December 24, 2004

Social capital and end-oriented networks

Posted by John Quiggin

I’m just about to knock off for Christmas1, but I have to get ready for a conference at Queensland Uni of Technology early in the New Year where Larry Lessig will be the main speaker. I’m giving a very short presentation, and struggling to improve my understanding of all this, in particular the relationship between the technology of the Internet and notions of social capital. I haven’t come up with anything earthshattering, but I have had some thoughts on which I’d welcome comments.

Lessig stresses benefits of networks with intelligence at the ends rather than the centre. I want to think about this with respect to the network as a device for disseminating innovations, as well as for communication. In a centralised network, any innovation is automatically available to all users, so the dissemination problem is trivial. On the other hand, only those in charge of the network have the capacity to innovate. In a network with intelligence at the ends, innovators may need to take positive action to share their ideas, and can easily conceal innovations or restrict access if there are incentives for this.

This is reflected, I think in the disappointing returns to the vast amounts or money and effort poured into customer-service innovations of various kinds during the dotcom era. Everything was proprietary, and often patent-protected2. So even if the parts of a good idea were all there, they were unlikely to come together.

To make innovation work in an end-oriented network, we need the kind of social capital, and also the kinds of technical protocols that encourage and facilitate sharing. What these are is a topic for further research.

1 Despite work intensification and so on, Australia still shuts down almost completely for a couple of weeks.

2 I had my first experience of One-Click shopping at Amazon the other day. I’m sure I’d get used to it in the end, but I must say I found it rather unnerving. One click to place the order and one to confirm it suits me rather better.

Posted on December 24, 2004 03:03 AM UTC
Comments

How interesting … this has echoes of Odlyzko’s paper http://www.dtc.umn.edu/~odlyzko/doc/pricing.architecture.pdf, that discussed what amounted to an end-to-end argument in pricing.

And, oddly enough, it has echoes of the end-to-end argument itself.

Posted by Chet · December 24, 2004 03:46 AM

Lessig stresses benefits of networks with intelligence at the ends rather than the centre.

Indeed. I’m trying to imagine the converse - perhaps a 1980s Telecom Australia that’s more like the ANU rather than the nation’s largest sheltered workshop.

To make innovation work in an end-oriented network

Depends on the innovation. The blogosphere is an example of an exchange of ideas without a pricing model. Does this fall under your definition of social capital, John?

Posted by Syd Webb · December 24, 2004 04:42 AM

The blogosphere is very much my paradigm example. In fact, if anything I run the risk of relying too much on it as a model, so if people can point out how it may be atypical, that would be good.

Posted by John Quiggin · December 24, 2004 05:19 AM

Kenneth Arrow relevant here? I was thinking of writing a piece called “God and Kenneth Arrow” in response to the 2004 election fiasco (basically, Republicans have top-down hierarchy based on, first, God; second, Party; third, GW—while Democrats have “majorities in minority a majority of the time”—something reflected in rather poor showing of blog/meetup hype of this years’ Internet campaigns; God of course, is ultimate dictator/arbiter of goods/wants—if we all follow God’s desires rath our own, no Arrow/Voters Paradox)—but probably won’t. What I’m thinking is this: given complexity AND logical structure of distributed goods on your network, any way in which you CAN arbitrate what gets distributed may just be a tremendous PLUS (also: isn’t this just why traditional media/editorial power hasn’t been undermined much by inet?).

Posted by joel turnipseed · December 24, 2004 05:48 AM

Reminds me of Thomas Malone’s recent book, “The Future of Work,” in which he discusses four kinds of decentralized production networks: loose hierarchies, democracies, internal markets, and external markets. He recommends different mixtures of each, depending on the task you’re trying to accomplish (in a discussion of this kind, for example, the blogpost/comments form a loose hierarchy, where the post forms a nodal point for [ideally] intelligent input). It does seem to me that effective sharing in a truly decentralized network requires the services of a “librarian” of some kind, even if it’s only a machine aggregator.

Posted by James Brink · December 24, 2004 06:07 AM

In a centralized network, everybody is directly able to access innovation in one step.

In a network like the web, there might be several steps (as there are in networks of who knows who.) You might be able to relate diffusion models to size of the network, degrees of connectivity, and the numbr of novel ideas integrated in an invention. Then you could predict rates of innovation as those three variables change.

Posted by Mike Huben · December 24, 2004 01:10 PM

Not all innovations may be desired by everyone in the network. It seems the more useful an innovation is to a wider audience the faster it will be adopted in an end oriented network. And it will be adopted first by the people who need it most.

In a centralized scheme I assume innovations are imposed top down rather than adopted bottom up. There seems like there could be some potential fall backs to this. Some innovations appeal only to specific groups. Others are better off with out being required to learn this new innovation that really doesn’t help them.

The lost time in mastering this new little widget could actually cost some people in the long run in a centralized network when established ways of doing things would have worked just as well.

Posted by Graham Heffern · December 24, 2004 05:35 PM

Paging Ronald Coase and his theory of the firm yet again? Of course you need a mix of different cooperation styles.

Someone whose views I respect on this argues that the end-oriented Internet as idealized by folks like Lessig is long dead, what with NAT and firewalls and vpns and QoS and whatnot. And treating networks with a bunch of internal intelligence as completely end-oriented can be as bad as thinking that market institutions are frictionless and don’t matter…

One of the issues here may be more about physical capital than about social capital — in many other post-cataclysmic booms (the post-ww2 manufacturing and electronics boom is pretty much paradigmatic) you had enormous amounts of surplus stuff lying around, already paid for, for anyone smart to make use of at very low cost. This time around, Moore’s laws has substantially reduced the utility of much of that surplus stuff, with the possible exception of some of the expertise. But whether that social capital is stable enough to stick around and be shared isn’t immediately clear to me.

Posted by paul · December 24, 2004 11:05 PM
Followups

This discussion has been closed. Thanks to everyone who contributed.