Dr. Jordan B. Peterson has created a new series that could be a lifeline for those battling depression and anxiety. With decades of experience helping patients, Dr. Peterson offers a unique understanding of why you might be feeling this way. In his new series, he provides a roadmap towards healing, showing that while the journey isn t easy, it s absolutely possible to find your way forward. If you re suffering, please know you are not alone. There s hope, and there s a path to feeling better. Go to Dailywire Plus now and start watching Dr. B.P. Peterson on Depression and Anxiety: A Guide to Feeling Better. To find a list of our sponsors and show-related promo codes, go to gimlet.fm/OurAdvertisers. We are always looking for listeners who are willing to talk to us about their favorite products, services, or anything else related to their lives. Please take a few minutes to fill out this brief survey. The results will be featured on the next episode of Dailywireplus. Thank you so we can keep bringing you quality, high-quality episodes. If you like what you hear, please consider pledging a small monthly subscription! We appreciate the support and share it with your friends, family, colleagues, and your fellow listeners! We look forward to hearing from you! Timestamps: 0:00:00 - What are your thoughts on this episode? 1: What would you like to hear from me? 3: What do you think of it? 5: What are you looking for? 6:30 - Is consciousness non-computational? 7:00 8: What is your favorite piece of advice? 9:00s? 11:00e - Is it possible? 12:30s - what would you believe that you would like to see me do it better? 13:00a) 14:40s - does it matter? 15:00d 16: Is it a good thing? 17:00p=1st & 11:10s 15, what is your answer? 16) Do you think it's a better than a better idea? +3rd? ? 17) 15) Is it better than you can trust the rules? 14) What is it possible to trust in the rules I think it s better than that?
00:00:01.000Hey everyone, real quick before you skip, I want to talk to you about something serious and important.
00:00:06.000Dr. Jordan Peterson has created a new series that could be a lifeline for those battling depression and anxiety.
00:00:12.000We know how isolating and overwhelming these conditions can be, and we wanted to take a moment to reach out to those listening who may be struggling.
00:00:19.000With decades of experience helping patients, Dr. Peterson offers a unique understanding of why you might be feeling this way in his new series.
00:00:27.000He provides a roadmap towards healing, showing that while the journey isn't easy, it's absolutely possible to find your way forward.
00:00:35.000If you're suffering, please know you are not alone. There's hope, and there's a path to feeling better.
00:00:41.000Go to Daily Wire Plus now and start watching Dr. Jordan B. Peterson on depression and anxiety.
00:00:47.000Let this be the first step towards the brighter future you deserve.
00:00:57.000I'm Stephen Blackwood, and I have the great honor today to be here with Sir Roger Penrose and Dr. Jordan Peterson.
00:01:19.000Let's get right down to it. Jordan, I know you have questions you're keen to pose to Sir Roger. Over to you.
00:01:25.000Yeah, well, I've wanted to talk to a theoretical physicist for about 30 years, and so I'm pretty happy that you're the theoretical physicist that I get to talk to.
00:01:42.000A colleague and friend of mine is an AI engineer and a computer engineer, and he's built a lot of the world's great chips, iPhone chip, and first 64-bit chip, the Alpha, back in 1985.
00:01:57.000And we were having a conversation. I said I was coming to meet you, and that you, and I don't want to put words in your mouth, believe me, but that you believe that consciousness is in some fundamental sense non-computational.
00:02:11.000And I asked him what he thought about that, and part of the reason I asked him is because he's, of all the people I've ever met, and maybe of all the people in the world, he's the person who's done most to build arguably brain-like algorithmic systems.
00:02:25.000And so I asked him if he thought that there was a distinction between the algorithmic computation of cognition, per se, and whatever consciousness might be, and he thought it was algorithmic all the way down.
00:02:42.000And I understand that you don't believe that. I also went with him a couple of times to a consciousness conference in Tucson where Hameroff spoke.
00:02:51.000So we got familiar with that line of reasoning. And I also understand, I believe, that part of the reason that you think that consciousness is necessarily non-computational is because of Goodell's theorem.
00:03:06.000And so maybe we could enter there. I'm very curious about your proposition that consciousness, per se, is non-computational.
00:03:16.000And I'm curious about why you came to that conclusion, and if you think that's a warranted conclusion, what do you think about that in relationship to these complex AI systems and also in relationship to Goodell's theorem?
00:03:30.220Well, I've never seen the argument refuted. I've just talked to people who've never really understood it, as far as I know.
00:03:36.240No, the argument goes back to when I was a graduate student, and I was doing pure mathematics, algebraic geometry, and I went to three courses, which were nothing to do with what I was supposed to be doing.
00:03:49.400One of them was a wonderful course by Herman Bondi on general relativity, which had a big influence on what I did later on.
00:03:56.520One was a talk by the great physicist Paul Dirac, and that taught me about quantum mechanics.
00:04:02.480And their third one was a course by a logician called Steen, and he taught me about Turing machines, the notion of computability, what it is and how you understand that, and the Goodell theorem.
00:04:17.900And I had heard vaguely about the Goodell theorem previously and had been rather worried because it seemed to show that there were things in mathematics that you couldn't prove.
00:04:27.520What I learned was that it's not like that at all.
00:04:31.620If you lay down the rules of what you call a proof, and if those rules are such that they could be checked by a computer, checked whether they've been correctly applied by a computer, so computation rules in that sense,
00:04:46.700then you can construct a sentence, this is what Goodell did, which by the way it's constructed, you can see that if you trust the rules, let's say, if you believe that the rules do, if they say, yes, you've proved it, tick, then you believe it's correct.
00:05:08.840But, let's say, if you have trust in the rules, that trust extends beyond the rules.
00:05:14.040In other words, you can see that a certain statement is true by virtue of your belief that the rules only give you truths, yet that statement is underivable, unprovable using the rules.
00:05:27.260That statement of faith about the rules?
00:05:35.940You understand the rules, you check them, you say, yes, that's okay, if that rule is correctly applied, I agree, it does, you know, it's a lot, it's a, it's a rule which is within something that you believe to be appropriate.
00:05:50.540And these rules, it's built up out of things like this, which nobody would dispute.
00:05:55.380You say, okay, if you follow those rules, and it says, yes, that's a proof, then you believe that the thing that it says, yes, it's a proof, is actually a true statement.
00:06:05.340So, does a proof really mean that it's true?
00:06:08.500If you believe that, that conviction, that the proofs actually do what they're supposed to do, gives you something beyond the rules themselves.
00:06:20.320Okay, that, that's, sorry, that's what I was referring to with, with the word faith, is that the statement of belief, well, I shouldn't maybe use that word.
00:06:27.700Well, I guess I'm wondering, what, what do you think it is that constitutes that belief?
00:06:34.300Okay, and why the word understanding specifically?
00:06:37.380Because that's the thing in some sense that's outside the system, the understanding?
00:06:40.900Yes, it is, because you can see it is, because it's the understanding that the rules give you only truths that enables you to understand that this Gödel statement is actually true.
00:06:57.140And so, is that belief in that truth of that proof, that is one of the things that Gödel pointed out, would be necessarily outside any system that's both, what is it, formal, logical, and coherent?
00:07:11.480It shows, it shows that, I mean, I read it in this particular way, I don't think he said it quite like this, but I read it in the following way, that understanding, whatever that word means, is not computational.
00:07:27.060Okay, okay, okay, so that is what I...
00:07:29.280It's not the following of rules, it's something else.
00:07:32.120Okay, so let me ask you a question about that. So, this is a three-pronged question, let's say.
00:07:41.220It seems to me that there's a high probability that the future is actually indeterminately different than the present and the past, that it's actually unpredictably different.
00:07:52.380Oh, this is a different question. Now you're talking about determinism.
00:07:55.800Yes, yes, but I think it seems to me that it's tied to this idea that computation can be complete and algorithmic. I don't think it can be, because if the future differs in a fundamental manner, an unpredictable manner from the present or the past, then a deterministic algorithmic system can't maintain a grip on the horizon of the future. And I have another part of that question.
00:08:21.080It's a different question. So, I think it's important to distinguish these things.
00:08:29.000I was talking about rules, well, just yes or no. I mean, it's not a question of maybe. I mean, it isn't even talking about the laws of physics at this stage. That's the second step, if you like.
00:08:43.260I guess I looked forward to something like the potential necessary function of consciousness.
00:08:49.380So, because one of the things consciousness seems to do from a neurophysiological perspective, for example, we tend to become conscious of our procedural errors.
00:08:59.420And so, consciousness becomes alerted to the errors and then zeroes in on the source of the error in some sense and corrects it.
00:09:07.200And so, it looks to me like it's something like a correction system for underlying algorithmic systems.
00:09:12.840So, for example, if you practice a motor routine for a long time, you build specialized algorithmic machinery in your brain that runs it.
00:09:22.020But maybe you've put in an error. You're playing a difficult piano phrase, for example, and you stumble over a note.
00:09:28.380You've automatized that. You play it and you listen and you hear the anomaly, which is the error.
00:09:34.300Your consciousness focuses in on that.
00:09:36.140In fact, a large brain area will activate as a consequence of becoming aware of that error.
00:09:42.380Then when you practice the new routine that's corrected, the brain area will shrink and shrink and shrink until it's a small part of the brain, usually in the back of the left hemisphere.
00:09:51.740And now you've built another automated machine to play out that phrase.
00:09:56.140And consciousness, I think it was Whitehead who said that at least the purpose of consciousness, although he might have used thought, was to increase the number of things that we can do without consciousness or thought.
00:10:09.220But it seems to be this horizon phenomena.
00:10:12.380And the reason I was asking about the indeterminacy of the future was twofold, is that if the future is deterministic, then an algorithmic system could, in principle, adapt to it.
00:10:23.660But I don't, it doesn't seem to me that the future can be predictable.
00:10:28.500And I think that that might be grounded in something like quantum indeterminacy, because there isn't a fundamental determinism that propagates all the way up.
00:10:37.940So, well, you see, I mean, we have to, the things I was talking about up to this point were not to do, they weren't even to do with the laws of physics.
00:11:09.900If it's just indeterministic, it's not connected.
00:11:16.900You see, the kernel argument is to do with things where you have definite rules, you can check whether these rules have been followed or not.
00:11:25.480And the question is whether it coincides with your understanding about what things are true or false in mathematics.
00:11:37.120Now, you see, you can question how you move from that into other aspects of what consciousness does.
00:11:44.040And also, the question you were referring to is whether something is automatic and your pianist can play things.
00:11:49.380And obviously, where the little finger goes next is not something that he or she decides to do.
00:11:56.240It's all largely controlled by the cerebellum, probably, which, as far as we know, is entirely unconscious.
00:12:02.060So the greater number of neurons in the brain, which are in the cerebellum, seem not to be acting according to conscious actions at all.
00:12:14.380It's something completely unconscious, as far as we can see.
00:12:16.460That's a very strange thing that, you know, people make the case as well that there's some simple relationship between neuronal function and consciousness.
00:12:23.280But as you pointed out, the cerebellar activity doesn't seem to be conscious at all.
00:12:27.820And then there's a tremendous amount of neurons in your autonomic nervous system distributed throughout your body.
00:12:33.680And there may be some consciousness associated with that, but it's not particularly acute.
00:12:38.380And most of the time, it's entirely unconscious.
00:12:41.140And the autonomic nervous system is running your digestive system and your heart and all of these inner automated systems.
00:12:47.380And it's interesting, too, because often becoming consciously aware of a highly functional unconscious system actually impairs its function rather than improving it.
00:13:11.560May I just jump in to ask Sir Roger if you would say a word or two more about why it is that consciousness cannot be reduced simply to mechanistic processes?
00:13:21.620Well, you see, I'm very careful to say I'm not talking about consciousness in all its aspects.
00:13:28.220For example, I mean, I have nothing to say about the perception of the color green, for instance.
00:13:33.740I mean, sure, there's something going on which makes green have a certain impression on one.
00:13:39.940But this is not what I'm talking about.
00:13:41.840And probably most of the things that we think about when we talk about consciousness are not what I'm talking about.
00:13:47.740So I'm only talking about a very specific part of what consciousness does.
00:13:54.000And the argument is that if this is something which is not a computational process, then it sort of sheds a question mark on the whole thing.
00:14:05.820But it's only very specific to the question of understanding.
00:14:55.100And so there's a huge amount of life that is like that.
00:14:58.320But to then ask the question about why this is a goal or why this is worthy of being a goal or what would make it worthy of being a goal or what would make that worthy of being a justification for that to be a goal,
00:15:12.440the kinds of thinking that you have to engage in in order to reflect upon the nature of the ends and purposes is distinct from the kinds of thinking you engage in to calculate your way to a goal.
00:15:24.760And that seems to point towards the realm or a kind of thinking or awareness that is clearly distinct from a simply mechanistic calculation.
00:16:08.560So I'm certainly not saying that's an indicator of consciousness.
00:16:13.180I mean, I'm saying it is something which requires consciousness.
00:16:17.140But I completely accept that there are all sorts of other aspects of consciousness which are going on all the time and which are much more important.
00:16:27.520But it's just that if you can find something in what consciousness seems to do, which is not which is demonstrably not computational, that's saying something.
00:16:39.080And that's the limited little thing I'm trying to say.
00:16:41.840Now, you started working with Hameroff, as I understand it, to try to provide something approximating a localization or a neurophysiological account of what this non-deterministic process might be.
00:16:55.720But I didn't say that's usually non-deterministic.
00:17:23.220But usually one talks about randomness there.
00:17:25.720You say that the theory does not have a complete description of what it tells you happens in the future because there is a random element in it.
00:18:15.560I don't think that gives you anything in the way of establishing results which seem to be a non-computational process, like with the Gödel thing.
00:18:29.000So there's an evolutionary answer to the problem of emergent randomness.
00:18:33.880And the answer is, so a mosquito, a mosquito is a good example, or a fish, any animal that lays a tremendous number of eggs that could conceivably march to maturity.
00:19:11.260And it's actually a consequence, I would say, of events that are actually manifesting themselves in some sense at a quantum level.
00:19:17.160Because at least some of the mutations are caused by solar, by radiation.
00:19:21.880And so there's disruption at a molecular level.
00:19:24.140And so evolution seems to be able to use the admixture of randomness into structure as a means of dealing with the determinacy of the future.
00:19:34.140And to some degree, it does that through death, right?
00:19:36.840Because of those million mosquitoes, on average, only one manages to propagate itself to reproduction.
00:19:44.740No, I understand what we're saying, yes.
00:19:46.480But what I'm trying to say is that what's going on with consciousness is different from that.
00:19:50.680Because I don't see how this, you know, you putting randomness in the way you're suggesting, and clearly that is an important aspect to evolution and so on.
00:20:03.640I certainly wouldn't deny that at all.
00:20:26.040So, well, because I'm very curious about what you do mean.
00:20:29.220I mean, this is obviously a tremendously important distinction between the computational algorithmic domain and something that's in some sense outside of it.
00:20:40.480And I'm struggling to understand at the most detailed level, let's say, how you envision the structure and function of consciousness or maybe just the function.
00:20:54.800It's not producing mere random variance.
00:20:57.520And that can't be because random is too widespread.
00:21:01.780So at least, at the very least, so for example, if you study creative people, we've done a lot of this, there is in some sense more randomness in their speech.
00:21:12.260Because imagine that if you utter a given word, there's a certain probability that another word will emerge in the field around that.
00:21:20.940The creative people use lower probability concepts and words in their approach.
00:21:29.500They go farther out into the word association field.
00:21:32.600And that does help them generate more creative solutions.
00:21:36.260But that's not, if that becomes unconstrained to too great a degree, you get, well, maybe like a manic creativity that's counterproductive and too random.
00:21:46.820People are jumping too much from disconnected point to disconnected point.
00:21:50.620And so consciousness doesn't seem to be, creative consciousness doesn't seem to be a mere random walk.
00:21:58.280So that's a psychological take on that.
00:22:00.360But so what do you think is, what do you think, I'm still struggling to understand what you think consciousness does.
00:22:08.480You see, I think probably you're trying to make me be more specific than I can be, because I don't know what it is that, how to make a device that can understand something.
00:22:24.880So I'm just trying to say that whatever understanding is, it's not a computational process.
00:22:41.080I mean, when we say it is not non-computational, does that mean, or does that not mean, by definition, that consciousness in some deep level is free?
00:23:09.840And the trouble is, I think it's not a concept which people appreciate usually.
00:23:14.400So I can give you examples of non-computational things.
00:23:17.900And one of the examples I often give is if you take, imagine a pattern of squares, equal squares or just a normal square array.
00:23:29.360And you can consider a finite shape made out of squares.
00:23:33.340I think it's called a polyomino, shape made out of squares.
00:23:36.660And if you're given a finite set of these polyominoes, and the question is, can you cover the plane with those shapes, only those shapes, no gaps, no overlaps?
00:23:50.640Now, that question, the answer, yes or no.
00:24:01.560It's shown mathematically that there is no algorithm which can tell you, yes or no, whether these shapes will cover the plane.
00:24:10.840Okay, so when I was talking to my brother-in-law, I was talking to him about these AI systems that learn how to recognize, let's say, caps from photographs.
00:24:22.440He told me there is no way of algorithmically determining the program that the machine learning systems will eventually apply to the problem of identifying cats in a photograph.
00:24:37.200But if you let the AI neural networks run and then you analyze their output, you often get something that resembles an algorithmic program as an output that you could have hypothetically calculated if you could have specified the search space.
00:24:53.040But there's no way of doing that without letting the program do its walk through the domain of cat photographs with its differentially weighted neural network architecture.
00:25:08.780Yeah, well, I still don't think it's the same thing.
00:25:14.140I could certainly give it different shapes and you could say, tell the machine, you know, which of these will tile the plane and which won't.
00:25:22.160Now, will that learn to give you correct answers?
00:25:28.320I suppose once it's tiled, you could formalize the process by which it was tiled, right?
00:25:34.720Because you could describe the mechanisms or the order in which the tiles were located and the rotation of them.
00:25:42.760You could specify it after it had all been laid down.
00:25:46.560And I think that's analogous to what the AI systems seem to be doing when they're learning to perceive.
00:25:50.940Well, the trouble is that it's not, the methods whereby you can tile the plane are, I mean, just the theorems tell you that you can't put them on the computer.
00:26:04.780I mean, you might get the thing which works most of the time.
00:26:42.840I mean, there was an earlier result which showed that if it were true, that any way of tiling that plane with these shapes, with, say, some given set of shapes, finite set of shapes, if it tiles the plane, you can do it periodically with a repeating pattern.
00:27:00.380If that were true, then there would be an algorithm.
00:27:03.500But it's not true, because there are certain ways of tiling a plane which do not have repeating patterns.
00:27:18.260Well, the fact that you're, that the tiles are, they're essentially, you're essentially mapping an area with a, with a predetermined concept in some sense.
00:37:49.740I was afraid I was a very bad correspondent.
00:37:51.560It took me a little while before I got back to him.
00:37:53.560But I showed him what it was based on.
00:37:56.660And on the basis of that, he produced, I believe, with his last watercolor, maybe for his last picture, as far as I know, a thing called Ghosts, which is based on this.
00:38:08.640It's the only tiling, as far as I know, that he ever did, which is what's called non-isohedral.
00:38:16.220You see, usually he did periodic ones, but they're periodic in a strong sense that if you find a shape, the next time you see it, it has the same relation to the pattern as a whole.
00:38:27.340So you could move this one into that shape, and the whole pattern goes with it into itself.
00:38:33.600But the one I showed him was what's called non-isohedral, that you can have different instances of the shape.
00:38:39.300So this one has a different relation to the pattern as a whole from that one.
00:38:44.300And so if I move this one into that, I can't bring the whole pattern along with it.
00:38:49.640So you have two different roles that the shape plays.
00:38:53.700And the last one of his pictures showed this.
00:38:57.700So I'm curious, too, about two things now.
00:39:03.680I'm interested in why you're so fascinated by the relationship of a geometric shape that can be arrayed in a variety of different manners to this underlying problem of mapping.
00:39:16.320So you're reducing or establishing a relationship between the problem of mapping a large terrain to the utilization of very stringently defined, what would you call them, representational systems?
00:39:36.540What is the geometric form conceptually in relationship to the problem of mapping?
00:39:44.660Well, you had a shape, and then you have certain rules about which pieces will fit next to it.
00:39:52.020But there's certain freedom in that rule.
00:39:53.780You could put this one that way or another way, you see.
00:39:55.680And, you know, if it's a shape which very clearly has to fit that way next to it, then it just repeats, you see.
00:40:05.380But if there's some freedom as to what the next one will do, then you might have to make that choice.
00:40:10.720And certain choices will run you into difficulties later, and other choices maybe will allow you to continue.
00:40:16.960Is there a relationship between that and what composers do with music?
00:40:20.720Because, I mean, there's a certain repeating determinacy in music, but obviously a composer just doesn't take a pattern and repeat it indefinitely.
00:40:30.880They take a pattern, and the pattern seems to allow for some choice in movement from that pattern forward.
00:40:57.480And then as it moves towards purely unpredictable, it becomes indistinguishable from noise.
00:41:03.340So there's some place in between there, and you could probably move on that place,
00:41:07.600where you get some ultimately harmonious relationship of predictable form.
00:41:13.420And, well, something like the play of novelty that seems to me to be analogous to that possibility of shifting the shapes in this tiling problem.
00:41:21.480I mean, I think music is tiling something.