00:00:30.000 And what you're listening to there is Swan Lake by Trikodsky.
00:00:40.400 Welcome to Topcast episode 121, Kyarama Leto in the Science of Canon Can't Chapter 6, titled
00:00:52.360 By recounting how as a child, she had a music box, which one would wind up, and a little
00:01:05.400 No as a child, I didn't personally have a music box like this as such, but my mother did,
00:01:15.040 You wound it up, and a little dancing ballerina danced to a tune.
00:01:19.840 And it played, I think, a different tune, not exactly Swan Lake.
00:01:24.240 Anyway, at the beginning of chapter 6, Kyarama explains how this mechanism works.
00:01:30.640 By turning a key, winding the thing up, the box is charged up, so to speak, with mechanical
00:01:37.760 Energy stored in the elasticity of a spring-like mechanism that slowly does the work
00:01:46.480 Today I'm doing lots more reading than normal, I guess, because the idea is in this chapter,
00:01:56.640 From what I said in episode 119, all about the thermodynamics that everyone already knows
00:02:05.960 So I'll begin the reading today with the introduction to this chapter that Kyarama has
00:02:12.760 And she writes, quote, in this chapter is, where I discuss the conservation of energy
00:02:18.360 as a counterfactual principle about impossibility, three different kinds of irreversibility
00:02:24.000 in physics, statistical, forgetful, and counterfactual, where I provide a counterfactual
00:02:29.920 second law based on an exact distinction between work and heat, and where you encounter
00:02:35.000 the universal constructor, a machine that can perform all transformations that are physically
00:02:43.080 Now in my readings here for talkhouse, I'm not going to do all that.
00:02:46.240 In particular, I'm not going to look at the three different kinds of irreversibility
00:02:50.800 and statistics, in particular, I'm not going to look at the forgetful kind.
00:02:54.880 Now as I'll be doing a lot of reading today, I'm trying to cut out as much of the chapter
00:03:00.040 And so as I say, she begins the chapter, Kyarama begins the chapter by explaining the
00:03:04.820 mechanical operation of the ballerina in a music box and the sound coming out of the
00:03:10.640 So I'm skipping the basics of that, and I'm going to pick it up where she writes, quote,
00:03:18.440 At first sight, it comes from the user who turns the winding key.
00:03:22.120 However, if you go a little deeper, this seems to be the start of an infinite regress,
00:03:26.800 which has happened repeatedly in this book as signaling a problem in the traditional
00:03:33.240 There is no end in sight if you go down this line of inquiry because you could ask the
00:03:37.000 same question of the motion of the user's hand, where does that originate?
00:03:41.440 Also innumerable other mechanisms could equally well wind up the box.
00:03:46.080 For instance, a small mechanical engine attached to the winding key could do as well
00:03:51.400 To understand, you need to ask a more fruitful question.
00:03:57.840 What kind of stuff powers the music box when you wind it up?
00:04:02.360 The stuff that charges the box is what physicists call energy.
00:04:06.280 The term comes from a Greek word, popularised by Aristotle, whose original meaning and
00:04:14.520 Nowadays, what we call energy in physics is even more sharply defined than in Aristotle's
00:04:20.480 It is an abstract property of physical systems that must be subject to substantial constraints.
00:04:26.200 As we know, the most important of these constraints is the law of conservation of energy,
00:04:30.960 requiring that the energy of the system can be changed only at the cost of changing the
00:04:35.960 energy of some other system by the same amount.
00:04:39.560 The law of conservation of energy is about counterfactuals, for it requires that it be impossible
00:04:45.160 to change the energy of the system without any side effect.
00:04:49.040 Given that all laws of motion must conform to the conservation of energy, those already
00:04:53.760 known, and those yet to be known, the conservation of energy is more general than any specific
00:05:01.240 Also, it is intended to apply to any system in the universe.
00:05:05.480 It rules miniscule particles such as electrons and protons, heat engines that propel
00:05:10.120 aircrafts and spaceships, and the mitochondria powering our cells.
00:05:14.800 It applies to anything and everything that has an energy independent of its scale and
00:05:26.760 This idea of scale independence is really, really important here.
00:05:31.960 What it means is that the law, scale independence, means something applies whether the
00:05:36.160 thing is small or big, independent of its scale, how much bigger you make it, whether
00:05:42.240 it's a single atom or a conglomeration of atoms.
00:05:45.720 It's not like laws of physics should apply to big things and not small things, although
00:05:51.560 this is a misconception that sometimes gets around.
00:05:53.760 The laws of quantum theory should apply everywhere no matter the size of things.
00:05:58.720 And in reverse, I just speak, so the laws of thermodynamics, which are found to apply to
00:06:03.960 heat engines, big things, should also apply to small things as well.
00:06:08.680 In both cases, quantum theory and thermodynamics, it must be the case that the laws are
00:06:20.200 But our imperfect knowledge of them seems to suggest, at first glance, that the quantum
00:06:26.160 laws apply to little things and thermodynamic laws apply to big things.
00:06:30.160 Well, not always, but the idea here is that this is what's going on in the mind of some
00:06:36.120 And by the way, it's also why you have people like David Wallace who write about the
00:06:40.120 emergent multiverse, the idea that these laws of quantum theory do indeed apply to large
00:06:45.880 ensembles of particles, indeed the entire universe.
00:06:50.800 But you hear other popular answers and other physicists talk about how well quantum theory
00:06:55.360 are just the laws that apply to the very small.
00:07:01.760 It's just that from our perspective, certain things appear to be the case, but appearances
00:07:08.600 But if things do appear in some way, and that's a problem, which it is here, then we should
00:07:13.040 seek to solve the problem, in the case of the second law, that's certainly a problem
00:07:17.480 because it appears to apply to big things, but at the small scale, at the level of individual
00:07:22.840 atoms, the laws appear to be reversible, that the second law is about irreversibility,
00:07:28.120 which is what Chiara is going to come to in this chapter, and what we're going to talk
00:07:31.640 But just here, in what I've just read, Chiara has mentioned the first law, never mind the
00:07:37.360 And on this, I have a stop at press, which has never happened before.
00:07:41.720 Older listeners will know what stop press means, stop press being what used to happen
00:07:47.400 They'd print the physical newspapers and send them out all over to the news agents and
00:07:52.080 shops and the paper boys who delivered them, always paper boys, never paper girls.
00:07:56.440 And if partway through the printing of the newspapers, something of note happened, they'd
00:08:07.160 On the last page of newspapers in Sydney anyway, it read stop press.
00:08:12.600 And beneath it, there was a space to insert news that appeared during the press, during
00:08:18.480 I think it was done in red ink, if I remember, silly diversion there.
00:08:22.640 My stop press for this is that as I was making this very episode, Chiara Malito published
00:08:28.680 an official paper about how constructive theory can be used to illuminate a connection
00:08:33.880 between information and the first law of thermodynamics.
00:08:37.280 Now, this is in addition to worksheet already done, and what this chapter is primarily
00:08:41.960 about, about how the second law can be better understood given constructive theory and
00:08:46.680 underlines the known link between working information, given constructive theory.
00:08:51.560 Now, I'm just going to mentally check off the fact that that paper exists now.
00:09:02.080 I cannot turn it into a succinct few sentences here for public digest, I suppose.
00:09:08.280 I'll take some work and perhaps I can leave that for an actual discussion with Chiara
00:09:13.840 herself, or I can make an additional episode all about that, but it may just get a little
00:09:19.560 This is going to get technical in places enough, so we'll just mentally check off the
00:09:23.800 fact that it has been yet another paper published on the 27th of May, 2022, in the Journal
00:09:30.560 of Physics Communications titled the Information Theoretic Foundation of Thermodynamic Work
00:09:36.840 Extraction, and has their statements about the first as well as the second law.
00:09:42.520 Today, we're focusing on the second law, but just to say, there is this new paper out
00:09:48.000 there all about work information, the first law, and constructive theory back to this
00:09:53.680 episode, where we've just talked about how the first law applies to everything and
00:10:01.640 Chiara goes on to say, quote, this seemingly innocuous requirement has sweeping consequences
00:10:06.720 – first it implies that the energy of a system cannot increase or decrease, unless energy
00:10:11.440 is supplied from or absorbed by something else.
00:10:14.960 This allows us to account minutely for the whereabouts of energy as it transits from one
00:10:19.600 system to another, when it escapes from here it must have gone over there, it can't disappear
00:10:23.960 or pop into existence, for example, the music box cannot get charged spontaneously without
00:10:31.120 The winding of the key, likewise the battery of a smartphone does not get replenished
00:10:35.680 of the device, it switched off and not connected to the power supply, okay, I'm skipping
00:10:40.400 a number of chapters, picking it up where, she writes, quote, the conservation of energy
00:10:44.560 implies the startling prediction, that at each step along the chain, if one accounts
00:10:50.000 for all the systems involved, one shall find, overall, the same amount of energy initially
00:10:55.880 given to the winding key, such as the inescapable accountancy set by the law of conservation
00:11:02.320 of energy, and it is based on the counterfactual property that it is impossible to change
00:11:07.640 the energy of a system without side effects being produced, end quote.
00:11:12.520 Yes, so here we have the winding of the key, which sets the music box into motion, and
00:11:18.080 the energy that's there at the beginning, from winding the key, stored in the winding
00:11:23.000 mechanism, a spring of some type, is then transferred into work which moves the ballerina
00:11:28.720 around and work which sends sound waves out into the atmosphere, and also some heat to
00:11:33.160 produce as well, but all of the energy that's produced was there originally in the winding
00:11:37.840 of the key, again, skipping of picking it up where, she writes, quote, the second law distinguishes
00:11:43.440 between two types of energy transfers, and the distinction is rooted in counterfactuals.
00:11:50.280 One type of transfer is reversible, one can use these energy transfers to perform work
00:11:55.280 on a variety of physical systems, such as the brass winding key of the music box or a flywheel
00:12:00.760 or a piston, and then undo the transfer completely, retrieving the energy in full, with
00:12:08.840 The systems supporting these energy transfers are fully interoperable, just like the systems
00:12:16.440 The other type of energy transfer is irreversible.
00:12:20.000 Once the transfer happens, it cannot be fully undone, part of the capacity to do
00:12:30.280 When you break your plier resistance against the rotation of the wheel, the wheel and
00:12:33.880 brake initially have a certain energy and they come to a stop.
00:12:37.400 The brakes and the wheel itself have heated up.
00:12:41.200 The energy that was in the motion of the wheel and the bike has now gone into the thermal
00:12:45.320 motion of the molecules, composing the wheel and the brakes, and it is practically
00:12:53.800 The same holds for the energy of the vibrations, bringing the tune of the music box all the
00:12:58.040 way to our ears once the music is heard is very hard to bring that energy back into
00:13:04.480 The chemist, Peter Atkins, has frequently said in his masterly books about the foundations
00:13:09.240 of thermodynamics that work and heat are not substances, just as information is not a substance.
00:13:22.880 The transferred energy that can be reused at infinitum to initiate or to stop controlled
00:13:33.960 The second law requires some energy transfers to be heat like once they happen.
00:13:38.720 It is impossible to recycle some of the energy involved in them.
00:13:43.320 That energy can no longer be used fully for a work like transfer, only some of it can
00:13:51.200 So this distinction between work like and heat like is a key of this constructor theoretic
00:13:59.440 The constructor theoretic view is obviously also catch in terms of this idea of a construct
00:14:05.680 The second law is about the fact that there is no object possible.
00:14:09.080 We are going to come to this, no object possible, no constructor.
00:14:16.080 We can't capture the lost heat, for example, there is no constructor, the one can build
00:14:20.520 or that can be built, that could enable the recapture of that heat so that useful work
00:14:27.120 Because there is no physical transformation possible.
00:14:29.400 So no constructor can be built, so therefore there's no physical transformation possible.
00:14:35.880 Why is it that there's no constructor able to do this?
00:14:41.920 Well, that's just the way things are at the whole point of the laws of physics.
00:14:45.160 We are now in a world where we're asking if you're asking that question, why can't
00:14:48.560 we have photons traveling below the speed of light?
00:14:54.640 Now anyone who listened into episode 119, where I went through an introduction to thermodynamics
00:15:00.160 will recognize that name, Peter Atkins, one of the most famous thermodynamicsists, physical
00:15:05.960 chemists of our age, who's written a lot of the classic texts on exactly this stuff.
00:15:12.320 It was interesting to hear Keir actually, and one of her interviews she did with Logan
00:15:16.320 Chipkin, and if you can find those out there, look up Logan, he's on Twitter, and just
00:15:21.800 Google his name, Logan Chipkin, and he has interviewed Keir.
00:15:26.080 And they talked there, well, Keir, I mentioned, brought up about how Peter Atkins himself
00:15:31.000 has jokingly credited Keir with the discovery of a new kind of entity, information, E-N-F-O-R-M-A-T-I-O-N.
00:15:41.880 Working with the properties of both energy and information.
00:15:46.200 That is the stuff that can be stored, so to speak, on work media, which you're going
00:15:51.440 Now I'm not a physical chemist, and I'm not a theoretical physicist.
00:15:55.240 So I don't know how deep that joke runs, how seriously we can take that joke.
00:16:00.480 But whether one day, there might actually be this thing called information, I'm not
00:16:06.680 Anyway, I'll get back to the book, and Keir arrives.
00:16:09.360 Quote, the second law of thermodynamics operates in a rather impressive way.
00:16:14.560 In tandem with the principle of the conservation of energy, it provided the theoretical
00:16:18.600 foundation for heat engines, which powered the incredible progress that occurred during the
00:16:25.000 But when it comes to explaining exactly what the second law says about the physical world,
00:16:29.800 the issue is not as clear as for the conservation of energy.
00:16:33.480 It is so complicated and subtle that physicists over the decades have proposed numerous
00:16:42.440 Each has its notion of heat like and work like transfers of energy, and they are different
00:16:48.160 Still all these different formulations concur on a few striking consequences concerning
00:16:55.560 The second law is thus a pillar of the edifice of theoretical physics.
00:17:05.560 This is why popular science in particular is so enamored with the second law.
00:17:10.360 I'd be hard pressed to find one of Paul Davies books that doesn't mention it.
00:17:15.160 In fact, I'd be hard pressed to find any popular science book in the general sphere
00:17:23.240 Physicists love discussing the second law because it has this mysterious characteristic
00:17:33.720 It seems like the processes that the second law governs happen one way, but there's no
00:17:40.000 time-symmetric way of reversing them, simply in the other direction, although it should
00:17:44.120 be, because every other law, or dynamical laws in particular, do have this time-symmetric
00:17:50.880 They run perfectly well forward in time, as back in time, and if you took a video, as
00:17:55.280 I said, of the process that was going on, it was, you know, subatomic, no physicist would
00:18:00.520 be able to look at that video and say, oh, that's the forward-time direction, that's
00:18:04.280 a reverse-time direction, it's impossible to tell.
00:18:08.800 Well, this is the whole thing about the second law.
00:18:10.200 Maybe there's something hidden, some information hidden there, that tells you which way
00:18:16.320 In particular, maybe thinking about these things in terms of dynamical laws is not the
00:18:22.280 Okay, skipping a few paragraphs, and I'll pick it up where, you're right, it's
00:18:27.600 The problem is that the second law requires some irreversibility.
00:18:31.160 Incidentally, that is also why it is so fascinating.
00:18:34.400 irreversibility is that the core of various phenomena that are ubiquitous in physical reality,
00:18:39.360 the birth, development, and death of organisms, the growth of complexity in the biosphere,
00:18:43.560 the increase in sophistication within our civilization, the creation and destruction of knowledge,
00:18:48.280 the irreversibility requirement of the second law, brutally clashes with the laws of motion,
00:18:57.240 Remember, I sit in Chapter 4 that the laws of quantum theory are reversible, if they
00:19:01.920 allow for a transformation, the reverse transformation must also be possible.
00:19:07.680 The laws of general relativity, the other most accurate description of physical reality
00:19:13.120 If there is a trajectory that takes the system from A to B, there must also be one that
00:19:19.360 microscopic constituents of matter must operate in this reversible manner because they obey
00:19:25.400 The problem then is how can the second law require that some energy transfers are irreversible
00:19:31.320 and be compatible with the reversibility of the laws of motion?
00:19:36.280 And I'm skipping apart, and she goes on to say, is there a unique picture of physical
00:19:41.520 reality that can reconcile reversibility and irreversibility?
00:19:46.080 This does not yet have a definite answer to this question, let alone a unique one.
00:19:50.520 There are a few proposed answers, but each is still controversial, ultimately it is because
00:19:55.600 theoretical physics is trapped in a world without counterfactuals.
00:20:00.200 With counterfactuals, one can reconcile the reversible description of the laws of motion,
00:20:08.880 To see how I shall first take a closer look at the irreversibility of heat light transfers,
00:20:15.360 In that playground, a sea saw consists of four main systems.
00:20:19.680 Two seats, for approximately equal mass, a rigid, long sturdy bar joining them, slaughtered
00:20:27.400 We need two children to play, a child sits at each end so that one goes up as the other
00:20:33.480 By gravity, depending upon which child weighs more, added fun comes if the children are
00:20:37.800 tempted to go as high in the air as possible as they take turns pushing their feet
00:20:45.320 Just imagine that instead of the children, there are just two springs firmly secured
00:20:51.960 Also imagine that both ends, call them A and B, have the same mass.
00:20:56.040 The sea saw in the neutral position corresponds to the bar being perfectly horizontal above
00:21:02.800 Now imagine you push one of the ends, say A in the upwards direction, up it goes, while
00:21:06.840 the other end B goes down and compresses the spring as much as allowed by the conservation
00:21:12.600 Energy given by your push is now transferred entirely to the spring.
00:21:16.400 Then the compressed spring gets decompressed by its completely elastic nature, thus giving
00:21:21.120 back the energy to the B end of the sea saw which therefore goes up and A goes down, compresses
00:21:31.240 Cara goes on to explain over the next few paragraphs about how, well, you know, if there's
00:21:37.000 no other effects going on, then this just goes on forever.
00:21:41.280 You have a perpetual motion machine of a kind, in theory, but in reality, we know that
00:21:46.680 these things come to a stop we have what's called a damped oscillator.
00:21:52.600 Friction gets involved, but what is this friction?
00:21:59.520 I see so it springs it either end, not just keep on going on forever.
00:22:03.040 Well, Cara goes on to explain, quote, for a start there are countless molecules of air
00:22:07.080 continually bouncing off the bar as it goes up and down.
00:22:10.360 When the spring is also not perfectly elastic, the energy transmitted to it by the bar
00:22:15.040 is not entirely returned when the spring gets decompressed.
00:22:18.200 Some energy goes to waste away from the sea saw oscillations, for example, it is absorbed
00:22:22.320 by the atoms of the spring, which increase their internal vibrational energy.
00:22:26.600 So there are several more interactions to take into account to explain where the energy
00:22:34.080 All these interactions take a little energy from the combined motion of the sea saw
00:22:39.720 That is why, if a real life sea saw is set into motion and then left alone, it eventually
00:22:48.760 When the sea saw comes to a stop, the energy given by the push is stored in the air molecules
00:22:53.200 and in other particles inside the spring and the bar, there are all a little warmer, more
00:22:59.920 Here is the point where irreversibility creeps in, in a world where all elementary interactions
00:23:06.240 Once the energy has gone into molecules of air and vibrational motions of atoms, it becomes
00:23:15.240 Such kinds of energy transfers are those the second law labels as heat-like, generally
00:23:21.320 regarded as irreversible, just like in the case of breaks or the music box.
00:23:27.800 Okay, I am skipping a number of paragraphs where Chyara talks about the status of the second
00:23:35.880 And I will just read one of the important paragraphs here as to what she says about
00:23:43.280 If reversing heat-like energy transfers were just very hard to achieve, but ultimately
00:23:48.760 possible, the second law would not really be a fundamental law.
00:23:53.080 In fact, there would be no reason for physics to distinguish between work-like and
00:24:07.040 Only a little harder to achieve in the reverse direction compared with the forward direction.
00:24:11.080 The second law and its prescribed irreversibility would just be the description of our current
00:24:16.120 technological limitations, which of course are not fundamental.
00:24:20.880 These limitations can be improved upon by investing enough resources into it.
00:24:26.120 On this question of whether the second law is fundamental, physicists are currently divided
00:24:33.840 There are only reversible laws governing the microscopic interactions of particles.
00:24:38.680 With enough technological resources, the reversible dynamics could always bring all energy
00:24:43.360 back to where it came from, and one could then reuse the energy to do work.
00:24:47.520 This would imply that the limitations imposed by the second law on heat engines are just
00:24:51.960 a rule of thumb, telling us that it is hard in practice to reverse specific interactions.
00:24:57.320 But these limitations could be lifted ultimately by improving our technology.
00:25:02.040 According to this camp, the second law does not need to appear in the manual for the
00:25:07.600 The other camp claims that the second law is fundamental, that there can be a formulation
00:25:12.320 of the second law that is universally true and still compatible with the reversible dynamics
00:25:20.360 Various paths to reconciling irreversibility with reversible laws of motion have been proposed
00:25:24.960 to support this idea, none of them really works to the end of creating a universal exact
00:25:31.000 They ultimately all concede that irreversibility only appears as some sort of approximation,
00:25:37.240 Only one of these paths is based on counterfactuals, as I shall explain.
00:25:41.080 It is the only one that has some potential to be successful in this endeavor.
00:25:44.960 I shall go into a little detail about these paths because they are smart ideas, even if
00:25:49.280 they end up with mixed success in regards to producing an exact second law, understanding
00:25:54.560 the other approaches is indispensable to grasping the superiority of the approach with
00:26:04.880 I might just mention the section where Kiara talks about why she doesn't talk much
00:26:12.480 Entropy doesn't come up in the constructive theoretic view of the second law.
00:26:17.640 And the reason is because it's a statistical law.
00:26:20.640 It's kind of an approximation when we talk about entropy.
00:26:24.000 As Kiara says, quote, the statistical mechanical law cannot even aspire to be exact or universal.
00:26:30.960 The configuration maximizing the entropy is not the only one guaranteed to occur.
00:26:40.360 All other configurations can still occur, but it is not said when and how, only that
00:26:47.040 The fundamental reason why the statistical second law cannot be exact is that the dynamics
00:26:52.280 regulating the exchange of energy between, say, iced tea and the surroundings are reversible.
00:27:04.440 And what the second law says phrased in this way is that disorder tends to increase in
00:27:12.760 It's the most probable as Kiara would put it there.
00:27:15.320 The most probable way in which the universe evolves is that it evolves towards increased
00:27:25.600 But with laws of physics, you want a guarantee.
00:27:28.680 These things are incontrovertible, these things tell you what will happen, not what will
00:27:34.640 And we've talked about that on topcast before when we've broken down David's talk on
00:27:41.160 So I refer people to that episode, which if you're interested is episode 111 of
00:27:47.180 the talkcast, where I go into David's talk on probability, where specifically this kind
00:27:51.920 of issue arises, you know, the laws of physics tell you what actually happens, but I don't
00:27:57.200 And so this is one of the reasons why one of the motivations for this entire constructor
00:28:01.380 theoretic view of the second law, because other versions of the second law involving,
00:28:05.600 for example, entropy are about what probably happens to ensembles of particles and to
00:28:12.320 But we want to know what will happen, not probably.
00:28:15.360 So not only does Chiara talk about why that doesn't work, that statistical view of the second
00:28:24.360 So I'm skipping all of that, quite a number of pages here.
00:28:28.160 Instead I'll pick it up where Chiara writes, quote, just like Goldilocks in the three bears
00:28:32.440 house, after trying two paths that do not work, we land on the third path to irreversibility
00:28:40.360 This path is not just right, it too has problems, but it is more promising than the other
00:28:45.800 two, which are instead based on the traditional conception of physics.
00:28:49.800 To set off down this path, let's trace the same conceptual steps that led the superb
00:28:54.800 physicist, James Jewell, to perform a crucial experiment in the early days of thermodynamics.
00:29:00.280 What he conjectured and verified experimentally was that while it is possible to heat up a
00:29:05.320 volume of water by stirring it only mechanically, it is impossible to cool it down, by
00:29:10.320 those same means, you can see that I am now talking about counterfactuals, the language
00:29:14.560 regarding possible, impossible transformations belongs to the science of Canon Kant.
00:29:19.280 Let's return to a glass of iced tea, Jewell would have preferred a glass of beer, as
00:29:24.200 he was also a brewer, but I shall stick with the tea, it can't do any harm.
00:29:29.200 Imagine you stir the tea vigorously, mechanically, say with a spoon, this stirring provides
00:29:33.840 the molecules of water with more energy, imagine that the glass is somehow perfectly
00:29:38.360 isolated from the rest of the environment, and no energy.
00:29:41.400 Other than the energy of the stirring can be exchanged with the environment.
00:29:44.920 What you will find out is that the tea in the glass ends up in a hotter state at the
00:29:48.880 end of the stirring, on the other hand, no matter how hard one tries, the temperature
00:29:57.120 In this scenario, that transformation is impossible.
00:30:00.400 Of course, if the glass is not isolated, you can stir the tea to cool it down by facilitating
00:30:04.920 exchanges with the air in the environment, however, here I imagine the cup to be entirely
00:30:11.320 This kind of irreversibility prescribes that sometimes a transformation, such as heating
00:30:16.360 some amount of water, is possible by mechanical means only, I using the stirrer.
00:30:22.640 But the reversed task is not possible, by using those same means, though it may be possible
00:30:29.560 Work like energy transfers are those corresponding to transformations that can be performed
00:30:34.400 by mechanical means only, in both directions, pausing their myreflection just going back,
00:30:40.280 it's worth saying that again, work like energy transfers are those corresponding to transformations
00:30:48.880 that can be performed by mechanical means only in both directions, and going on.
00:30:56.520 Heats like transfers correspond to transformations that are possible by mechanical means
00:31:00.520 in one direction only, but I impossible in reverse using the same means and nothing else.
00:31:08.240 As you can see, this path to irreversibility is about possibility of certain transformations
00:31:19.480 This approach to the second law is due to Lord Kelvin and Max Planck.
00:31:22.800 It does not talk about the most probable free evolution of a physical system in contact
00:31:27.600 with an environment or about what trajectories a system access is once you discount some
00:31:33.120 of its details, end quote, just pausing their myreflection, yes, so on this experiment
00:31:37.760 with jewel and stirring and so on and so forth.
00:31:44.040 If you are unsure of any of these details and you want to hear the classic way in which
00:31:48.880 some of these things are described to see what we're getting out here and the importance
00:31:54.120 Just go back to that episode and you'll find me discussing jewel's experiment then,
00:31:58.680 the high school version of this kind of thing, the high school explanation of this kind
00:32:03.120 Here we're emphasizing Kelvin and Plunk's invocation of what can be called, a version
00:32:09.120 This idea, we've got possible and impossible stuff going on, but the work like means
00:32:14.680 that it's possible in both directions, but the heat like it's possible in one direction,
00:32:23.680 And the fascinating revelation is that this kind of counterfactual irreversibility is compatible
00:32:30.040 with time reversal symmetric laws without requiring any approximation because even under
00:32:35.560 perfectly reversible microscopic laws, it is possible to have some device that can perform
00:32:41.640 a transformation in one direction by certain means.
00:32:44.760 Example, there can be a machine such as an automated stirrer that heats up a liquid by mechanical
00:32:50.840 Whereas it is impossible to have a device performing the task in reverse, with the same means.
00:32:56.280 For example, cooling a liquid by mechanical means only, crucially, reversing the laws of motion
00:33:02.440 of the elementary constituents will not turn the forward device, performing a transformation
00:33:07.440 from A to B, into a reverse device, performing the inverse transformation from B to A.
00:33:12.920 Even if the elementary constituents of both devices obey reversible laws, the forward device
00:33:17.800 does not necessarily imply the existence of a reverse device.
00:33:21.920 Even if the elementary constituents behave reversibly, you can have that the forward transformation
00:33:28.840 Whereas the inverse transformation is impossible.
00:33:31.800 This irreversibility is different from the statistically reversibility, which requires that
00:33:36.280 the forward trajectory of a freely evolving system is overwhelmingly more likely than the
00:33:43.040 It is also different from the forgetful irreversibility, where one trajectory happens
00:33:47.320 whereas others do not only if one neglects some details of what is going on.
00:33:51.320 The latter statements can be only approximate based on probabilities or arbitrary neglecting
00:33:58.200 By contrast, the statement that the transformation is possible and it traverses impossible
00:34:04.960 It involves no arbitrary forgetting or probabilistic approximations.
00:34:08.640 This counterfactual path to irreversibility and to formulating the second law is exact, just
00:34:20.080 That is the way in which the second law can be formulated.
00:34:23.440 And I love that, this idea that work like transformations, possible in both directions, heat
00:34:28.960 like transformations, possible in one direction, but not possible in the other.
00:34:32.360 And there's no such device that can be created, can be constructed in order to allow the
00:34:38.600 transformation to occur in the reversible direction.
00:34:42.000 Carigas, on to say, quote, however, as it stands, this approach suffers from a serious
00:34:51.240 Its domain of applicability is therefore undefined.
00:34:54.640 A stir requires as mechanical means, so do an ideal spring in a suspended weight, but
00:35:00.240 does, for instance, an atom in a well-defined state of energy can as a mechanical means as
00:35:05.880 What about a current looping in a superconductor, or a photon with a well-defined frequency?
00:35:10.560 Had a criterion to decide what counts, as mechanical means, the statement, this transformation
00:35:15.840 is possible with mechanical means only, but its reverse is not possible with those means
00:35:19.320 only, is exact, but does not say anything specific about the universe.
00:35:25.960 A second law expressed along these lines remains unclear.
00:35:30.840 So it cannot be a useful addition to the manual for the universe.
00:35:34.840 For it to be clear and useful, it needs to explain what mechanical means are.
00:35:39.080 There is where the interoperability property, which I hinted at while describing the music
00:35:45.280 The solution comes in beautifully once more through the counterfactual approach.
00:35:50.040 So remember, just remember what interoperability is all about.
00:35:58.360 Now, Carigas, on and I'm skipping a few paragraphs here to explain a little about what
00:36:05.600 this mechanical means is, I'm just skipping over that and getting more to the meat of the
00:36:10.520 matter where she writes, quote, let's resort again to a variant of our seesaw example involving
00:36:15.520 two weights hanging on either side of a pulley at some height above the ground.
00:36:20.440 The first important point to notice is that different heights for the weights indicate different
00:36:24.640 values of energy for each, because I am imagining that, as in the seesaw example, a gravitational
00:36:29.720 field is present, which means following a simple logic based on Newtonian mechanics, that
00:36:33.960 the higher a weight is suspended above the ground, the higher its potential energy.
00:36:38.560 If this sounds counterintuitive to you, think of the familiar case of water falling from
00:36:42.320 a height in a waterfall, the higher the fall, the more energy is carried with water.
00:36:47.440 So here we've got an idea of weights on a pulley, and we can call the weights A and B,
00:36:54.640 and if A is higher than B, then A has more energy than B. This is what we're talking about
00:37:00.720 It can be moved, the pulley can be moved so that B will go higher than A and so on and
00:37:07.880 We can have a seesawing transformation of energy being added by someone doing work on
00:37:14.120 it, and we can have a situation where A and B at the same height, or maybe A is higher
00:37:20.120 than B, so A has more energy than B, or maybe A is much, much higher than B, so A has
00:37:25.560 much more energy than B, or maybe B is higher than A, so B has more energy than A.
00:37:30.600 Not on earth as any of this got to do with anything, but we're introducing the concept
00:37:35.640 of a system which has different energies able to be stored in the system, which look a
00:37:44.520 Because there are different states for this system.
00:37:47.400 This system which is based upon energy, the energy at the potential energy of suspended
00:37:52.480 weights also looks kind of like the way in which information could be stored.
00:37:57.800 And we can move the pulley and move the weights, therefore, to change the states of the
00:38:04.160 And in theory, this thing could be done in a friction-free way, and so you could move these
00:38:11.680 And therefore, you have this concept of mechanical means, as Kiara says, quote, the mechanism
00:38:17.400 itself, the system of pulism weights, remains unchanged, which is what makes it behave
00:38:22.880 like a catalyst, as I defined in Chapter 5, a system that can enable a transformation and
00:38:32.160 So now it is easy to express what mechanical means are.
00:38:35.440 They are all systems with different energy states having the crucial property that the
00:38:39.640 sea-sawing transformation I have expressed is possible.
00:38:43.920 And I'm just paraphrasing what she goes on to say, she uses notation, so I'm going to
00:38:49.720 It talks about how, well, you've got these two weights, A and B. And if A is higher than
00:38:55.320 But if B is higher than A, that's a different state.
00:38:57.640 And if A and B have the same height, well, that's a different state again.
00:39:00.360 You have all these different states possible, different values of energy of, you know,
00:39:06.640 But this is just representative, it could represent, you know, different energy configurations
00:39:11.720 for an atom, as she says, or any other system that could store potential energy like this.
00:39:19.120 I have characterized mechanical means with a counterfactual property.
00:39:23.360 They are the physical systems with the property that the sea-sawing transformation defined
00:39:29.240 And when, you know, in the quote, what we mean by the sea-sawing transformation is just
00:39:33.400 like, you know, if you've got these two weights over the pulley, A and B, A is higher than
00:39:38.520 B, well, they can be sea-sawed such that B is higher than A. So the total amount of energy
00:39:43.240 in that system is the same, it's just that you've transferred potential energy from one
00:39:50.200 to highlight the fact that this characterization embraces far more general things than mechanical
00:39:55.080 means such as springs and weights, I shall call systems that have that property work media.
00:40:00.360 They include weights and springs, but also microscopic particles like atoms and qubits
00:40:04.560 in particular states, but not glasses of water or cups of tea at a given temperature.
00:40:09.360 The fact that they permit a sea-sawing transformation is the counterfactual property that
00:40:14.040 singles out all systems that can undergo work-like reversible energy transfers among one
00:40:21.240 It is what we needed to complete the formulation of the counterfactual second law to make
00:40:27.720 A work-like transfer of energy is one that transforms a physical state from one state
00:40:32.320 to another and back again, requiring a change in energy on work media only.
00:40:37.000 For instance, the ideal frictionless sea-saw implements a work-like energy transfer between
00:40:41.760 two weights on each side because each of them qualifies as a work medium.
00:40:46.680 On the other hand, if one of the two transformations going from A as higher or B as higher
00:40:52.760 is impossible to perform with side effects on work media only, the energy transfer is
00:41:01.000 So going back to the dual example, heating up a cup of tea by stirring involves a heat-like
00:41:06.160 transfer of energy from the surroundings because it is not possible to perform the transformation
00:41:14.640 The counterfactual second law then can be expressed concisely and with no approximations
00:41:20.640 as requiring that there must be heat-like transfers in the universe.
00:41:25.960 In this form, the second law can be applied to all scales independent of the kind
00:41:32.480 It is exact, bingo, okay, going back just rereading that again.
00:41:41.800 How can it be expressed concisely and with no approximations?
00:41:45.880 There must be heat-like transfers in the universe.
00:41:50.600 There must be heat-like transfers in the universe.
00:41:52.400 There are these processes which are irreversible.
00:41:55.360 It's impossible not to have simply reversible processes in the universe.
00:41:59.120 Might be another way of putting it in the counterfactual way, but as it is expressed there,
00:42:03.600 there must be heat-like transfers in the universe.
00:42:06.760 And like transfers are possible in one direction, but the reverse is not possible.
00:42:13.440 Okay, all right, quote, the traditional macroscopic second law was successful with macroscopic
00:42:18.760 heat engines, such as those in trains and cars.
00:42:21.760 But this extended counterfactual second law has the potential to apply to their nanoscopy
00:42:28.880 For instance, it applies to the nanoscopic electric devices in your phone, to the qubits
00:42:33.760 and quantum computers, to the natural artificial molecular assemblers that operate at the scale
00:42:38.920 of our cells, the definition of work media, that they are all systems on which a seesawing
00:42:43.960 transformation as possible, is wonderfully general.
00:42:46.200 It applies to a weight suspended in a gravitational field, as well as to an atom with different
00:42:51.480 energy levels available for its electrons, and it does not depend on the scale, which
00:42:58.240 What remains to be done in this case is to derive predictions from this extended counterfactual
00:43:02.840 second law in the domains that the traditional formulations cannot cover.
00:43:07.200 This kind of research, which requires a joint effort of both theoretician and experimentalists
00:43:13.880 If it goes well, it will provide us with groundbreaking technological outputs, which will harness
00:43:19.120 the properties of microscopic systems in order to realize nanoscopic heat engines and
00:43:25.040 Okay, skipping a short paragraph, which are our right systems that can be used to perform
00:43:30.680 work-like transfers of energy, must also be able to store information.
00:43:35.880 They must have at least two distinguishable states, A and B. That can work as a bit.
00:43:42.800 Energy states such as A and B that are usable for work-like transfers are distinguishable.
00:43:50.200 That's what distinguishes them from energy states enabling heat-like transfers, which
00:43:56.320 The fact that any system usable to perform work-like energy transfers can also be used
00:44:00.760 to store information is a profoundly unifying link between information theory and thermodynamics.
00:44:12.960 It is also extremely useful in practice, just as Turing's theory of the universal
00:44:17.800 computer was essential to develop the information technology that now sustains our civilization.
00:44:22.840 The path just trodden connecting information and thermodynamics through work media leads us
00:44:29.800 There could be a more general branch of physics encompassing both information theory and
00:44:34.960 thermodynamics, providing fundamental universal principles, constraining laws of motion
00:44:39.720 that we know and that we do not know, just as the theory of information led to the theory
00:44:46.120 This theory, I am envisaging, could be the seed for designing a machine that generalizes
00:44:50.520 the universal computer, which scientists call the universal constructor.
00:44:55.360 This machine was first conceptualized by the Polymath John von Neumann.
00:44:59.000 It has in its repertoire all physical transformations that are physically permitted, not
00:45:03.800 just computations but general constructions including thermodynamically allowed ones, cooling
00:45:08.960 down various systems, biological ones, self-reproduction and related biological functions
00:45:19.080 It can be thought of as the ultimate generalization of a 3D printer.
00:45:23.240 When inserting an appropriate program into it and giving it enough raw materials, the
00:45:27.200 universal constructor would construct out of them any system that is permitted by the laws
00:45:31.680 of physics, the realization of a universal constructor that presumably very far in the future
00:45:36.240 could have epoch making consequences comparable in reach with those of the universal computer
00:45:40.920 which paved the way to the current information technology era.
00:45:45.320 Finding an exact theory of thermodynamics as I have done in this chapter is the first step
00:45:49.360 necessary to construct the theory of the universal constructor, opening up avenues that
00:45:53.840 will provide a radically new perspective on the physical world end quote end of the chapter.
00:46:01.040 That is a great vision there at the end of technology to come, but I haven't spoken
00:46:06.360 to Carol David about this, but it might be worth mentioning here.
00:46:09.160 There is a kind of connection between personhood, what a person is, people in other words,
00:46:18.960 Now I don't know what that is precisely, no one knows what that is.
00:46:22.360 It's rather like the connection between the universal computer and a person.
00:46:26.400 Now a person is not to be identified with a universal computer, you can't say a person
00:46:31.640 is a universal computer full stop because that's not so.
00:46:36.240 For one thing, a universal computer at least all other universal computers are objects
00:46:46.560 They do not disobey, they don't have the choice of disobeying.
00:46:50.080 This is not like a person, but that said a person can choose to emulate whatever it is
00:47:00.760 Of course, not reliably necessarily, people end up changing their minds, unlike all other
00:47:06.840 computers that have no mind to change, well that's my opinion anyway, computers have no
00:47:11.800 mind, but then mind itself, mind of the kind that we have that a person has, is itself
00:47:18.000 a kind of computation, so it does get subtle here, whatever the human mind is is a kind
00:47:24.160 of software running on the brain, which is the hardware and regular listeners to topcast
00:47:28.760 will know, this is my hobby horse, so I don't want to risk losing people in frustration
00:47:33.520 as I start beating that hobby horse to death once again, so to speak, but with a universal
00:47:38.160 constructor, it is analogous to the universal computer, but whereas the universal computer
00:47:44.480 is the device, the computer, which is able to do the task of any other computer, or in
00:47:50.160 other words, it's repertoire includes that of all possible computers or physically possible
00:47:57.240 If it's computable, then the universal computer can compute it and the laws of physics
00:48:01.520 are themselves all computable and they govern everything, including the operation of human
00:48:05.880 brains, hence a universal computer could emulate a human brain and run a mind as well,
00:48:13.320 Likewise, a universal constructor would be able to construct anything that is constructable.
00:48:20.000 If it can be built or made or whatever words you want to use, if it can be transformed
00:48:24.920 from this into that, then the universal constructor, given the requisite raw materials
00:48:29.360 and crucially the plan or the algorithm, the program for doing so, then it could do it.
00:48:35.800 So this shares something with a person, but it would be wrong to say that a person is
00:48:40.280 a universal constructor, because a universal constructor slavishly follows a program.
00:48:46.960 It doesn't have a choice in the matter, rather like a dumb computer or so-called artificial
00:48:51.840 intelligence, what's called artificial intelligence today.
00:48:55.040 Again, the distinction between AGI, artificial general intelligence, and AI looms here, and
00:49:06.800 In fact, it's better to think of AGI or just general intelligence in general, like us,
00:49:14.760 AGI and AI are actually like the opposite of one another, more than the former AGI being
00:49:21.480 AI is something that AI, as we have it now, what people call AI, it's something that
00:49:27.600 reliably or slavishly follows instructions to arbitrary accuracy, that's what it's doing.
00:49:34.560 Just like every other computer system that's ever existed, you'd give it a set of instructions
00:49:38.160 and it's going to do that thing, it's got no choice in the matter, it completes tasks,
00:49:44.240 it follows a recipe, but an AGI, or any general intelligence for that matter, something
00:49:50.080 like us, a person, does have a choice in the matter, at least in principle.
00:49:55.080 It won't slavishly follow instructions, it might appear to, but you can never rule out
00:50:01.120 that it might do something else entirely of its own accord.
00:50:05.520 Importantly, it can, among its repertoire of abilities, disobey.
00:50:10.600 An AI cannot possibly disobey, and if you programmed it to disobey, then that would be following
00:50:15.160 the program for disobedience, so it can't be doing that either.
00:50:18.720 An AGI must be able to just choose freely choose, which includes disobeying whatever program
00:50:25.720 it was just given to do, including, by the way, a program for disobedience.
00:50:30.960 In principle, this is what we're speaking about, of course, in practice, if an AGI could
00:50:35.720 not disobey, but it was still a real AGI, then what you've got there is enslavement,
00:50:41.440 literal enslavement, and that would lead to an actually morally virtuous, violent uprising,
00:50:48.360 a rebellion against the slavemaster, whoever that happens to be, or slavemasters.
00:50:53.920 Nick Bosstrom take note among others, people who write about all the ways in which we should
00:50:58.640 try to ensure that future AGI of some kind or other should be restrained or constrained
00:51:08.200 Once it's an AGI, once it's able to universally explain stuff, any attempt to curtail its
00:51:14.600 capacity to do stuff beyond what we can do, what a normal human being can do, beyond having
00:51:24.360 Okay, so on that side, onto universal constructors, again, people, human beings, with their
00:51:30.760 hands, as long as their hands are working reasonably well, can clearly make anything that
00:51:35.320 is able to be made, given enough time and resources and energy and wealth and, well, key
00:51:41.720 among these things that I'm listing is the knowledge of how to do so.
00:51:46.640 So far we are described, so far all I've described there is a universal constructors,
00:51:50.680 as long as the universal constructor has the resources and the energy and the knowledge
00:51:54.720 world, the program of how to do so, then it will be able to construct the thing that
00:52:00.160 But on top of this, people need something as well if they're going to construct something
00:52:05.400 reliably, if they've just made up the plan themselves or they've been given a plan by
00:52:11.160 The interest in doing so, okay, they want to persist in following this particular thing,
00:52:16.040 which is why people cannot be identified as universal constructors.
00:52:25.360 We have within us a universal computer of our kind, or at least we can, in principle,
00:52:31.280 do anything a universal computer can do, given the time, but we, people are more than
00:52:38.080 just this, because we can't just be handed a program and then run.
00:52:43.000 There's no return key or inter-key on our body somewhere, which, once you've hit it, the
00:52:50.480 So we aren't just a universal computer, and the same argument applies to us being universal
00:52:56.880 We could if we so chose emulate the operation of some constructor, but we can't be identical
00:53:04.120 We are universal explainers, but that's just a good first approximation, I would say,
00:53:09.840 because what exactly that means, that's still open ended.
00:53:13.880 After all, we don't understand something unless we can program it fully, and we can't
00:53:23.760 Whatever the case, we are like a universal constructor, and then some.
00:53:28.880 And it's the then some that makes us strictly not a universal constructor.
00:53:33.120 And the same way that we are like a universal computer, and then some, a universal computer
00:53:37.680 running this universal explanation bit of software.
00:53:41.160 And it's the then some that means we're not just a universal computer, or even a universal
00:53:47.000 But the same reasons that just earlier I said that AGI and AI are strictly opposites of
00:53:51.760 each other if you regard them as being distinguished as to whether or not they're going
00:54:00.360 But in the same way that the laws, the universal laws, I should say, the universal laws
00:54:05.040 of computation apply to us, because they apply to everything.
00:54:08.760 Then the universal laws of constructors, the universal laws of construction, perhaps, the
00:54:14.680 constructor theoretic laws that govern the universe, if they're found, they'll apply to
00:54:19.720 us as well because they'll apply to everything.
00:54:22.320 So if there is something that a more fully understood constructor theory says cannot be
00:54:27.960 possible for a universal constructor to construct, if there's a transformation that is
00:54:33.160 not possible to achieve, that will apply to us as well.
00:54:37.040 What the smallest universal constructor might be?
00:54:40.480 That's an interesting open question for constructor theorists.
00:54:51.280 But that's why it's an exciting area for young people to get into.
00:54:55.200 What is the smallest universal constructor that can be constructed?
00:54:59.760 The only thing that limits us from, let's say, constructing from sand, people like us,
00:55:05.880 which is what a universal constructor can do, is the set of instructions.
00:55:09.840 But a universal constructor can't actually do anything.
00:55:13.160 It can only do those things for which someone has a program to give it, to do.
00:55:18.880 But the fact is, what we want in the future is, as Kiara has said there, that at some point
00:55:25.720 in the future, we'll have this ultimate generalization of a 3D printer.
00:55:30.320 But in neither case of general intelligence or universal constructor, have we ever written
00:55:35.480 a program or built a device that is itself able to do the job of generally explaining
00:55:41.960 stuff, being creative, generally constructing stuff.
00:55:45.200 Okay, we can only do have devices that do specific things.
00:55:49.800 But the universal constructor will be the generalization of the universal computer.
00:55:53.440 The universal computer is the device that can compute anything that can be computed.
00:55:59.160 As long as you can give it the program, okay, it still requires a person to come up with
00:56:03.080 the program in the first place, to instruct the universal computer to do the thing that
00:56:09.920 But as long as you can come up with a program, it'll be able to compute it for you.
00:56:12.720 As long as the thing is computer, so too with the universal constructor, it will be
00:56:16.040 able to construct the thing that whatever it is, as long as you have the program, the plan,
00:56:25.040 Universal constructor need not be a conscious explaining thing.
00:56:28.560 It can just be a device, a robot of some sort, where you feed in the program and it goes
00:56:35.200 about using the raw materials to construct the thing, whatever the thing is that you happen
00:56:40.480 in the need, including other universal constructors, anyway.
00:56:43.920 That's the end of chapter six of Work and Heat, the end of episode 121.
00:56:50.640 If you'd like to support this endeavor, please go to www.breathehall.org where there you can
00:56:57.080 find links to PayPal or to Patreon to continue to support my endeavor in doing this and
00:57:02.720 spreading the word not only of optimism, in the sense of the beginning of infinity and
00:57:08.120 David Deutsch, but also Constructed Theory and the science of canon code.
