This is the second post on Clay Shirky’s talk at LSE, looking at some of the same issues raised, but in the context of decision making and crowd sourced wisdom. I hinted at some of my thoughts in the previous post (Mass Collaboration Snow Joke), and JP has also blogged about it, based on Clay Shirky’s talk at the ICA the day after. 


In his post JP shares some thoughts about systems for decision making. The ideas are interesting (and have been debated in other contexts). It is worth remembering that government’s influence on our lives goes far beyond spending from the state wallet. Government sets policy and makes laws too. Thinking about recent anti-terror and surveillance legislation, along with proposals in the Digital Britain report, arguably, policy and law affect our lives the most.

It is possible to build an on-line voting system to provide access to every policy decision, but as Clay noted in his talk, the results aren’t always the utopian ideal we would hope for. Controlling policy directly may not be a good thing. It becomes easy for a well organised minority to ‘out-influence’ a quieter, less galvanised majority. To avoid that problem would require compulsory voting, but do you want people forced to vote on issues they don’t care about or that don’t affect them? 

The same issues exist for social decision making tools used in an organisational context. While “Voting” has become popular for making some decisions, generally we don’t run companies as democracies. Why not? Because we (or more specifically the business owners) prize expert decision making. Some of the larger companies I have worked with do have town hall meetings. These are loose approximations to the early Greek ideas of democracy, soliciting feedback and dialogue, but they are the exception rather than the rule. Businesses are, at least notionally, meritocracies. People gain authority based on their ability to make good decisions and to use authority well.

Several times in his talk, Shirky made the point that the democratic franchise grew up based on the ideal of one person one vote – actually one man one vote, but that’s another issue. Democracy requires a strong grip on identity. I must be sure of who you are before you vote, in order to enforce one person one vote. However, the voter’s opinion itself is afforded anonymity. I know who you are, but not how you voted. In the UK, more so than in US culture, most people’s voting intention is an intensely private matter, expressed in an intensely private ballot. 

Dalai Lama (not) on Twitter

The Internet grew up as a very different type of franchise. Via NFSnet and FIDOnet (and communities like The Well) anonymity was accidentally implicit, if not deliberately and explicitly so. The systems had no way of knowing who someone was, in the sense that we would understand identity management. People frequently used synonyms, and even when they used a ‘real name’, verifying they were actually that person was a non-trivial exercise. As a side note, Twitter has been experiencing the same fun and games recently, with people grabbing Twitter accounts and masquerading as celebrities (from Tony Benn to the Dalai Lama).

Back to votes and opinions for a moment. In the on-line world we often know a lot about what someone thinks. There is anonymity of identity, but not of opinion. A mirror to the democratic franchise. Interestingly, from my own work with Wikis and from other academic studies, I have noticed that sites where people can post completely anonymously get significantly more contributions that those that don’t.

So, how can we make companies more democratic, and how can we make government more participative, in the social media sense? The answers come not from technology, but from understanding the nature of democracy itself. The art of an effective democratic system is to defend factions from each other. Tony Benn, articulates it well in this clip from “Big Ideas That Changed The World”, you might not agree with his views, but his argument is an informative one, if you are new to the concepts:

Incidentally, the video is also a good counter to Shirky’s statement that Democracy started in the UK with the Magna Carta, signed just up the road from where I am sitting right now now, in a a field in Runnymede. Democracy has evolved mechanisms to deal with working at scale. I can exchange a little information with a lot of people, or a lot of information with a few people. Information exchange doesn’t scale to both ends at once. Democracy tackles that problem by the use of elected, professional representatives as intermediaries. A vote is a small piece of information from a lot of people. A consultation process is a lot of information exchanged within a smaller group of people. Familiar mechanisms that tackle the problem.

What came before democracy was tribalism, which JP’s post alludes to with the description of the open source community. Projects are often run by a ‘tribal leader’ and rings of followers creating a social structure supporting them (see Dunbar and other anthropological studies). That structure does not work with anonymity of opinions. Visibility of allegiances is required to allow the structure to sustain itself.

Here is an apparent paradox: Anonymity promotes extremes of views, by taking away the moderating effect of social influence. We are compliant creatures by nature, and social pressure pushes us towards moderate, or normative, views. We adapt our views, based on our perception of other people’s views. It is a socially useful behaviour, since it makes it easier to form coherent groups. Johnnie Moore and Mark Earls, author of Herd, gave some great demonstrations of this during their session on social influence at NESTA yesterday, and Mark’s book is probably a good place to study it more.

In designing social decision making systems, one must take account of identity, anonymity and accountability. Systems must also balance the desire to have everyone participate, with the need for informed expert opinion. During the NESTA session, Johnnie Moore made an astute observation about organisational design: business design is about balance the need for efficiency, and the desire for full participation.

These are all thorny issues. In a representational democracy we vote for someone we believe is able to represent us. At least we should. In the workplace this is expressed in the form employee councils and so on. Can we place these things with social software? An old IT/programming adage springs to mind: Don’t mess with something unless you understand why it was that way in the first place. We need to apply new technology, with the benefit of understanding old ideas. Yet another thing to add to the list of important, but non-trivial tasks.

There are usually trade-offs and compromises to be made. No system is perfect. However, experience shows, from systems to products, that it doesn’t take perfection to win. In the early days of Cisco Systems, a group of consulting engineers got in to terrible trouble for having a T-shirt printed that said “Cisco – We suck less”. The positioning goes a little against the grain of modern marketing techniques, but it rings true. As Darwin would have put it, survival of the ones that are the best (least bad) fit for their environment.