 What I'm going to talk about today is a piece of my research here at the Berkman Center. One of my projects is digital institutions, and I hope that by the end of this you will see at least my own rather odd and idiosyncratic take on digital institutions, but understand why it's so fascinating to me and why the digital world is an important place for this work I've been doing. And I'm calling it modeling cooperation for first and second lives, suggesting a general case. Those of you who are Berkmanites probably don't need to be told that second life in this context is a metaphor for the digital world rather than a literal reference to the second life digital platform. I've had a few other people not out of the digital world be confused about what did I mean by that second life, so that's why I'm here. And what I'm going to be working towards, I hope, is suggesting a general case, a general description that will help us to understand what are the underpinnings of cooperation in a way that then will be useful to us in lots of realms. Talk outline, it's useful to have a roadmap. I'm going to talk briefly about why we should study cooperation. Those who don't believe me perhaps can talk with Yokai because he is indeed doing some of this as well. But in general and in the digital world, why is this a problem? Why is cooperation a problem? We do it all the time. Why should this be a problem conceptually? Well it is, and I hope you'll understand why. I'm making a better description which will get us out of some of that, making the limit the general case as the approach and talking about mechanisms and institutions. What does this look like? What does this look like? Suggesting a toolkit, the tools in the toolkit for constructing institutions and cooperative mechanisms. Talking a bit about where these are located, that is an interesting subtopic which is particularly relevant here in this kind of context. Then looking at, I was, I have two applications and then I added a couple more once I got to that slide so we'll actually probably have four applications as a result. But the Koch Machine, the digital company and a couple of other applications as well. Well, why should we study cooperation? Why should we study cooperation? Well generally it's a key element of our existence. We are able to sit in this room and talk into microphones and work on machines and do all the things we do because of cooperation. Cooperation is a key element of our existence. It makes possible the gains of trade, that holy grail of economic output and development. It makes possible specialization. It makes possible scale. All of these things which we do every day are dependent upon us as humans being cooperative. It's also a key element in biological thinking. It's not just sort of economics and human behavior but also in biology has also tackled this as a key element in the growth of life, in the development of life from single cell, simple organisms to multi-cellular organisms to cooperative organisms to social organisms. These are all in a sense subsets and subcases of the problem of cooperation. So it's widely practiced. It's ubiquitous in human life. It's widespread everywhere else. Why should it be a problem? Well it's just poorly understood. It's one of those things that we do every day and that has not been terribly well modeled. Partly because we do it every day we take it for granted. We don't worry about modeling it. Partly is when we try to model it. We've hit some pitfalls which I hope I'll explain. Neoclassical economics has an interesting contradiction in this. They both assume that cooperation is easy, i.e. property and contractings like that as being just things they assume when they start out their day. And they have a model which denies that it's possible. Their rational actor model when parsimoniously and not very intelligently applied tends to suggest that cooperation won't work. An economist will spend half their day telling you it's easy and we just assume it and the other half their day is saying it can't possibly happen. So anyway, there's this inherent contradiction. It's one of the reasons why it deserves better study. The biology has had a similar kind of trap in that it's gotten located in around some of the logic of the selfish gene, that famous if now somewhat outdated book by Richard Dawkins that summed up this notion that somehow or other selfishness was the order of the day and we had to trip over things like altruism, cooperation, all of the things which then we see in fact occurring. So we needed a better understanding of this. We needed an understanding that cooperation is a requirement to be simultaneously solved along with output. This is one of the things I'm suggesting is a take home message. In economics, they assume cooperation and then work to solve output. In classical economics in any rate, not in more progressive economics. What we need to also understand is that that's true, but at the same time, the predicates of cooperation, the predicates of stable interaction in these ways must also be solved as a simultaneous problem involving a somewhat different logic than the logic of output, as we'll get to in a little later on. So the rationality which is the rationality of output decisions is not the same as the rationality of cooperation decisions. And part of the problem has been that things like Nashian analysis have assumed that there was an identity between those two. And I'm going to argue that that is not the case. Yes? Can you explain what output decisions are? Yes, output is basically efficiency style decisions. Things like, how do you get to greater efficiency in output? The allocation of goods and services through rational trade leads to an increase of the output of welfare as it sometimes called as an overarching term applied to that which we'd like to have happen. But the way that things happen, making things we want to have happen, happen requires at the same time a simultaneous solution of the structures within which that happening occurs. And it will help us create better institutions. When we understand better the fundamentals of institutions, we hopefully can create better institutions. And I argue it helps us to understand traits like fairness, honesty, and other values in human psychology. The kinds of things that classical economics has had a hard time and struggles with that actually that becomes relatively self-apparent why these things might exist if you take this kind of approach towards cooperation. Now, for digital worlds, why do we need it? For digital worlds, well, I'm accounting on all of you to help inform me. One of the reasons to give this talk early is honest. I'm hoping that I will spark some interest from some of you and that you will come and tell me, yes, you need to take this into account. And so that is part of what we'll do. But we think about transactional platforms as a place where this is the case. eBay, a massive exercise in cooperation that is in terms of classical economics nonsensical, but we see that it works at least to some degree. It's not perfect, obviously, but works to some degree. Well, I'm being rhetorically overstating. My classical economist is going to correct me and he's right. He's right, there we go, there we go. But anyway, they'll be puzzled. Why might that be? Because the interactions between strangers in a context where being taken advantage of is very easy, should at least in some relatively simple-minded versions of classical economics, there we go, I've reestablished my credentials. Be viewed as so defection-prone that it will collapse. Why is it that that transaction platform exists in the face of the potentiality that we're being taken advantage of all the time? So the eBay folks are very cognizant of this and have used a bunch of structures, which I would argue are fit within the model I'm putting forward. Exactly, neoclassic economics of a certain kind, there we go. I hope that I dropped the footnote that institutional economists think beyond this and are thinking beyond this. But at least the neoclassic, the straw man I'm setting up for neoclassic economics, there we go. They are very imperfectly understood. There we go, so I'm being corrected here. Good thing I brought somebody who can keep me honest, here we go. You can affiliation in here. There's so many affiliations. Work projects, work projects in the digital realm, wikis, software projects, that kind of thing. Again, massively cooperative in ways that sometimes work and sometimes don't. So again, I'm hoping you will give me better examples we go for. Now, why is this a problem? Why is this a problem? And here we come to the work of Nash, which I'm going to give you a cartoon version of, which is to really understand that many of the opportunities for cooperation come in defection prone packages. They become in defection prone packages. Packages where somebody is frequently in cooperation, not always, but frequently in cooperation, there is some mutual contribution that must be made, and that that mutual contribution can often, one of the simple ways for a defection to work out would be for that mutual cooperation to collapse back on itself, because someone just takes advantage of it. Let me give you a very simplified example. That's a lovely mug there. That's a lovely mug. Let's presume that's your mug, okay? And would you sell me that mug for $1? Thank you so much. Now, note how I've just defected. I waved a dollar in front of her and I walked away. Many, many transactions, not all, but many, many transactions are subject to this form of predation, this form of defection. I can just not do it. And what is her reaction now? Okay, she now knows that I'm not to be trusted at all. And any transaction going forward will become harder. The basis of trust that we worked here, we'll get to later on. But I use this to illustrate why it is that although we spend everyday transacting stuff, basically it's easy to predict. And why don't we do it all the time? The model often suggests we should be doing this all the time. Now, that was in the comments she was going to sell me. So she's predating on the comments to sell to me. All right, the propertizing out of the comments. So we see this. Another example might be what sometimes a hunting game described. Let's suppose that it's an old-fashioned kind of hunt up in Vermont where I need to push the game out of the woods and someone's got to wait and shoot the game, right? Well, I run around in the woods all day pushing game out and when I come out, the person who shot them is just gone. Again, defection prone. That piece of cooperation will not happen unless I have a reasonable expectation that I'm going to get my share of the game at the end of the day. And we can extrapolate this into all kinds of economic relations. Okay, well, we see then that there was this, the notion in Nash, the way Nash explained this was to say that each one of these games, as he called them, each one of these interactions had a dominant solution, had a solution which was the one that made sense for the players. And some kinds of game forms have dominant solutions where they collapse back. Predation is simply likely to happen because people, an actor, I shouldn't say people, actors acting in their rational interest will take advantage early on and will not do it. This has been the problem, this has been the stumbling block for lots of people. Now, not A, we should realize, not all possible cooperative games have a natural dominant solution for defection and predation. There are game designs that don't. The prisoner's dilemma game is one that does. This hunting game is one that does. But these others do not necessarily. There are other game forms you can construct where A, there is some constraint on the defection that makes it costlier or be some impossibility of defection that makes it unlikely so that we can create the surplus by cooperation and have the sharing reasonably occur in a way that we will all be willing to play the game. The rewards are decent on the table and the likelihood we'll get our reward is good, we go play. If the reverse is true, we don't. So what I'm suggesting is really quite a simple step which is to turn the limiting case into the general case and say that cooperation is likely to occur in circumstances where it is the dominant game strategy. Very simple-minded reversal of these terms in a sense. Do you mean the dominant game strategy? The dominant strategy is the best thing for me to do regardless of what you do. Right. Whereas a Nash is that I'm doing the best I can in anticipation of what I may do. True enough, true enough. So which do you mean? Which do I mean? I'm being clarified. I would mean in some ways probably the Nash. It would have to be Nash because you're playing Nash. There we go. Sorry, I'm less restrictive. Okay. I will amend my statement appropriately. So we're not... Here's the second step in this which is we're not stuck in bad games. A competent actor, whether human or other, is not stuck in a bad game that indeed we can choose and shape the games we want to play in. Part of being a highly competent social animal like humans is that we can assess games. We can advance and say, that's a prisoner's dilemma. I'm going to be very unhappy at the end of that day. I won't go there. It's like quicksand. It's social quicksand. You know, if we want to identify where that is, we're not going to walk there. We're going to take the path around it. So what we're going to be doing instead is searching out these... What I'm using is the better description. We're going to be searching out those cases where, whether Nash or dominant, we're going to expect the strategy when it all plays out to give us our reasonable share of the reward. And where we can't find those games, we're going to restructure the rules of the game to turn them into those games. We're going to restructure those. So we both choose and we shape. We restructure. And this brings me to a principle called mechanism design. And mechanism design is a field of economics. In fact, fortuitously, just yesterday, the Nobel Prize Committee announced its winners and the winners were for their work on mechanism design. So it is a field... No, I didn't. I was not pressing it. This was an edit today. I put a Nobel Prize in there. It won the Nobel Prize. Now, the mechanism design approach looks at creating game structures that make certain end dominant or Nash equilibrium likely, but likely along one of these kinds of lines. And what happens is, I'm suggesting, is that we have a slightly different notion of reality, sorry, of rationality when we are choosing and shaping our games than we necessarily do when we are in them playing short term. So we can step back and say, I could predate, but I can also create something where predation isn't going to happen. And if I'm expecting the other side is going to understand those as well, avoid those nasty Nash quicksand traps, we will be able to cooperate together in the designing of the mechanism of cooperation we're then going to undertake. So we can have negotiated mechanism design. We can have deals. We can have settlements. We can have institutions. I'm going to come to that in just a moment. Or we can even have, I argue, evolved mechanism design that essentially evolution of cooperative structures can be seen as an evolutionary means of mechanism design. You said different notion of rationality. We're not taking those short term sort of must defect opportunities. There is this sometimes expressed notion that rationality assumes that you will defect whenever it is in your short term interest to do so. And what I'm saying is that I face a game. I could do that with the game. I could try and play it with you and defect. But if I understand you're not going to play that game with me, then we come into this design of rationality. Okay, what is it we're going to design together to make a structure within which we can get to a better end and avoid these defection traps and therefore work with each other to capture all those nice gains of trade, specialization, et cetera. I like to call it a different rationality. Perhaps I'm being overstating the case. But I think that what the neoclassic sort of narrow minded, may I malign them down the road here at the law school and one economic center rationality is frequently this kind of how do I best defect rationality? And there's a different mode you can think in, okay, I could defect or we could work and construct. And so that may be at least, usefully described as a different kind of rationality. It may well be exactly, exactly, exactly. I was a deal lawyer. There we go. What can I say? I'm a transaction lawyer at Hart. And I want to construct these things. Although actually a good litigator should also be able to understand where, because the litigation should be about actually using a mechanism and perhaps even negotiating to a better mechanism to find a mode of cooperation. But not a bad analysis. All right. So what is the toolkit then that we see in this mechanism design process? And there are various lists of these tools. I have a list that's in the paper that I've got there of reciprocity, punishment slash strong reciprocity, hierarchy, dual key approaches, partnerships, contract, property, fairness, I could add in there, honesty, a bunch of word keeping, a bunch of things like that. Martin Nowak, who's here at Harvard and is one of the leading scholars on cooperation in the world at this point, has a lovely paper, I believe, in science last year in which he, from a biological standpoint, discusses five possibilities and also sets out his own description of something of a general case as well. He talks about kin, direct reciprocity, indirect reciprocity, network reciprocity and group selection. In the interest of time, I'm not going to go into these in much detail. I will go into one or two of them as I go forward. If any of you are interested in further discussion, I hope you will be because I look forward to talking with you further about them. So what does it look like? Another key point I want to put out for you is the notion of locating these institutions and mechanisms in various places in the world. They can be physical mechanisms. Let me mention the dual-key lockbox briefly. The dual-key lockbox is a small safe that two partners can use if they don't fully trust each other or even if they do. They put the receipts into a slot in the top during the day and at the end of the day to collect the receipts, each one of them has a key and they have to show up and both turn the key. There is actually an analogy to this with sexual cellular replication meiosis which guarantees that each of the two cellular parents get half the genetic content of the offspring. It's an analogy, not a direct physical cognate, but it's in that circumstance. Which then leads me to the genes. Genes actually carry some of these things. We see cooperative mechanisms in animals that are presumably not culturally based and again some kind of genetic instantiation of these solutions can be half. It can be in psychology. Values, the paper I have here argues that values can be understood as a psychological mechanism, a psychological institution if you like, which allows us to make productive game strategies with each other without having to rely on other means, just out of our commitment in our psychology and a commitment to each other that we can recognize. So you are willing to try and sell me that if you owned it because you thought I had sent, you trust where these signals, I'm your friend, there's a set of things. Actually, I didn't want to ruin my example. I didn't want to answer, I don't own it. There we go, okay, there we go. But we see then, and the paper goes into this, that in fact one of the roles that values can play in understanding economics is in this kind of institutional economics context where we can view values as psychologically instantiated institutions that we can use to form the basis of productive interaction. It can be in culture, in lots of places in culture, it can be in our customs and mores, how we just do things, how we get up, who does the dishes every day, could be a customary thing, that there could be a customary rotation. We can instantiate it in more, slightly more formalized cultural institutions, companies, clubs, associations. We can have the rules of the game. Baseball, baseball played in the back lot is a version of our understanding of an institution bounded by rules not always played without an umpire. An umpire is useful. An umpire helps with some of that punishment aspect. It also helps, we might think of it as culture when I didn't put the three little I in in culture three, law. Law is a clear case of mechanism design, at least from my standpoint. If we start to approach it as saying, it's one of the places we can put this kind of institution nominating the game so we come out ahead, it's a way of thinking about it. Finally, digital worlds, digital worlds. Digital worlds are also having this. Digital worlds can put it in code. Code is law, law is code, all of that. That's partly saying that code is institution. Code is a solution. Code is mechanism. Code is a bounding way to create these things. If the dual key lock box can work, so can a dual key computer program. So we can create solutions in the digital world that are culturally based, that are physically based, that are psychologically based, all of these layerings that we see. And indeed many of our solutions are not one dimensional. Partnership. A partnership. Psychology of commitment and values. The dual key lock box. Trust but verify. The law. The law which requires a kind of dual key approach to many decisions if it requires unanimity among partners in certain ways. Each of these is a version of the same principle that would... But it's located in a different place in the composite, which complicates the whole thing. Applications. A couple of applications. This is my last substantive slide. The Coke machine. The Coke machine. I think of the Coca-Cola machine, particularly in a tough place like a dormitory, a college dormitory Coca-Cola machine. You know, as a college kid, I was reasonably honest, but there were times that you really wanted a soda, right? And I didn't have a quarter or 50 cents or a dollar or whatever he's gotten to $1.50. And you know, the Coke machine is a transaction waiting to happen that is made reliable to the Coca-Cola company through physical armor. Physical armor. There's the big lock on it. There's the cage built around it. There's the, you know, you can't get into the thing and get the thing out. So there's physical armor making it reliable to the Coca-Cola company. On the buyer's side, what makes it reliable? The buyer's side, it's reputation, signaling the big Coca-Cola sign flashing away at you. It's signaling this is Coca-Cola's reputation, repeat transactions, all of those things which can be in that domain to stabilize. It's just a machine, but it's a machine that uses armor in one direction and reputation location in the other. eBay. eBay, I mentioned eBay works. Why? Because we've got this reputation thing. We also created its values. The whole, we're a village. I'm the mayor of eBay thing, which meant to create, draw in the psychology of values and community. It was a relatively sophisticated, if probably not fully thought out system of keeping it honest. My understanding is that eBay is redesigning that system as we speak and is trying to make it even stronger and better. They are, in a sense, an exercise and mechanism design in the digital world. Google and YouTube see today's talk later today. I was overhearing a bit of an interview, which may or may not be in today's later talk, suggesting that YouTube may go to a system where they design a mechanism that allows copyright owners to opt in to give permission while also opting in to give advertising. So they've created an upside together with the permission, which is an effect, a kind of mechanism design that makes it worth the while of the copyright owner to play the game. Again, I'm a guy with a hammer, so everything's a nail, but I view that as an exercise in mechanism design viewed in this kind of thing, where we're attempting to create a cooperative outcome, a dominant outcome, a Nash outcome where people are going to give permission to use this stuff on YouTube against the background of copyright law. Well, what are the things we have to put into the game to make that probable that they will actually play and cooperate? And finally, the digital company. The digital company. I'm working with the state of Vermont to work out some new legal frameworks which will make it possible to run companies, recognize companies entirely through digital means, not just using email as a good telegram, but using software in place of agreements and all of that kind of thing. Why is this important? Because one of the things that the digital world provides is these opportunities for massive cooperation on projects in ways that we didn't see or were hard to do in a non-digital world, the kind of wiki kind of approach to things. Well, what is the governance structure of that? And I'm grateful, particularly for Berkman and David Johnson, working with Berkman, for bringing this problem to my attention and saying, well, how do we do that? You have to have some other form of stabilizing. At the moment it's done mainly through kind of reputation and values approaches, a legal approach would be very helpful. And so we're trying to craft a law that would make either through the corporation or through the LLC, a management structure that will give the force of law to these mores, to these customs, to this cultural milieu of how these kinds of wiki-style projects are run currently. So that is the end of my talk. I hope now you come to see, first of all, that the cooperation problem is one we are making progress on. It's one that progress is coming from somewhat non-traditional sources in economics, in biology, in other ways of thinking about the world. That as we make that progress, we can begin to see into mechanisms, we can see around some of the obstacles that were thrown up by some of the more limited traditional economics and biology about how we make decisions. And also we can begin to have a way to study them and to suggest them and to design them as well. I'd be happy to talk further with you, as I know we will now, but also individually and would hope that you would be willing to come forward and help me as I work on how this applies, how this elaborates and how it applies in the digital world. I have many thanks. I support Grutter Institute, John Templeton Foundation, Vermont Law School, Berkman Center, UCLA, the Ann Gordon Getty Foundation. I want to give, although a lot of people helped me on this thinking, I want to give a particular help to, thanks to Carl Bergstrom's help, who is the one, he's a theoretical biologist in Washington who has given talks here at Berkman and is the one who directed me towards mechanism design. And the paper is available. I had hard copies in the middle, which seemed to have all gone, but you can visit us, or on our website, www.gettyfoundation.com, abstract 9-3-3-0-1-2. Thank you very much. Questions, comments, thoughts? Let's get someone who hasn't. Yeah, yeah, right in the back. Yeah, I know that you're, I guess, waiting into the digital world, but does your theory talk about why people would cooperate in a structure like Wikipedia, which has created its whole set of more roots and customs and rules, but doesn't have any obvious economic benefits any of us who does participate? to this is that classical monetary returns are only one of the coin that we like. There is the opportunity to show expertise. There are a whole series of different things that will motivate us in this structure. And so the payoffs, when I talk about payoffs, do not have to be monetary. They can be prestige. They can be psychic in various other kinds of ways. What is important is, I think, to understand those and build them in in a way. So my theory doesn't require that the payoff structure be monetary. Now what's interesting is that sometimes these two payoff structures can be somewhat incompatible. One of the reasons why I think it is often the case that we want to keep these large cooperative structures unmonitorized is that the monetization will unravel the moral basis of why it is we do. They've been stabilized through morality and values and non-monetary paybacks. And if you monetize them, you may destroy them. So while I'm looking forward to having a structure that can monetize them, I'm also very leery of the fact that if we've used values and prestige as the coin to stabilize them, we may undo them if we go to money. I should answer that question. Okay, yes, this man is probably better. I should mention that Andrew works fully in this field and in some ways is a more qualified speaker on it than I. Well, not in all ways, but in some ways. And so he can give good answers. He teaches in this field both at Tuck and is now doing research in it at Harvard University. That's a large question. I mean, the rest of the people in the room, and you know this already, there's a literature on why do people contribute to open source software. It's the same problem. You're giving your intellectual property away. Rights management in this context. So, on the one hand, it's a rights management institution in code to hold us to our bargains about how we will use copyrighted works. And that is alleged to facilitate transactions and cooperation that would otherwise not exist. On the other hand, DRM is hoisted upon us as, you know, because we are collectively unable to, we're unable to act collectively to refuse it when it would be an overall cooperative interest to refuse that argument and demand open source, open access. But since we've offered it on a one by one basis, each of us individually would accept that conflict. And part of what, there's a concept called individual rationality, which is in this, in this mechanism design thing. And the individual rationality would tie into this, which says that we don't accept the best deal we could design, but we accept that part of the deal that we can get to provided that it's better than the deal that would otherwise be available. If we see what I'm saying in that. A participation in one of these things. I drew a bright line between quicksand patches and the nirvanas of cooperation. And of course, it's continuum. And that there are places that, you know, not great to cooperate, but better than what would happen otherwise. And there are places where it's not great to, where it's great to cooperate, would be better to cooperate. And so what we're making is these trade-offs. One of the ways to look at politics, I suggest, and particularly out there on a limb view, is that it is a system of negotiation of the rules of these things at a societal level. That you can use politics to renegotiate the distribution rules and renegotiate these kinds of allocation aspects, like what you're talking about. And politics is a way of accumulating that individual voice into a bigger voice. But I think your point, again, as the man with the hammer and liking nails, I think your point fits perfectly into my analysis, so I like it. Thank you. Two words which you may have used or may not, and I may have missed it, but I'm just curious to hear your thoughts on them. One is generosity, and the other is relationship, and how those play into what you're describing here. That would be in my realm of psychology, psychology of values. Generosity, one of the things, let me step back a moment. One of the things that happens is that when you start thinking like this, you start thinking mechanistically about some of the most noble human attributes, which is a dangerous thing to do, dangerous on two levels. One is that you come to be dismissive of them if you're not thinking clearly. And secondly, other people come to accuse you of not having them, because clearly if you're thinking about them. Let me go back to love for a minute. Love probably plays a highly evolved biological role, but it's also splendid. And if you don't give yourself over to love, you miss out on all kinds of stuff that's part of the best stuff of being human. I would explain generosity in terms of, it's a piece of long-term reciprocity. If you're within a certainly in-group generosity is easily understandable in this kind of way, as being a long-term, both stabilizing from a fairness standpoint, you have to have certain redistribution. Also establishing leadership roles can be done through generosity. People look to people who were generous as leaders. It's not always disinterested in that sense. There is this instinct of disinterested generosity that's just splendid. And I'm not going to dispute it. In fact, what I'm trying to do is say it actually has a place in this economic world. And we should just embrace it for that reason and not go into the sort of neoclassical pathway and say couldn't exist, couldn't exist. Must just be a delusion. I've had bad neoclassical economists tell me that generosity was just a delusion and I should get over it. And what I'm trying to say is no, it's who we are, it's what we are. It has a place in this of a functional kind. But the minute you think you can dispense with the human feelings of it, you've already lost its benefit. What I'm wondering is whether or not, if we were to diagram what you're talking about, these two domains, the splendid and your theory, whatever characterizes that, is this event diagram that touches but does not cross? What is a place in the other? I'm trying to suggest that I think there's a place for the whole thing in this theory. Now again, I'm the man with the hammer so I like it. But essentially my starting point is that generosity is either a gift from some divine almighty, in which case then we just say thank you, or it is the product of a set of material processes that we can understand. And if it's the latter, it's the product of a set of material processes that we can understand, then we have to come to grips with it in that kind of context. I act on the former and believe the latter. I have a question that follows up about the question actually. There are studies of cooperation in the, let's call it offline world, opposed to the online world. We show that the key norms of the capacity and crucial parts are more likely to arrive where there are very close-knit networks. So the more close-knit the network is, the more likely you are to cooperate because you are more likely to, you know, have these functions that you basically expect. So I was wondering how do you transfer these mechanisms into the digital world because on the digital world you are very unlikely to know at least face-to-face who are the other people in the network, for example on eBay. You just have a combination of them and you don't actually know them. So your perception is that you are not likely to basically get sanctioned which you expect. So I was wondering if generality is maybe one mechanism that might make up the problem or how do you explain such a digital world. What a good question because when you're designing digital institutions like an eBay, I wish I were more expert on eBay. It just makes me nervous. I'm not a very good eBay-er because it just makes me nervous. What you're doing is you're combining a number of points. First of all, there's a reputation system that may be possible to spoof and do things with but there's a reputation system we can use. There is also, there's the repeat transaction aspect. The transaction histories are posted if I recall correctly. And so you can get some comfort from the fact that if someone's done 100 eBay transactions, eBay is important to them, they probably aren't going to mess you around for the hundred and first. Okay? There's the repeat game interaction. You note that on the NOAC list there was network reputation, if I recall correctly, which is again exactly some of this notion of stabilizing through that mechanism. There's the code. If you're eBay, you want PayPal, right? Because PayPal is a little bit like the lock box at least at that level of the transaction. There's many steps in the eBay-style transaction, including, you know, is the money actually going to come from one spot to another? Well, we put in the intermediary of PayPal and it comes clear. Another thing you can put on the list, in fact, is a trusted intermediary, is a way that things can be made to work. I don't think eBay allows you to send the money to a third party and the good to a third party and then have the exchange done. Do they have that mechanism in place? Because that's a way of doing that. It's closing an escrow, as it were, in the legal world. For large deals. Yeah, there are. Okay, for large deals. Exactly. You would expect it to be there. Why? Because it's one of the ways we can do it. If she doesn't trust me anymore when I buy her cup, maybe it's your pen. Is that your pen anyway? Does that belong here? It says Institute for Information Law. There we go. It's not her pen, but it's yours now. Again, if I want to buy it for $1, she knows not to trust me. Colin, we both trust, right? So I give Colin my dollar, you give him the pen, and then he acts as the intermediary. Again, these are... You should keep both. Who is it who stands to lose big if they aren't trusted intermediaries? Lawyers. Lawyers stand to lose big. That's part of what they do. They have to close an escrow with a lawyer. They're reputational. In fact, their reputation is how they make their living, which is odd given the low reputation lawyers as a class have. But their individual reputation... Also, commitment. The three years of expensive law school that many people think is unnecessary from an educational standpoint, from the standpoint of spending a ton of money to stabilize your commitment is a very effective signal of what you have to lose. Why do banks, classical banks, have marble in their architecture? Because it is a physical instantiation of the commitment they have made to the amount they would lose if they stole your money. And so a legal education can be seen as the equivalent of the marble in a bank. From that standpoint of setting these up. So how do we put that into the... How do you do that? Well, again, an eBay as a big exchange is interesting because they at least have a lot at stake in keeping the rules honest. And you're willing to go to eBay as opposed to, you know, Joe's bargain online exchange company because you think you're going to get a better deal. Yeah. Go ahead, Gene. I was wondering if I could go back to Wendy's question and then apply it to her and just ask if you were designing... If you're doing the mechanism design as between, for example, record companies and all customers in the world. And, you know, maybe DRM is the answer, but I'm curious if there was a solution that you could imagine that plays on everything that you just described. The record company was a mechanism. It was a mechanism that allowed a great deal of extraction of the benefit of the exchange by some relatively unscrupulous individuals. It was a mechanism that allowed artists and their audience to interact in a certain technological world. Part of the problem the record companies face, I think, is the technology is essentially destroying their role completely. Their role was as a kind of a middle broker in a circumstance that's no longer necessary. So they're desperately hanging on to an intellectual property position to try and maintain what was, once upon a time, a sensible economic position, a sensible mechanism. So would your model predict, for example, that Radiohead will succeed with their experiment or that any band will succeed or only a band with the kind of community that Radiohead has? Good question, good question. And again, when we're laying into the psychology, again, some of this is just an orderly way of thinking common sense, but the things you were raising are exactly the right thing. What psychological mechanisms of respect of community, of all of these mechanisms that are in our heads that are part of our value system, can a band like Radiohead use? Can they use it directly? I mean, what's interesting is the record companies, at some degree, set this up for themselves, right? Because they start to enlist the bands to say, come and tell your fan base why they still need to make us rich. You know? Oh, so there are all these ads about, you know, you shouldn't copy. Well, hey, if I'm the source of respect, I can be the direct source of collecting. And if the distribution channel now is available, so I can be both the source of respect, if we're using respect instead of property to stabilize the interaction, then a very different dynamic and a very different set of players come to the fore. David and then Sam. So, excuse me, I understand that, I'm correct if I'm wrong, I understand that this is a correction to an economic view of cooperation that you find. Well, I set up a little bit of a strawman for which I apologize, but only a bit of a strawman. I know nothing about economics, but you totally got away with it. Well, I'm not going to say that. I'm not going to defend myself. I'm actually asking someone a different question. You put this, as I understand it, as you're posing a correction to a traditional, maybe, strawperson economics. But this is obviously, obviously you have to acknowledge a very broad topic of cooperation that can touch on many, many fields, some of which you've talked about. So my question is, I'm not sure what domain you're proposing this within. Is this an economic model? And if it's not, if it's something else, if it's across multiple domains, then what is it competing with? What do we have to stop believing in order to accept this? Two or three things, two or three things. One, economists will tell you they don't study money, they study behavior. And so in a sense, it's a behavioral model. The behavioral traditions that were most rigorously formalized tended to be economics and biology, in my experience. You know, you've got traditions, behavioral traditions coming out of sociology, anthropology, psychology, a bunch of things like that. They tended to be, with some exceptions, less seriously formalized. The serious formalizations that most, up until recently, went forward in economics and biology both didn't get how any of this worked in a sense. Now, both of them have gone forward better than that. I mean, the institutional economists have done a lot of this. The game theoretic economists, although game theorists got kind of stuck in some of this too. And the biologists, the classical behavioral biologists, talked about the puzzle of cooperation. And people like Martin Nowak here are undoing a piece of that puzzle in biology. Some very good work. Again, I'm building on the shoulder. I'm standing on the shoulders of a lot of folks when I'm doing this. And my one, perhaps original contribution, really, is to turn it around, rephrase it in the mechanism design, and then turn it around to say, we actually are looking at, if we only knew to look for it, we'd find a general case here. And then if we worked from a general case back into specifics, we can see that instead of listing the five ways it can happen, we can list, here's an overarching principle for us to understand, and then the sub-cases will make sense as sub-cases rather than as things. Yes? Is there some amount of order if that's okay? I mean, you can... Okay. I have a question that sort of takes an abstract look at some of the questions that we had about DRM and the piracy in the record company. I'm interested in sort of legal theory more than sort of a specific law. One of the things you talk about is using law as a vehicle for mechanism design, for creating these sort of cooperative games or cooperative scenarios. Now, one thing that I've studied a lot and I'm very interested in is American legal realism and sort of the... some of the ideas that were very prevalent in legal scholarship in the early part of the 20th century that law, in fact, is sort of a... it's a mechanism of coercion and that it invariably involves differential treatment of different actors in society, so you find actors bargaining in the shadow of the law. So, you know, the law gives to A and takes away from B, and there's always going to be a differential relationship between the two, and that's sort of one of the things that policymakers and people that are creating the structure for markets have to take into consideration and to recognize that there is no free market and some sort of very early ground that was broken in defeating some neoclassical economic arguments. So, I guess my question is, does your theory... how does your theory view law in the sense of is law something that can bring about good outcomes for everyone? Can it lift all votes? Or is it something that is always involved with choice, involves a difference, and a sort of, you know, that sort of those hard political, you know, are there hard political questions that have to be answered in order to develop a cooperative... In a mechanism... That may need to be a law professor, actually. In a mechanism structure, there are people who get better deals and people who get worse deals. The question of a mechanism that is put in the mechanism design literature is, is being part of that mechanism individually rational, i.e. am I going to be better off in that game than I would be out of the game? That is a different question of, would I redesign the game if I had the opportunity to make me better, better off? Do you see what I'm saying? So, yes, there's allocations that can occur, but there are, and this is part of the fairness part of the paper that's there that I really didn't go into today, is what does the value of fairness do in this kind of mechanism stabilization? And I equate it with this notion of individual rationality and also the notion of negotiation and politics, again, as renegotiating the distribution. Suppose, let me go to my hunt game again, I'm from Vermont, we like that kind of stuff, right? I'm the guy with the loud shouting and getting the game out so that the person with the gun can call them and take them home. Am I better off, you know, if all day hunting, all by myself I could get, you know, trying to chase things down, I could get one. And if an all day hunting with this paredup thing, there's a pair that can get ten, if the person collecting them at the far end gives me two, I'm better off. Now, I may not be better off enough to make, to like it so much, I may argue about it, but if there's some reason I can't affect the game, then I take my two. Is that sort of tie into what you're saying about it? And I may not like it. I may then sit around and try and figure out how I can get the leverage to re-denominate the game so that I can get three, four, or five. And as I said, I think that's politics to some degree. Politics is something about how we... Because in some ways it's a question of who gets to decide. Like, who gets to restructure the game, right? Yes, yes, yes, yes. Exactly, exactly. And what they're interested in, because it's not... And so if you unionize that you can do a ton better, exactly your point about, you know, dealing individually, strength versus individual. It can be individually rational at a relatively low level of reward unless you can organize and re-denominate the game in some way. Seek us a secondary mechanism of cooperation which we bring to the primary one to redirect how that is going to work out. Amalgamated, hunter's helpers of the world unite, you know, kind of thing. I'm not trying to figure out what's new. What's new? Sorry. Because to some extent you're sounding like early Douglas Nord. There are problems in exchange and institutions arrive and they fix the problem. Later on he said, well, maybe they don't. And it seems to me that you... I think you're saying three things. I'm trying to figure out which of those three things is new. Here are the three things I think that are... One is you're saying there are multi-period... basically the way an economist used multi-period games. The rules get set, then the players play, and then outcomes occur. But the notion of mechanism design is just the rules get set and that's been looked at in cartel theory. And it's not... maybe you're broadening it, but it's not really... The technology piece seems really new. This notion that... because indeed there's this battle in the literature, you know, between the people who say, well, organizations occur because of some kind of technological reason and no, because they solve the transaction problem that would otherwise occur and it's got nothing to do with technology. You're saying, well, actually, technology has a constraint on where you can put the restrictions. Can you put it actually in the device in the field? A constraint and an opportunity. And that seems really new and really important. I'm grateful. Three. The third one is that... and you started out this way. I thought you were going to talk about the ultimatum game. And then you didn't talk about the ultimatum game. But then when somebody started talking about Wikipedia giving away music for free, you started sounding like you were talking about the ultimatum game. And to some extent, what that sounds like you're saying is these non-rational cultural restrictions can work in a really surprising place even when people don't have face-to-face exchanges and even when, you know, there isn't a long history that's about that culture. That also seems really new, which you'd have to prove it to me. Okay. On the first... You guys know the ultimatum game? No. The version, and correct me if it's the only thing I've... Okay, I'll give you a dollar. I'll give you a dollar. I'll give you a dollar. Your job is to then offer some of the dollar to him and he can either say, I'll take it or neither of us gets any of it. And even the best possible world, it wouldn't even be identified as me. You're giving it to someone. Okay, so this is, again, one of these elustered games. These games, economists love these games. I enjoy them as well. They're clearly little abstractions. They're not what we do, but they're, nonetheless, a lot of what we do in the world plays into these. So, you can have this dollar, you make an offer to him of how much money you get it and you have to share. If I reject it, you get nothing. Now, the rationality, the short-term rationality... What? Just think about it. What's the short-term rationality in the ultimatum game? Denominated in pennies. If you get a penny, you're in a penny. That's right. She should offer... She should take 99 cents and offer me a penny. But... But, otherwise, I can offer you 50 cents. You could offer me 50 cents in which case, but the question is, the economists would say it should be 99 of pennies. The interesting thing is, and you probably know this result, is that it differs in the developed world than in the undeveloped world. And it differs in cultures that are more integrated and cultures that are less integrated. How do you... No, how does it differ? In one case, it's more 50-50, and in another case, it's more 99-1. Which one? I've been telling you, what do you think? I would say in the underdeveloped world, it's more 50-50. No, it's actually the other way around. And his arm-length transaction thing that he started with is exactly that. If you think about it, how many times you've ordered shoes online, and they come, that's kind of amazing. Right? I mean, you just have this... But we can write you out, right? Pick them up. There's this possibility that the other person is going to behave responsibly and things are going to come back. And in my class, I run this year after year, and I'm telling you, the students most often give 50-50. And if it's not 50-50, a student in class will reject the split, which means they're throwing away money. And I do it with a hundred bucks. So they're often throwing away a hundred. Did they actually get the money? It's physical. He's throwing away the money. And they'll reject the split, and I'll keep my money. Which is astonishing. That's the cultural piece that might make... like giving away your music for free, that might make open source for free, but that shouldn't work there because there's no long term. That's really cool too. Yeah. But again, we can instantiate these things in different places in this continuum of technology, culture, expectation. All of these things build on each other, which is part of why it's fiendishly difficult to pace apart, but also very interesting. And good digital institutions will cooperate with the psychology of the culture in which they're embedded to make things happen that would otherwise. It's a skyhook. It's pure skyhook. You put it up there in the sky, and you can hang transactions off it. It's the most gorgeous thing possible. But the minute you stop believing in it, it all crashes and none of it works. I just wanted to pick up on something that Andrew just said about the second piece about that maybe one of the things you're adding here is looking at how the technological affordances actually give you the ability to create this kind of system, whether it be trust or otherwise. And going back to a comment you made about the instances where money kind of can sometimes undermine these other values. I'm curious if part of it is not so much the monetization of it as it is the rational counting of it, and that one of the maybe shortfalls of the reputation system that's an eBay that we can eventually overcome through further technological design is that that reputation system is numbers-based? And one of the things that I'm doing now is a kind of a fairly experienced eBay is I no longer just look at the numbers, but also look at who are you. I'm actually in some ways trust you more than if you're one of the big companies that just turn products all day. And so there's kind of these other richer forms of signaling that we have that goes beyond just simply numbers. And I'm wondering in part what the technology that we're striving for is something that's richer and deeper in terms of how to reflect how we do this in face to face. Good for you, good for you. And how to put some of this layered psychology more and more into that. There's a whole literature on, for instance, impersonal versus personal exchange. Kevin McCabe who's come up and talked a couple of times is one of the key writers in this field. And one of the interesting problems in designing big markets is do you want to strip away all of that communication and somehow mechanize, make firm really strict rules that will allow players to come in without the disadvantages. Because once we start using those personal cues again, are we back into who's my relative, who looks a lot like me, all of those cues which get in the way of a broader society and a broader participation. So there's a downside to reloading the personal transaction indicia into what is an impersonal world. But maybe that's how humans are set up. This is one of the arguments that Kevin McCabe and his colleagues have done a lot of interesting thinking about this. How do we mediate back and forth between these? And is the digital world really an impersonal world? Or is it a personal world? And which of those is going to be the modality or both or when? Now to come back to Andrew to you for just a moment if I may. I'm not sure that Douglas North, I would agree with you, I've taken a bunch of ideas from Douglas North, I'm not sure that he goes the next step to the mechanism design piece of it. I may be doing him disservice because I've hardly read it all, but I don't think he goes that next step. He sets the solution up, but doesn't have, I think, the pieces this way to describe it at any rate that I would hope to bring to it. Again, all of these, this is one of those ones where a lot of the pieces are sitting around and if I'm doing anything, I think it's just putting the puzzle together a little bit more thoroughly and saying, yes, this is corner A, this is corner B, and we can fill in the middle of it. I had a question much like Andrew's, but just to take it in a slightly different way for the question, which is I think around this table or this virtual table and those listening on webcast and so forth, you probably have most of us convinced in your core critique of classical economics anyway, and that we've all seen different traits expressed by people in the online environment that suggest other kinds of motivations, the work that Yokai and Terry are doing in a class, for instance, they're teaching in the spring. It seems to me that the rubber in a way hits the road in two ways. Andrew's second point is plainly where it seems most different and interesting, at least to me, which is what is it about the digitally mediated environment that causes human beings to act differently or the same way in just pulling it out in new ways? It seems like how you would have me convinced and where I think the great, fertile area for further work, at least from the sort of digitally driven employer approach would be two. One would be what are the strands that you can pull from the examples that have undergirded your critique in the first instance? You could pull forward to say this is what's different about a digitally mediated world or the way in which people are acting in a digitally mediated world, being clear to stay free from technological determinism an important point. David and I are teaching in the class in spring which is called the web difference and one of the things we're puzzling over is what's different and then what can we derive from that? So similarly answering that same question that says the web is different or we act in this way and the web is different and that's both why I think the critique works and then helps look forward, right? The second one, this is you and I have had conversations over two years about this where to me, I need more I know it's there but I don't I am in need of convincing is what is that institution that you want to build? What problem you want to solve and what's the institution? And I think of it as use cases as one example or I get the taking versions of the eBay story, the open source story, the Wikipedia story, each of which of course are different sideways and somehow the extrapolation of here's how this wonderful synthesis and this game theoretic improvement and these things that are different about the web turn into something that is a brighter future for all of us. It's that piece where I think you can knock the ball out of the park to mix the metaphors hopefully. I take your challenges as indeed that. I wish I could come and say ah, ah, you know here it is. Two things one is the approaches in my head and where I'm the first thing I would say is that if there are folks here around the table and in the ether who are willing to chat about this as we go forward, I would love opportunities to interact. I'd love the problems. This whole question about the distribution of cultural product is an interesting one. It might be a spot where some of that could happen. The one I'm working on is this business one. And what is inspired in me is not the notion that I could design what the perfect sort of wiki company would look like but rather to try and figure out a constraint space within which we could use the techniques of contracting and combined with computer and software and that kind of thing to create something where people can then redesign something better. I won't design the best one right up front. I don't know enough I don't do it every day. But can I help create a legal framework in a state which is friendly to rather than inimical to it's like saying how would we design intellectual property law around respect and generosity rather than around property. It's answering Sam's question to say law can be number one not number two. I'm sorry, it's answering Sam's question to say law can be number one not number two in his construction. Yes, yes, yes. Law can be the enabler in all this. That's right, that's right. Where we can create the space. That's been in some way it's the genius of the law. If you think of law as a mechanism design space it is both has set mechanism design. The rules of property are a relatively constrained set mechanism which solves a ton of problems in this game theoretic landscape. Contract is this astonishing thing where we can design our own mechanisms and the sanction of the law will come in if we've done the thing formally correctly. Can we use those in ways now that will be friendly to the opportunities which grow out of these two these differences so when people actually discover them they've got the laws now in a position to say, yeah right on this template we've given you a very very loose template you program it. So just one very specific example that you might consider taking up. At the end of the third case study that David Risco over here is working on an interoperability one of the things we're talking about is we think that we could improve the likelihood that greater interoperability in the Web 2.0 space could lead to more innovation and better things for society if indeed there were a standardized set of commitments on the part of those offering open contracts. So this seems to me exactly your contract problem which is we could or at least fairly early on in the process construct an environment for Web 2.0 as sort of an operating system on which people are building if we were to get the set of contracts right by now it's an ad hoc bunch of contracts which I fear could have houses of cards falling on one another but this seems like a very clear example where we could try to apply your differences and create structural stuff through contracts that actually would be meaningful and useful and test your theory. So anyway that's a... No no and remember back on my list another thing is that you can stabilize things is hierarchy so you can have also and it's not to be I mean of all stuff is nice and when it can work but there are times when you just need the government or the boss or whoever it is to come in and say you're all going to do it this way and the problem with that is that the technical potential solutions almost always are less than when they come a little more than the bottom up we end up with a bad television standard or a bad VHS standard or whatever it is because the standard setters often go for something that isn't the best the French, remember that French thing that was all about that what? Minitel? Or the thing where you went to the post office and it was like the internet only it wasn't yet Minitel, Minitel, Minitel There we go, there we go The French Minitel In a sense, in a sense the gorgeous aspect Three centralized French things related to what you were just saying It's crazy So the centralization was a bad thing there and a much more let the flowers bloom approach was better It was good And there are real problems with the totally let all the flowers bloom which is 5,000 people build applications on facebook but with no decent contracts you undergird what they're doing we're pouring all this effort and personal data and we've got thousands of flowers blooming at google and Yahoo! and other places there may be lots of innovation but it may not be in the long term a particularly good solution absent some form of even bottom up driven contractual standardization I hear you but what I think becomes useful and if there's again maybe a little bit of a payoff for thinking about these things this way is that we come to understand that the hierarchy and the contract and all these things are in the same toolbox Often we treat them as though they were oh hierarchy never Well you know you shouldn't think about that way you should say I've got if I want to build a stadium I use concrete sometimes and wood sometimes and I'm going to hang things and I'm going to put things up and you know if I get used to the idea that all of these are part of what I can use they have their pluses and minuses then I'm less dogmatic about I've got to be all free or I've got to be all ordered or whatever it is Any last words Do you want to Just an answer In the top down countries there's a new institution called authority of regulation of technical protection measure which is going to regulate conflicts between technical protection measure and what we have instead of fair use list of exceptions and also a conflict of interoperability so the head of this authority which you want to be connected they just started the work early September and last September he wrote to me that unfortunately he didn't have internet connection at the office That would be great if we connect David Nothing like something actually to come out of a lunch discussion and lots of thought provoking ideas I'm totally looking forward to see how this evolves over the course of the year Thank you so much