David Deutsch Podcast Transcript

Original video: The Fun Criterion, Inner Conflicts, and How the Self Resolves Them — David Deutsch

Intro

Edwin:

I'm delighted to be joined today by David Deutsch. David's books, The Beginning of Infinity and The Fabric of Reality have been a primary inspiration for my work. My own curiosity has always been about understanding how the world works so I can base my own values and strategies for living a good and productive life on that understanding. When I first came across David's work, he introduced me to this binding element behind it, which is Popper's epistemology—namely a theory of knowledge which gives objective criteria for what constitutes a good explanation and how explanations are created and improved upon by humans.

Of course, David's work goes far beyond epistemology. He describes foundational explanations, notably the many-worlds interpretation of quantum mechanics and how the multiverse is needed to explain many parts of reality, more than we think. The often underappreciated significance of human creativity in the universe and the rational optimism that follows from that insight. Also the role of rational and anti-rational memes in shaping cultures and their social norms and also ideas about morality and parenting rooted in non-coercion and taking all parts of the mind seriously.

And today we're going to discuss fun, knowledge creation and human minds and I'm expecting to learn a lot from this conversation. David, thank you for joining me and welcome to the show.

David:

Thank you for inviting me. That's all very pleasant to hear.

Can we measure Fun?

Edwin:

The first topic I'd like to get into is the fun criterion. In short, it's the idea that the absence of fun is a warning signal. But that also brings me to my first question, which is, can we reliably determine when we are having fun or can we only really tell when we're not having fun?

David:

No, generally we can't reliably discover anything. So in order to put that in the reverse direction, we can always make mistakes and in fact there's no limit to the number or size of mistakes that we can make. So trying to be sure, for example, that we're having fun or that we are making progress or anything like that is the wrong framing. Even that framing has a strong tendency to produce the wrong irrational answers and non-workable answers.

So following Karl Popper, I advocate regarding lack of fun or apparent lack of fun or apparent lack of progress as a criticism rather than a hard stop to what we're doing. As long as we don't have a problem we don't have to answer questions like that. But if we do have a problem then our only recourse is to conjecture how we might change it and conjecture whether it's really true that we have that problem and so on.

Edwin:

Yes, I see. I think what I'm trying to get at is, when we're having fun, are there proxies or indicators that can maybe signal that? For instance, sensations of play or curiosity or interest or other emotions. Would you say there are useful proxies or correlates or how do I look at that?

David:

Yes, but proxies and correlates are no good in particular cases. Like, we can say in general that when you're having fun, usually when you're having fun, you're not in pain. Usually when you're having fun, you're not bored, you know, and so on. But it's not a reliable thing to use in particular cases.

So if you are a keen mountaineer or marathon runner, then you will be in pain when you're having fun. The thing is, you know why as well. So you can answer the question, hey, are you sure this is fun? You know, the question from yourself. Is this really the right thing for me? And then you can say, well, yes, I'm trying to break the record. So let me think about breaking the record and not whether I'm meeting a certain criterion. Meeting the criterion is a description of the process. It's not something you should be thinking about explicitly when deciding between ideas.

Edwin:

Yes, exactly. So it would be a mistake to follow or try to follow a specific feeling or sensation and then say, okay, now I feel this, I'm having fun. That in itself is a mistake.

David:

That's right. Yes.

Thwarting creativity

Edwin:

I think the most astute thing to look out for is the sensation of thwarting—that we are actively coercing our minds. Would you agree that's the case?

David:

Yes, I agree, yes. And in general, conflict. Conflict can be a very good thing, but again, conflict where you don't understand what's in conflict and it feels bad and you don't understand why and that sort of thing is a clue that there is something coercive going on, and therefore the different parts of your mind are working against each other using their creativity against each other which is bad, I mean it's the bad thing.

Edwin:

Yeah. And I guess it can be tricky to determine when you're merely experiencing a problem or an undesirable situation that you want to get out of, and when you are actively thwarting or hurting your own problem solving process.

David:

Yes, well, as I said, if it doesn't seem problematic to you, then you have no purchase on the issues that might in fact objectively be bad. If you're not aware of a problem and you're not aware of anything bad happening, then you can assume that there isn't anything bad happening because anything else would be irrational. Again, you'd just be making things up to frighten yourself.

Can mundanity be Fun?

Edwin:

Yes, exactly. Does it also mean that when we're doing something quite mundane, but it's not problematic—for instance, we're in the grocery store, we're doing groceries, we don't mind doing it, it's just something we do—are we then technically in a state of mind where we could call that fun, since there's no thwarting going on?

David:

No forcing going on is, if you like, a necessary but not a sufficient condition. I think it is possible to enjoy grocery shopping and I think it's possible not to enjoy it. And I think if we're not enjoying it, we will feel things like boredom or tedium or impatience and those are symptoms of something going wrong and one can conjecture what that might be and conjecture improvements to it. There's always a way of making one's life better. That's true even if one is in ecstasy.

Edwin:

Yeah, I definitely agree with this. I just think it could also be interpreted that one must always try to make everything enjoyable or fun, whereas maybe just an emotionally neutral state technically is also unproblematic and completely fine to be in. When you're not bored, you're not problematic, but you're also not necessarily happy or you're just doing your thing. I'm just wondering if then still fun is the right label or would it be better to use different terms there?

David:

Well, I think such a state is not optimal. I think that there are people who look forward to going shopping and enjoy themselves while they're shopping and shop in a different way each time, so that each time they're improving on their experience of last time. And that's a better way of shopping than of being emotionally neutral.

What does "follow the Fun" mean?

Edwin:

Sometimes you also describe people opting for a choice that is more fun. For instance, in a podcast with Lulie, I believe you mentioned Ginger GM, the chess player—he was debating and choosing to follow an inexplicit idea because it feels more fun and leads to a more beautiful game. What does more fun in that context mean? Is that referring to a thought process or to a feeling?

David:

It's referring to a state of mind which includes, in the case of—it includes all sorts of complicated things, including thought processes and feelings and curiosity. And there is the desire to win, but if one has only the desire to win, then it's very dangerous from the point of view of fun, because when there are two people playing, one of them is going to lose or at least one of them is going to not win.

And I thought, and I still think that Ginger GM is an example of someone who, for whom chess is not just a matter of winning, chess is a matter of enjoyment, of beauty and curiosity and so on and he likes all those things. And the example I gave is that quite frequently—I'm quite a fan of his, I watch his videos—and quite frequently, let's say once every three or four videos, he will say something like that so-and-so is probably the best move but it's boring. So on the other hand this other move might make me lose, but it'll lead to an interesting game. He usually—well, he makes a joke. He says, I can't stop myself. I can't stop myself. But what he really means is that he wants to do that. He would prefer having a good game to winning a boring game.

Edwin:

Yes, and also I think the point there is he has, I guess, two options or multiple theories or strategies that he could take and criticisms of the options that might lead to a win are indeed that they're boring and that they're unfruitful but one has this excitement to it or this beauty component to it.

David:

Yes. That's right, yes.

Edwin:

I think it's important to keep stressing that it's not about pursuing a specific feeling, but interpreting the sensations in your mind as criticisms.

David:

Yes.

Fun and inner alignment

Edwin:

And in the fun criterion, you also describe fun as arising when explicit and inexplicit and unconscious ideas influence one another and are in alignment. I'd love to unpack that a little bit. What does influencing one another mean?

David:

Well, if you have an explicit criterion for the choice you're about to make, if choosing which one it is just involves evaluating, say, an objective function, so that you don't know in advance which has the greatest chance of winning or chance of achieving the objective, then fun doesn't really come into it at that level. It's a question of whether the whole thing is fun or not that comes into it. But for choosing between the options, fun in that case isn't relevant.

So the place where it becomes relevant is when there are different kinds of conflicting impulses in your mind. For example, your desire to win, your desire to not be bored, your desire to look good in front of your audience, where none of those are mathematically quantifiable. Well, maybe some of them are, but not all of them are mathematically quantifiable. So then you have a conflict between your inexplicit and even unconscious ideas and your explicit ones. And if there is a problem there, you will be feeling it rather than working out that it's a problem. And that's why fun has a connotation of being about feelings. But as you said, it's not specifically about having good feelings, it's about having non-conflicting feelings.

Resolving conflicts

Edwin:

So specifically unconscious ideas make themselves known via our feelings, especially when they're in conflict with one another—an unconscious idea with an unconscious idea. But I think also an explicit idea and an unconscious idea, if they're in conflict would also signal through their feelings. And then you have to create I guess a third theory, which was trying to approximate what this conflict may be and this is what you can work with to try to conjecture candidate solutions for your conflict. Is that also correct?

David:

Yes, yes, that is correct. And that's, by the way, true for explicit conscious theories as well. Because although working out a particular thing, you know, does this strategy lead to checkmate? That's a thing that can be worked out mechanically without any feelings. But even in something like chess or mathematics, which line of thought to follow, which is the one that is best for you to follow, already involves inexplicit and unconscious thoughts and feelings. All explicit thinking contains an inexplicit component. But vice versa is not true. There are inexplicit and unconscious ideas which don't have an explicit component.

Edwin:

Yes. But if the explicit ideas also have an inexplicit and unconscious component, do you mean by that that the explicit idea has naturally included also the inexplicit and unconscious sensations within it? Or is there really like an explanation complex maybe going on where it's just like a set of explanations?

David:

Well, yes, when there's an explicit and an inexplicit idea that implicitly contains an inexplicit idea—dear, this is rather complicated terminology, but I hope you understand what I mean—then it automatically invokes the inexplicit ideas.

An example, but not necessarily an inexplicit explanation. I think that's the term you used. Because, for example, when we're speaking, especially when we're speaking in our native language, we have a lot of grammatical knowledge that is unconscious. And we don't necessarily know why a certain phrase is grammatical and a similar phrase is not grammatical and why we are using one and not the other. Of course we can choose to use the non-grammatical one for all sorts of reasons but generally when we're speaking we're obeying a lot of rules which we don't know and they can become problematic and if they're problematic then we can conjecture what they are.

And we enter the same kind of process as when we're conjecturing anything or trying to solve any problem. We may be wrong, or we may be partly right. We may partly solve the problem of what the rule is, but have in mind only an approximate version of the rule. And that approximate version might be usually right, but wrong in this particular case, and so on. There's no end to the possibility of error.

Must ideas align?

Edwin:

Right. Yeah, so really the point is approximating and it will always just remain approximating. We should be aware that as all our knowledge, this is also fallible. But maybe I'm leaning into it too much—that ideas, the explicit, inexplicit and unconscious ideas must influence one another and are in alignment. Does it mean that they must not be in conflict with one another or does it mean something else as well?

David:

So, if we take the grammar example, when you're speaking you have some purpose in mind and you maybe have different purposes in mind in different situations and at different times and so on. And generally speaking, the process that delivers the right word or the right turn of phrase is unconscious. You don't know what that process was. You don't know what possibilities your mind rejected.

If I want to say "very," I don't know in general, I don't know whether my mind rejected the word "extremely" when it was thinking about what to say for that word. But we do know because of the generality, because of the universality of Popper's epistemology, we do know that the process, even the most unconscious process, that delivers something like, or that solves the problem, or attempts to solve the problem of what is the right word in this context, consists of—if it's rational—consists of conjecture and criticism.

And so there will be, there will have been unconscious conjecture, unconscious criticism, unconscious explanation, all going in to produce that word "very" or "extremely" and every single word you say has undergone that process.

Are all thoughts creative?

Edwin:

So every utterance and thought is really a part of the creative process. There's no such thing as something that just automatically pops up or you deduce in your mind.

David:

That's right. Well, there is... So you can do like an LLM. You can say, what's your unconscious mind or your conscious mind? You can say, what's the most likely thing that I will say? Okay, stop there. Now, I know probably without much conjecture and criticism that the next word is likely to be "next." But in fact, in what I just said, it wasn't—because I interpolated an explanatory sentence before the word "next." And that was created by a creative process. If I hadn't done that, then I could have just worked out that the next word is likely to be "next" without much creativity.

Edwin:

But—okay. Yeah, I understand the fact that it can be creativity. I'm just trying to wrap my head around the fact that it's always creativity when we're uttering, or usually creativity.

David:

It's usually creative. Well, when it isn't, then it's more along the lines of... Often the thing that is being brought to our attention and transmitted to our mouth and so on is not just a single word, but it could be a phrase or it could be a stock argument where we've been here before, we know what we're supposed to say in this situation, so we say it. And in such situations, we might be actually thinking about something else and just speaking automatically, and that can sometimes be embarrassing.

Edwin:

Exactly. Or like when you're trying to recite a multiplication table—if I asked you what is five times five, it probably will just be a straight thing we would just memorize.

David:

Yes, yes.

Edwin:

Interesting. This is maybe a bit off topic, but in our minds, is all knowledge that we have, would you say explanatory or are some ideas non-explanatory?

David:

Some of them are non-explanatory. I hesitated because conjecture and criticism are explanatory processes. And so our manipulation of ideas always involves some kind of element of explanation. But the ideas themselves, you know, if I remember something like what I had for breakfast this morning or what year the Battle of Hastings happened in—the memory may not be explanatory but as soon as it enters into a problem it will become explanatory. It'll become the subject of conjecture and criticism with the intention of solving a problem, using the method of finding a good explanation and so on.

Edwin:

What I also struggle with—I think that all knowledge that humans come up with is indeed from the creative process. The only process we have to come up with ideas is by forming explanations and now some of those may be bad explanations, I think you would call them, where they're just justified to believe or just memorized and there's no logic behind it, no explanatory mechanism that we refer to. But they all came about creatively. So in a sense, you could say that they're all, their origins are explanations. Would that be accurate?

David:

Yes, we do have inborn ideas as well, so their origins are evolutionary.

Edwin:

That's true. But even those ideas when they get surfaced to the mind, I think that's still always an interpretation process, right?

David:

Yes, yes, yes, and quite right, and that is explanatory.

Different modes of creativity & learning

Edwin:

Fascinating. Alright. We did touch also on briefly the different modes of creativity. So you mentioned unconscious conjecture and unconscious criticism and inexplicit conjecture and inexplicit criticism. I was wondering if we could maybe try to go through those all and then picture how it would work and maybe if there's some examples in daily life we can look to for those.

David:

Yeah, sure. Although note that this isn't a rigid taxonomy. This is just in a very approximate way of trying to understand what's happening in the mind. The most obvious differences between different kinds of ideas. It's almost on the level of Sigmund Freud's, you know, the ego and the id and the superego and so on, just because we don't understand the mind in any detail. But yes.

Edwin:

That's actually a good addition—the fact that, well, we can approximate them by knowing what is explicit and what is inexplicit and unconscious, but there's no quantized knowledge particle in our mind that we can say draw a boundary of. It's just an approximation.

David:

Yes, that's right. That's right.

Edwin:

Yeah. But nevertheless, I think explicit conjecture and criticism is the more straightforward one.

David:

Yeah, well, explicit is just expressed in a language. If I'm thinking when I'm writing a book, I'm thinking what the next sentence should be. I'm thinking what the criticisms of it might be and how to reply to those criticisms. And for all those things, I'm creating words and sentences in my mind.

But when I'm walking along the road and walking along the pavement and see that there's an obstacle ahead that I might trip over, the word "trip" might not appear in my thought at all. It might all be inexplicit. And I'm just walking along and I recognize the problem and I solve it and I step over the obstacle and all that is happening without my—without the word "step" or "obstacle" or so on ever entering my mind. So that's an inexplicit process.

But there are also unconscious processes because what I just described is an example of an inexplicit but conscious process, because somebody could say to me, why did you avoid—why did you make a large step there instead of a small one? And then I could remember that it was because I saw there was an obstacle. But there are some things which one does unconsciously. Maybe a simple example is, you know, scratching your head and you don't notice that you scratched your head and somebody will say, why did you scratch your head and you'll say, I don't remember doing that. Because it was unconscious.

Edwin:

And what would unconscious criticism look like?

David:

Well then, so criticism and conjecture come when there's a problem. If I, perhaps this is a forced example but we may as well stick with the example, if I was going to scratch my head and I found that I couldn't reach, then I would think that there's something wrong. It might be that my shirt has caught on a thread of the armchair. And so I can't lift my arm and then I will be thinking, again, it might not be explicit. I might just pull on the thread and break the thread and then carry on without really knowing that I've ever done any of that. But it might be that I think explicitly, what the hell is that? And then I turn and I think, it's not that I've suddenly got a neurological disease, but it's just that there's a thread attaching me to the chair.

Edwin:

Okay, that's quite clear. When I think of unconscious conjecture and criticism, I imagine like, let's say I moved to England, I think I would pick up in a matter of months, some slight accents or just particularities of the language, but I would not be conscious of it. Would you say that constitutes also unconscious conjecture criticism?

David:

Yes. Yes, well, people who want to get really, really good at a foreign accent usually have to bring to bear all their weapons. They have to be consciously and unconsciously engaged with getting the right accent. And one very rarely wants to do that, though. In real life, you just do it automatically without ever being aware of it. And I have known people who have spent relatively short time in another country and come back with an accent and then they're not aware that they acquired an accent there.

Edwin:

I think it's an added mode of learning. You could of course try to learn a language by staying in your home and just reading explicitly the rules of grammar and memorizing the vocabulary. But you would ideally also add an inexplicit mode of learning where you're just actually talking. And the third layer might indeed be just immersing yourself within an environment where unconsciously it will also seep in.

David:

Yes, yes. And the knowledge you're trying to acquire that way includes all three types of idea and therefore it stands to reason, I think, that it's going to be more efficient if you allow yourself to be acquiring all those simultaneously.

Edwin:

Yes, interesting. Have you also given thought to the instantiation process of learning? There's of course the error correction and getting an improved theory of, for instance, grammar, but how that gets ingrained or stored—is that also different across the three modes of knowledge and learning?

David:

Well, I don't know. This is like the implementation of knowledge in the brain is something that I think very little is known about. So we know it's a neural net but that really tells us nothing, literally nothing, because the neural net architecture is universal. So it could be doing anything.

Edwin:

Yeah. I guess we can point to some differences that there are. For instance, if I just learned your name today and you correct my pronunciation of the name, I would probably remember that very quickly. But if I had mispronounced it for, I guess, years, it would take a while for that to be fixed. I guess maybe that also—

David:

True, true. Usually, yeah.

Edwin:

Usually true. Yeah, maybe I get a very rude awakening. I'm very aware, acutely aware that I would never make that mistake again.

Evolution of ideas in the mind

Edwin:

Another thing that was also in the fun criterion video was the fact that explicit ideas can directly confront each other through logic and experiment, but more broadly all ideas, so every one of the three types we have been discussing, evolve together. I would love to understand what this evolving together means. Is it like, do you mean that whenever new ideas are created, they reshape a sort of pool in the mind, similar to how there's a gene pool for genes?

David:

That's right. So they affect each other either via the outside world, so that, for example, if you find yourself doing something and you can't explain it, or you think you shouldn't be doing it, or you think you should be doing more of it, then your explicit thinking about that will be part of the environment in which that behavior is evolving.

And again, if you're having fun, if different strands of your thinking are consistent with each other and are helping each other and so on, then when you—usually, again, should say usually in front of everything I say in this context—usually when you decide you want to do more of a thing, then you will find yourself doing more of it. Not only when you think, well, now I will do the thing, but also when you're not thinking that.

And vice versa, if you think I shouldn't do that, then usually, again, if you're having fun, if you're like playing a video game and you don't want to get caught by the bad guy that lurks behind the pillar, then next time you encounter that pillar you will not be so vulnerable, not just because you have decided, well, or rather, the fact that you have recognized explicitly that there is a danger there will be also affecting your inexplicit behaviour of your fingers when you approach such a pillar.

Edwin:

Right, so a change in any of the types of knowledge that you hold can cascade into, or likely does, change everything. So if you adopt a certain value, then that might create new conflicts, but it might also resolve other conflicts in the mind.

David:

Yes. Yes.

Edwin:

Yeah, that's a very good example. I wonder, the mechanism of evolution itself, have you given thought to what may be causing that to happen in the mind? For instance, like what specific selective pressures there would be or...

David:

Well, it's got to be conjecture and criticism. Criticism is always a problem. It's to solve a problem. So there are at least two conflicting theories. And then the criticism involves trying to modify one or both or invent a new theory to replace one or both. One is also implicitly appealing to a criterion because one may be mistaken that they conflict. It may be that that's not where the problem is at all. It may be that the two ideas that one thinks are the problem aren't the problem. And then one would have to change the criterion by which one judged there to be a conflict.

Edwin:

Okay. So, it's the evolution in our minds always then through creativity—because of course, when I make the analogy with, I guess, genes, there's also just selective pressures of which one survives better. That's more mechanical obviously.

David:

So that's entirely mechanical, but when we're talking about ideas in the mind being changed within the mind, so I think the lesson of Popperian epistemology again is that all rational ideas are conjecture or criticism. There is no other kind of rational thought such as was thought, for example, confirmation or induction or direct observation. Those are just like if one relies on those one is being irrational. But even when one is being irrational some rational process has chosen that path even though it's a wrong path and is unlikely to lead to the growth of knowledge. But so—all rational thought is either conjecture or criticism and in this extended sense one can say that all human thought is either conjecture or criticism. Even if one doesn't know that that's what it is.

Edwin:

Hmm. I'm just wondering if there's some underwater component to our mind where maybe there's some more mechanical churning going on there.

David:

The process must at the atomic level or at the cellular level consist of non-creative processes, must consist of purely mechanical, automatic processes. So, creativity is an emergent property, an emergent high-level property of non-creative, low-level processes.

How (poorly) we understand biological evolution

Edwin:

I recently looked into how well we understand biological evolution and these more mechanical processes and evolutionary algorithms. And how poorly we understand biological evolution—I was really quite surprised by how little we seem to understand about evolution in practice. It seems that the only thing that we have been able to model are knowledge in the sense of optimization strategies, like finding economical solutions or efficiency gains and specializations. But they never generate new forms of organization or organs or body plans or biological processes. Only if humans encode or hard code some of those guardrails or rules, then those come.

David:

Yes. Yes, so as I said in one of my books, I forget which, we can't even simulate biological evolution on a computer, even though we think we know everything about how it works and what it does. We still can't do it, so it would really be remarkable if we could simulate human thought or execute human thought on a computer when we don't even know how to do evolution.

Edwin:

Do you think it's the same problem that is the reason we don't have AGI and the reason we aren't able to create evolutionary algorithms properly?

David:

I don't know. I don't know the answer, I'm afraid. It might be that there are two problems. It might be that one of them is necessary to solve the other. Or it might be that there are two problems and either of them can be solved independently. I don't know. I don't know. It's just too hard and just too few people working on it. If I may complain a little.

Edwin:

Yes, unfortunately. No, that's good. We need more people on it. But I think a core component of that is like the non-randomness of that also seems to be inherent even in biological evolution. Like that it's more so targeted, right?

David:

Yes. Biological evolution is in some sense exhaustive search or not exhaustive search, random search in the total space. That has got to be misleading because the chemicals, the DNA for example, already constrains, it already places a structure on the space of possible codes for proteins. The DNA, searching randomly in DNA space is different from searching randomly in protein space, for example. So the DNA, it might be that the DNA has got some property that we don't understand over and above coding for genes.

Edwin:

Yeah, there must of course be something else going on and that might be more targeted than we give it credit for.

David:

It may be. Well, "targeted" sounds a bit supernatural or something. It's not targeted. I mean, whatever it is evolved itself. It evolved. But it may be more than just information because the mode of storage privileges certain types of information over others. All methods of information storage privilege some information over others. So DNA does, presumably.

Edwin:

Okay, yeah. When you discuss creativity in humans, you often also describe it as the opposite of randomness. It's really specific to solving a specific problem.

David:

Yes, but that's at the high level, that's at the very emergent level.

Creating knowledge vs. learning knowledge

Edwin:

That's different. Yeah. This may be related to this—is there a clear demarcation between new knowledge and learning existing things? As humans, we use our creativity to form all our understanding of both just learning existing knowledge, but also to create new knowledge.

David:

Yeah, so fundamentally, at the level of epistemology, there can be zero difference between them because there is only one way of acquiring knowledge, conjecture and criticism. Learning Einstein's theory of relativity requires creativity in the same sense that Einstein required it to create the theory of relativity.

If we learn relativity from Einstein, we're encountering a different problem from the problem that Einstein—Einstein wasn't acquiring it from someone else. Einstein was acquiring it from the problem situation. And his problem situation was different from our problem situation because ours includes Einstein.

But what isn't different is that acquiring this knowledge 100% involves creativity. It is not a process of downloading the theory, the contents from Einstein's brain into our brain. It's a process of us guessing what Einstein is going on about and in particular what his problem is.

Edwin:

Yeah. I think the reason I ask also is you also are fond of work on the foundations of physics. But it might be good to be able to demarcate or at least try to approximate like what is the difference then that regular science does versus indeed what foundational science does. Is that possible to approximate?

David:

It depends on your point of view. I mean, there is ultimately—science shades over into engineering when you're thinking how can the laws of physics be applied in a certain situation and then you might actually—let's say you're trying to work on controlled fusion. You may know some equations about how plasma behaves, but in the course of your research you may want to try a different equation and say maybe that isn't the best way to describe what plasmas do, because we don't have a—we can't deduce it from first principles because it's too complex.

So I suppose what I can say is that there are certain kinds of research that are more mechanical than others. But there are none that are purely mechanical. Well, I suppose those that are purely mechanical are the ones that LLMs can do.

Knowledge is information with causal power

Edwin:

You occasionally describe knowledge as information with causal power. First of all, is that still the definition that you'd like to give for knowledge? Have you updated your position on it?

David:

No, that's my current definition. As you are indicating, I have gone through several definitions and they're all, you know, they all kind of describe the special property of knowledge compared with other kinds of information, but they none of them completely capture what knowledge is. And I think if you were to completely capture what knowledge is, that would be equivalent to solving the problem of AGI and so on.

Edwin:

Then I think within knowledge there's the category of explanatory knowledge and is that then where the causal power is expressed in a code?

David:

Yes. Not in the sense of code being explicit. The knowledge might have causal power, might have very powerful causal power without being expressed in language.

For example, a great tennis player might know how to win, might know how to win in a given situation, might know what is a mistake in a given situation and so on and yet be unable to express that in words, you know. When the great tennis player writes his memoirs there'll be a chapter called "How I Did It" and the people who read that chapter will not become great tennis players as a result. So it's knowledge and it's explanatory knowledge because he can guess what it is. He can say, you know, always follow through. The tennis coaches say, you always follow through. You can have a theory of why following through works. It does seem to work. But ordinary physics can't be telling you that the behavior of the racket after it's hit the ball can have an effect on the ball and yet following through does have an effect.

Edwin:

I guess expressing it in language is even not necessary. Before we had language, I assume we were still explaining concepts to each other, inexplicitly with gestures or small movements. But this is not necessarily conveyable explicitly.

David:

Yes, yes, and for that matter when we do convey it in words there's still a component that isn't in words.

Edwin:

Yeah, it's always that.

The Self and how the parts of the mind are organized

Edwin:

There's one particular tweet that I really love of yours, I hope to unpack it. And that is: "The self is the collective term of the creative institutions of consent among the multiple strands of creativity and criticism that constitute a mind."

Like every word there I think is interesting and deserving of its own attention. For instance, the creative institutions of consent. With "institution," do you have a specific meaning?

David:

Yes, again, yeah, I'm using it as usual in Popper's sense, where an institution is not necessarily a building with a large marble entrance and so on, and it's also not an institution in the sense of a law book which defines the purpose of such a building and so on. What Popper means by an institution is a kind of idea, it's a kind of theory or idea or habit or practice which governs the way that other ideas within the mind interact with each other.

So these creative institutions—which I forget what the wording was—but those are kind of ideas whose job is to manipulate other ideas and make sure that, or try to make sure that impediments to the growth of knowledge get corrected.

Edwin:

So institution is just a type of knowledge that is a facilitator of creative processes, maybe a set of rules or tradition, custom, yeah.

David:

Yes. Probably the most important ones are inexplicit. It's so hard to describe. The scientific mode of thinking is not inborn, and it is a mode of thinking that helps to avoid both errors and conflicts between ideas. And it can be internalized, meaning that it can be made inexplicit and unconscious when it starts off as conscious.

So usually we want to go the other way, but in the case of science, we want the explicit definitions that, you know, how would you know if that was false or how could you find out if that was false? That kind of idea should be coming to mind and it should be coming to mind more often than our genes tell us to have it come to mind. The scientific frame of mind is better than our natural frame of mind, the one we evolved with.

Edwin:

So I can interpret creative institutions to also contain any modes of thinking and different modes of criticism. Those are also encapsulated in these creative institutions of consent.

David:

Yes. Yes, and modes of conjecture and criteria. Yes, all of that.

Edwin:

And you also describe in a comment to this tweet, you described a mind as analogous to an anarcho-capitalist society, where all exchanges are consent-based and voluntary, as opposed to coerced. Would you still think that's the case?

David:

Well, so it's—it has a lot of truth in it, but I wouldn't, you know, it can't be an exhaustive description because otherwise an anarcho-capitalist society would be a mind. And it's not. And also, it's a bit more like a—I hesitate to say a government—the self is not just a neutral coordinator, it also forms judgements. So when you're thinking, "that isn't me," then you're forming a judgement. It really means that that shouldn't be me. I'm not that mean, or something.

Edwin:

Indeed. There is also just a single "I" in this case, a single self.

David:

In a rational mind, there is a single self in that conflicts get resolved and when they get resolved then at the resolution there's only one idea left so that there are no chronic hangovers of the rival theories left at the time when one acts or does whatever the problem was about.

But people often report having conflicting voices in their mind, either, you know, really voices that sound like voices, but also conflicting ideas that present themselves and won't go away, even when other parts of the mind would like them to go away. And so, you know, there's something irrational going on, something anti-rational going on there.

Edwin:

Oh, interesting. When you hear multiple voices, that's a symptom of the rational process being disrupted in a way. Going wrong. There's lots of theories about is the self real? Is there a single self or distributed?

The way I think of it is that I think in our consciousness, whenever we are making a judgment or a theory, there's always just one "I" at that moment.

David:

Well, some people report that there can be more than one and they're both yelling at each other and so on. Note that we're universal so that we could host several selves. If we couldn't do that, we wouldn't be universal. But it's normally—usually, I have to say again, usually—having several selves which can't resolve their differences is a sign of something going wrong.

But let me hasten to add that the picture of the mind as being—a rational mind as being normally completely consistent so that every part of it is consistent with every other part and that the object of rationality is to locate any deviation from that and eradicate it—that's very wrong.

My picture of the mind is that it is a vast collection of possibly conflicting ideas. They become problematic not because they logically contradict each other, but when there is a problem in which more than one idea is used, for example, as a criterion and they conflict with each other. And then that causes a problem where one has to invoke a creative process. And that's a good thing in general. Because that's where new knowledge comes from. What's bad is if reaching that state is being thwarted by something.

Edwin:

Right. So the normal healthy state is our minds are very diverse. We have different preferences, opinions, and they can be seen as all like independent parts of the mind and the way they interact with one another or do the conflict remediation is through the "I," the self that is arbitrating this process in a healthy way.

David:

Yes. Yes. Although as I say that in that tweet, I also don't want to think of the "I" as being a thing that lives in a particular part of the brain and it gets invoked when there's a conflict like a policeman. It's just a descriptive thing. Conflict resolution, problem solving happens and the aspect of problem solving that leads to unanimity or that aims for unanimity we call the self.

Edwin:

That's a great way to put it. I like that.

Focus and attention

Edwin:

One thing I also think about quite a bit is like the properties of our minds of focus and attention. So sometimes we just feel more lethargic and have less attention and sometimes we feel energetic and have more attention. Is that loss of focus in your view better understood as like a software property or more like a hardware energy source depletion? How do you think of things like focus and attention?

David:

I don't know. It might be a hardware thing, it might be a software thing. This is all, as I said, I don't think we really know how thinking is organized. But I know that there are people who can walk and chew gum at the same time. There are people who can play the organ with both hands and both feet contributing and they're singing at the same time and at the same time they're composing their shopping list from when they've finished their organ practice and they can do all that simultaneously.

Edwin:

It's true. If you read into the scientists that describe focus, they often equate it to random access memory. So you have a finite amount of things you can keep in mind. But as you just said, that is not universally the case. There's quite often outliers and it's also a trainable skill. As a Popperian, what would be a good way to look at focus? Is it just a type of knowledge that we can hone like any other skill or should we treat it more as a bodily process?

David:

Well, there's certainly—the organist example shows you that it is trainable, is learnable to some extent, but to what extent this is constrained by hardware properties of the brain? I don't know, because I don't know how it works. I don't know how the sort of natural organization works.

But obviously we can simulate parallel processing with serial processing, if we want to, but that will slow it down. But there's no kind of thought that we can't reach because of hardware limitations, but there might be kinds of thought that are faster because of hardware or slower because of hardware.

Edwin:

Because I think it's quite a common occurrence that if we're sick or if we just had a bad night of sleep, our creativity definitely feels like it's hampered or at least severely more difficult. So there's definitely an interplay of hardware.

David:

Yes. Yes.

The roots of inexplicit knowledge

Edwin:

In a conversation with Lulie, you guys also discussed that all explicit knowledge is rooted in inexplicit knowledge. Do you recall this particular conversation?

David:

Yes, yes. As we've already said, explicit knowledge is expressed in words, but the meaning of the words is not expressed entirely in words. If it were, I mean, it can't be, because if it were, that would be an infinite regress. Dictionaries could never define anything.

That again is a problem with LLMs because they don't have a root level of contact with reality. They only have words and sentences in terms of other words and sentences. Once we have LLMs in robots, where they have some contact with reality as well, then they might have the equivalent of unconscious ideas.

Edwin:

That's interesting to think about because also in an LLM it doesn't know what it's doing so in a sense it could also be seen as inexplicit knowledge.

David:

It is explicit. Everything about the LLM is explicit. It has an architecture. I suppose, yes, I suppose that its architecture contains inexplicit knowledge or explicit knowledge that was put in by the programmers, but the idea is to free it from that as much as possible, especially if they're hoping to make an AGI—in vain, I think. But they're trying not to let the hardware determine what it can think.

Edwin:

Alright, and when you also say that explicit knowledge is rooted in inexplicit knowledge, that also—can I just think of it as a different type of coding language? So explicit knowledge is of course expressed in a set of symbols like English, but that—yeah.

David:

Yes, so the state of the brain must at some level be expressible as a large integer, ultimately, and a language. Yeah, but it's probably not useful. Yeah, even the inexplicit things. But that's not what we mean by explicit.

Edwin:

No, no. But if we would have to approximate a coding language for inexplicit knowledge, it would not be universal, I would assume. Or would it?

David:

Well, I would guess it would. I mean, I don't know again, but I would guess that the low level machine language of the brain is itself universal already. That's not necessarily the case because it might—in any case, a different level of universality happens at every level of emergence.

We know that we are universal because we can simulate a Turing machine in our minds, in our imagination. We can simulate a Turing machine up to some level of complexity. So, but when we do that, we're not using the universality of our hardware to make Turing universality. We're re-implementing Turing universality after many, many levels of emergence.

Edwin:

You have these cliche examples where they call it the "je ne sais quoi" or the example where they say, "I cannot define pornography, but I can tell you when I see it." I think these are forms of inexplicit knowledge that we can't really pin down very well with words.

David:

Yes.

Physical sensations vs. emotions

Edwin:

Is it meaningful to distinguish between physical sensations happening in mind and emotions? Like are those two categories meaningfully distinct and should we treat them maybe differently when it comes to the mind?

David:

Again, I don't know. It seems reasonable that they are different because although due to universality we can manipulate either of them, it seems not only that—so, manipulating physical sensations to make yourself say not feel pain when an external stimulus would normally give you pain—first of all, it's very difficult. One can't just decide to do it. One has to exert a huge amount of creativity in order to do that.

But more importantly still, it seems to be a different kind of process from, let's say, when one is sad, comforting oneself. That also requires creativity, but it seems to be a different type of process. So, they might be different, they might be at some underlying level the same. I don't know.

Edwin:

I think the physical sensations are largely innate, maybe even hard coded in the sense that we really can do very little about them. Whereas emotions are much more constructed. So we create knowledge and that knowledge may be a fear and that fear then signals a sensation.

David:

Well, yes, but yeah, that's plausible. Sensations have to be interpreted before they're experienced. You know, one can train oneself to acquire perfect pitch and then musical tones sound different from then on. One can train oneself to distinguish more shades of colour than one naturally did and then one becomes more competent as an artist or whatever. And certainly in many situations one can train oneself to modify sensations. But it requires creativity to do it.

Edwin:

True, so modifying the sensation itself. I understand the concept of our reaction or response to sensations being malleable, for instance, like learning to enjoy the pain of running the last five kilometers of a marathon or something like this. But that sensation itself, I'm not sure. Yeah, I'm trying to understand if we have similar access or malleability to instruct that from happening or not—for that pain signal to arise or not—as we would with an emotion.

David:

Well, yeah, so what is much more common, as you are pointing out here, is that we can modify whether we enjoy the situation, whether we enjoy the sensation or not, without necessarily altering what it feels like. Though I suppose that whether we enjoy it is a component of what it feels like.

But I think one can also—there are stories of soldiers in battle who get shot or have their arm blown off and so on and they don't notice until the end of the battle.

Edwin:

Yeah.

David's upcoming book

Edwin:

I do have a closing question about your upcoming book on reason, but I don't know if you feel like going into that.

David:

Well, it depends what the question is, so ask it.

Edwin:

Okay, okay, great. I won't ask you how far along you are or when it will be finished, but I was wondering whether you could share a few of the problem situations you plan to explore and address in this book.

David:

I could give a glib answer and say all of them.

The idea of the book is not to distinguish between rational and irrational types of thinking because at least that's not really central. I can only say a few things about that because only a few things are known about that. But it's about how to think about—can I call it—the human condition in the universe, which is a condition of being a thinking being.

Like I start off, at least I start off the relevant section by talking about the hierarchy rule, which I've sometimes spoken about in public, which is that for the rest of the universe, big, massive, powerful things affect small, non-powerful things, but not vice versa. Whereas when there is knowledge present, it's the other way round. It can be regarded as a—when we say knowledge is information with causal power, we really mean its information with the power to violate the hierarchy rule.

So that's another way of looking at what knowledge is and explanatory knowledge. Although explanatory knowledge has not yet had as much effect on the physical world as biological knowledge, it is in principle infinitely powerful, whereas biological knowledge is in principle finitely powerful.

So that's a bit about what it's about.

Edwin:

Well, thank you very much, David. I really enjoyed the conversation. I did learn a lot.

David:

Yeah, fun conversation about fun.

Edwin:

Yeah, great. Thank you very much, David, and have a nice day.

David:

Same to you.

Outro

Edwin:

I hope you enjoyed the conversation with David as much as I did. If you haven't read his books yet, I highly recommend The Beginning of Infinity and The Fabric of Reality. They are exceptionally rich books, the kinds of books you can re-read for years and still keep finding new ideas.

If today's conversation resonated, especially understanding the nuances of how the mind works and how that can guide how we think and act, you might like my upcoming book The Four Acts. It's a four-stage framework you can use on any problem in everyday life and today's topics directly feature in it. For example, the fourth act, learning, unpacks the three modes of learning we discussed today—explicit, inexplicit and unconscious learning.

The book is nearly finished with a release planned for 2026. You'll find a pre-order page in the description. And if you want more in this vein, subscribe to the channel to stay up to date.

Thank you for joining me today and take care.