Yochai Benkler is giving a talk about his new and wonderful book, The Penguin and the Leviathan. (I interviewed him about it here.)

Yochai begins by pointing to Occupy Wall Street as teaching us much about cooperation and collaboration.

On Oct. 23, 2008, Alan Greenspan acknowledge to Rep. Henry Waxman that his model of the world was wrong. “I made a mistake in presuming that the self interest of organizations…was such that they were best capable of protecting their own shareholders.” We live in a world built around a mistaken model of human motivation, Yochai says. The basic error is not that we are sometimes self-interested, for we are. The mistake is thinking we could build our systems assuming that we are more or less uniformly self-interested. We’ve built systems that try to get incentives right, or that try to get punishment right. But now scientific selfishness has retreated, and we should model our systems on this new knowledge.

In 1968 Gary Becker said that we could model crime by thinking it of a pay-off model: the benefits of the crime vs. the cost of the penalty. So, we get Three Strikes laws. In another domain, the Jenson and Murphy paper on incentive pay for top management assumes that every level of the enterprise will try to shirk and put more in their pockets, so (the theory goes) you should increase the stock options at the top. But that hasn’t worked very well for companies in terms of return to stockholders; you get misalignment from this model. This model is like Becker’s: it’s about getting the incentives and penalties right. Yochai tells of a mother trying to get her three year old into a car by threatening to take five cents off the child’s allowance. “This model penetrates everywhere,” he says.

This intellectual arc is everywhere. Evolutionary biology has moved from group selection to selfish gene through kin altruism and direct reciprocity. Economics also: strong assumptions of self-interest. Political theory, from Downs, to Olson, to Hardin: all assume the inability to come together on a shared set of goals. Management science and organizational sociology: From Taylor to Weber to Schumpeter through Williamson. Although there are counter narratives in each of these fields, selfishness is the dominant model.

And yet on line we see how easily we cooperate. “Things that shouldn’t have worked, have worked.” He draws a 2×2: market based and non-market based vs. decentralized and centralized. In each, there have been huge successes of social production. This is in fact a new solution space.

In each of the aforementioned disciplines, there is now a development of more complex models that take account of cooperation. E.g., evolution: indirect reciprocity; cooperation emerges much more easily in the new models. Economics: shift to experimental and modeling away from self-interest, and the development of neuroeconomics. Political: Eleanor Ostrom on the commons. Management science: Work on team production and networks; high commitment, high-performance organizations.

The core insight of all of these fields is that the model of uniform self-interest is inadequate. Then there’s debate.

Yochai compares Dawkins in The Selfish Gene (1976) and Martin Nowak (2006). Dawkins says we are born selfish. Nowak says: “Perhaps the most remarkable aspect of evolution is its ability to generate cooperation in a competitive world.” It’s an old debate, Yochai says, citing Kropotkin vs. Spencer vs. Boaz vs. Margaret Mead. The debate is now swinging toward Kropotkin, e.g., neural research that shows empathy via brain scans: a partner’s brain lights up in the same way when s/he sees the other person undergoing pain. He points to the effect of oxytocin on trust, and for the first time in Berkman history makes a reference to monogamous voles.

Why does this matter, Yochai asks. He refers to an experiment by Lee Ross et al. Take a standard Prisoner’s dilemma. All predictions say that everyone should defect. Take the same game and give it to American students, Israeli fighter pilots, etc., and told them either “You’re going to play the Community Game” or “The Wall Street Game.” The former 70% opened cooperatively and kept cooperating through the 7 rounds. The latter opened at 30% cooperative. The 30% in the Community Game represent a significant segment that has to be dealt with in a cooperative system. But there’s a big middle that will one or the other depending on what they understand their context to be. So, concludes Yochai, it’s important to design systems that lets the middle understand the system as cooperative.

So, we move from tough on crime to community policing. That changes all sorts of systems, including technical, organizational, institutional, and social. Community policing has been widely adopted because it’s generally successful. We see that we have success with actual practices that depend not on reward and punishment and monitoring, but on coperation. We’re finding out about this online, but it’s not happening just online.

Yochai says that he’s just at the beginning of an investigation about this. There’s a limit to how much we can get out of evolution, he says. It’s hard to design systems on the basis of evolution. Instead, we see a lot of work across many different systems.

But we still want to know: Won’t money help? The answer is what’s called “crowding out.” We care about material interests, but we also care about fairness. We have emotional needs. We have social motivations. What if these interests don’t align? The Titmuss-Arrow debate 1970/1 about the motivations for donating blood. A 2008 study (Mellstrom and Johannsesson) paid people money to give blood. When you allow them to give the money away, it increased the number of people who gave blood. Adding money can suppress an activity more than it increases it. That’s crowding out. It’s not uniform in the population. Designing systems is much harder than coming up with a material reward that appeals to people’s self-interest. We do not have full answers here

Think of cooperative human systems in three vectors. 1. Conceptual: from rationality as univeral self-interest to diversity of motivations. 2. Design: Cooperative human systems designed on behaviorally realistic, evidence-based design. Politics: We cannot separate out incentives from fairness, ethics, empathy, solidarity.

Yochai points to a number of factors, but focuses on fairness: of outcomes, of intentions, and of processes.

Outcomes: What counts as fair is different in different cultures, especially when you move outside of market economies. In market societies, 50:50 is the norm for fairness. Once it gets to 30:70, people will walk away. But you can change that if you change the framing, e.g., “You got lucky.” But there is no single theory of justice. Yochai looks at a study of the cement trucking industry. It turns out that there are large pay disparities. They also differ in what they say they pay for: performance, or equally time. They don’t always do what they say, though. But when you look at real performance measures, you have fewer accident and out of service events if the company is accurate in what it says, no matter what it says.

We don’t have an agreed upon theory of justice, he says. This explains the 99% vs. 53% debate around the Occupy Wall Street. This is a debate over basic moral commitments without which a system cannot function. There is no way to resolve it either through neutral principles or by efficiency arguments.

Intentions also matter to fairness. When you Where bad intentions excluded (e.g., it was just a roll of the dice), then there’s much less negative reciprocity.

Processes: Tyler (2003) showed that procedural justice correlated with internalized compliance. Yochai points to the militarization of the police as they deal with the OWS. The image projected to the crowd is one of lack of regard for process. He compares this to a massive demonstration of Israel in which the police stood a good distance away, and a different relationship was fostered.

We can see a revival of the “sharing nicely” idea we teach our children. In science. In business. Science is beginning to push back against the assumption of selfishness. It turns out that we aren’t universally self-interested. Different people respond differently, and each person responds differently in different contexts.

We need a new field of cooperative human systems design that accounts for the diversity of motivation, and that takes seriously the issue of “crowding out”: adding incentives can result in worse outcomes.

And, Yochai concludes, we need a renewed view of our shared humanity.

Q: Fascinating. But: The passage from evolution to the social sciences has long been discredited. Also, it’s too simple to say that the solution to the banking problem is that we need more cooperation. The banks are supported by a set of interests bigger than that.
A: You say sociobiology has been discredited. That’s true of the early to mid 1980s but is no longer a good description. The social sciences and anthro have been moving to evolutionary models. Economics too. What was in the 1980s was resolved, now, especially in the social sciences, is unresolved. Second, sure, bankers self-select and control the system. The real answer is that it’s a lot of work. When you have a system optimized for money, and money is the social signal, it self-selects for people driven by that. We need long-term interventions to increase cooperation. E.g., the person who can work with Open Source at, say, IBM, is different than the person who can work her/his way up a hierarchy; the company therefore has to train itself to value those who cooperate.

Q: I just went through MIT’s tutorial that instructed me how my ideas would be licensed. I said that maybe there should information in your office about how to contribute more openly. How do systematize open, collaborative forms across the entire educational system?
A: Lots of people in this room are working on this problem in different ways. We fight, we argue, we persuade. Look at university open access publication. We use our power within the hierarchy of universities to raise a flag and to say we can do it a new way. That allows the next person to use us as an example. After I released Wealth of Networks for free on the Web, I got emails from all sorts of people wanting to know how to negotiate that deal for themselves. Universities should be easy.

Q: What are the burning policy implications of this shift in the way we rule the world? What would you change first?
A: I should note that I don’t address that in the book. We need an assessment of community policing and the big board [?] approach. The basic question is whether we continue to build a society based on maximizing total group, or one that trades off some growth for a more equitable distribution of outcomes. The point is much broader than open access, patent, copyright, etc. The deregulatory governance model is based on an erroneous model of interests. But all of my work is done on the micro level, not the level of organizations. But we know that the idea that musicians need the payoffs afforded by infinite copyright is false; we have empirically data about that. So there are places where the relation between the micro interests and institutional interventions is tight. But I don’t talk about that much in the book.

Q: I’ve looked at pay inequality in Japan and the US. The last thing that matters to the level of compliance with regulations is the gap between CEO and workers. The deterrents are very effective in the US, explaining [couldn’t hear it]. Compliance is much better in the US because the penalties are effective deterrents.
A: First, once you’re talking about the behavior of an organization, we don’t have the same kind of data on what happens within a corporate decision. When people see themselves as agents, there can be conflicts between the individual and the organization. For that you need external enforcement.
Q: Jail time makes a huge difference.
A: Then how do you explain the findings that amount of tax options predicts probability of tax fraud. Same baseline enforcement, but whether you had stock options predicts tax fraud. Adding money and punishment certainly has an effect on behavior. But it depends on whether that intervention has better effects than other interventions. But we only have a little bit of data.

Q: If a high school principal came to you who serves many interests and types of people, how could your ideas influence her or him?
A: My mother founded two schools and a volunteer organization. The lessons are relatively straightforward: Higher degrees of authority and trust, structure with clearly set goals, teamwork, less hierarchical distance between students and teachers, less high-stress testing.

2 Comments »

How proteins fold over themselves has a lot to do with how they work. Envisioning such folds is a hugely complex problem for computers that human brains with eyeballs attached happen sometimes to be able to do better. The FoldIt game supplies humans with protein models and asks them to fold ’em.

According to a post by Alan Boyle at MSNBC.com: “Video-game players have solved a molecular puzzle that stumped scientists for years, and those scientists say the accomplishment could point the way to crowdsourced cures for AIDS and other diseases.” The post is about an article in Nature Structural & Molecular Biology by Firas Khatib et al.

Way to go, human brains!

(I talk about FoldIt in Too Big to Know, which has now gone to press. Ohhh, irrevocably ink-stained paper!)

Be the first to comment »

Carl Zimmer at The Loom points to Rosie Redfield’s blogging of her lab work investigating a claim of arsenic-based life forms. It’s a good example of networked science : science that is based on the network model, rather than on a publishing model.

I find open notebook science overall to be fascinating and promising.

Be the first to comment »

Randy Scheckman, the new editor of the Proceedings of the National Academy of Sciences — online and free — explains how the a new journal will work: scientists will edit for scientists, there will be rapid turnaround, and the journal’s acceptance rate for submissions will go way up. He positions it as more scientist-friendly than Public Library of Science.

The fact that this interview was (admirably) published in Science magazine has some significance as well.

(By the way, the authors of a report on obstacles to open access have left a hefty and useful comment on my post.)

In my continued pursuit of never getting anything entirely right, here’s a comment from Michael Jensen: “Not quite right — the new editor of what I think is a still-unnamed OA biomedical journal was announced, but Randy Schekman currently edits the PNAS, as I read it.” I have edited the above to get it righter. Thanks, Michael!

Be the first to comment »

Fascinating essay by Kristi Dykema Cheramie at the Design Observer Group blog about a 200-acre hydraulic model of the Mississippi River begun in 1943.

A snippet:

3,000 German and Italian POWs began construction on a 200-acre working hydraulic model. The ambitious model would replicate the Mississippi River and its major tributaries — the Tennessee, Arkansas and Missouri Rivers — encompassing 41 percent of the land area of the United States and 15,000 miles of river. [11] It would reflect existing topography and river courses throughout the Mississippi Basin, using the best data drawn from hydrographic and topographic maps, aerial photographs and valley cross-sections.

The prisoners cleared the site of a million cubic yards of dirt and rough-graded the land to match the contours of the Mississippi River Basin. To ensure that topographic shifts would be apparent, the model was built using an exaggerated vertical scale of 1:100 and a much larger horizontal scale of 1:2000. While the existing topography offered a close approximation of the actual Mississippi Basin, some areas required significant earthmoving; the Appalachian Mountains were raised 20 feet above the Gulf of Mexico, the Rockies 50 feet. An existing stream running east-to-west provided the model’s water supply. The streambed was molded to take on the shape and form of the upper reaches of the Mississippi, and a complex system of pipes and pumps distributed water throughout the model; it was regulated by a large sump and control house sited near what would become Chicago, Illinois. To simulate flood events, Reybold needed to introduce large volumes of water over short periods of time, so he designed a collection basin and 500,000-gallon storage tower system at the model’s edge. Small outflow pipes at anticipated data collection points channeled excess water to 16 miles of storm drains

1 Comment »

The National Park Service site has a fascinating article about the discovery of a very early talking doll made by Thomas Alva Edison. This was apparently the first commercially-available phonograph recording ever.

The artifact is a ring-shaped cylinder phonograph record made of solid metal, preserved by the National Park Service at Thomas Edison National Historical Park. Phonograph inventor Thomas Edison made the record during the fall or winter of 1888 in West Orange, New Jersey. On the recording, an unidentified woman recites one verse of the nursery rhyme “Twinkle, twinkle, little star”. The voice captured on the 123-year-old record had been unheard since Edison’s lifetime. The recording represents a significant milestone in the early history of recorded sound technology.

To “play” the recording, scientists at the Lawrence Berkeley National Laboratory did a 3D scan of the grooves and reproduced the sound without having to touch the physical material.

Each phonograph was made live, rather than reproduced from a master, which adds just a little more of thrill to listening to it. You can hear the recording here.

[Tip of the hat to my brother Andy for the link. And see the excellent article in Science magazine.]

3 Comments »

Ever since I ended my paper-based relationship with the Boston Globe, I’ve done my breakfast reading in front of a monitor on our kitchen table. Today, I spent a little longer, and read some excellent articles:

• Stephen Metcalf writes about Robert Nozick’s legitimizing effect on Libertarianism, and the philosophical weakness of the case he made for it. I have not heard anyone else make Metcalf’s critique of the Wilt Chamberlain argument (so far as I recall).

• In Scientific American, John Horgan defends Stephen Jay Gould from the charge that his brilliant example of personal bias affecting scientific outcomes was itself based on Gould’s own biases. It’s actually a weak defense, with Horgan instead defending a related point SJG was making: “Maybe Gould was wrong that Morton misrepresented his data, but he was absolutely right that biological determinism was and continues to be a dangerous pseudoscientific ideology.” By coincidence, a couple of days ago I came across my old copy of an anthology titled “The Sociobiology Debate” that was compiled in 1978 when the idea that evolution shape ours our social behavior was not just controversial, but to many of us (including me) seemed quite threatening: it implied a lack of free will (which I no longer care about) and it was sometimes used to give a sheen of inevitability to the most conservative and even oppressive of social behaviors. Unfortunately, Horgan’s argument against sociobiology consists of the following single paragraph:

  “Biological determinism is a blight on science. It implies that the way things are is the way they must be. We have less choice in how we live our lives than we think we do. This position is wrong, both empirically and morally. If you doubt me on this point, read [Gould’s] Mismeasure [of Man], which, even discounting the chapter on Morton, abounds in evidence of how science can become an instrument of malignant ideologies.”

• Also in Slate, I disagree so sharply with Jack Shafer’s criticism of Jose Antonio Vargas that I think I must be missing something obvious. Jose is the former Washington Post and Huffington Post journalist (and Pulitzer-prize winner, by the way) who came out as an undocumented immigrant in an article in tomorrow’s NY Times Magazine. Shafer declares himself to be an “immigration dove”: “I believe in open borders and detest our current laws and their enforcement.” If you hate the law’s enforcement, how can you also get in a snit about someone who lies to evade that enforcement? Or perhaps it’s only journalists who shouldn’t lie to their employers about their immigration status because there needs to be a special bond of trust between the editor and the journalist. So, which jobs does Shafer think do not require trust? Or is this just journalism dealing with its self-esteem issues again? Jose didn’t lie about his credentials, and he didn’t lie in his stories. He lied about the thing the bad laws Shafer “detests” made him lie about, just as forty years ago he likely would have had to lie about his sexual preferences. If Shafer thinks Jose’s admission makes him unreliable, then go through his work and find where this lack of reliablity manifests itself. If it doesn’t, then salute Jose for his honesty and courage. (Disclosure: Although I haven’t talked with him in a year or two, I count Jose as a friend, beginning in his pre-Pulitzer WaPo days. He has struck me as an honest, open-minded, and impassioned inquirer. I like him a lot.)

3 Comments »

At the final panel of the conference. Judith Donath is moderating.

Charlie Nesson asks: “When we talk about our space, who are we?” In Jeff Huang‘s presentation, it seemed like he was given the perfect hypothetical — a desert — to build a public and private place. “In cyber terms, we are people of the Net. What then is our domain? It’s the public domain. And if you are to build the public domain, then I believe the wisdom to follow from a lawyer’s point of view is the same wisdom that has more or less informed the world of real property. If you want an orderly world of real property, you build a registry. If you want an orderly world of bits, you build a registry.” This is Charlie’s new project: a registry of the public domain. They’re starting with IMSLP.org: a musical score library. It has 93,000 musical scores in the public domain., exquisitely put together.

The Net divides into two domains, says Charlie, one that is free and one that is not. Free means free of copyright and other encumbrances. Charlie wants to build our domain on a foundation solid in law. The registry he’s building identifies works as public domain, with links to the registrars attesting to this. He wants it to be populated by librarians with public domain collections. But, the problem with registries is litigation risk, i.e., the threat of lawsuit. “So the essence of this idea is to couple the registrar with a pro bono commitment of legal service from a law firm of repute to defend litigation based on infringement.”

Where do you find the institutions that want to protect privacy, asks Charlie. How about libraries, he suggests?

“I’m tough on privacy, Judith,” says Charlie, in response to a question. “I’ve never liked it.” He explains it’s so often based on fear and looks backwards.

Martin Nowak looks at cooperation evolutionary term in which a donor pays a cost and a recipient gets a benefit. He explains game theory’s Prisoner’s Dilemma. Why do people cooperate? “Natural selection chooses defection,” rather than cooperation. In a mixed population, defection becomes increasingly more popular. So, natural selection needs help to favor co-operation. Martin categorizes the factors into five mechanisms: kin selection, direct reciprocity, indirect reciprocity spatial selection and group selection.

Direct reciprocity (I help you, you help me). If you play the Prisoner’s Dilemma several times, the economics changes, as The Folk Theorem shows them. Martin quickly summarizes Axelrod and Rapaport. [Too hard to live blog. Read Ethanz. Really. Now.] Errors turn out to ruin cooperation, so you need a process that allows for forgiveness. Martin’s doctoral dissertation showed that if everyone plays randomly, the right tactic is to always defect. A tit for tat strategy corrects that, and generous tit-for-tat (I may still cooperate even if you defect) provides a math model for the evolution of forgiveness and cooperation. There are always oscillations; cooperations are never stable. We need structures that rebuild cooperation quickly after it is destroyed because it always will be destroyed.

Direct reciprocity allows allow for the evolution of cooperation if there’s a prospect of another round. Indirect reciprocity (I help you, someone helps me) leads to cooperation if reputation matters. You need natural selection to care about reputation, so to speak. “What you need for indirect reciprocity is gossip” to spread reputation. For that you need language. “You could argue this is the selection process that led to language.” “For direct reciprocity you need a face. For indirect reciprocity you need a name.” (David Haig) Our brain has both capabilities. If interactions are completely anonymous you run into problems. Also, you need gossip to be relatively honest.

Spatial selection = neighbors help each other. Martin flips through some graphs that shows that it selects for coop if you have a few close friends. Likewise, evolutionary set theory says that people wanting to join particular groups can also lead to coop.

Judith: What about strong vs. weak ties?
Martin: We assume equal ties. There’s a trade-off between wealth and vulnerability.

Nicholas Negroponte asks himself a question every morning: Is he doing something that normal market forces would do anyway? If so, he stops. He wants to do that which market forces will not do.

There are now 3M One Laptop Per Child laptops in the hands of kids. This isn’t huge since OLPC would like to get laptops into the hands of about 500M kids. Before that, people assumed computers teach by imparting content. Instead, you want to see children teaching. 20-30% of the million Peruvian kids with OLPC machines are using them to teach their parents how to read.

Nicholas goes through some points he made in a talk at the UN recently. Among the points: Measurement is overrated. You only measure when the changes are so small that you can only see them by measurement.

Judith: When we see well-off kids sitting side by side looking into screens, we think it’s a nightmare of anti-sociality, but when we see your adorable photos of third world kids in the same position, it looks desirable?
Nicholas: I don’t see the well-off kids that way. And why don’t we make OLPC’s available in the US? Because the issues are deeper than that.

A: Talk about anonymity…?
Jeff Jarvis: It’s foundational to democracy. It’s getting a bad name because of trolls. But it must be protected.

Q: This discussion is soaked in privilege. There’s much inscribed in the language that affects how people act. When you idolize the public space as a place where all can share their ideas safely, it feels really far away for me.

Q: (Charlie) Nicholas, you’ve said that Uruguay has given all 500,000 of its kids OLPCs. Given your position on measurement, what change will we see?
Nicholas: Their curiosity, the way they approach problems, the way they look at things…I think you’re going to see a nation that is far more creative than many other nations. Nicholas tells a story of kid whose homework got 100K hits.
Martin: Who teaches them how to use it?
Nicholas: It’s genetic :) We’re going to do a scientific experiment in which we drop OLPC laptops out of helicopters onto remote villages and come back in a year and see how many have learned how to read.

Q: (urs gasser) One vision says build a great tool and see what happens. The other is to study human behavior scientifically. (Nicholas vs. Martin). How difficult is the translation from findings from science about human behavior to adapting them to technology?
Martin: I’m fascinated by mathematics, but we do apply it to practical issues. In the field of cooperation, we’d like to bring the models closer to human observations. For example, many cultures like punishment, but I think it doesn’t work well to create cooperation because it creates complications. Reward seems better. So, we study that. We do the same experiment in multiple cultures. In Romania, for example, people differentiated between public and private outcomes, because they lacked faith that public engagement had positive outcomes.

Q: (zeynep) The Net has let the cooperative side of human nature be more manifest. Does your work in evolutionary biology take account of this?
A: The coop we see in the animal world must rely on direct observation. Humans can communicate. We don’t have to rely on our personal experience with another to decide whether to coop. The Net can help us to evaluate others quickly.

2 Comments »

I’m thrilled that I’m going to be writing an article for Scientific American on big data models — models that cover some huge swath of life, such as the economy, the climate, sociopolitical change, etc. What’s the promise and what are the challenges? How far can such models scale?

So, who do you think I should interview? What projects strike you as particularly illuminating? Let me know in the comments, or at selfevident.com.

Thanks!

3 Comments »

Esquire’s article (by Tom Junod ) about Eric Schadt and non-reductive molecular biology would have been chapter fodder for 2b2k if the book weren’t (I hope) done. Fascinating.

(See the brief but interesting discussion at metafilter.)

4 Comments »