Richard Dawkins and Steven Pinker predict the future of religion!

Thank you. Thank you. Thank you so much for being here for another sold-out event here at the Royal Institution Theater. Every single show we announce just gets filled up so quickly. So, thank you for being here.

Again, everyone working behind the scenes tonight is a volunteer, and the show's only made possible by the LGFG Fashion House, who are sponsoring tonight's event. So, please join me in thanking all the volunteers as well as LGFG for making tonight's show possible.

[applause]

I have the absolute pleasure of joining you in the audience tonight because I'm filling in for a producer, someone on ticketing, uh, someone doing audio as well. So, I've got many hats on tonight, but that means I get a wonderful break to sit down and enjoy Steve's and Richard's conversation. I'll join you again for the Q&A in just over an hour's time. But until then, please put your hands together and give a massive warm welcome to Steven Pinker and Richard Dawkins.

[applause] [applause] [applause]

[applause]

Well, thank you very much. I was hoping to see a clock, but I can't see a clock anywhere. I don't know. Somebody will no doubt signal when we're when we're running out of time.

Um, well, we're supposed to be looking at the future and at the past. The future is a bit of a risky thing to do. I'm going to read a couple of of um, bad experiences from the 19th century. Lord Kelvin, father of thermodynamics, said radio has no future. Heavier than air flying machines are impossible. X-rays will prove to be a hoax. And uh, another uh, pessimistic writer, William Henry Pickering, around the same time said, "The popular mind often pictures gigantic flying machines speeding across the Atlantic and carrying innumerable passengers in a way analogous to our modern steam ships. It seems safe to say that such ideas must be wholly visionary, and even if a machine could get across with one or two passengers, the expense would be prohibitive. Another popular fallacy is to expect enormous speed to be obtained." He goes on to do some authoritative calculations to show that the effect of air resistance would be to prove that an airplane could never travel faster than the express trains of his day. So we may get on to the future, but with due caution.

Um, but meanwhile, to the past and the present, I thought what I'd do, Steve, is take a bit of inspiration from each of your books. Um, all of which I've read. Um, starting with *The Language Instinct*. I think it's fair to say that the thing that really singles humanity out, probably, is language. Uh, would you agree with that?

[clears throat] Certainly the most conspicuous trait that we have that uh, that other animals don't. And from it flows so much. The standing on the shoulders of of giants, in Newton's phrase, um, depends upon language. The whole of culture, the whole of um, engineering, the whole of technology. I'm curious to ask you, as a linguist, some questions about it. First, what's your current thinking on when language started?

So, um, language is, as I mentioned, it's most conspicuous, um, zoologically unique trait. Uh, although I don't think that language was kind of grafted onto the brain of a chimpanzee. That uh, we wouldn't have any use for language unless we also had a a hyperdeveloped ability to model the world. That is something to talk about that. And and here I'm drawing from a a uh, brilliant paper by my late friend John Tooby on um...

Who just died, didn't he?

Who died just a year ago? Yes. On called "The Cognitive Niche." That being the um, way of surviving and reproducing that our species, or our ancestors to our species, entered. And this includes language, but it also includes the ability to uh, model the environment, to engage in uh, cause and effect reasoning, to do mental simulations, to have intuitive theories of physical objects and living things and other minds, which gave us the advantage of being able to uh, outsmart other animals and indeed plants. Because, as we know from from your writings, every living thing uh, is in competition for it its own life and uh, has numerous defenses against being uh, eaten, which is what we want to do with them. Uh, and so our ancestors developed ways of um, trapping animals, of ambushing them, of driving them, of detoxifying the poisons in plants, a lot of acquired know-how, which then uh, gives us something to share via language. That is, we don't just jabber, but we we teach, we learn. And uh, knowledge makes language that much more valuable because it uh, because information is a non-rival good. That is, unlike stuff where if I give it to you, I no longer have it. If there's an idea and I give it to you, I still have it. Uh, which also, and to to pick up a theme of of uh, *The Selfish Gene*, it means that it's an ideal um, basis for reciprocal altruism. Because knowledge is something that confer can confer an enormous benefit on someone else at a small cost to oneself, because in sharing knowledge, you don't give it up yourself. Uh, so uh, and that introduces the third part of the uh, I think of it as a triad in the cognitive niche, although it's many other features of our of our lifestyle as well. But but language, cognitive know-how, and sociality. We are also zoologically unusual, as that's again a major uh, a major theme of *The Selfish Gene*, in how much we cooperate with um, organisms that are not biologically related to us. So it can't be explained by shared genes, but rather by exchange of uh, of of uh, benefits, of which knowledge is the ultimate trade good. And um, so the fact that we cooperate, which means we are quite literally on speaking terms. That is, we uh, are willing to engage in conversation with with uh, others, which itself presupposes coordination. When you have two opposing coaches prior to a football match, they don't engage in in chitchat. They have no one want, neither of them has anything that they would be prepared to share with the other. You have to cooperate in order for uh, for communicate, for at least for um, language to uh, to to be useful. And uh, the know-how gives you something to share. So I think of the cognitive niche as um, at least those three. They're probably other unusual features of human.

But there are aspects of that that other other species have. But language is unique in having a huge vocabulary and also in having recursive grammar, which enables a sort of indefinite expansion of ideas in a way that no other species even approaches. And what I'm wondering is whether um, I I've always been an ardent gradualist in evolution and I've never believed in any kind of evolution by jerks.

Yes.

Um, [laughter] a little private joke. Uh...

Um, this was this was coined by John Turner to describe the theory of punctuated equilibrium by Steve Gould. Now...

Yes.

Um, but in the case of recursive grammar, this does seem to me to be a possibly all-or-none, sudden jump. Because if you look at computer languages, the computer language either can do recursion or it can't. And it's not a gradual, you don't do sort of half recursion or quarter recursion. You either can do it or you can't. It's a sudden jump. And I'm wondering, could that have been a macro mutation?

Yeah, I I doubt it. And and and and I'm I'm sorry that I I never got around to answering your question of when language have emerged.

Uh, seems almost certainly before the um, split with Neanderthals. So, um, that puts it back to anyways somewhere between half a million and a million years. But I suspect prior to that, um, you quite possibly with with uh, around the time of Homo erectus, with at least the the rudiments.

Do you think spoke? Do you think spoke?

I guess they they don't show the sign of um, huge [clears throat] variety in culture as far as we know. Since their c- the cultural artifacts they left us, such as the hand axes, were very stereotyped.

We didn't really take off until about 50,000 years ago in that respect, did we?

Well, true. There is an explosion there for. And of course, it's debated what h- happened in what Jared Diamond called the Great Leap Forward.

Yes.

Um, 30 to 50,000 years ago in Europe, but probably earlier in Africa. And that may have been a co- there there was a theory that there was a great cognitive leap. Um, Richard Klein, the anthropologist from Stanford, suggested that. But there another factor may have been sheer population size. So that if you have more people u- [clears throat] coming up with ideas, there's just a a richer source. Um, so we don't really know. But there the um, even before the uh, explosion that in Europe we associate with Cro-Magnon, um, there there were um, going back, I think, 80 to 100,000 years, um, a variety of forms of tools that hadn't been seen before and and cultural variation that at least plausibly could have depended on transmission of information among individuals. With so, so I'm I'm now kind of pulling back on on the possibility that Homo erectus had language, had a rudiment of language. It's possible, but we don't see as much evidence for it in say, cultural variation that would have um, been a sign of individually transmitted uh, information.

But just going back to whether recursion requires a macro mutation, which I think both of us are skeptical of, uh, just because [snorts] uh, any big change is unlikely to lead to anything particularly adaptive, simply because it's of sheer improbability. Whereas a sequence of, and here [snorts] I'm feeding back to you the argument from uh, *The Blind Watchmaker*, that one wouldn't ex- even if large mutations are possible, large m- uh, changes in adaptive um, uh, function, that is, an organ, a body part, a cognitive ability that has many working parts, since it does need many working parts, it's very unlikely that one random change in the DNA could have put all of them in place. Whereas if it was more of a hill-climbing procedure, set of mutations, each one of which >> improved it in some way, then it's it is uh, within the realm of possib- That's the general argument why why one is in favor of gradualism. But but um, in this particular case...

So it's hard to think of it as a halfway stage.

So I think, you know, I I think in in recursion, if [snorts] truly recursive sentences are actually fishly difficult to understand. And Noam Chomsky, back in the '60s, had had a set of them such as that that that he was sad is unexpected is uh, surprising. Yes.

Now, that's a grammatical sentence, but it feel it sounds like word salad, presumably because we don't have the equivalent of the computational architecture that handles recursion, that is a subroutine stack with a set of addresses, each one of which you can pop back up to when you've fully...

Is only seven deep or something like that.

Yeah. So there is uh, and it may even be that [clears throat] it's not just the depth. Um, that is, in a in a computer, you can have any, you can have a stack of any number of uh, symbols and and uh, peel them off as you complete each task to to recover where you you left it off. Human short-term memory probably isn't just a set of slots that the content matters. So it's almost as if this is an exaggeration, but if you can imagine every um, if every concept that we have learned, every word we have, has a kind of representation in the brain [snorts], then and when we hold it in memory, it kind of lights up or there's a tick mark next to it. If there are multiple instances, they tend to overwrite themselves, and so we get confused very, very quickly. So probably what we have is a rule system that is can be extrapolated to be recursive. That is, it has all of the, if you can say, you know, that that he is sad is regrettable, that is perfectly intelligible. As soon as you have that, you have all the tools necessary to embed one sentence inside another inside another. But we don't really have the cognitive um, capability of of handling it, equivalent to popping a stack in a computer architecture.

You you might have quoted the letter that Luria got, "Dear Professor, I want to tell you that I want to tell you that I want to tell you that I want to tell you that I want to tell you," as an indefinite stack, but never never...

[clears throat]

Yes. So like the uh, recently at uh, at Harvard, there was an anti-semitism uh, initiative, and there someone had an argument against that, and so that was the anti-anti-semitism uh, initiative. And then then someone wrote an argument against that, so it was the anti-anti-anti... So these these do become difficult, especially those aren't so bad when you have what's sometimes called tail recursion. That is, if you just hook something on onto the beginning or onto the end, we can handle that. But when you um, embed a phrase within a phrase, especially a phrase of the same kind, a noun phrase inside a noun phrase, a sentence inside a sentence, um, then that start to starts to tax us. That is when you have uh, a a clause or a phrase, and then you embed it so that there's material at the beginning and material at the end, and then material at the beginning of that and material at the end of that, that become that that's one of the definitions of uh, clumsy prose, if there's too much embedding. Anyway, to get back, I'm sorry, this is a a bit of a shaggy dog story because I have to to answer your question. Could recursion have, does does our ability to handle recursion suggest that there was a macro mutation? And I suspect not, because we don't really have pure recursion. What we do have is combinatorics. We have the ability to um, connect uh, a subject and a predicate. We can embed uh, a modifier inside the predicate. With that system, it has the mathematical capacity to generate a a recursively explosive uh, language. But I don't think we ha we actually have the ability to handle the recursion. And so what we do have, namely combining two things, that is much less surprising as an as an ability.

There have been suggestions in my field of animal behavior that um, animals without speech use some kind of recursive organization in their behavior. The organizing of a hunt by hyenas, or something like that, can be seen. I'm not sure how plausible that is. It it has been suggested.

Yeah. So I think any complex goal-directed behavior might have a combinatorial structure that in a sense is recursive. So if you >> a global goal and then you open a sub-goal and in the service of that, and then a sub-sub-goal, etc., and then you wind a way out of it.

Exactly. Yes. Exactly. Now, the what where that becomes difficult, even for us, is if the goal and the sub-goal and the sub-sub-goal and so on are are qualitatively the same. Then we start to lose track because the thing about handling a goal hierarchy is that when you execute the sub-sub-goal, you got to remember, okay, now it's time to go back to the sub-goal. And when you finish that, then you have to go, or the sub-sub, then you have to go to the sub, and so on. >> That's something that that requires the stack, and that is difficult. So, just to be concrete, because we're talking very abstractly, uh, you know, if you're making breakfast, that's the goal. And a sub-goal is, well, I want to make the coffee. Well, if to make the coffee, coffee, first you got to get the the grounds. To get the grounds, you got to grind the beans. To grind the beans, you got to take them out of the the freezer, and so on. And after each step, you have to remember where you are in the sequence as a whole. Uh, and so that does have a kind of combinatorial structure, which quite plausibly could be the neural biological basis for the recursive ideas we have, which in turn are expressed by recursive language.

I want to move on to another book, um, *The Blank Slate*. Um, you would be regarded as a hereditarian as opposed to an environmentalist...

In the sense that I believe that the role of heredity is greater than zero.

Yes.

Yes. Yes. Exactly. Um, what I find so irritating is the way, and and this comes to another thing. One of your things is my-side bias. Um, you are okay. So you agree with Donald Trump, you know, I mean, it's it's um, it's you you are tarred with the brush of somebody else who thought the the same thing in that one narrow respect. Do you find that is a problem?

That's a problem throughout intellectual life. Yes. I don't know if Donald Trump has opined on nature.

No, I shouldn't say. I mean, that's not No, I didn't.

But yes, that style of argument, you know, didn't so and so say something that reminds me of that? Therefore, you must be in the same camp as so and so. Dog lover. Okay. You agree. You agree with Hitler. You don't...

Yes.

Right. And and you want to engage in, you know, felinocide or something. It must mean you hate cats, too.

Yes.

No, I think that's a pervasive um, impediment to any kind of understanding, that is sorting people into camps, especially by association or by association with association, in how how the mind works. I suppose you you are one of the leading um, evolutionary psychologists. And this is a a discipline which comes in for a lot of stick. I mean, I was once traveling with a philosopher, and and we we getting on fine with him, until I mentioned evolutionary psychology and Steven Pinker, and he started bristling and bubbling over with... What is the problem with it? I mean, isn't it obvious that that evolutionary psychology is going to make sense?

Yes. And and indeed, the reason it makes sense is that it's one of a number of levels of analysis. And here I'll I'll uh, name check your advisor, Niko Tinbergen, who who suggested that to fully understand any kind of behavior, you have to understand um, its neurobiological basis. You have to understand how it develops in the individual, that is, you know, in our case, kids. Um, how it evolved in the species, its phylogeny and its adaptive function. That is, if it is improbably complex, what were the selection pressures that brought it into being? And so evolution is, I don't think, explains everything, but it is one part of a satisfying explanation for any aspect of our phenotype, including our cognitive phenotype. I I think the bristling does come from the chain of associations that I discussed in *The Blank Slate*. In fact, the reason that I wrote *The Blank Slate* was after I wrote *How the Mind Works*, I was surprised at how much of the reaction was politically driven, and I thought it deserved a book of its own. And the the associations, I mean, the heart of *The Blank Slate* was why is the blank slate, which is scientifically utterly uh, implausible, that because blank slates don't do anything, and human brains do do stuff, do a lot of smart stuff. So therefore, they have to have some kind of structure. Why should this seemingly innocuous idea that the brain has structure, and we can see that it has structure under a microscope, we know that most of the human genome is expressed in the in the brain, why should this be such a uh, an inflammatory idea? And I said, partly because of reactions to *How the Mind Works* and other books uh, making similar arguments, identified four uh, touching points. One was the the fear of inequality. That is, if we're blank slates, one politically appealing um, aspect of the blank slate is it means we're all the same. Zero equals zero. And so the whole question of whether some people or some groups are are smarter than others or or more conscientious than others can't arise if there's nothing for which we differ. Now, I therefore, if you are an egalitarian, you must believe that we are blank slates. Now, I mean, and what I argue is that that's a non sequitur for several reasons. One of them is uh, there could be a universal human nature. So there could be a very rich innate endowment, but all ordinary humans have it. Um, but even if there uh, is variation, and anything that is genetic inevitably will admit of some variation, the the commitment to equality is not the empirical hypothesis that we're clones, but it's the moral hypothesis that all people, by virtue of being sentient, of being um, uh, uh, responsible, have equal rights and deserve equal respect. And that moral principle shouldn't hinge on the empirically dubious dogma that we're blank slates or that we're indistinguishable. Another um, politically touchy point about the very idea of human nature is, if we have human nature, then maybe it incorporates some things that we're not very proud of, like dominance and vengeance and lust and and and greed and selfishness, and that would seem to stand in the way of social and political improvement. Uh, uh, wouldn't it be better if we started out as blank slates, which would mean that our culture, our parenting practices could grow perfectly nice children who would grow into perfectly nice adults, and so we could have a a kind of utopian society of perfect cooperation and harmony? Uh, and I I call this the fear of imperfectability. That human nature, the idea of human nature is kind of a nuisance if you are a utopian, because it means that um, you can't just grow the the as they used to call it, the "new Soviet man," or as Mao said, that uh, on a it is on a blank paper that the most beautiful poems are written. Now, it's not a coincidence that I brought up those tyrants, because one could turn the argument around and say the fact that there is a human nature means that there isn't an opening for totalitarians to program the human of their design specs. That you have to respect humans with their flaws, with their their pleasures and pains and and wants and so on. And perhaps ironically, I later, more than a decade after writing *The Blank Slate*, wrote the ultimate book on human progress and human improvement, *The Better Angels of Our Nature*. The title itself taken from Abraham Lincoln, making the point that even if human nature encompasses some rather nasty traits like like vengeance, it's a big complex system. And that's not all it it it houses. It also, we also have a capacity for empathy. We also have the power of self-control. We also have cognitive processes that can engage in problem-solving. And if our uh, our institutions, if our norms bring out our better angels, and we see it as a problem to be solved, how do we live in harmony despite the fact that we are self-centered? We have a long list of of of um, flaws and sins. Then not only is progress possible in theory, but it's happened. And so I I reproduced 100 graphs in that and the follow-up book, *Enlightenment Now*, showing that of all things, I the hereditarian um, and often attacked for being Panglossian and utopian, so on, and saying that progress has happened and it can happen. You are attacked for being overoptimistic. And and the attacks, the attacks come from people's intuition, and and um, yet you have the data. And the the intuition is um, not reliable in this in this respect. But have you been at all shaken since *The Better Angels* was published? I mean, is the world getting worse?

So so I did, yes, I have um, I tried to keep keep track of the data. I got to grit my teeth and sort of uh, take a deep breath before looking at the uh, the data for the for the the last year. Um, and um, and the answer is is is complex, as you would expect, because I I don't actually believe in progress as a force um, as a uh, as uh, when [clears throat] the humorous friend, Louis, once said, "I don't believe in anything you have to believe in." So I I don't believe in progress as a as a force. Quite the contrary. Excuse me. I think that the forces of nature tend to kind of grind us down. I mean, things fall apart. You, second law of thermodynamics applies. Things rot, things rust. Um, there is such a thing as human nature, and it's it's got some some nasty bits. Um, the progress that we enjoyed came from specific accomplishments like, you know, antibiotics and vaccines and democracy and the rule of law and [clears throat] uh, and markets. If those are disabled, the prediction is that we should go backwards, simply because progress is a matter of cause and effect, not a force of nature. So what are what do the [clears throat] data show? Um, well, during the pandemic, we went backwards. There's just no question. And in in the data [clears throat] on longevity, on poverty, on on various good things, uh, we definitely went backwards. But to my, I guess, to my relief, we we've [clears throat] recovered the lost ground, and we're actually, in in measures such as longevity, um, child mortality, extreme poverty, we, this the last year for which we have data was a record in human history. So the say, the graph of say, longevity, shows an unmistakable notch for Co- but we recovered, we rebounded from the notch, and now we're we're a little higher than ever. For other things, we're worse. Democracy has definitely gone backwards in the last years, and uh, and and worse is to come, at least in in in my adopted country. Um, certainly uh, and I I say this not as a snarky remark, but there are organizations that track democracy. Uh, and you can see the curve. We're still, despite the regression of the last 10 to 15 years, which is, you know, wiped out maybe 15, 15 years of progress. We're still way better than we were in the, you know, the '90s, the '80s, the '70s, the '50s, and so on. So the democracy curve, it it went down during the era of fascism and communism. Then it shot up in the um, after the fall of the Berlin Wall, and it's kind of leveled off and sunk a bit. We're still, by historical standards, you know, not bad, but we have gone backwards. For war deaths, there, um, the decline in war deaths really only is really a post-World War II phenomenon. Prior to that, it would be hard to argue that there was any decline in war. But after 1945, after the world kind of decided, well, let's not do that again. Uh, [laughter] at least so far, we haven't. Uh, and it's something of a roller coaster. There absolutely has been in the last two to three years an upward tick because of, now everyone thinks, oh, because of Gaza, but actually, as as horrific as Gaza was, that's actually only a small part of it. There's also Ukraine, but also some enormously destructive wars that we don't read about in uh, Sudan, South Sudan, and before that, in in Ethiopia. So again, we there we've wiped out, you know, maybe about 30 years of progress. We're back where we were in the uh, in the the early '90s. So better than we were in the '70s, '60s, and '50s, but we're not not as uh, good a position as we were when I wrote *The Better Angels of Our Nature*. So there we have gone backwards. Intuition tends to be pessimistic. There are reasons why people respond. You you've given examples of this, reasons why we tend to respond to bad news rather than rather than good news. And and you put out the good news in the form of data, tables, and graphs and things. Um, why do we respond to...

Yeah, bad news?

A couple of reasons. One of them is um, a combination of a cognitive bias and the um, very nature of journalism. The cognitive bias was named uh, the availability heuristic by Amos Tversky and Daniel Kahneman, and that is that we judge frequency, probability, risk by um, basically by anecdotes or narratives or images. That is, if we can imagine it, a particular instance, then we overestimate its its frequency. Um, now, what is journalism? It's anecdotes. It's stuff that happens, not stuff that doesn't happen. Uh, if there's a a year uh, without a war in a particular country, that's not a headline. [laughter] Nothing happened. Uh, likewise, it tends to be about sudden events rather than incremental events. Uh, and so it's easy for a bad thing to happen all at once. It's very rare for a good thing to happen all at once. I mean, you know, the Berlin Wall came down, that's one example, but there aren't very many like that. Whereas terrorist attacks, explosions, school shootings, wars breaking out, those are uh, are newsworthy. [clears throat] Um, if something improves by a few percentage points a year, indeed, even if it compounds, then it can transform the world, but it's [clears throat] never news. And um, in fact, it's even more perverse than that, because if there is a trend that say keeps going upward until a year where it goes down a bit uh, and then keeps going upwards for 10 years, and then there's a bit of a backtracking. Well, the backtracking is news. The 10 years in a row are not news. And so you hear about every backtrack. And you might think things are getting worse and worse and worse, because every time you read about something like longevity, uh, like uh, car accidents, uh, [clears throat] like like anything that that can be me- anything that can be measured, you can come away with the impression that things are getting worse and worse and worse, even if they're getting better and better and better. Um, Max Roser, the uh, the the founder of the uh, invaluable Our World in Data website, said the papers could have had the headlines: "137,000 people escaped from extreme poverty yesterday."

Yes.

"Every day for the last 30 years."

They never ran the headline. And so a billion people escaped from extreme poverty, and no one knows about it.

Yes. So those are that combination is one of the reasons why we're more pessimistic than the data would uh, would would entitle.

Oddly enough, 1984, the Ministry of Truth, which was the ministry of lies, of course, um, was constantly putting out um, "Production of steel has gone up," and so you [laughter] know it was...

Well, so that's another feature of journalism. Ex- there's exactly that. And I've been told this by journalism journalists that precisely because of that, which is based not so loosely on the Soviet Union with the five-year reports and, you know, pig iron is being reproduced at greater greater rates than ever before. Journalists are very on top of just the built-in bias. There's also a negativity bias that's part of their their ethic. Uh, as one journalist put it, "We were taught that um, uh, bad news is journalism. Good news is advertising." Yes. And so there is a feeling that any that good news is supporting the status quo, the establishment, the man. And that if you if the role of journalism is to speak truth to power, then if you say that anything is good, um, then you're kind of propping up the establishment. You're you're not you're not doing your job. So there is the, together with another psychological bias, is the negativity bias, which is that we that the negative does have a bigger psychological impact than the positive. That is, we are hurt by criticism more than we are buoyed by praise. Uh, we uh, there are more uh, negative emotions than positive emotions. There are more negative words than positive words. And when it comes to the past, it's not so much that we remember um, negative events more than positive events, but we forget how negative they were at the time. So there is a a kind of nostalgia built into memory that uh, we don't, people look back and I see this all the time in the papers, nostalgically to the 1970s in terms of it being almost like a a laborer's paradise. Um, at a time when there were energy shortages and uh, life life expectancy was 10 years uh, shorter, there was less, you know, anesthesia in dentistry, we didn't have the internet. There's a lot of, there was unemployment, double-digit unemployment and double-digit inflation at the same time, even though those are supposed to trade off, but we had both of them in the '70s. It was a pretty grim time, but it's often held up, especially to people who weren't there at the time, as u you know, much better for the for for the working class, which it was absolutely not. But people tend to forget, both in their own autobiographies and in retrospective accounts, um, how bad things were at the time, and the negative emotion declines in our in our memory. That's yet another reason why we are uh, kind of wired for nostalgia.

Yeah.

A couple of quotes from uh, *Rationality*. "We evolved not as intuitive scientists, but as intuitive lawyers."

I that resonated with me. [laughter]

"That I I think I I that came from um, I think from from Philip Philip Tetlock. We're intuitive lawyers and theologians." Um, but but yes, so the uh, and there is an argument, a theory, I think goes too far, from Dan Sperber and Hugo Mercier, that uh, reasoning in the sense of step-by-step persuasion evolved in service of uh, winning arguments, not of establishing the the truth. And so we are lawyers in the sense that we take on a defense or prosecution, and we spin doctor or suppress or exaggerate whatever evidence we can to make that the case, rather than having the most objective ascertainment of reality uh, possible. We're probably all guilty of that. We have to fight against it, and scientists have to fight against that as well. They have a a pet theory. They they have to train themselves to to look for evidence against it. And that, of course, is part of the the ethic of science, to to do that.

Indeed. And um, uh, so I think there's a, I think it goes too far to say that reasoning uh, evolved only to, you know, to browbeat, to hoodwink, to to to to spread propaganda, because we do have to, it does have to be aligned with reality enough that we don't just blow it off. But I do think there is a very strong bias, you know, as you say, to advance our pet theory. Not coincidentally, the pet theory shows us in in the best light, not only for having discovered it, but often theories tend to be aligned with the interests of those who hold it. But it's precisely the institutions like science, like um, adversarial system in the court of law, like free speech in a democracy, that [clears throat] are in a sense workarounds for that, some that set of biases, the my-side bias, it's sometimes called, or the our tendency to be um, intuitive lawyers. The thing about being a real lawyer is there's a lawyer on the other side, uh, which means that both of them have to have their their day in court. And one of the reasons that science depends on peer review, open criticism, the absence of a prestige hierarchy as a as something that commends the the veracity of a statement. You, an undergraduate, can criticize a Nobel Prize winner. At least in theory, we that that's the norm we we ought to promote. Um, the reason that's necessary is so that the collective can be more rational than any of the individuals making it up. And and this resolves the paradox with which I began *Rationality*, which is, since Tversky and Kahneman began their writings in the in the 1970s, many educated people um, are aware of the long list of biases and fallacies the human brain is is vulnerable to. I think the Wikipedia page for biases and fallacies now lists 200 of them. [clears throat] But then it leads uh, to the paradox, well, how did we, you know, invent smartphones and antibiotics and vaccines and get to the moon and discover DNA and and and neural neural networks if we're so saddled with biases? And and the answer is we form institutions with norms such as free speech, uh, open criticism, demand for empirical testing, >> peer review, >> peer review. I mean, for all their editing, fact-checking, for all of their flaws, and we're all >> double-blind trial, which is a remarkably simple idea, but brilliant. I mean, and and quite recent. I don't I'm not sure when it was invented, but but it's very low-tech, but but it's absolutely essential.

>> Indeed. That's right.

In a in a book by Steven Johnson called *Extra Life*, alluding to the fact that the increase in life expectancy from about 30 historically to more than 70 globally and more than 80 in the developed world. So he lists the innovations that led to the, oh, sorry, so that his his title refers to the fact that with the doubling of life expectancy, we have been granted an extra life. It's like we have at least two lives compared to our ancestors. And he lists the innovations. They're obvious ones like sanitation, uh, antibiotics, vaccination, but one of them indeed is the double-blind clinical trial, which distinguishes various kinds of of of uh, of hokum and flimflam from actual effective uh, cures and treatments. But yes, that's that's a great idea, great great example. But the o the overall story though, is that we're rational to the extent that we um, submerge our egos to institutions that are designed to make us collectively rational. And maybe the exception that proves the rule is that with social media, which [clears throat] this will get go back to the original theme of how hard it is to predict the future. When they were introduced, I don't think anyone could have predicted what what we have now seen. The the the explosion of of fake news and conspiracy theories. But it's precisely because giving everyone their own megaphone without the mechanisms of editing, peer review, uh, and so on, [snorts] you know, has led to this explosion of of, uh, of flimflam and nonsense. We have laws against libel, slander, but they have to damage somebody. And I keep wondering why we can't, why nature can't sue...

[laughter]

Um, for I you have to prove that it damages somebody. If if you if you tell a a straight lie about the world, the the you say that the world is flat, that's actually pretty harmless. But but um, unless you can prove that that actually you've somebody has made a financial loss because of that...

Has to be tortious, then a harm, of course.

Yeah. You know, when when sheep are abused by sexually frustrated shepherds, [laughter] um, they are um, I believe that they are represented in court. Um, [laughter] I want to see lawyers stand up and represent nature [laughter] in court um, against just straight lies.

Yes. I mean, the problem there, as a, you know, putting on my other hat as a free free speech advocate, is that since none of us can know with certainty what is true or false, we we we get closer and closer, I believe there is such a thing as truth, and that none of us has it. I mean, I think those are not contradictory. Uh, there is the the danger of a a premature consensus about the truth um, being used with overly broad libel laws to stifle heretics who turn out to be correct in the long run. So you do want to give you a long leash to to expression. And and of course, America and Britain, America and the rest of the world differ in terms of the latitude that we that the United States gives to to free speech. Something, one of the things that I think the United States gets right.

Yeah, I agree. Um, I want to come on to my-side bias and and um, you quote the most depressing discovery about the brain. Um, do you remember what that is, and and can you explain it?

It was an experimental demonstration that um, when u- an experiment is done, the the way people see flaws in it depends upon their political bias. Yes. This is some studies done by a legal scholar, Daniel Kahan, at Yale University, that that inspired this headline, "The Most Depressing Discovery About the Brain Ever," which is that if you uh, give kind of fake data from a hypothetical but plausible study, and uh, the data are arranged that you have to think twice uh, before seeing what the data show. That is, you have to look at relative rates instead of absolute numbers. Something that a literate person, it's kind of second nature to, it's almost what literacy consists of, is that don't be fooled by the numbers on the page, but think about the the the proportions, the ratios. Um, and so it um, uh, the the experiment used the same data in um, a hypothetical study that was either politically fraught, in this case, it was gun control in the United States, or um, utterly um, uh, unexceptionable, efficacy of skin cream in treating rashes. So in in one case, the data showed that gun control either did or did not reduce crime, and and for the other half of the subjects, the skin the skin cream did or did not treat rashes. Um, and it required, again, it required a bit of literacy to see what the data actually did show. And the other manipulation was an independent test of whether the test-takers themselves were numerate or not, and their political orientation. So it was a 2x2x2 design. What he found was that uh, for the politically fraught uh, study, that is, does is gun control effective at reducing violence? Then, of course, people who were uh, innumerate misinterpreted the data as one would almost by definition, that's what numeracy means. Among people who were numerate, it depended on their politics, and it depended on whether the study was of a politically fraught issue. So the leftist numerate subjects, despite their numeracy, misinterpreted the results of a study that went against their pet convictions. In the case of the American left, that would be believing that that gun control is effective. Whereas the right-wing numerate subjects spotted when the when the study in fact did not support the efficacy of gun control. For the study on skin cream, all that mattered was numeracy. So even so, the what was depressing is that numeracy is no protection against cognitive fallacies when the study confirmed seems to confirm or disconfirm a hypothesis that one was uh, uh, embraced in uh, to begin with. So even statistically trained people will be too charitable to bad studies that support their >> prejudices. Uh, whereas they will have a gimlet eye and will be suitably skeptical about studies that refute their their pet beliefs.

It is depressing. I I'm mindful we we're supposed to be talking about the future. So, ah, damn.

[laughter]

The uh, there are optimists and pessimists about the future. Optimists both about the um, the desirability of things that that may happen, and also about the likelihood of of things happening at all. Um, there are sort of contrary to the quotations from Kelvin that I read earlier, that um, there are people today who are optimistic about the indefinite power of human technology. I think of David Deutsch, Arthur C. Clarke, Freeman Dyson. Um, David Deutsch said, um, I haven't got an exact quote, which I think he says something like, um, "If it's possible, if it doesn't contradict the laws of physics, then it then it can be done." And...

With the right knowledge.

Something like that, yes.

With given the right knowledge.

That's right. Given the given the right knowledge.

And Freeman Dyson has these extraordinary these gigantic possibilities for what what could be done. Um, do you feel yourself an optimist? You're an optimist about the past and the present. Are you optimistic about the future? Not in terms of its desirability, but in terms of the power of human technology?

In terms of the the the power, um, so yes, we don't, yes. I mean, one proviso is we don't know, we don't understand the laws of nature well enough to know what is impossible. There may be things that the laws of nature do rule out, and this was Deutsch's uh, proviso. Uh, and another is that something may be possible, but um, have exponential costs as you approach it. So even if it is hypothetically possible, it may not be practically attainable. So, those are just a little reasons why I'm most of the way with David Deutsch, but I I hold back a little bit. I do think there's tremendous um, there is tremendous promise in uh, I think a lot of things that um, could very well be u- achieved within the lives of people in this room. I think the possibility of prevention or or treatment of of Alzheimer's disease, of many forms of cancer. I think a lot of dull and dangerous work could be automated. I think there's potential for um, enormous uh, much greater abundance because of automation and artificial intelligence. I think there's the capacity for there to be less less violence, still less um, crime, violent crime, still, that is a...

A lot of the world, uh, I think Francis Fukuyama coined the term "getting to Denmark." Um, you know, Denmark's not a utopia, and I've never been to Denmark, and maybe this may be based on some stereotypes, but it seems like a pretty pleasant place.

I tell you, I, I, I had a friend who [clears throat] lived in Denmark, and he, he left Denmark, but it was too perfect. He couldn't bear it. Yeah, we should all have such troubles. But, you know, if, if, if Somalia became like Denmark, that would be a really good thing. Yes.

And that it wouldn't be utopia because I'm, you know, I'm sure people in Denmark had their aches and pains and their romantic breakups and and and heartache and all the rest. Uh, but, you know, could the world get to Denmark? Um, well, Denmark got to Denmark, and and and they descended from, you know, bloodthirsty Vikings. So, uh, so I think there is tremendous upside. Yeah.

How about, but, you know, just to, to, just round out the thought, um, uh, the, as you mentioned at the, the beginning there, predictions that are comically, uh, um, that is pessimistic predictions that turned out to be comically false. There are also optimistic predictions that turn out to be comically false, such as that we would all be by now, um, having, uh, uh, electricity too cheap to meter, nuclear-powered, uh, cars, colonies on the moon, and so on. You know, after the moon landing in 1972, it's now been 53 years, and we haven't had another one. Even though it's technologically possible, but they're all social, economic, political factors that can be be breaks on on uh progress. So I guess I wouldn't predict which breakthrough would happen next, but that there is an uh the ceiling is is very high, that there's a lot of things we could discover.

And again, going back to to Denmark, um, here, the, you know, the world went massively backward when, uh, Putin invaded Ukraine, um, three years ago. Um, prior to that, it was envisionable that war, and I, I think it's still envisionable for, for whatever that's worth, that war could go the way of, um, burning heretics and, um, um, throwing virgins into volcanoes and laughing at the insane and all kinds of practices that were quite, uh, slavery that were common for [snorts] millennia. Uh, and then, uh, pretty much disappeared. Uh, could war be seen by our descendants as a barbaric custom like human sacrifice or child slavery? Um, I think we were moving in, in that direction. We've in the last three years, we've lurched backward, but I don't think it's utopian in that, um, many countries, France and Germany, Germany and Poland, uh, Canada and the United States, despite some, uh, trolling by our, our current president, show that countries that used to be at, at each other's throats can achieve very long periods of peace, perhaps, uh, indefinite. I don't think France and Germany will go to war in the lifetime of anyone in, in this room. Could that happen globally? It's not impossible. I don't think it's crazy, even for someone who believes in the dark side of human nature.

How about human understanding? I think you and I are both on record as saying that, um, there, there are limits to what the human brain can cope with. And I, I find it remarkable that, that at least some of us can cope with quantum theory and relativity. Um, and, um, there are people, I think David Deutsch, again, is another one, who thinks that there is no limit to, to what the human brain is capable of understanding. Um, whereas I feel, um, humble enough to say that there probably are things about the universe, as, as J.B.S. said, not only queer than we suppose, but queer than we can suppose.

Um, I think you agree with that. >> Yeah. Um, there are, um, uh, at least certainly in, in the sense of the intuitive understanding that makes us satisfied that we really understand something. I think even there may be things that the human brain can't grasp intuitively that we can kind of appreciate through other means, such as equations.

"Shut up and calculate."

"Shut up and calculate." Yes. Uh, and quantum mechanics being an example. Doesn't sit well with us, but, you know, we do, we understand it well enough to have invented, uh, many technologies that that depend on it. Um, and the, the most famous, uh, being, um, human consciousness, uh, according to a theory sometimes, uh, referred to, uh, pjoratively as the mysterion theory, um, a name that actually came from a one-hit wonder rock band in the 1960s, question mark, of the Mysterians. But this is the notion expressed by many thinkers over, over the, the, uh, the centuries, David Hume, and more recently, um, Colin McGinn, and Noam Chomsky, that they may be, because the human mind is a gadget, it works in, in certain ways, maybe since, and it's not, we're not angels, we're not miracles, this is just an organ, um, there's an enormous, uh, kind of remitch to what we can conceive, but it isn't everything that is, that that could exist, and perhaps the solution to the so-called hard problem of consciousness, namely, why there is subjective experience, um, not just access to information, but that it's there's something that it is like to be alive and awake and and aware. U, since that seems so eternally paradoxical and, um, unsolvable, maybe it's just there is a solution, but we can't grasp it. Now, our late friend Dan Dennett, who, who we lost about a year ago, would say that this whole line of, uh, argument is misconceived. There is no hard problem of consciousness. It's, if you think there is, you're confused. And, and that is another attack one could take. And that would be more consistent with the idea that there aren't thoughts that we're incapable of thinking. That the thoughts we're incapable of thinking aren't real thoughts. At least they're not coherent thoughts.

Yeah. We ought to talk about artificial intelligence. Um, it's very much on everybody's lips at the moment. Um, I've been quite astonished with my various experiments. I, I want to read something. Um, I presented ChatGPT with the, the last paragraph of your book, um, on rationality, and asked [snorts] it to write a poem about on the same theme. So, um, the paragraph is the following: "The power of rationality to guide moral progress is of a piece with its power to guide material progress and wise choices in our lives. This is me. This isn't you. Our ability to eke increments of well-being out of a pitiless cosmos and to be good to others despite our flawed nature depends on grasping impartial principles that transcend our parochial experience. We are a species that has been endowed with an elementary faculty of reason and that has discovered formulas and institutions that magnify its scope. They awaken us to ideas and expose us to realities that confound our intuitions but are true for all that."

"Please write a poem," I said. Two seconds later:

"By reason's lamp, we chart our course,

Through moral realms and matters force.

Its power, one and undivided,

In life and science long has guided.

From cruelest voids where stars ignite,

We tease out good. We seek the right.

Though flawed, we strive to treat each other

As equal kin, not foe but brother.

We lift ourselves from tribal clay

By truths that hold the self at bay.

Impartial rules by reason known,

Transcend the seeds that we've been sown.

This mind endowed with thoughts first spark,

Has learned to push beyond the dark.

It forged the tools, designed the frame,

To stretch its grasp and stake its claim.

Such insights jar our native sense,

Yet prove their truth through evidence.

They wake us from our inward view,

And show what's strange, yet deeply true."

Not, not bad.

I think I think that passes the Turing test.

Well, actually, we had this conversing test in the sense that...

It's too good. It's too fast.

You know that it wasn't written by a human. So, so it did not fool you.

That's, that's true.

Yes, that's true. I, but I'm just... And you have it in another genre.

I mean, there's another, there's a set in the style of Shakespeare. There's one...

I'm, I'm just...

...one in the style of Roger Kipling.

What about... But I'm disappointed you didn't know how to do a dirty limerick.

Anybody got a laptop? [laughter]

Um, anyway, um, what are we going to do about this? I, this is astonishing. I, I'm, I'm, I'm simply bowled over with with admiration and, um, apprehension. What's your feeling?

Uh, it is astonishing, and, and here I have an interest, not so much in the, uh, future, which I feel quite incompetent to, to predict, but as a cognitive scientist, devoted his life to studying natural intelligence, um, and natural stupidity, the, uh, looking at artificial intelligence and trying to understand what the trick is, I find flabbergasting, and I, I admit that until I saw it, I would not have predicted that a system with that architecture...

Precisely. It's unbelievable. Is the, the trick itself doesn't seem up to the job, yet it is.

Yes, it is. So that the, and, and this does get into issues in cognitive science that I have been writing about for, for decades, that there, and, and our, our late friend Dan Dennett, I mean, there are roughly speaking two ways of traditionally of achieving intelligence in, in artificial intelligence. One of them is somewhat like logic. There are propositions, and then there are rules of inference that deduce new propositions from old ones. Sometimes called good old-fashioned AI. This is also the start. This is just how, how, um, computer algorithms, uh, work. It's how Chomsky's, um, theory of grammar works by, by rewrite rules. Um, the other is, uh, artificial neural networks, where you have computational units that are kind of roughly reminiscent of neurons. They aggregate, um, analog information, and then either fire or don't, if the sum weighted sum of inputs exceeds a threshold. But there are many, many, many of them, uh, feeding each other in layers, and then trained with, uh, typically with, uh, the input to the, a problem and its correct output, and, uh, dynamically strengthens the, a, the weights that aggregate the inputs to approximate the output. And, um, with enough training, uh, uh, can often succeed. There's a debate that went back and forth. If I was really on the side of pointing out the limitations of neural network models, lacking, uh, anything that we would call, uh, propositional logic, that is, things that can be true or false, statements about, uh, who did what to whom, what is where, those don't explicitly exist in neural network models. And that was what was behind the, uh, the, the [clears throat] failure of them to really advance for many decades, including even the ability to generate the, the English past tense, to go from a novel verb like to you, to, to sme, well, I smeged. Um, but the artificial neural networks for decades couldn't do that because they generalized by similarity to what they were trained on. If you then surprise them with something outside the training set, they, um, they flounder. They put together bits and pieces. Now, so that would have led me up to just two and a half years ago to predict that ChatGPT should not, uh, and its, its brethren should not have worked as, as, uh, as astonishingly as they do. What I think where my intuition falls short is just not being, and everyone's, because, uh, I think I'm not the only one who is surprised, is that not being able to intuitively fathom what sheer scale could do. That is, if there are not just one or two layers between input and output of these interconnected units, but a hundred. If there's not just several hundred thousand parameters that weight the inputs, but a trillion, and if it's trained not just on a few million exemplars, but on trillions, on the entire worldwide web, then that quantity leads to a kind of quality where it can do things like, like what we just witnessed. At the same time, I think the shortcomings of, uh, neural network models are still there, and they're called hallucinations or confabulations or blends. And it's precisely because their intelligence comes from, um, generalizing based on similarity to things in the training set and blending things that, uh, of the general kind that tend to go together, even when we now know they never did go together, and hence you get the hallucinations. Several years later, still a problem. And as a, you know, semi-regular user of, uh, of, um, generative AI, I'm still surprised at the hallucinations that come about. And they come about for a systematic reason. Namely, there's nothing in there corresponding to a proposition to the capital of France is, is Paris, or, um, um, uh, so-and-so did such-and-such at such-and-such a time. There are blends of many things that tend to co-occur in the training set, resulting in output that is always plausible, but not always factual.

But you can learn why you can just look it up in Google and if...

So [clears throat] there are now hybrid models that will, before they produce the output, they'll kind of look them up on Google. Uh, and not surprisingly, Google itself. I mean, that was what Gemini was originally designed to do. And I better tell my, my story. Um, uh, I, I happened to be asking ChatGPT about a geneticist called Helen Spurway, who, who must have been, who died some years ago. She was married to J.B.S. Haldane. And it said, "After Haldane died, she married Richard Dawkins." [laughter] And...

So, so I said, "No, she didn't." Um, [laughter] and it said, "Oh, sorry about that." Um, actually, she married Ordis Huxley. That was also...

Okay, so that you put your finger on it. These hallucinations that come from there's sort of statistically probable because your name occurred in association with Haldane and evolutionary biologist and so on. And so each of the bits was plausible. You put it together. Uh, it's not implausible, it just happens to be false. And that happens a lot still.

Yeah, we better switch to audience questions. So, Steve, thank you very much. Thank you. [applause]

Thank you, Richard and Steve, for a wonderful discussion. I can see Richard reading Friedrich Nietzsche's book there. I'm not, uh, [laughter] I don't approve. It's not, uh, it's not my choice. Uh, please raise your hand if you have a question for Steve or Richard. You can field your question to either of them. Please keep your question short and tell me who your question is for at the beginning as well. We'll start at the front here, just with the red shirt. And we'll take the second question at the top left there, the young lady in the white.

Amazing. Thank you both. Amazing to see the conversation. Question for Richard, if that's okay. Um, how do you see religion influencing the future of humanity?

That was one of the things that was listed on the, on the crib sheet. I rather avoided it. [laughter] Um, I think it, I'm looking forward to it going extinct.

[laughter] [applause]

Can we bring Steve on the same question?

Yes. So there's two. I, I, it's a question I did take up in, in my book *Enlightenment Now*. So I'll just, which is, you largely a book driven by discussion of of graphs and trends. There are actually two trends that are going in opposite directions. There, one of them is that people tend to lose religion. I mean, that's the, the overwhelming tendency that when people do, more people kind of convert, often not to atheism because the word is kind of scary. Um, but in the United States, the word is, uh, "none," not the, the ladies with the habits, n-u-n-e, that as you ask people what their religion is, they say "none." And that, that is growing, and it's a very strong generational effect. It's not actually so much people losing individuals losing their religion. It's old people who have religion dying, and young people, um, not replacing them in terms of their, uh, their religious belief, having, or having any religious belief. Um, going in the other direction, it, with every, uh, in every denomination, the more religiously orthodox, the more babies people have, or, as, as someone unkindly put it, religious people have much more unprotected sex. So this is the, the, uh, growth rates for ultra-orthodox Jews, evangelical Christians, um, orthodox Muslims...

That presupposes that children inherit the religion of their parents and that...

And that may change. True. True enough.

Yeah. A question at the top here.

Thank you. So I've got a question about what it means to be human versus what it means to be AI. So we know what it means to be human. It means we possess intelligence and we are biological beings. Whereas before we were saying that AI is not human, but now that AI has intelligence, but also now we can, they're becoming more bi, they're infiltrating biology as well, like we can print genes, we can edit DNA with AI. So what, what are your thoughts on this?

Steve, do you want to kick with me?

Yeah. So actually, this this connects, um, to one of the philosophical problems that came up in this conversation, namely the, the so-called hard problem of consciousness, consciousness, the problem of phenomenal or subjective experience, because one of the answers to that question might depend on what is necessary to have pleasures and pains and consciousness and so on. Uh, we can imagine that a lifelike robot, uh, there may be no one home who's actually experiencing anything. Uh, that it may just be circulating bits and and patterns in electronic circuitry, but no one actually feeling the pain. Even if you step on its toe, it says, "Ouch." Um, and so this imponderable philosophical question, namely, what does, what is necessary for subjective experience? Where does it come from? Do you need to be made out of neurons, or can silicon chips doing the same computations give you that subjective experience? Turns out to have massive ethical consequences. Namely, is it immoral to unplug a robot in the same sense that it's immoral to murder a human? Um, and, and, you know, I think we really don't have a coherent way of answering that question. And maybe we never will, or, or, maybe some unborn genius will will tell us how to do it. But I would certainly say that morally, what makes humans, as, as best we can tell, uh, different from even the most advanced AI is that at least we think we feel stuff, and they don't. And, and of course, if you're a first-year philosophy student, you can then say, "Well, how do I know that you really feel something? Maybe you're a robot." But I'm, I'm pretty confident that everyone in this room, made out of the same flesh that I'm made of, is as conscious as I am, even though I can't experience your inner life, um, uh, directly with first person. I'm much less certain about a lifelike robot.

Richard, do you want to come in on the same question?

Well, I think that, um, there's a Darwinian question to be answered, which is why pain has to be so damn painful. Um, because theoretically, it could just be a, a little flag that goes up in the brain and says, "Don't do that again." Um, if you don't, if you've done something that causes bodily injury, the point about pain is, is it's a warning of death. It's a, it's a warning that if you do that again, you may, in, uh, injure your life expectancy. The problem with the little flag going up is you can rebel against it. And I think it may be that, um, because an animal is conflicted between different drives pushing it in different directions, there has to be some kind of discipline, um, to stop it, um, going all hedonistically in favor of one kind of reward rather than, rather than another. But it, it is a profound problem as to why, why subjective feelings are so, so powerful, rather than just, just robotically.

I think I would put it slightly differently. Uh, that it's not that they're powerful because that's perfectly intelligible, namely that it, um, it's a, you know, it's a, it's a big bright, you know, waving flag, and, and you can't help but, uh, uh, obey it or respect it. Stop doing what you're doing. So, it's not so much the intensity, because that could just be the magnitude of a signal, how big the flag is. But why there's any subjective dimension to it at all, even if it was faint, even if it's like a slight scent of lemon, why isn't that just, you know, a little lemon flag, as opposed to, well, kind of smells like lemon.

Well, torture wouldn't work if, if it was a flag. I mean, you could, you could...

Well, sure, it, it could, in the sense that it would be the, you know, the whole system would be designed that if the, uh, if the damage was that severe, uh, it would mobilize every part of the system to react against it. It would have the strongest possible tendency to avoid it, minimize it, neutralize it. Now, here again, our friend, uh, from, from the grave, our friend Dan Dennett, would say that's all that pain is. Uh, that is, it's just the absolutely unignorable, maximally avoidable, uh, state with the red flags and all of their, uh, their consequences in, inside the brain. And that when you ask the additional question, why does it feel like something, you're asking a question that can't have any answer because it's not a sensible question. Here, I'm, I'm channeling Dan from the grave. Uh, I don't, it's not the way I think.

No, me...

Because I know what pain feels like. And...

You know, and as you say, it does have this...

...irreducible subjective aspect. Yeah. But the paradox being, and here's where there's an opening for, for, for Dan, is that as soon as you try to grasp it, study it, make a testable hypothesis, you've now departed from its very subjectivity, and you've pointed to the objective aspect, such as something happens in the brain, you withdraw your hand, you avoid it in the future. All of that is no longer the subjective dimension. That's what makes it so fish. Talking of pain, I'm worried this lady's hand in the heavens might fall off if she holds her arm up for any longer. So, there's a question from the shadow up there, and then we'll take another one, maybe from this side. [snorts] Just after, if we take a question just over here.

That works. Yeah.

Hi. Um, yeah, I, I've tried, so Stephen, I've tried to, um, parent my children in line with some of your ideas from *The Blank Slate*. So, um, taking on board the fact that parenting doesn't matter and I should just build a positive relationship with them and do what I can in that respect. But I do find it hard to believe, despite everything I've read on what you've said and spoken about, that if I was to take my kids out of school so they could watch video games all day and eat fast food and not get them orthodontic treatment or whatever it might be, I find it hard still to believe that the outcomes would be exactly the same. I don't know if there's anything that you could say to convince me further.

Yes. So it's the, the, um, so the generalization from behavioral genetics on the, um, effects of, of parenting are that, that, that parenting within the normal range doesn't change the child's, uh, intelligence or personality and many of their behaviors. Um, so that you won't make the parenting practices, as best we can tell from the fact that children in the same home are no more similar than their genes would would predict. And parents, and, and children who are adopted are, um, have no resemblance to their adoptive siblings, but do have resemblance to their biological siblings, plus a lot of random noise that's neither parenting nor genetics. So this doesn't mean that, that [snorts] access to information doesn't matter. So that if a child is never exposed to, uh, the English language, they obviously won't end up speaking English. If they don't go to school, they won't learn what's in school. These tend not to be picked up in the studies because they are, uh, true of the entire society. So that among your sample of adoptive parents, for example, or of twins, you know, all the kids go to school. Um, and so that kind of, uh, of, uh, contrast, which isn't so much one parent from another in a given society or neighborhood, you know, that obviously matters, and that they can't be miracles. You can't learn something, uh, without exposure to the information, uh, necessary to, to, to, uh, to learn it. It tends to be not what we mean with parenting practices, but, but I mean, that certainly would be a case where, where, um, uh, upbringing matters. Also, of course, physical health and safety. Uh, clearly, since parents are, um, are the main providers of, of, uh, sustenance and safety, they have a huge effect, uh, in, uh, in those regards.

I've never had so many people trying to catch my own one go for questions. So, I'm going to try and get through as many as we can. We just have one over here, and straight afterwards, we're going to jump one to the center at the back in the heavens again.

Uh, so Stephen, this is a question for you. Um, regarding what you were saying with, like, I guess you could call it, like, the Denmark-ification of other countries. Could that, who decides that that's good? Because that could end up with, like, a loss of culture. So, like, I think you used the example of, like, Ethiopia turning into a country like Denmark. Well, Ethiopia has its own rich history. You said, like, Denmark comes from Vikings, whereas Ethiopia has its own thing. So, would that still necessarily be classed as good?

Yes. So by, by that, um, Fukuyama and I, don't, don't mean that Ethiopians would become Danes. Uh, because even, um, so Denmark, say, compared to another affluent Western democracy, like, uh, um, you know, like Italy or, or, or Britain or Germany. Um, you know, Italian food isn't the same as Danish food, and it needn't be. But both Denmark and Italy have, um, pretty low rates of violent crime. They haven't gotten into any wars lately. They have, um, they're, uh, uh, pretty affluent. Um, so you can, there could still be Ethiopian culture, but Danish affluence and, uh, and, and health and longevity. So I think that that's all that's meant. That needn't, um, involve a flattening of cultural differences. Um, so who decides? It would be the Ethiopians. That is, do they want to be richer? Do they want to live longer? Do they want to have less disease, less war, and so on? And I suspect they do.

Okay, we got a [snorts] question from the top, and right after, a preference for a question for Richard.

No, no, don't.

No. Well, the gentleman with the fantastic beard in the middle will be asking the asking the question just after at the top. Uh, thank you very much. And this is a question both for Richard and Stephen.

Where am I looking? Sorry.

Sorry, I'm here.

Okay. Oh, there you are. Okay. Yes.

Hello. Thank you. Uh, so my question was is that Richard mentioned a case of sexually frustrated shepherd and a sheep, which was very funny because presumably because we believe that sheep should be represented in court because this is something bad that is happening to them and this is immoral. Uh, while we are actually doing something much worse every year by taking their children and actually eating them. So, this is a provocative, uh, start. But my question is, uh, um, do you believe that, uh, uh, the trend on lowering of violence against animals is something that is actually happening, and if this will continue, because well, our morals evolved quite a bit since medieval times on this topic, and do you personally, uh, try to participate in this trend, and do you believe in it in the future? Thank you very much.

I think if we, we were to ask what, uh, people a century from now would look back on moral questions, just as we look back on slavery as, as the great evil of our, of, of [snorts] their time. This is probably the answer that, that, um, um, we need to broaden our feelings to, to humanity, to all sentient beings. Yeah, I think that that's plausible. In terms of, um, the, you know, what, what, um, whether there's been progress, what direction we're heading in, in this, in the sense, no, in the sense that probably the sheer number of unhappy chickens is greater now than it used to be. Yes, in the sense that it's being raised and discussed and, and on the agenda. A greater number of, at least affluent countries, are passing laws for protecting the welfare of, uh, farmed animals, especially factory-farmed animals. U, large numbers of consumers are, are insisting on more humane treatment of the animals that, that, uh, contribute to food. There, there are a number of, um, technological initiatives to, for, uh, synthetic or, and lab-grown meat that I suspect will, will, uh, uh, expand. And so Richard's prediction is not implausible, and it's certainly morally, uh, uh, morally causal, morally impeccable. This is also a great case, by the way, in which reason can guide, um, moral judgments, in that for many years, and in certainly my parents' generation, it just was barely an issue. It's like, well, they're animals. Um, and it's, uh, and, you know, I really intuitively don't feel I should, but just in terms of my gut empathy, I don't feel as bad for chickens as I do for, for humans, even though intellectually I have every reason to believe that a chicken is is capable of suffering. Um, and so I can't defend factory farming, uh, thanks to my, the ability for my ethical reasoning to be fed by my, um, rationality, even though my own sense of empathy wouldn't have gotten me there. But I tend to think that, that, that can happen. And the passage that ChatGPT turned into a sonnet was basically making, came in, in a chapter of rationality that argued that this would not be the first time in history that intellectual arguments drove moral progress, because something somewhat similar happened in the case of slavery, where the circle of empathy of slave owners just didn't, didn't include slaves, and they were kind of, uh, many countries were argued out of it, and, and there were debates on the immorality of slavery that hinged [snorts] on arguments such as, patently, there's no conceivable argument for denying humanity to, uh, the people that were enslaved. And then there are other cases that, that I talk about, uh, and, and I think that the, the case against cruel, um, uh, farming practices would, would, uh, would fall into that category.

Well, last time I had you on a show, Steve, you said you were practicing reducetarianism, where you, you said you're having a steak once a year for a special occasion. Is that something you still...

Yes. So, I'm an immoral person and that I'm not a vegetarian. Um, but I am a reducetarian. Yes. Which is, uh, you know, if I started out eating twice as much, and I, uh, and I became, then ate half as much, it's morally equivalent to someone who ate half as much who became a vegetarian. [laughter]

Some funny math going on there. At least for, at least on utilitarian grounds. I think more significance though than any individual becoming a vegetarian. I think that's a case where there is some moral value to it because it, you know, somewhere there is a signal that results in fewer animals being raised under, uh, cruel conditions. I think far more, uh, powerful would be advocacy for lab-grown meat and synthetic meat that would satisfy our carnivory without the concomitant suffering.

Okay.

Presumably, it'll be possible to grow human meat in tissue culture.

Whatever you're into, Richard. I'm so, I don't think there'll be much. Just in terms of predictions, there's lists of comically false predictions, of which, like, like by Kael, but one of them was Winston Churchill in the 1930s predicted within a certain number of decades, um, that there would be, that all meat would be synthetic, that there, that it seemed ridiculous to him that you'd have to grow a whole animal and, you know, throw away the, the bones and the gristle, whereas there's no reason you couldn't grow it in a test tube. So, he was premature, but...

I confused the hell out of the production team when I said a question from the gentleman in the beard, because I know there's a load of people with beards there. [laughter] So, pick whoever you choose there. Um,

Pick best beard.

Pick the best beard. The person in the heavens can pick another person up here, and then we'll have a question right front center there, and then we'll see what we're, how we're doing for time.

Um, thank you for the opportunity. Now, we're coming back to the AI, uh, thing. I was listening to this podcast by Anahad O'Connor, and she mentioned that how consciousness is might be a very well-ingrained in the universe. And when we talk about AI not being conscious at the moment, you know, how AI is deceiving us in terms of what Richard's got the marriage now, in, in a sense, um, how AI tells us, and how experiments like when you split the left and right brain hemispheres in a person, and then where we tell them to go to that door. Well, they would walk to the door, but the, that would be because of one hemisphere of the brain. The other hemisphere of the brain, which would allow you to talk, will tell the person, if they asked later on, why did you walk to the door? Something else, like, "Oh, I want to go out," or "I was thirsty, I was going to get water." Um, is this like, kind of deception we can see in AI that is cropping up? Um, another thing that was recently, OpenAI service, I read that they declined to shut down, rather did more calculation onwards when they were asked to shut down. Is that like a sense of, I don't know, like saving themselves from something? Um, I mean, I'm coming from a very materialistic point of view. If, if we, if we say that we are material and we are coming from the brain, well, everything. Okay.

Yeah. So that kind of confabulation, such as in the, in the famous split-brain experiments with, with, uh, done by Michael Gazzaniga, where you flash to the right hemisphere, uh, command like "get up," "get, walk to the door," and then you ask the whole person, including the left hemisphere, why did you walk to the door? And it makes up some reason that we, the scientists know is bogus, because we know why he got up, namely, the command was flashed to the right hemisphere, "stand up and go to the door." But he says, "Well, I just felt I was hungry at that moment," or, and, and we know that that's false. So, the question is, does AI do it? Yeah, all the time. Namely, it confabulates reasons, uh, because it, even more than a person with an intact and connected left hemisphere, has no explicit access to any of its beliefs. It's just smeared over hundreds of billions of little fractional numbers. And there isn't one proposition that says, you know, I, I, I was hungry. Time was up. Uh, you know, I wanted to go. And so it, it does a lot of that, uh, in terms of self-preservation. And this, this gets to the heart of, uh, the nature of evolution, as, as explained in, in so many of Richard's books. Um, you know, I, and a whole topic we didn't get to in terms of the future of AI is whether we are doomed because AI will inevitably do to us what we have done to all those species that we extinguished or that we, we dominate, like, like factory-farm chickens. Will AI do it to us? And I think the general answer is, if AI is engineered, there's no reason that it should, because goals are independent of [snorts] knowledge and intelligence. And there's no reason we couldn't have fully superhumanly intelligent but self-sacrificial or neutral, uh, AI. The danger would come. It's an argument that, um, Martin Bucci, a Belgian philosopher, has has made, that if we were to allow AI to evolve by [snorts] natural selection, then we could be in trouble, because then survival would be, uh, built into the very source of its intelligence, would be, uh, as a means to the end of survival, and then it could [snorts] subordinate us as a, as a goal. But as long as, and so that would be a very foolish way to enhance the intelligence of AI, namely, have many, many AIs, uh, compete for, for finite resources. But as long as the criterion is, can it produce poems? Can it drive cars? And so on, um, it would do whatever the goals are that were programmed in.

Have we found someone at the top with a question? I can just see...

There we go. My my question for both of you is, given that the cost of intelligence is decreasing with the rise of performant AI, how do you envision this, um, access to low-cost intelligence to change society or change the way we value intelligence on a societal or even a species level?

Not necessarily. Because, um, I mean, a crude analogy is, our machines have been faster and stronger than us for, for a long time. Um, but we still admire athletes, gymnasts, um, basketball players, skaters. And so, I, I think we would, and because the, we outsource to AI certain tasks that, that we don't do, what we, what we ourselves do, um, because of the fact that we do it, because we reserve those tasks for ourselves, I think we'll still be in a different category of evaluation. Um, so, you know, we still admire Usain Bolt, even though, uh, a cheap Toyota can be go faster than Usain Bolt, but it just doesn't impress us in the same way.

I wonder what order we will become redundant in, and whether, I mean, I would have thought poets would go last, but, um, um, it'd be nice to think that solicitors would go first. There is, you know, there is some intuition of, I mean, I think the reason that we, no one would, no one would at least present, and maybe this will change, and maybe my intuitions will fail me here, but I, I don't think anyone would go and buy a book of, of ChatGPT poetry for the same reason that you might spend an awful lot of money on Jimi Hendrix's guitar, but you wouldn't spend the same amount of money on an identical guitar that Jimi Hendrix didn't play. That is, we have a intuition of authenticity, of provenance. Why a fake work of art, uh, is radically different from a real Da Vinci, even though it might be indistinguishable from it. Our intuition of what a thing is, is inherently bound with, uh, where it came from. Uh, Paul Bloom wrote a book on, on this, and, uh, Bruce Hood, over here, here in the UK, called it Bruce, called it Super Sense, uh, and he used the Latin word "heckyesness," thisness, that it really matters what something is made of, where it came from, what it was designed to do. Now, maybe we'll outgrow that in the case of intelligence, but it is...

That has been true, but I think, I mean, my intuition is that something that was produced by an artificial intelligence would be, I would think that has great collector's value, except it can be reproduced at any time, of course, but, but, um, the thing about the painting is that it's, it's a one-off. But, um, but I think a, a symphony by a computer, something that I would be very interested to listen to. And, um...

So, yes, but although I suspect there'd be a difference between someone discovering in a rare books library and an as yet undiscovered Mozart sonata,

Yes.

Versus ChatGPT creating a sonata in the style of Mozart. Even if note for note, they were identical, I suspect our reaction would be very different. Yes.

We're going to take this question from the front and center. We've probably got time for one more question afterwards. I'm being, I always, I never trust myself to ask the question of the person waving frantically, but let's see how that one goes at the very end. [laughter] We got two thumbs up there as well. That's even more of a concern. But let's have one at the front here.

Um, thank you both for the great talk. Um, probably a question predominantly for Richard. I wondered if, um, to expand the timeline slightly into the evolution of, um, humanity, and taking into account our ability to diagnose mutation, potentially to engineer DNA, and to, um, uh, treat mutations often as, uh, illnesses that need to be fixed. Do you see that we have put a stop to human, uh, evolution, and would that be replaced through with genetic engineering or transhumanism and the integration with technology?

Brilliant question.

There are two ways in which you can guide evolution. One, one way is by selection, the other way is by mutation. And, uh, we've been doing it for centuries, uh, with, with selection for domestic animals. And we've, we've turned wolves into Pekingese and poodles and, and, and wild cabbages into, into kale and Brussels sprouts. So, um, this is something which, um, has been done for centuries. We never done it to humans. Hitler tried briefly. Um, so suddenly we have the power to do it by the other part of the equation, which is, which is mutation. Uh, and, uh, this, this may happen, but the fact that we haven't done it by the selective route so far suggests that we're not actually overeager to modify humanity. Um, so I'm not, I'm not so sure about that. Um, transhumanism, I'm never quite sure what that is. Um, but, but it, it certainly is possible by using biological technology to change the way, the way human evolution goes. Um, if you wanted to win the Olympic Games in 500 years' time, you could do it by selective breeding. Um, nobody's done it, I don't think. Um, and you could also do it by, by mutation.

Our final question from the very enthusiastic...

Question. I, if you can save it for a moment. Thank you. Just take a seat for a moment. Thank you.

Thank you for taking my question. Hopefully, it's not too crazy despite my waving hands. Um, so if I understood correctly, you gave the example of quantum physics as being maybe too hard to understand, you know, in one perspective, but when we have structure, and the structures that we have in society, including math, we were able to understand. And I think that was a point you made. And one of the things I've always struggled with is particularly consciousness and sentience. And is there an equivalent in the same way that we're able to understand quantum physics? Is there an equivalent structure that may one day allow us to understand specifically sentience, i.e., that ability to feel something from the firing of a neuron? Because obviously, that gap still exists at the moment. Is there an equivalent of a mathematical equation or some structure that will allow us to ever understand that? Do you think we'll ever be able to do that?

I, Yeah.

It's a very interesting question.

Yes. Yeah. I mean, there's a sense in which we already do, in that there's the so-called easy problem of consciousness, which is a bit of an in-joke because there's nothing easy about it, but just, what is the neurobiological basis of the distinction between conscious and unconscious processing? Why did we evolve with such a distinction? Uh, what is the flow of information that's in the conscious pool versus in the unconscious pool, and how did it go back and forth? You know, those are all difficult but tractable scientific problems. We've made a huge amount of progress. I, I think I think they're quite plausible theories now as to where consciousness in that sense comes from. Uh, and one could say, as, um, some people have said, like David Chalmers, and I think it came up earlier in one of the questions, the so-called theory of, of informational panpsychism, that wherever there's information, there's a bit of sentience. Now, this has some counterintuitive implications, such as, your thermostat is a little bit conscious, and so on. If you can swallow that...

Well, then, okay. [laughter] So, it's a big if. Uh, then that would be an answer to your question, namely, you swallow the identity of information and sentience. You show how consciousness in the human case depends on information flow, and therefore your mind can at some level entertain or, or juggle, even if it never feels intuitive, because you do have to swallow that highly, um, unintuitive premise that wherever there's information, there's consciousness, sentience.

Richard, you jumped in there saying you weren't a fan of panpsychism.

Well, no. I mean, but I, the panpsychism I've, I've met is people think rocks are conscious, and, and the sort of deep hat chop.

So, this would be not exactly like that, because the rocks wouldn't process information.

They don't.

So, rocks, no thermostats, maybe.

Yes.

Okay. [laughter]

I know everyone's got heaps of questions they want to ask. I've got loads myself, but I'm really sorry that we've run out of time to ask questions.

I do apologize. There's lots of people with questions, including myself. I appreciate that.

I appreciate that. There's lots of people with questions, but I've been...

I've seen lots of hands. I do apologize.

Please, please join me in thanking everyone who made tonight possible. [applause] As well. [applause]