March 2009 Tyler Cowen - Robin Hanson Bloggingheads.tv transcript

The audio is here (tiny mp3 link) and below, and the video is lost. Lots of us haven’t heard this discussion, so I made a transcript; I used Descript, then cleaned it up and added links. Unclear bits are in [brackets]. Post-discussion comments on Robin’s page, on Tyler’s page, on Arnold Kling’s page. Enjoy!

+Transcript of Tyler’s discussion with Robert Wright, June 2009

+Transcript of Tyler’s discussion with Will Wilkinson, July 2009

Robin Hanson: Hello, Tyler. 

Tyler Cowen: Hello, Robin. 

Robin Hanson: And hello, listeners. I'm Robin Hanson, Associate Professor of Economics here at George Mason University. And my co-Bloggingheads.tv companion here is Tyler Cowen, my mentor and fellow colleague here at George Mason University. 

I should start out with a word of deep praise for Tyler. Tyler hired me to come here to George Mason, for which I'm extremely grateful. And I have to say that although Tyler and I often seem to disagree about some things, whenever it comes down to something very concrete about why somebody's doing something and the motives and reasons for what people do, Tyler's almost always right.

Tyler Cowen: That means I almost always agree with you.

Robin Hanson: And I've come to rely on your opinion on these things. If I don't have an opinion and you do, I accept it. And when you've disagreed with me, I've learned that you are right.

Tyler Cowen: I defer to you on virtually all matters quantitative and anything to do with physics. I think the question is the areas where we disagree, why we still disagree with each other. And I think there's a systematic pattern to where we find those other disagreements.

Robin Hanson: Well, let's figure out what they are. Sometimes I'm not convinced that we do disagree. Sometimes it seems like we can say some words that kind of sound like they disagree, but I'm not sure that we do. So what's a clearer area where we disagree?

Tyler Cowen: I think anything at the meta level, we are bound to at least appear as if we disagree. So if we talk about meta-ethics or if we talk about very general ways of thinking about science or economics, I think we will systematically end up with different positions.

I take a kind of pluralistic “things are messy” position, which you find unsatisfying. You take what I call a position of logical atomism. I'm not sure that you view your own position that way. 

Robin Hanson: I don’t think so.

Tyler Cowen: We find this pattern of disagreement repeating each time in many different areas. But again…

Robin Hanson: I've heard you say that, but can we find some more concrete examples of how it plays out?

So for example, on disagreement we both have meta level concerns about disagreement, or both at the meta level bothered by the fact that people disagree and find it somewhat troubling and suspicious at least. So I don't know that we disagree at the meta level there.

Tyler Cowen: Correct. 

Robin Hanson: So where do we, aside from general slogans or something, is there something…? 

Tyler Cowen: I think you're more likely than I am to see a one-to-one mapping between the model and the world. So if you're doing Bayesian analysis on why people disagree, you'll look at a bunch of reasons why they might be justified in disagreeing, rule them out, and then assume that the only possibility left must be the true one.

And I think you're taking the model too literally. I'm more likely in my view to view the model as a construct, and then if we can't figure out what's going on, I'm more willing to just say, well, the model is failing to capture some aspect of reality, the reasons why people disagree are messy. There's not any neat way in which you could always agree with other people, and therefore we're just a bit stuck.

Robin Hanson: So I was accused of the same thing yesterday. 

Tyler Cowen: By whom?

Robin Hanson: That will be evidence on your side. I gave a talk at an artificial general intelligence conference on the economics of artificial intelligence. And people accused me of having assumptions in my models. I guess there's a question of how much you rely on your best models. 

So clearly we agree that you shouldn't just throw your hands up and say, nobody could know anything. You should – I presume we agree – take your best shot at finding the models that seem to fit best with the phenomena at issue, and using those conclusions to rely on somewhat.

So then the question is, how good are our best models? Do we really disagree on how good our best models are? 

Tyler Cowen: Probably not. Maybe where we disagree is my willingness to use messy, pluralistic intuition to override those models a bit. Maybe you think less of that than I do. Is that possible?

Robin Hanson: I might rely less on my intuition, but that's not so much a meta level disagreement.

Tyler Cowen: I would say you're relying on your intuition in other ways. You have intuitions packed into the model. 

Robin Hanson: Sure. I prefer mediated intuitions, for example, that are revealed in the structure of some sort of model rather than direct intuition.

Tyler Cowen: That's right. I think that may be one of our most fundamental disagreements: that I'm more willing to go with what you're calling maybe unmediated intuition. I think of it as very mediated. I think it's mediated by taking a lot of information from a lot of different areas, trying to process it objectively and then applying something called judgment… [bad audio]. Occasionally, it's called common sense, though I don't think it's very common.

Robin Hanson: Okay. Well, the question of course is how much you can articulate where this judgment comes from. So the more you can say that the judgment is composed of certain components, the more it looks like a model. 

Tyler Cowen: Sure. 

Robin Hanson: Then the question is, if there's any difference between unmediated or mediated intuition, it must be in the amount of detail in the explanation.

Tyler Cowen: I’m not sure it's the amount of detail. I think it's the number of dimensions. So the ways you use to mediate your intuitions, I think tend to have quite a small number of dimensions. The ways that I use tend to have a much larger number of dimensions, which can be either good or bad.

So to you, it looks like what I'm doing is harder to articulate, but I don't see that myself as a fundamental difference.

Robin Hanson: Well, we've probably lost half our audience by now, so let's try something concrete and see if we can play it out there. 

Tyler Cowen: Very concrete. Okay. What do you want to do?

Robin Hanson: For example, do our differing intuitions lead us to differing conclusions on cryonics? Are you going with just a knee-jerk intuition that it doesn't make sense, or is there a structured model behind your opinion there? I'm picking it somewhat at random, but…

Tyler Cowen: No, picking cryonics is not at random, I keep on thinking for the last week when we had this scheduled that it's not Bloggingheads.tv, but Bloggingfrozenheads.tv. Somehow I can't get that phrase out of my mind.

Robin Hanson: You got it in. It was a good one. I applaud.

Tyler Cowen: I don't reject cryonics per se, but I think of the people who adhere to cryonics as primarily concerned with signaling. And interest in cryonics is very strongly correlated with interest in science fiction. I don't think it's correlated with what an economic model would predict interest in cryonics should be driven by.

Robin Hanson: We agree.

Tyler Cowen: So people want to signal some image of themselves — 

Robin Hanson: Certainly.

Tyler Cowen: As being in a particular way rational. And thus they're attracted to cryonics and they overrate the chance it will succeed. I don't think it's crazy to spend $200 a year on cryonics, because I don't think it's crazy to just burn the money either.

Robin Hanson: I gotta say it's kinda crazy. 

Tyler Cowen: I don't feel I'm missing out on any opportunity in not doing it.

Robin Hanson: I gotta say, burning $200 sounds a little crazy to me. Certainly burning $1000 would be. 

Tyler Cowen: Think of it as a form of charity. Prices will fall, if only stochastically, and you confer an externality on other cash balance soldiers, what's wrong with that?

But if we're gonna have grand dreams of coming back, I'm much more interested in theories of the universe being infinite, and the notion that right now there are copies of me out there doing things and I can live on in that way. 

And the extra boost of imagination I get from cryonics and thinking about my head being frozen and maybe someday thawed out in some form, to me, it's a very small increment, so it's not as if I'm giving up a chance at immortality. I think I already have that chance, but at some point you just say, enough is enough. I'm gonna signal something else now. And that's what I do.

Robin Hanson: Do you have the same attitude toward medical science? 

Tyler Cowen: A lot of it, sure. As you know, we agree on this. But I'm not obsessed at extending my life at the expense of all other values.

Robin Hanson: No, but you do spend some money on health insurance, right?

Tyler Cowen: I don't spend any extra. Obviously it comes with my salary.

Robin Hanson: Okay. So you're pretty skeptical about medical science overall, in terms of direct instrumental value. 

Tyler Cowen: I think what works is [setting bones], some number of pharmaceuticals, attitude, high social status, diet and exercise and sanitation. If we want to call that medical care, that's fine. Those to me, are the things that work.

A lot of extra visits to the doctor, I tend to think have a zero or maybe even negative marginal return. I think we agree on this.

Robin Hanson: Okay. I think I've noticed that you use seat belts in the car, right? 

Tyler Cowen: That's right. Is that a mistake?

Robin Hanson: It's a little extra time and trouble. You must think you're getting something for it. Is it just signaling?

Tyler Cowen: I feel more comfortable with the belt, but I've been assuming that it makes me safer. I've never studied the issue. I'm willing to be talked out of it, but since it's also a law and there's no cost, I don't mind.

Robin Hanson: It probably certainly makes you safer when you're riding with somebody else who doesn't notice whether you put your seatbelt on.

Tyler Cowen: This is true. 

Robin Hanson: If you're adjusting your behavior to the seatbelt, that might be different.

Tyler Cowen: And you incur a lot of social disapproval.

Robin Hanson: So, I'm not sure we're finding a substantial disagreement here in a direct way. I guess you're not really saying that you'd care much about actually living longer or not.

There's not very much you do that makes you live longer or not. You don't do very much because it would make you live longer.

Tyler Cowen: No, I exercise more than I would, I eat healthier food than I otherwise would, though I'm unsure what the return is there.

Robin Hanson: But you don't give the signaling argument about those things. You don't say, why should I exercise more? Because I'm already signaling in enough other ways I don't need that signal.

Tyler Cowen: That's right. Because exercising actually has a good chance of making me live longer. Whereas I think cryonics has a very small chance. And I'm also unsure as to what would live longer would in fact be me. That's a matter of degree.

Robin Hanson: Sure. But it's not clear whether the guy who exercises more is you either, is it?

Tyler Cowen: It's me enough from my point of view. It's me in a way that an upload isn't. What I find funny about your view is you're a skeptic about medical science, about almost everything – except freezing your head. You think that's the one thing that works.

Robin Hanson: Well, it's a matter of the chances. I think if there's at least a 5% chance it works, it's a good deal. 5% chance of medical science working is a pretty low opinion of medical science by most people's opinion, right? 

Tyler Cowen: But don't you read about freezing and crystallization and think what will come out of this is simply in no way me, my brain is gone, it's destroyed. What if you burn your brain to ashes and put the ashes in a jar? There's still in some [unclear] sense, maybe a way you can be reconstructed. There's some chance of that, right? But I think the chance of cryonics is about the same.

Robin Hanson: About the same as somebody reconstructing your ashes? That’s a lot less than 5%, right?

Tyler Cowen: Well, if you believe Frank Tipler, it's more than 5%.

Robin Hanson: Yeah, but we don’t, right?

Tyler Cowen: I'm just saying cryonics to me is in that class of events where the chance is something like the chance of completely revising your whole view of the world and what's possible. I'm not sure what number to give that, but I think that's the kind of move that needs to be made.

But once you're then in this Pascal's wager sort of game where very strange things with high payoffs can happen, it's hard to also apply a lot of standard decision theory. No? Or do you think you can just do EU max?

Robin Hanson: I don't think we're anywhere near that strange world. This seems to be a very simple sort of hypothesis, very close to ordinary standard physics and biology.

It's not imagining there are angels or aliens hiding in the bushes. It's just assuming or believing that you are the structure of your brain. That what's worthwhile about you is all encoded in the relationships of the major structures in your brain. And if we could save those we can put you back together again.

Tyler Cowen: Well, what do you think of the recent literatures that talk about how much of your brain in a sense is in your gut and how your brain and body interact and how the brain is dependent upon hormones? When I read all that put together, it's just hard for me to see freezing the head as saving that much of you.

I think the best you could do is get some reproduction of you within a computer that would mimic some of your thought processes, but not really be conscious. And I would value that pretty much at about zero. 

Robin Hanson: So it's the consciousness part that you’re skeptical about, or you're skeptical about reproducing the input-output behavior as well?

Tyler Cowen: I'm most skeptical about consciousness, but how consciousness even differs from reproducing the input-output behavior is a tough question. 

But I actually think the chance of angels is higher than the chance of cryonics. I'm a non-believer, but if I assign belief, say probability 1 in 20. Maybe I'm wrong about the argument from design. And then given that there's a deity, what's the chance of angels? To me that's more likely than cryonics in a scientific way.

Robin Hanson: So, is your skepticism about cryonics based on physics, or some more metaphysical relationship of consciousness to physics, or skepticism about current brain science, or skepticism about the social institutions required to make this work?

Tyler Cowen: I would say all of those, but just that it's really hard. Look at our Bloggingfrozenheads dialogue. It took you hours to set this up and I couldn't even do it. And we're relatively smart people. So to set up a Bloggingfrozenheads dialogue takes hours, and then I'm supposed to think that my head can be frozen and brought back with even a one in a million chance? So I’m just not there.

Robin Hanson: Okay, so it's just a pure surface implausibility. It's not based on the structure of the pieces of the argument, right? 

Tyler Cowen: Absolutely. That's exactly right. 

Robin Hanson: So this does come back more to our initial hypothesis for our disagreement, which is I prefer to have some structure where I rely on the conclusions of the structure, even if my intuitions will influence which pieces I choose for the structure.

And you will more have just a reaction to the overall conclusion – that even if your subconscious is having a structure, you're not very aware of that. 

Tyler Cowen: Well, I think I'm somewhat aware of it, but you're always wanting to give me a number. What's the chance you'll be frozen? What's the chance you'll be thawed out? What's the gain of being thawed out? Convert it to an expected utility equation. Again, I pretty quickly put it in the class of cases of extremely unlikely. Payoff might be high. By Pascal's wager, it gets murky. But in practical terms, I'm inclined to think I should evaluate the prospect in terms of what kind of signal does it send.

And we all know the great enemies of cryonics are, I would say, wives, right?

Robin Hanson: Well, one of the major blocks, I suppose.

Tyler Cowen: One of the major blocks.

Robin Hanson: Since most of the customers tend to be men. I guess an interesting related subject is that there are many topics that most academics consider silly, and they also are reacting to these topics at a sort of surface level.

Just the very idea seems silly. I think that you don't share that reaction in many areas.

Tyler Cowen: No, I don't think thinking about cryonics is silly at all. I think it forces us to get at very real questions as to what we actually care about and how we evaluate [the proposal], as we've been trying to do.

Robin Hanson: But at the end of the day, your judgment is based on just a holistic evaluation that seems to say this has a really low chance that isn't based on some structured calculation of the chances of each parts or something. It's just: overall it looks unlikely, and that's it.

Tyler Cowen: I don't like that word “just”. In my view, your evaluation is as holistic as mine, but it fits into a holistic picture with robots and science fiction and big radical transformative changes, which you find it appealing for holistic reasons, and that's fine, but I'm less that way and I think of yours as a kind of -polism, cloaked in logical atomism.

Robin Hanson: So we should both agree, and I will certainly agree that the vast majority of my thoughts are not accessible directly to me. 

Tyler Cowen: Sure. 

Robin Hanson: So that whenever I think about something, my mind goes through an enormous detailed calculation and it only leaves me aware of a tiny fraction of it.

Tyler Cowen: That's right. 

Robin Hanson: So I certainly grant that. Nevertheless if I am skeptical and doubt this vast calculation, I feel a little more in control of — or at least holding it accountable, if I can pin it to some sort of an analysis structure — some set of parts. And I guess you don't feel that's as necessary or useful.

Tyler Cowen: Maybe one of our core differences is we have different ways of wanting to feel in control, and that leads us to different intellectual and emotional positions to achieve those ends. And then our self deception works backwards and leads us to the idea that somehow we came upon these ideas because they're more rational.

That's what I'm inclined to think. So my notion of feeling in control is a certain comfort with not being in control in a particular way – I think, to a greater extent maybe, than you are. Whereas your notion of feeling in control is to have worked through simpler models which are very foundational. And even if at some level you admit that's not all there is, that's still what is giving you that [bad audio] feeling.

So relative to my point of view, you're overvaluing those models.

Robin Hanson: Well, I think more likely I'm trusting my intuition less than you do. My unstructured undisciplined intuition.

Tyler Cowen: I'd say it's exactly the opposite. I think you're trusting your intuition more because you're making it look like the results are coming from the calculations, whereas I'm more bluntly facing up to the intuition.

Robin Hanson: Well, for fear that we've lost another half of our audience... 

Tyler Cowen: That's the goal here, isn't it?

Robin Hanson: Not entirely. We'd like there to be more than one guy left at the end. Maybe we can ground this in another of our concrete areas of disagreement. Do you have a favorite?

Tyler Cowen: Well, the more I think about it, the more I think that we tend to agree. We have lunch with Brian and Alex.

Robin Hanson: Yeah. Whenever it gets concrete…

Tyler Cowen: Whenever it gets concrete, usually you and I are allied against Brian. Unless it's meta-ethics. And when it comes to Alex… Yeah, I think of the four of us, you and I tend to agree the most, actually. 

I just think there's this extra layer on top of you, which I'll call broadly the transformational side. There's this extra layer on top of me, which you could call the messy pluralism humanities side. And that gives us one level of disagreement. But other than that, I think it's actually remarkable how little we disagree.

Robin Hanson: We must be able to find something concrete here. Or maybe we can elaborate what it is we agree about that other people disagree and how sure are we?

Tyler Cowen: Can we talk about a movie? I bet we disagree about movies a lot. Can you name a famous movie and tell me what you think about it and we'll try to figure out why we disagree. I predict we will.

Robin Hanson: Well, the last movie I saw was Watchmen.

Tyler Cowen: That I haven't seen yet, so we won't get too far there.

Robin Hanson: I like Doubt. That was the last movie I saw that I liked a lot.

Tyler Cowen:  I was convinced I wouldn't like it, so I didn't see it. I felt it was going to manipulate me by trotting through standard cliches about, like, uptight nuns and people taking stands for justice. And then can we really know? And a lot of just toying with the audience and not delivering much substance, am I wrong?

Robin Hanson: It did a good job of putting people on the edge where they should have been on the edge, but they all fell to one side or the other pretty much, of deciding if the man was guilty or not. That's what I was impressed with.

Of course that's a manipulation job. That's what a movie does. What's a movie you saw that you liked a lot?

Tyler Cowen: Well, I'm trying to think of a movie you would know. Not too long ago I saw The Godfather series again, I and II. Do you know those?

Robin Hanson: Yep. No, I haven't seen that again.

Tyler Cowen: Okay, how about Blade Runner?

Robin Hanson: You haven't seen The Wire either, have you?

Tyler Cowen: I've seen about six episodes of The Wire. I thought actually The Wire had phenomenally good production values, but I was more or less unable to follow the narrative, the reasons which I ascribe to my own cognitive defects. So I didn't actually enjoy it, but I can see why people think it's such a good show. You like it more?

Robin Hanson: It's my favorite, I think, TV series that I've seen but it does focus less on characters and the character interaction than the setting and the whole setup. It's a very believable, realistic setup to the extent you can have there. But that often does come at the expense of the characters.

The opposite extreme is Battlestar Galactica, which both you and I like. Which has a really silly setting, not very believable at all. But it has great characters and character interactions.

Tyler Cowen: I think what we both like about a lot of science fiction is the notion of alien being – that you're forced to confront ways in which people could be different or could have evolved differently, and you need to think through some evaluation of that. And then you need to develop mental models of how, say, the Cylons will make decisions. 

And that's a lot more interesting than a lot of what is on TV. You could say the same about Blade Runner. So you ask the question, who in the movie is a replicant? What does it mean to be a replicant? How does it matter? In what ways are Cylons actually different? 

And there's some conceptual bite to that, which I enjoy, that I don't quite get in The Wire. The Wire to me is actually too much messy, pluralistic detail. The Wire makes me feel like Robin Hanson. 

Robin Hanson: Good!

Tyler Cowen: I see The Wire and I want to visit your homepage.

Robin Hanson: I was gonna say that I have more of a love-hate relationship with literature and fiction than you do. I'm human and I can't help but love it, but I also fear that it distorts my thinking in severe ways. 

And so that comes up in this topic we have right here, and the topic of my conference yesterday, which is that a lot of science fiction or other stories focus on the dynamics of a few characters and interaction because that's what people are most interested in. But often they pretend to talk about some larger society and some larger arc of history.

But they usually do a lousy job of showing how that really works because they're so obsessed with acting as if it mattered on the choices of a few people and how they were interacting in some small social network. 

Tyler Cowen: That's right. Fiction will overrate decisiveness, I think, otherwise there's not that much of a novel. I've often wondered what would it be like to write a novel about Hayek’s Spontaneous Order. And it's a stupid question in a way, but it's interesting to try to think through how it would actually work, what would be on the page. It's not obvious to me. Is such a novel simply Debreu’s Theory of Value, is that what the novel looks like? 

That's what Debreu himself thought. He saw himself as a kind of a successor to Proust. So general equilibrium theory was another way of thinking about the human condition where the individual is of [measure size zero]. And that's what this quote unquote story then looks like.

But I'm more skeptical about fiction than you are because I tend to think what's there already is, in any case, a distortion and fiction is far more likely to help you limit distortions than to create a new one. But you don't think that way?

Robin Hanson: I mean, it's possible that for certain people in certain situations that certain fiction could help them reduce their distortions, but on average over all people and all fiction and what they consume, it seems overwhelmingly the other way.

Tyler Cowen: Well, but I'm talking about what you consume. So if a woman reads a romance novel and develops an overly high expectation from her would-be husband, that's a simple example of a distortion. But even that might be in your self-interest to have that distortion. 

Robin Hanson: Yes, it might.

Tyler Cowen: Because your ability to receive that distortion comes from somewhere. Just because it's evolutionarily efficient, it doesn't mean it's good for your happiness, but it could be.

Robin Hanson: Here you're being a reductionist of the sort that I usually am and I would fear you would accuse me of.

Tyler Cowen: Okay, I'll accuse you of it. 

Robin Hanson: Of saying that my presumption is that people evolved this tendency to love fiction because it benefited them personally, And not just for some grander insight into the world or contribution to social cohesion or something. But that sounds awful reductionist, doesn't it?

Tyler Cowen: It's reductionist, but clearly it's true. But you could say the same in a more direct and obvious way about food and sex. But there's still a cognitive element to eating or having sex, or for that matter, sleeping, and the fact that they're fun for obvious Darwinian reasons.

It doesn't make us that skeptical about the cognitive element, does it?

Robin Hanson: Sure. Obviously people's cognitive elements regarding sex are pretty well [disturbed], don't you agree? I would've thought that would be one of the very widely accepted things. For stories, I think the best stories focus on – if people can't help but want to focus on a few characters and how their lives interact with each other, then the best way to be realistic about that is to have these characters not be very influential about the course of history around them. 

I respect a story where characters are placed in an alien environment – maybe alien to me, but that really existed in some time or place, or could exist – but I respect that a lot more if they are just small ponds in a vast world. But, say, Battlestar Galactica, that's not acceptable, right? It makes the story more interesting if these characters you're focused on not only have interesting relations with each other, but somehow are pivotal with respect to the world they're sitting in.

Tyler Cowen: Sure. But don't be so suspicious of decisiveness. Maybe we need decisiveness to learn. So just going through life, it's easy enough to see how you can learn from that on net, even though you'll pick up some distortions. So if what fiction is doing is simulating a course of life, for the simulation to be efficient, you need a lot of decisiveness, right? It's just like an economic theory. You manipulate the variables that matter. That doesn't mean that all variables matter. 

So fiction is doing the same and you need to be cautious, but with some basic core elements of caution, isn't it just a cheap way of getting other people's visions of how things work and then evaluating those visions and taking in more insights?

Robin Hanson: It could in principle be that. I don't believe that's what it actually is, and I'm not even sure you believe that. So since this comes down to the issue of how large a fraction of human behavior is signaling we both seem to agree it's large and probably larger than most people attribute.

But when we come to fiction, I'm tempted to think that the main thing you get out of fiction is the ability to show your associates the kind of emotional reactions that you have to the fiction. The kinds of ties you feel to it. The notions of loyalty and attachment that you have to the characters are important signals to the people around you of the kind of people you would like and become attached to, or your beliefs about how you would act in a world is an important signal to other people around you. Whether you think you would save the damsel from the castle or the dragon. Whether you would stand up in the war, or run away. We’re trying to show the people around us at least how we think we would act in those situations.

Tyler Cowen: I think it's also important that you show people how you like fiction, not simply what you like. So if you just announce, “I love Tolstoy”, you can’t infer that much about a person. Maybe they're educated, or from Russia, but when people see how and why you like Tolstoy, it's as if it creates a kind of map for getting people in touch with each other. 

But again, in cognitive terms, I think there are a lot of pretty simple stories where, yes, a large degree of it is about signaling, but it still has cognitive benefits. You learn something about the world and that's partly why it's a good signal. You're showing people what you are capable of learning about the world, and you want to construct alliances and affiliations with other people capable of learning the same things. 

So in some ways it makes me more optimistic about fiction. I think if it's for signaling, there must be something real in there to make the signal hold up. It’s not just arbitrary attachments like with high school football teams.

Robin Hanson: The question is what it is you're signaling. If you were signaling your ability to learn things about complicated situations and to master them, yes, then you would be learning things as you signaled your ability to learn things.

But if it's more signaling that you think you would not run away in a war, or signaling that you would feel attached to your wife and never leave her, that you find despicable the rake and his behavior – those sound like trying to reassure people of your alliances and your loyalties.

Tyler Cowen: Sure. But keep in mind, even that can be a good thing. And I'm not defending the view that all of fiction is valuable, but that there's a relatively small set of works, sometimes known as the canon, which are intellectually very valuable.

And to me, the best argument for them, it's a bit like the best argument for economics. There's plenty you can say against it, but when you meet people who don't know any of it, you realize there's really quite a bit to it. And it's basically my view with fiction. People who have very little sense of complex fictional narratives have a hard time conceptualizing and articulating a lot of social facts.

And you can argue back and forth, what's cause and what's effect? But that the correlation is real, I think, is undeniable – and some of it is cause and effect.

Robin Hanson: Well, I guess we could touch back on economics and I'm sure neither of us are eager to talk about the current financial crisis. But the related issue to what we were just talking about, of why people don't like economics: it's not just that it seems boring or complicated – it seems evil.

Tyler Cowen: Well, maybe it is evil.

Robin Hanson: Is it? 

Tyler Cowen: As I understand economic science, the abstract microeconomic level, saying there's nothing special about the notion of loyalty or affiliation. And that is considered by many people an evil message, even if it's not directly perceived as such. But this notion that is just another value weighed off against all other values is exactly the opposite of what loyalty's supposed to be. So economists are seen as evil – but I'm not ruling out the fact that maybe we are evil.

Robin Hanson: Not ruling out is a pretty weak statement. I'd come down and say economists aren't evil. I would say that loyalty and affiliation have more local value than global value. And that's an insight that economists contribute to the world.

Tyler Cowen: That's possibly the case. But if you need local groups to produce public goods to do things of great global value, then loyalty having local value might actually be globally efficient. And I think that's the kind of mechanism that you can make sense of within economics, but a lot of economists tend to undervalue. 

So if you say the word “participation” to most economists, they roll their eyes and they get upset.

Robin Hanson: “Community.”

Tyler Cowen: “Community” is another example. We could go on with these words.

“Need” – I bet that one. I remember Walter Williams mocking “need” when I took his class in 1982, saying all demand curves slope down.

Robin Hanson: It does a good job.

Tyler Cowen: Does a very good job. But I'm not sure that on these issues, common sense morality is wrong. I think it has a kind of deeper evolved wisdom that isn't quite right as it's presented, but there's some synthesis between that and the economic way of thinking that we haven't quite yet made.

Robin Hanson: Nevertheless, since you're an economist, I'm pretty sure that you actually think the economist take isn't too far wrong. That if we were gonna take the standard economist position and the sort of standard naive human position, the economist is more right at least about what they're claiming. Perhaps not about what the ordinary person misunderstands them to say. 

Tyler Cowen: I'm less sure that's true if you adjust for IQ, income and a number of other factors. So economists are gonna look good because on average, they're smarter, or they've had more education. 

Robin Hanson: Okay, adjust away. 

Tyler Cowen: But again, holding all that constant, if I ask myself: typical economists – not people who I have chosen to affiliate with through the mechanism of being a colleague – but they're not such a thrilling group to me. Are they to you? I don't think so.

Robin Hanson: Well, they're the community I've chosen to join. So compared to the others, I was attracted, impressed.

Tyler Cowen: Well, it might have been a mistake since you're talking about how attraction can mislead you, right? This has to be one option. So they have models with a small number of variables that are relatively atomistic and relatively transparent. So maybe you fell in love with the models, and just as romantic love can be distorting, it gave you a distorted picture of economists. What actually about us is so admirable? Can you explain that?

Robin Hanson: I think we do a decent job of trying, at least, and often achieving some sort of neutrality in figuring out what is good overall – all things considered, weighing the interests of different people in a moderately neutral way.

So then the question is, do we actually neglect the virtues of loyalty and affiliation? Or at the scale we're interested in, do most of those virtues cancel each other? Sure, a local work group ought to feel some loyalty and affiliation with each other, but most economists who study team production would acknowledge that and accept that. 

It's the much larger scale affiliations and loyalties that seem to most economists to have a relatively little net value: loyalty to your city, or to your country, or to your profession. Those are the kinds of affiliations and loyalties that don't seem very instrumentally useful to economists. And nor do they seem persuasive at some sort of aggregate moral value. But do you have a counterargument? What are we missing?

Tyler Cowen: I think we're getting into territory of disagreement here, which is a good thing. I don't view economic methods as very neutral at all. They might be extremely useful. I view them as a kind of tool or mechanism for generating new ideas, but the kind of new ideas they generate are extremely biased. But I don't mean that necessarily in a negative way. 

But I think economics tends to breed the way of thinking that all values are based in ordinal preferences and can be expressed by willingness to pay and be paid. And that's highly controversial, but in any case, it's extremely non-neutral and it's certainly not what most people think.

Robin Hanson: Wait, the idea that values can be expressed as ordinal preferences is non-neutral? 

Tyler Cowen: Absolutely. 

Robin Hanson: Because neutral values have non-ordinal preferences?

Tyler Cowen: No, I'm not sure what neutral values mean, but I think economics has a pretense of neutrality. And most economists, including yourself, believe it, but I think it's just plain, flat out wrong. And economists are not any more neutral than [Platonists]. You could argue they’re correct.

Robin Hanson: Who are we non-neutral in favor of and who against? Where's the division that we're…

Tyler Cowen: The very way you asked the question, in terms of who – so all values must be reduced to the preferences of a single lone individual. It's clearly not how most of the world thinks about philosophic values.

Robin Hanson: So you're saying we're not obviously neutral against any one person or group. It's certain kinds of values that we’re not neutral regarding.

Tyler Cowen: That's correct. Which may or may not be linked to particular individuals.

Robin Hanson: And the values of affiliation and loyalty are prime examples of the kind of values that we're not neutral regarding.

Tyler Cowen: Possibly aesthetic values would be another instance. 

So I'm happy to say that for substantive reasons economists are often making very good value commitments, such is the relative cosmopolitanism of economic science. But it is very much a substantive ethical commitment and we don't want to view it as such. And I think it makes us collectively into a kind of coward, where we hide behind science and think we're being value neutral when we're absolutely not being value neutral.

Robin Hanson: Is it clear that we actually have an option that we're rejecting? So one standard defense would be to say we have better elaborated tools and methods for considering the kinds of usual values we have.

We've tried to consider analyzing the values of loyalty, affiliation and aesthetics, but we just didn't get very far, or when we did get some ways, it didn't seem to make much difference. We did what we could. Is that wrong? Have we really neglected something we could have done more with?

Tyler Cowen: I think that's a more reasonable and modest attitude. But when you say “we” could have done with – if the “we” is economists acting as economists, the fact that we can't do more is maybe not so persuasive to other people. So if a philosopher says he can't get very far with a downward sloping demand curve, we hardly take that as a great defense of the philosophic method. We think, my goodness, something is wrong, these people can't even handle the downward sloping demand curve.

Robin Hanson: Can we be more concrete? Which is a recommendation economists would make based on their analysis, which is likely wrong because economists didn't consider, say, aesthetic issues where some expert in aesthetics would've recommended a different choice.

Tyler Cowen: I'm not sure there's such a thing as an expert in aesthetics.

Robin Hanson: You're saying economists were all right given what they considered, but that was like asking a philosopher what he could do with the demand curve. If there’s somebody else who could do more with it, then we should listen to them, but if nobody can do more with it, how are we at fault for not doing more than we've done?

Tyler Cowen: I think a simple example is the literature on valuing human life, which I take to be an enterprise fraught with a lot of hazards and not quite philosophically sound. So to simply ask the question: let's compare a certain number of lives to a certain amount of money, and just think really hard about that. To me it’s probably a better way of proceeding than to pretend we're valuing a life by measuring the value of risk reduction. I don't think those two are the same enterprises at all. So that's an example where very often I would side against the economic perspective.

I don't have any illusions about how stupid the process is to sit down and try to talk about how much a life is worth.

Robin Hanson: Are you saying that ordinary choices to reduce or increase risks are not related to people's value of their lives? 

Tyler Cowen: Related is a very weak word. 

Robin Hanson: Or not very directly related. If I am willing to pay $100,000 to reduce my chance of dying by 10%, that doesn't say very much about how much I value my life?

Tyler Cowen: That's exactly what I'm saying. Unlike you, I am not so close to contractarianism. And I think that's another area where we tend to disagree. So although sometimes here you make the claim, if people behind a veil of ignorance could agree, they would agree to do something like this. And that's a path which I almost always reject as a way of solving an ethical problem. 

It's actually related to freezing your head. The question of what you have to do to the people to get them behind the veil makes them not those people anymore. So I don't think you're getting much information out of the thought experiment.

Robin Hanson: What I see is that you're saying that you disagree, but if I probe for the reasons for your disagreement, you don't tend to elaborate very much here. So you just say, I just have an intuition, or it just seems to me different.

Tyler Cowen: No, I'm giving the reason. The reason is that there's no logical argument from the premises to the conclusion. I don't have to have an additional reason. The conclusion is never established.

Though economists who have this approach, they always wanna know, what's your reason, why do you reject this? I reject it because it's never been established. That's what I would say.

Robin Hanson: Don't you have to have some reason for your alternative point of view? Like I was saying at the conference yesterday, many people had the opinion that economists made assumptions and that the problem with economists is they had models that had assumptions and therefore since the assumptions could be wrong, that economists weren't very useful.

But surely it's not enough to just reject the economic analysis because it makes some apparent assumptions. Don't you need some other basis to draw an alternative conclusion?

Tyler Cowen: Sure. But if the best we economists can do is throw the burden of proof back in other people's laps, that to me makes a lot of economics look fairly weak.

Robin Hanson: But that's the way it's going to always be in any discussion, right? You make an argument and if somebody says, I don't like that, aren't you – have some burden to make a counterargument, offer some structure, something?

Tyler Cowen: I think economists are far too likely to think that the default position is theirs, that anyone trying to argue for anything other than economic methods is somehow postulating some kind of strange value, totally devoid from individual preferences. And until the other person can establish the validity of that value, just somehow left hanging in the wind, and economics should march forward and proceed. And I think that's exactly wrong.

Robin Hanson: So let's go to the back of the value of life then. So we have some concrete, familiar ways that economists try to estimate the value of life in terms of choices people make that affect their chances of living.

If those are not very good indications of the value people place in their lives, what alternative structure analysis do you suggest? You can't just pick a number out of the air and say, that feels to me like that, can you?

Tyler Cowen: No. But what we do as a matter of practice is we create institutions which make these decisions. And the algorithm is not a theoretical one. The algorithm is in essence the description of how the institution works. 

So I think it's fair enough to say you can improve the working of the institutions by giving them a better understanding of economics. I'm all for that, but when people actually want the economic reasoning to displace the institutions epistemically, then I tend to think something's wrong and that we've overstepped our bounds.

Robin Hanson: If we can't use economic analysis to criticize the conclusions that an institution comes to…

Tyler Cowen: It's not what I said. 

Robin Hanson: Well, how can you criticize the conclusions an institution comes to? If I say, it looks like they're valuing life way too highly around around, say, garbage dumps - the value of life seems to be a billion dollars there - and valuing life too little with children's nightwear that burns, according to ordinary economic analysis, that would be a criticism of the political process producing those decisions. 

If you say, no, that's valid, the process is embodying things you don't know. How do we evaluate that institution if I can't use the sort of analysis to compare its outcomes with some independent standard?

Tyler Cowen: Sure, that's a perfectly valid criticism of the institutions. But economists, whenever they get in trouble, they wanna walk back to sticking with the dominance principle. They find some example where there's a way to make everyone better off and no one worse off by reallocating something. And those examples are all fine and well, but they don't describe that many real world situations.

And when you're just out there, you're sitting in Bethesda at the Consumer Product Safety Commission and someone asks you, should we spend $2 million to prevent a little kid from dying? Economics still doesn't have a good answer. I'm not sure that in the sense of a global ranking, there is a single correct answer, a single number for how much should be spent to prevent a kid from dying.

Robin Hanson: Wait a minute, they should pick a dollar cutoff, shouldn't they? If there's three children in front of them, shouldn't they use the same dollar cutoff for those three children to decide which – either they spend on all three, or on none of them, or on some of them, based on some characteristic that would reflect how easy it was to help that kid?

Tyler Cowen: Well, it's Adam Smith's point that we're not arranging society like pieces on a chessboard, and our moral theories should reflect that fact. So if you want to put all moral questions into a single global ranking and we need to make sure we're transitive and we can reallocate wealth however we want to get to the better outcome, then economics will look pretty good.

But when you get back to the messiness of how real world institutions work, then economics to me looks much weaker. And I think that latter comparison is the more relevant one, I think precisely because that's a messy pluralistic comparison that you are rebelling against considering it because you're like these cleaner, neater, fewer variables kinds of comparisons.

And this gets back to why on the meta level I think we disagree where we do, this is a very good example of that.

Robin Hanson: Well, when you say messy pluralistic comparison, you mean – what I see is some political process produces an outcome and you’re granting it the benefit of the doubt that it's somehow messy and pluralistic in the right sort of way you like. It could just be wrong, right? It could just be messed up, broken. How can we tell a broken process from something that's doing a good job of including messy pluralistic analysis, whatever that is?

Tyler Cowen: I think we have a highly imperfect politics. Economists can improve it locally in some ways. But when it comes to broader moral theories, economics doesn't get very far. That's how I would put it. I wouldn't say I always give politics the benefit of the doubt. If anything, the contrary.

Robin Hanson: Well, for example, in cryonics I'm using my simple value of life calculation to say, if you value years of life in the ordinary way and even have a moderately ordinary discount rate, how much should you be willing to pay for a chance to come back and live another thousand years?

I use that to calculate roughly a 5% cutoff, of saying for most people, a 5% chance should be enough to make it worthwhile. So that's exactly where you say, well, that's a crazy calculation to make because lives don't have values and lives in the future don't have values comparable to lives now, or at least not in any sort of way that any calculation could usefully inform. Is that what I'm hearing?

Tyler Cowen: That's a possible argument, but I'm making a slightly different argument. I'm saying that the value of ex ante risk reduction is not the same as the ex post value of a life. If you're using just ex ante risk reduction to value ex ante risk reduction, while I think that's highly imperfect, I think it's a lot more defensible than looking at how much people will, say, spend on smoke detectors and then deciding how much to value the life of the kid who's gonna die for sure. I think those are very different comparisons.

Robin Hanson: But we're almost never valuing the life of the kid for sure. We're almost always choosing between ex ante risk reductions, aren't we?

Tyler Cowen: I think very often we're valuing the life of the kid for sure. We may not know which kid, but I'm not sure that's morally important. We know policy X or policy Y will kill or save roughly so many people.

Robin Hanson: Ex ante.

Tyler Cowen: We may not know the exact number, but ex post for sure, some number of lives will be saved or killed. And again, I don't think risk reduction is the same kind of act of valuation.

But Chris Massey wanted me to ask you a question. I'm not sure I can phrase it properly, but he wanted me to ask, will prediction markets ever really take off? And if so, why haven't they up until now, or do you think they have?

Robin Hanson: This is one of those topics I think you and I largely agree on, and we both stated our opinions in writing, but we can clarify. So I think we both agree that from a simple instrumental management perspective they are a cost effective way to increase the accuracy of important organizational estimates.

In the test so far, they seem to substantially improve accuracy, often enough that their moderate costs would be justified on a ordinary corporate cost benefit criteria. 

Tyler Cowen: And they don't ever seem to do terribly. 

Robin Hanson: Right. So then the downside is relatively minimal. They don't seem to do much worse, and they often do substantially better. Yet we both agree that relative to this apparent large potential, very few organizations are very interested and the few that are are often more interested in being in on a fad, or doing something cute or ideological than the straightforward cost-benefit calculation that would compel me, say, to recommend it.

Tyler Cowen: That's right. And why are they not more interested?

Robin Hanson: I also think you and I agree that most managers’ job is not well described by comparing costs and benefits and making the decisions that benefit the organization most on a cost-benefit criteria. That management's positions are often more a matter of theater and organizing political coalitions. 

Tyler Cowen: And building up morale.

Exactly. Presenting an impressive image that people can get behind. Giving people the impression it's a continuing going concern that has good relationships with other important parties, and this appears to – and reasonably in many ways, often does – threaten that. 

Tyler Cowen: And how does it threaten it, or why does it threaten it? Why can't you use the prediction markets to figure out the best way of staging the theater?

Robin Hanson: Well, because theater isn't supposed to be staged in public.

Tyler Cowen: Well, it's always staged in public in a way.

Robin Hanson: It’s presented in public. The decisions about who will be at what part of the stage are made in private.

Tyler Cowen: That's true. 

Robin Hanson: And then the theater is presented in public.

Tyler Cowen: So you think it's this violation of the private and public realms that makes it so unappealing to so many people. Do you think it just makes it intuitively unappealing? Or do you think the bridge actually cannot be leapt over?

Robin Hanson: I certainly don't wanna make any sort of absolutist claim. I wanna put this in the company of many other innovations that just had a tough, long slog to eventually become accepted. But I certainly think it's possible and even likely that eventually this will become a standard accepted tool. And at that point it will have added a lot of value. 

But the question is, what value do the initial adopters get personally from their initial adoption?

Tyler Cowen: Can we think of institutions that reflect things that used to be done privately and now they're done publicly, and it's accepted that they're now done publicly? Are there examples of that?

Robin Hanson: I don't know about the private public, but there's certainly a lot of elements of organizations where, once upon a time, I'm old enough to remember when most managers would say with a bit of pride, “I don't understand computers.”

Tyler Cowen: I remember that too. 

Robin Hanson: It wasn't an embarrassing thing for them to say. They thought it spoke well of them.

Tyler Cowen: That's right. It's like saying now, “I don't bother to Twitter.”

Robin Hanson: Maybe something like that, right. 

Tyler Cowen: Do you Twitter, by the way? 

Robin Hanson: I have Facebook updates, but I haven't bothered to do the Twitter because I don't want to have to manage too many different sites that I'm supposed to update. But yeah, that's an analogous thing. Now, maybe in 20 years all managers will Twitter, or update their Facebook status lines often. But now it seems beneath them, it sends the wrong signal to be doing that sort of thing.

Tyler Cowen: But workplaces have moved to using wikis with mixed results, but you see quite a bit of it. And that's an example of information gathering. It used to be private and now it's public. I think a lot of that will stick and probably be expanded.

Robin Hanson: Moderately, yes. I don't think the key distinction is the public versus private. So for example, many managers use a wide variety of consulting and analysis techniques, but they prefer them to have a bit of a fudge, a few fudge knobs. They don't mind doing a spreadsheet and some charts, or even a statistical analysis, as long as – if the answer comes back not quite the way they wanted, they've got some knobs they can turn to make it better. 

And this is of course one of the things economists provide, our standard methods have a number of knobs we can turn to give people answers they more want to hear. But with the prediction market, you're committing to publicizing a number that you don't know what it's going to be ahead of time. Not only that, you're publicizing the fact that you didn't make the number. It wasn't your brilliant choice to choose the number at the last minute. It wasn't your brilliant choice of an approach or of a consultant that determined that number. It was something pretty much out of your control.

Tyler Cowen: So it makes leaders look less impressive maybe, is the key characteristic. And it embarrasses leaders at times. Because leaders can't be completely neutral on vision prior to the operation of the market.

Robin Hanson: Right. So those are some of the barriers: the inability to fudge the answer and the lack of some sort of tie to it. In most organizations, most managers, reputation is based largely on sort of the things they know that other people don't and the kinds of sources they can draw on that other people can't.

So they want some sort of comparative advantage in why should you hire them rather than hire somebody else who will ask them questions once in a while.

Tyler Cowen: What's the implied comparative statics prediction? Is it that you'll see prediction markets and firms where leaders don't matter, like maybe real estate sales? Or is the implied prediction you'll see these markets and firms where the leader is so strong, maybe a privately owned firm where the CEO and whatever is all rolled into one, that the leader can afford to have some slack?

Robin Hanson: This is exactly the sort of question that I would usually ask Tyler, because like I said at the very beginning, I trust Tyler to know, when push comes to shove in real organizations, what's really going on.

Tyler Cowen: I would think it's when you have a leader so strong that there's a kind of credible counter-signaling: “I'm so strong, I'm willing to turn this decision over to others and let them feel involved, and that will make me even stronger.” But even then, you're keeping it within some pretty well defined limits. That's just my intuition.

Robin Hanson: That sounds plausible to me. There's this old speculation about whether big changes are more or less likely in hard times or in really good times, or somewhere in the middle, which is a speculation we're testing out during this economic crisis. And which kinds of big changes are possible in these different times.

I recently blogged about the question of whether you could ever have a crisis big enough that people would jump to something libertarian. And I thought that seemed unlikely because that just doesn't seem to be the kind of thing desperate people grab for.

Tyler Cowen: They want more control. The feeling of more control at least. We have about two minutes left. I thought we'd give each other a parting question that didn't take too long to answer. How about that? My question for you is two-part: how long do you think the crisis will last for, and how is it changing what you do as an economist, if at all?

Robin Hanson: So, I did not get much of a training in macroeconomics, which was fine with me because I specialized in other areas. So I don't feel very expert in the area.

Tyler Cowen: Gives you an advantage perhaps. 

Robin Hanson: I hold to a principle that I should back off on subjects I'm not very expert on. So I've decided to back off on this subject and say, you know what, I don't know, listen to other people about this. So that's been my official and continues to be my honest opinion about the crisis.

I'm moderately skeptical about – many of the policies being enacted at the moment don't seem very related to the crisis itself. They seem more opportunistic, a chance to just do something different now that there's a crisis. 

Tyler Cowen: I agree. 

Robin Hanson: But beyond that I don't have a very strong opinion and therefore it hasn't much changed the way I do economics at all for that reason.

Tyler Cowen: Okay, fair enough. 

Robin Hanson: My question for you then will be two-part I guess, but very closely related. Do you see in other people a correlation between having more influence and having more conventional opinions? And do you see a correlation in yourself that way, across time? And are you happy with that?

Tyler Cowen: I think when people reach out to larger audiences, they promote more conventional opinions rather than less conventional opinions.

Robin Hanson: Is it because they just choose what they talk about, or do they actually change their opinions?

Tyler Cowen: It's a mix of both, but they actually change their opinions. Partly they have chances to earn more money, and partly they want to be liked more.

I think a lot of it is subconscious, people just wanting to be liked and not to be a pariah everywhere they're invited. So I think you have a lot of very smart people where this makes them uninteresting. So on that, to me it's tragic, actually. I would rather read weirdos, I would say.

Robin Hanson: So you see this correlation in other people, and you have an understanding of it, that it's causal in part from influence to conventionality.

Tyler Cowen: Correct. 

Robin Hanson: And then do you see this correlation in yourself, like across time or across the variation of when you've had more or less influence?

Tyler Cowen: I don't think so, actually.

Robin Hanson:  So you think you’re a counterexample to the general trend that you just happen to not have the correlation that others do?

Tyler Cowen: Well, I'm only 47. I would say when I was 19, maybe I believed weirder things than I do now, but since, say, I was 20, maybe in some ways my views have become more weird. So there's an initial age effect, which I don't think is driven by incentives. 

And since then, when I started blogging, my goal in part was actually to stay a bit weird. And then it was the problem. More people started reading it than I had expected. So maybe that's made me too mainstream and now I figure I need to do something weird again. But I'll manage, but don't doubt that.

Robin Hanson: Okay. That somewhat surprises me because it seems to me that over the time I've known you, you have become substantially more influential and you also seem to be substantially more conventional. But I don't have a data set unfortunately to analyze that. I'll have to admit less confidence in it according to my usual standards.

Tyler Cowen: Well, offline I'm curious to hear how I used to be weirder, but I think we're out of time now. If you feel the need to say it, by all means, go ahead. But I guess that would be the question.

Robin Hanson: I enjoyed this conversation very much. 

Tyler Cowen: So did I.

Robin Hanson: Thank you for talking, Tyler.