An Evolutionary Experiment

My summer reading last year was Eric Beinhocker's 'The Origin of Wealth'. It's probably the best non-fiction book I've read in several years. I was really struck by his description of Microsoft's strategy in the OS wars of 1987 and wondered how that same approach could be adapted to the formation of democratic, self-governing communities in virtual worlds such as 'Second Life'.

Beinhocker takes us back to a time when Microsoft's continued dominance of the PC operating system market was far from assured. In 1987 MS-DOS was coming to the end of its useful life and Microsoft faced competition from other operating systems and the potential to fail if it's preferred solution - Windows - did not find traction in the marketplace. IBM was developing its own multi-tasking operating system OS/2, AT&T was leading a consortium including Sun Microsystems and Xerox to develop Unix, Apple was also a threat and maintaining a reputation for innovation. What Microsoft did was to invest in six possible futures simultaneously. Firstly, they continued to develop MS-DOS; with a large installed customer base and the potential to make incremental improvements it was possible that users might prefer to stick with the operating system they knew and were most familiar with. Secondly, they worked with IBM in a joint venture to develop OS/2 - 'if you can't beat them, join them'. Third, they held discussions with the various companies developing Unix to consider joint working. Fourth, they bought a major stake in Santa Cruz Systems, the major seller of Unix systems on the PC. Fifth, they built themselves a major Macintosh application development division outstripping Apple as the major developer of software for the Macintosh. Sixth, Microsoft made a major investment in the development of Windows.

Beinhocker sees the economy as a 'complex adaptive system' akin to the brain, the internet or an ecosystem. In such a system evolution is the driving force for adaptation and change. If we think of the operating system marketplace as just such an evolutionary system, we can see that Microsoft was experimenting with a number of designs to prepare for what would be the best fit in the marketplace. This feedback from the system informed the process of selection so that projects could be upgraded, wound down or ended. Once it became clear that the Windows strategy had the lead, Microsoft could divert investment to the successful project. Having a portfolio of strategies operating at once meant that they could quickly switch horses depending on which outcome was working best.

One problem we have had in the Confederation of Democratic Simulators (CDS) is that, for most of our history, we have been the only democratic, self-governing community in Second Life. As a result we tend to attract all or most of the people who would be interested in developing systems of democratic self-government. There are many different ways we could organise ourselves and so this leads to endless debates as the parliamentarians argue their case against the presidentials; the strict separation of powers crowd battle with those who favour cabinet-style government; the direct democrats argue for referenda against those who support representative democracy; the federalists battle the republicans. The Constitution we have ended up with is a bit of mish-mash - the result of many compromises between these different camps. The history of constitutional development in the CDS has largely been driven by compromises between competing visions of what democracy in virtual worlds should look like. But perhaps there is another way, an evolutionary way to test these ideas?

What if a rich benefactor set out to test competing ideas by establishing six new democratic, self-governing communities in Second Life? Each one could adopt a different democratic model as its template. We could have a "direct democracy" sim where all decisions are taken by weekly vote at a meeting where all citizens may participate and make proposals. There could be a "presidential" sim where a strong executive and strict separation of powers prevails. We could try out a "constitutional monarchy" with a figurehead monarch and decision-making politicians. Another sim might have an elected "tribal council" which exercises all legislative, executive and judicial functions. We could vary the executive configurations to try a three-person "triumvir" instead of a single-person presidency. We could try small legislatures v. larger ones. I'm sure there are many more than six possible variations but six would be a reasonable number to test a variety of quite different systems. Once established, the six communities would be set running and the business of elections and running the governments set in train.

How would we measure success though? One way would be to permit the communities to expand and bring in new citizens. The most successful model would be the community that expanded the most. Or we could permit the citizens in these six communities to 'vote with their feet' by allowing them to move from one community to another at intervals; the successful models would thrive and the unsuccessful ones decline.

But is expansion or contraction by itself a sufficient metric? Perhaps success could be measured by assessing citizens' quality of life according to a number of measures? We could use a version of the United Nations Development Program's Human Development Index adapted for our virtual world. A successful community, by this measure, would be one in which citizens' satisfaction, government approval and involvement and activism rates were high.

Clearly any such experiment would not be entirely free of bias or ethical objections. Each community would have to have different people involved (and that could be a difficult rule to keep given the ease with which alts are created). It might be that one community just gelled better and got on better than another because of the interaction between the different personalities. The islands may have different natural resources due to geography, location and public infrastructure (unless they were all clones of each other and co-located to remove those variables). One could also object to the experiment on ethical grounds - is it really acceptable to treat Second Life residents like lab rats and devise an interesting cage for them to run around in? On the other hand, if SL residents are willing to participate is there really a problem? The opportunity to live in a self-governing democracy is arguably one that many would take given the option.

I hope and anticipate that further democratic communities will emerge on the grid in a more organic fashion in the future so perhaps my artificially-constructed social experiment will not be necessary. I should have thought though that Second Life would provide exactly the kind of space in which to conduct such social experiments at minimal cost providing ethical standards were maintained. If there are any political studies departments out there willing to sponsor, I'm prepared to take a year's sabbatical from my paid job to get this off the ground! :) (But I won't hold my breath waiting to hear from someone).