. They all jump out of the aeroplane. Now, on Aristotle's theory, the heavier that they
00:31:28.520 are, the faster that they fall, presume they've all got the same mass, so therefore they all fall
00:31:34.040 at precisely the same rate. Now, if they're skilled skydivers, they can direct themselves in such
00:31:40.360 a way that they're very close together, and they could literally put their arms out and hold hands
00:31:45.480 with each other. Now, do you regard that as a system of five people with five times the mass of any
00:31:51.480 one of them? And so when they hold hands, do they suddenly speed up, and then when they let go of
00:31:56.760 hands that they suddenly slow down? If they do indeed speed up when they hold hands, where is the
00:32:01.560 energy coming from in order to increase their acceleration? What's going on there? I'm not sure why
00:32:07.400 Aristotle himself didn't think of things like this. I don't know. Probably because he was thinking
00:32:12.680 of just so much stuff, that he didn't think too deeply about his physics. Aristotelian physics is
00:32:20.920 a, I guess, it's a first attempt. I was almost going to say it's a good first attempt. I don't
00:32:25.800 think it's a good first attempt. It's not precise in any way. I don't know that he did much in
00:32:32.680 the way of quantitative analysis, and he didn't do experiments as well, so that's not a good
00:32:39.320 scientist Aristotle. Fine at other things. Fine at other areas of philosophy. Good at virtue
00:32:45.240 ethics and that kind of thing. Bad physicist. Anyway, Kiara goes on to explain how Galileo, of course,
00:32:53.240 went on to test the actual prediction that he made against Aristotle's ideas by observation,
00:32:59.960 so he actually went and did the experiment. She doesn't mention the leaning tower of pizza,
00:33:03.640 maybe because that's an urban legend. It's a myth. Maybe he didn't actually drop things off the
00:33:10.760 leaning tower of pizza to see which one hit the ground first, but back to the book and Kiara writes,
00:33:17.720 explanations whose predictions are found wrong in an experiment automatically become problematic,
00:33:24.360 and they are usually dismissed in favor of alternative explanations. As mentioned in the first
00:33:29.800 chapter, Galileo's and Newton's explanations, and the resulting predictions have an important trait
00:33:35.480 in common. Their approach to explaining physical reality is centered on what is usually called a
00:33:40.840 law of motion. A law of motion, or a dynamical law, is a description of where a system such as a
00:33:46.280 sphere of planet goes, given that its motion starts at a certain point in space and time. Think
00:33:51.960 of a sequence of snapshots, each of which represents the state of that system at different times.
00:33:57.720 The law of motion provides the rule for how the snapshots are ordered. In particular,
00:34:03.480 there will be an initial and a final snapshot, representing the starting and ending states of the
00:34:09.080 motion, which in physics jargon are called initial conditions and final conditions. For instance,
00:34:15.560 in the case of a ball, fired by a cannon, the initial snapshot contains the ball sitting inside
00:34:21.640 the cannon, about to be fired. The final snapshot represents it when it lands on the ground.
00:34:27.720 Typically, any snapshot along the sequences explained in terms of its predecessor,
00:34:31.640 ultimately in terms of the initial snapshot. All sequences of snapshots described by
00:34:36.920 known laws of motion have a particular property. Each snapshot has only one predecessor and one
00:34:42.520 successor in the sequence. This property is something physicists call reversibility of dynamical laws.
00:34:48.920 Once you have gone all the way down the sequence of snapshots, you can go back without any uncertainty
00:34:54.200 because each snapshot has one predecessor. Unlike in a garden or labyrinth with
00:34:58.760 fourking paths, therefore, no bifurcations occur along the line. There is no ambiguity
00:35:04.680 in how to go back or forth. The explanation by laws of motion is the most traditional in
00:35:09.160 physics. It was first introduced by Galileo, then it became established with Newton's laws.
00:35:13.880 Today, the two most fundamental physical theories, general relativity and quantum theory,
00:35:18.840 are expressed by laws of motion 2. And so are all other theories that physicists
00:35:23.560 generally consider fundamental, like those governing electromagnetic fields and elementary particles.
00:35:29.480 The long-term success of the approach by laws of motion is remarkable. It predictions are extremely
00:35:35.080 powerful. Suppose, for instance, you are a general about to attack a city built by robust walls,
00:35:40.360 which you want to batter down. Newton's laws tell you exactly how to tilt the cannon
00:35:44.920 in order to maximize the impact of its projectiles by predicting their motion in every detail.
00:35:50.200 For example, they tell you that there are only two possible paths available to the same
00:35:55.640 point of impact for the cannonball with any particular speed, pausing as my reflection.
00:36:00.440 If you're interested in this, this is all this topic of projectile motion. A big topic in junior
00:36:08.280 level physics, there is this thing, physics education technology from the University of Colorado.
00:36:15.320 And there's this great little app that's on putting up on the screen now.
00:36:20.760 And you can play around with it in order to test precisely what Kiara has talked about there,
00:36:28.200 that you can either shoot something up really high so that it comes down and hits point x,
00:36:35.400 or you can shoot it at a low trajectory and also land at that same point x.
00:36:41.880 The calculation of this is simple and interesting because projectile motion brings together
00:36:48.040 various aspects of otherwise disparate areas of mathematics, algebra, trigonometry,
00:36:53.480 calculus, if you want to go down that road as well, back to the book and Kiara.
00:36:57.640 In both cases, the ball describes a parabola in the air, but with different maximum height,
00:37:02.200 depending upon the initial condition, the angle at which the cannon is tilted initially.
00:37:07.400 When the cannon is tilted at a high angle, the ball flies high and falls down beyond the walls.
00:37:12.920 If the cannon is tilted at a lower angle, the ball flies lower and it can if the angle is right,
00:37:18.680 strike the protective wall. In both cases, the description of what is going on is encapsulated
00:37:24.360 in the sequence of places the ball traverses as time goes by. This set of places is the ball
00:37:30.280 trajectory, its path, which is dictated by the laws of motion, Newton's law in this case.
00:37:36.600 In this approach, the explanation for why the ball hits the target at the end is given in terms
00:37:42.040 of the places the ball goes through. Ultimately, as I said, in terms of its initial position
00:37:47.560 and velocity, the initial conditions of the system's motion. Since the dynamical law approach
00:37:53.800 is so powerful, it is natural to wonder whether it could be extended to explain everything
00:37:59.080 that happens in our universe, including the whole universe itself. In other words,
00:38:04.040 would a physical theory of the initial conditions of the universe and of its laws of motion
00:38:08.440 provide a satisfactory explanation for everything in it? The answer, as you are about to discover,
00:38:13.640 is no. I shall point out that explanations in the form of laws of motion and initial conditions
00:38:19.880 are excellent for a special purpose, i.e. to make predictions about what happens on a sub-part
00:38:26.680 of the universe, like cannon or tennis balls, marbles and planets. But they cannot explain
00:38:33.400 everything in physical reality. In fact, when regarded as an explanation of everything,
00:38:38.200 they have serious problems. Problems are fruitful things in physics, as they are in life.
00:38:43.160 They hold the promise of improvement when they are addressed properly. These problems are the very
00:38:48.600 reason why we have to venture on our journey in the land of counterfactuals. As I said,
00:38:53.560 the dynamical law type of explanation looks like a sequence of snapshots. It has an initial
00:39:01.000 and a final snapshot. And there are all the snapshots in between, whose order is set by the laws
00:39:06.760 of motion. The explanation for something happening on an intermediate snapshot, for example the
00:39:12.200 cannonball is suspended in the air at the highest point of its trajectory, is in terms of what
00:39:16.600 happens in the snapshots before and after that particular snapshot. Now, if the initial snapshot
00:39:22.840 of the sequence reminds you of the dog in the cosmology I mentioned earlier in this chapter,
00:39:27.400 you are quite right. Why should one have a particular initial snapshot? And not a different one.
00:39:33.080 Surely there must be an additional explanation for that, but that explanation cannot be itself in
00:39:39.480 the form of initial conditions and laws of motion. It cannot be given in terms of another sequence
00:39:44.760 of snapshots. Otherwise, that explanation would just look like adding another sequence of snapshots
00:39:50.200 to the existing sequence placing it at the start of the latter. But that new sequence in turn
00:39:55.480 would require an explanation for its own initial snapshot and so on. In the dog-based explanation,
00:40:01.240 this would correspond to supposing dogs all the way down to explain the first dog.
00:40:07.000 The approach by initial conditions and laws of motion taken in its strict form is not a self-contained
00:40:12.520 explanation for the universe. Adding sequences to the first sequence or dogs to the first dog
00:40:18.040 does not help to address the problem. This problem is what philosophers call an infinite redress.
00:40:24.280 It is exactly the same problem that religious explanations for the origin of the universe
00:40:29.480 run into pausing their just my reflection. Wow, that's really important. As an all-encompassing
00:40:35.960 explanation, this dynamical laws and initial conditions thing, this thing that physics has
00:40:40.760 hitherto used so powerfully and so well. Nonetheless, ultimately, we'll fall into infiniter aggressive
00:40:48.280 at tries to be a theory of everything. Another reason that physics is, as it is, can't explain
00:40:56.440 everything, including physics. And Kerigos, on to explain that God suffers from the same thing.
00:41:03.640 If anyone says what explains the universe and they say, well, God, you know,
00:41:08.440 create the universe. And of course, reasonably, you can ask, well, who create a God? If they say,
00:41:15.400 and they usually do nothing did, well, why not just take God out and just say nothing created
00:41:20.520 the universe? Why have this additional assumption that doesn't actually explain anything?
00:41:26.680 It's just another unexplained thing. And two unexplained things do not create an explanation. Anyway,
00:41:32.600 back to the book and Kerigos. Quote, the issue of initial conditions is a serious
00:41:40.280 problem in physics, which has long remained unsolved. There are currently some viable proposals,
00:41:46.600 which constitute the branch of physics called cosmology. These theories incidentally
00:41:52.760 are not even remotely comparable in accuracy and sophistication with other existing successful
00:41:58.600 theories, such as general relativity and quantum theory. They also suffer from the impossibility
00:42:04.120 of testing some of their predictions. The reason is that they are all designed to agree
00:42:09.000 that the universe should look exactly as we see it now. And therefore, they are all confirmed,
00:42:14.360 but what we see now, but it is impossible to discriminate between them by considering their
00:42:19.320 predictions for how the universe should have looked at its origin, because it is impossible
00:42:24.920 to set up tests then. This does not of course mean that there could not be any
00:42:30.520 solution to the problem of initial conditions, but currently we do not have a satisfactory one.
00:42:35.560 We must therefore think of alternative ways of looking at the problem.
00:42:39.560 The science of canon can't provide the way because unlike the traditional conception of physics,
00:42:45.000 it does not rely on initial conditions or laws of motion, as its fundamental primitive elements.
00:42:50.920 When regarded as an explanation for the whole universe, dynamical laws are not self-contained
00:42:56.760 and in another important sense. Imagine for example a collection of pictures of the sphere
00:43:01.000 rolling down Galileo's groove taken in rapid succession, say one every second to cover the whole
00:43:06.840 motion. As we've seen, what a dynamical law does is put them in a particular order. For instance,
00:43:13.000 supposing that the pictures were scattered in front of you, you could use the dynamical law
00:43:17.720 to line them up in a row one after another, according to its prescription. So you would write
00:43:22.920 on the pictures, one, two, three, according to what the law tells you, meaning that time one,
00:43:29.640 the ball is at the top of the slide, at time two, it starts sliding down and so on,
00:43:33.000 until it reaches the end at some time end. So to describe an ordered sequence of snapshots,
00:43:39.240 one has to refer to an external sequence, a sequence of times, say, whose elements are already
00:43:45.400 ordered by labels, one, two, three and so on. We have found, again, an example of infinite
00:43:51.400 regress, the same problem about ordering the scattered snapshots reappears for the sequence of
00:43:57.560 end times we used to order the scattered snapshots and so on. In general, a dynamical law must
00:44:03.560 refer to some external entity, time, which is used to order all the events happening during
00:44:09.960 the motion so that they did not happen all at once. Yet the existence of time has taken as
00:44:14.760 axiomatic and never properly explained in terms of anything else. In addition, recall Galileo's
00:44:19.640 experiment in order to describe the motion of the sphere, he had to time it with a clock,
00:44:24.600 but in the case of the universe, this constitutes a problem. What is the clock to time its evolution?
00:44:31.320 The universe contains everything by definition. There cannot be anything external to it,
00:44:35.400 let alone a clock. These are the two faces of the problem of time, which affects all dynamical
00:44:41.400 laws when regarded as ultimate explanations. Incidentally, this problem also affects laws as formulated
00:44:47.080 in general relativity, where instead of a single external label, time, you have the set of labels
00:44:52.920 specifying a point in spacetime. The same problem presents itself as far as spacetime itself,
00:44:58.760 which is left unexplained. Here, I did not wish to expand on the solutions to this problem.
00:45:03.720 My point is just this. Whatever the solution of this problem may be, it cannot be given in terms
00:45:08.520 of initial conditions and dynamical laws. Otherwise, it falls into infinite address. It must be
00:45:13.800 given in terms of some other kind of explanation, some proposed explanations already exist.
00:45:19.640 If you are interested in reading about them, beautiful accounts are in Julian Barber's
00:45:23.320 Magisterial Treatise, the End of Time, and in Michael Lockwood's intriguing book, The Labyrinth of Time.
00:45:29.960 That Kara goes on to point out that this idea of initial conditions and laws of physics, where
00:45:39.160 given the initial conditions, you can then, given the dynamical laws, predict the trajectory,
00:45:45.720 to predict what it is going to happen to the system evolving at the time, is, in a sense,
00:45:50.760 a kind of bias, because you could easily choose the end of the evolution of the system
00:45:58.360 as your starting point and retrodict everything that happened prior to that. Or, indeed,
00:46:03.240 pick any point on the trajectory given the dynamical laws and be able to predict anything else
00:46:09.240 on that trajectory. This is well known, and we've talked about this before.
00:46:13.960 So I'm skipping that bit where she explains that, and then she goes on to write, quote,
00:46:19.720 dynamical laws cannot handle specific counterfactual features of systems appearing in our universe.
00:46:26.280 They cannot express them fully and adequately. First, there is the kind of counterfactual that
00:46:31.720 declares some transformation to be possible. Consider a specific transformation. Addition.
00:46:39.320 X and Y, two numbers encoded in some numbering system, must be turned into the number,
00:46:45.320 X plus Y. When we try to express the fact that addition is possible in the dynamical law
00:46:51.880 approach, we encounter a number of subtle and important problems. One way to express that addition
00:46:58.440 is possible is to say that an adder is possible. An ideal adder is a machine that, when given
00:47:05.640 any two numbers, X and Y, in input, it gives output X plus Y. And, mind you, it remains unchanged
00:47:14.440 in its ability to do that again with other numbers. The ability to work in a cycle guarantees that
00:47:20.920 the adder can add again if needed. An approximate adder is included in any smartphone calculator.
00:47:27.400 I say approximate, because after a certain number of years, the smartphone's ability to add
00:47:32.600 will wear out. And the precision of addition will deteriorate inevitably. Also, the input will be
00:47:39.640 encoded in a limited number of digits. Hence, achieving only a limited precision. Just pausing
00:47:45.480 that is my quick reflection on this. Yes, computers deteriorate over time. They are subject to all
00:47:51.960 the laws of physics, entropy and so on, just general wear and tear. And so, they will make errors.
00:47:58.280 And as time goes on, they'll make more and more errors. So, there's no such thing as any kind
00:48:02.280 of ideal program that the computer can instantiate. There's no ideal adder. Even adding up
00:48:08.760 simple numbers is not going to be perfectly able to be done by any computer over a sufficient period
00:48:15.480 of time. Anyway, back to the book. Carrot. The possibility of an adder, I just realized that
00:48:22.760 my accent probably makes it seem like I'm talking about the snake. The possibility of an adder
00:48:28.760 cannot be expressed fully if one wants to explain everything in the universe using only laws of
00:48:33.560 motion and given initial conditions as in the traditional conception of physics. For a start by
00:48:38.920 fixing a specific initial condition, the universe evolves along a single particular trajectory.
00:48:45.240 Set by the said initial condition, no ideal adder will ever appear on that trajectory.
00:48:51.240 On that trajectory, there will only be processes implementing approximate adders with limited precision.
00:48:58.920 Where only a fixed finite set of inputs is ever added before the approximate adder wears out.
00:49:05.080 Any particular instance of an approximate adder lasts only a finite time and will only ever add a
00:49:10.920 given sequence of input numbers. Otherwise, we would have a violation of the condition
00:49:15.720 that the laws are no design as I explained in the first chapter. That an adder is possible or
00:49:21.960 that addition is possible means far more than that. First, it means that the adder when given as
00:49:28.280 input to any two numbers can output their sum. This factor refers to any two numbers. Never mind
00:49:34.440 whether they are actually given as input to it in reality once the trajectory is selected.
00:49:39.160 Second, that an adder is possible means that there is no limitation to how well it can be approximated
00:49:44.520 by any of the approximate adders. But this fact, once more, cannot be expressed within the
00:49:49.320 traditional conception of physics because the latter cannot most say what happens to a particular
00:49:55.400 instance of an approximate adder if it occurs on the trajectory of the universe. It is going to be
00:50:00.040 true on each of the possible trajectories for each of the allowed initial conditions. So even the
00:50:04.680 enumeration of all possible scenarios that would happen if a given initial condition were to be set
00:50:10.200 would not express the possibility of an adder either. Another type of counterfactual that cannot
00:50:14.840 be accommodated in the dynamical law approach is the fact that something is impossible.
00:50:20.520 Think of the principle of conservation of energy which tells us that a perpetual motion
00:50:25.080 machine is impossible. In the dynamical law type of approach, one can only say that a
00:50:30.200 perpetual motion machine does not happen. That means that no point on the trajectory of the universe
00:50:36.360 contains one given a particular initial condition. But that a perpetual motion machine is impossible,
00:50:43.320 does not mean it does not happen under a particular initial condition. It means that it cannot be
00:50:47.640 built under any of the initial conditions and any of the actual dynamical laws. This statement is
00:50:53.720 much more powerful and categorical than any of the statements one can make about what happens on
00:50:58.360 a particular trajectory. Okay, I'm just skipping a little and then carrier goes on to write.
00:51:03.960 A final problem about the dynamical law approach is that it seems on the face of it to conflict
00:51:09.720 with the existence of entities that are capable of making choices, such as you and me,
00:51:14.760 every omniscient narrator knows this. The omniscient narrator is the entity that tells the story
00:51:20.440 in a novel in the third person. The narrator knows all the thoughts of the characters in advance.
00:51:26.040 Their ideas are set from the very beginning of the novel. Choices only look like true choices to the
00:51:31.640 characters, but in fact they are predetermined and fixed by the narrator's plans. Likewise,
00:51:37.560 the explanation based in terms of motion and initial conditions would seem to imply that so must be
00:51:43.240 hours. Our choices and everything else depending on them seem already to have been set in advance.
00:51:49.960 They are written in the dynamical law explanation and fixed by the initial condition of the
00:51:54.440 universe. The dynamical laws sequence of events fixes everything. It is given once and for all.
00:52:00.840 All your ideas are laid out there. There seems to be no room for them to be unpredictable as
00:52:06.200 they should be if they were truly free choices. We have just outlined what is called a
00:52:12.440 deterministic nightmare. The fact that there does not seem to be any room for free choice.
00:52:18.200 If one presupposes the existence of a fixed predetermined story for our universe,
00:52:24.280 which is the picture that the traditional conception of physics, in terms of dynamical laws and
00:52:28.520 initial conditions, seems to suggest. For example, whether tomorrow you will have croissants or
00:52:34.360 keepers for breakfast, has been fixed at the start of the universe in its initial conditions.
00:52:39.320 The same goes for the fact that you are reading this text right now instead of some other book,
00:52:44.440 or maybe doing something else altogether, such as watching your favourite show. All determined
00:52:49.320 at the beginning of the universe down to the precise words typos included. Unpredictability of action,
00:52:56.200 or free will, is therefore another counterfactual that the dynamical law approach does not
00:53:01.560 seem to be able to accommodate. We did not yet know how to accommodate exactly, free willing physics.
00:53:07.640 But that only means we have to think harder. This problem exists, but it is not insoluble.
00:53:13.640 It only appears to be so if contemplated from the narrow dynamical law type of approach,
00:53:19.720 pausing their my reflection, what will exactly? And then I would go one step further and say
00:53:25.640 only to some people. I've always kind of thought that even if, okay, here's a two until,
00:53:35.480 you know, coming to understand construct a theory a little better,
00:53:38.280 that even if we had this dynamical law approach to physics, even if that was true and
00:53:44.120 that was the only way to consider physics, we could still have free will. I'm kind of with
00:53:48.360 Daniel Denop for different reasons. Daniel Denop is something called a compatibleist.
00:53:52.280 I'm a compatibleist. A lot of people say, well, compatibleism is irrational, given we know that
00:53:59.160 the laws of physics determine everything that's going to happen. But do the does that mean that
00:54:07.080 everything is reducible to laws of physics? For example, the evolution of species. The evolution
00:54:12.200 of species happens because evolution by natural selection actually creates a certain kind of knowledge,
00:54:19.240 the knowledge of how genes can survive in certain organisms, in certain niches,
00:54:26.760 to enable to create different organisms and so on and so forth. In other words, I endorse a certain
00:54:32.920 kind of creativity, an emergent simplicity. And I think that free will is just a way of talking
00:54:39.880 about this emergent kind of simplicity. The emergence of simplicity in this case is the existence
00:54:45.400 of choices that exist in the world. Now, I think in defy the laws of physics, but at the same time,
00:54:50.840 the laws of physics don't explain everything. And what we're after when it comes to human
00:54:54.760 behavior is explanations, not merely predictions, that even if you could predict people's behavior,
00:55:00.920 you wouldn't have an explanation of their behavior. The explanation of the behavior would come
00:55:04.440 down to their personal creativity. And all of this is somewhere you have to do some linguistic
00:55:10.040 gymnastics in order to avoid this term free will. And if you do avoid the term free will,
00:55:14.280 you're still laughed with this deep, deep mystery. I've talked about this ad nauseum. I think
00:55:19.560 that free will is tied intimately to knowledge creation. Therefore, it's got something to do with
00:55:25.000 epistemology. And now, you know, I'm hoping that it appears we'll have a clearer understanding
00:55:30.680 of it, given constructed theory, given the science of can and can't. And what we're saying here
00:55:36.680 is that if we want to understand this more, why don't I understand this mystery of what a human is,
00:55:41.400 making choices, creating knowledge, having free will, we need to move beyond dynamical laws. That's
00:55:47.400 holding us back. It's leading to a lot of what I would say are misconceptions. Anyway,
00:55:52.520 Kyarrigos, I'm going to say, quote, fortunately, the dynamical law approach is not the only way
00:55:57.240 to provide explanations and predictions. Why should all good explanations look like chronologically
00:56:03.080 ordered stories which unfold from beginning to end? The fact that something has happened before
00:56:08.120 something else need not be the whole explanation for how systems work in physical reality.
00:56:13.800 End quote. The next part of the book is a many pages that I'm going to skip. It's about chess
00:56:21.960 and basically Kyarrigos is the analogy to chess to talk about how certain moves are possible
00:56:29.400 or impossible. And this can help to explain the evolution of the time of the game, rather than
00:56:36.120 simply understanding what will happen given what you saw happen during the game. Understanding
00:56:42.920 chess, in other words, means understanding that pawns can do this and bishops can do that,
00:56:47.800 but not these other moves, for example. Okay, back to the book. Kyarrigos.
00:56:53.160 Can this logic adopting counterfactuals be fruitful in physics other than in my elementary example?
00:56:59.320 Indeed, in physics and science in general, we already resort to modes of explanation
00:57:03.560 other than dynamical laws. Some of them adopting counterfactuals. What we have in place is, in fact,
00:57:09.480 a hybrid approach. For example, physics resorts to principles like the conservation of energy,
00:57:15.160 which, as I said, are about counterfactuals too. These principles are not in the form of dynamical
00:57:20.600 laws. They are statements that require certain things to be impossible, such as perpetual motion
00:57:25.560 machines, yet they can be as powerful as dynamical laws at generating predictions. A famous example
00:57:31.880 is the prediction of the existence of the neutrino, a previously unknown elementary particle.
00:57:36.440 This prediction is akin to the prediction of Neptune's existence, but this time,
00:57:40.440 it was a prediction of the existence of a subatomic particle, not a planet,
00:57:43.880 and the prediction was generated from a principle, not a dynamical law.
00:57:47.960 The prediction was obtained by reasoning that without that particle, the law of conservation
00:57:53.400 of energy, would be violated. It couldn't have been obtained from a dynamical law,
00:57:58.760 because the laws of motion of neutrinos were not known until much later.
00:58:03.400 Principles also appear in Newton's laws. Only the second law is a dynamical law,
00:58:07.800 relating the force of a system to acceleration and mass. But the other two laws are not really
00:58:12.120 dynamical in their existing formulation. The first law is a hybrid one. It says that it is
00:58:16.760 impossible for a system to change its state of motion when it is not acted upon by a force.
00:58:21.880 Hence, the system will continue in its given state of motion until some force intervenes.
00:58:26.120 The law does refer to dynamical laws via the concept of state of motion. But it mandates,
00:58:31.640 like a principle, that some transformations are impossible, specifically those changing the state
00:58:37.240 of motion of a system without it being acted on by a force. The third law is even closer to
00:58:42.040 being a pure principle. Informally, it requires to every action they must correspond an opposite,
00:58:48.120 indirection, and equal in magnitude, reaction. If, while on a walk in the park, your dog is pulling
00:58:53.640 you ahead via the leash, you are pulling the dog back with an equal and opposite pull.
00:58:57.400 This fact is necessitated by the principle included in the third law, not by a specific dynamical
00:59:02.920 law. Wouldn't it be wonderful if it were possible to take inspiration from these principles,
00:59:07.560 which relate to counterfactuals, and imagine an entirely different way to formulate the laws of
00:59:12.120 physics, one that takes counterfactuals as primitive, and the laws of motion and initial conditions
00:59:16.920 as derivative, one could even conjecture that this new mode might solve the open problems in the
00:59:22.760 dynamical law approach, as well as fill in the gaps in existing theories, while still covering
00:59:27.880 all their predictions, the kind of explanation I am imagining, which is that provided by the
00:59:32.760 science of canon Kant, is even more radical than the hybrid type of explanation which we
00:59:36.680 currently use in physics, it places counterfactuals at the most fundamental level. Then it
00:59:41.640 explains dynamical laws and initial conditions in terms of them. These laws of motion can then be
00:59:48.040 used to make testable predictions about cannonballs and electrons, much as they are now,
00:59:51.880 but their underlying explanations would be in terms of principles about counterfactuals,
00:59:56.200 and this could provide a solution to the infinite regress type of problems I mentioned earlier,
1:00:00.840 e.g. resorting to an infinite set of initial conditions. Just as the counterfactual properties of the
1:00:06.040 chess piece about what moves are possible and what are impossible, can explain draws on a chess board,
1:00:12.360 so counterfactual properties can explain why the universe is in a certain state, avoiding
1:00:17.320 mention of the initial conditions altogether, both statements about possibility and impossibility,
1:00:22.440 are equally important. You will see several examples of laws about possibility and
1:00:26.040 impossibility in the chapters to come. These may seem like bold speculations, and they are,
1:00:31.800 the first time I encountered the idea of reformulating physics with counterfactuals was in a
1:00:36.200 proposal by the physicist David Deutsch. At the time I thought it was fascinating, but crazy,
1:00:41.480 that was during my doctoral studies and Oxford when, to put it as Alice in Wonderland would,
1:00:46.520 I started trying to imagine as many as six impossible things before breakfast. That idea was one
1:00:52.040 of them, but within a few months David and I were working together on a paper developing this idea
1:00:57.560 and applying it to information theory, and after my doctorate I decided to focus completely
1:01:02.600 on pushing it further to try to address various unsolved problems in physics. By then I was convinced
1:01:07.720 its promise was enormous. My research today would the help of a few brave students and a handful
1:01:12.840 of other physicists has concentrated on putting this approach to the test. In the following chapters,
1:01:17.880 I shall explore the problems that this approach has solved so far, and it's potential to solve
1:01:22.600 further problems. It's time to journey deeper into the land of counterfactuals. End quote,
1:01:28.120 end of chapter two. That's fantastic. So this is a really optimistic, positive vision for the
1:01:33.240 future of physics in the future of science more broadly. I can't wait to continue reading this,
1:01:38.120 but this has been a longish episode if my, if my time is correct. And so I'll finish it up there
1:01:45.480 today, and look forward to seeing you in an episode four of my readings and reflections on
1:01:51.880 the science of canon cards. By Kia or Mallet, based upon work initiated by David Dornish. Until