Friday, December 24, 2004

Group Intelligence


How smart are you? Do you know? Did you take a test? Do you think you would do better on that test if you had someone else to help you, and would it matter how smart they were?

These are questions to which I don't know the answer, but which are the start of group theory. This url carries a radio transcript in which two people discuss this very phenomenon. For some tasks, a group average decision is reliably better than any expert, but then for others there are strong downsides, such as in the cases of lynch mobs, or perhaps more topically, racism and bigotry.

I would be fascinated to know more about this, as it must have all kinds of implications for economic and business theory. Is the monarchy model really the best way to run a business? Is a single CEO necessary, or might you be better off with two people at the top? For those of us who believe evolution suffers from hill-climbing problems (local maxima), maybe a little game theory could allow us to cross the divide towards a higher peak of efficiency.



Blogger Markel said...


12/24/2004 11:34:00 AM  
Blogger alex said...

This is a very interesting concept, I think. I recently had a discussion regarding a similar issue, in the context of Open Source - in particular, Wikipedia - and the ramifications of moving Authority from an individual entity to an ambigously constrained Group-Think.

I think that the tendency of humans to trust in singular authority in a decision or situation is a habituated aversion to, and protection from, the ambiguous penumbra of complex state of affairs. Plurality is frightening to those who see the world in blacks whites, and any mention of a "committee" or multiple, cooperating authorities reeks of communism and beaurocracy to them.

I think the exaltating of "experts" in these situations disregards the historicity of the expert's Expertise. Surely, expertise does is not born of a vacuumous unbeing, it is a Conditioning of Collaboration between the strands of influence in which it consists.

But, I return to the issue. Expertise, as a concept, is derived and bestowed by a certain "group average decision," and the Sort of Expertise is dictated by the group from which this average is taken. In a way, the Crowning of Experts is an abdication - and a necessary, and functional one, in a societal sense - of the Group, or rather a delegation to Represent the Group Average Decision.

As to whether a non-abstracted representation of his Group Average Decision - that is, a plurality of individuals doing the work from a Literal level - is beneficial, we have ourselves a very interesting question. In one sense, it depends on the Validity of how these Experts are dubbed, and the Forces behind such doing. Also, there is a human-relations element that always waits rear its divisiveness at this point.

Thus, it seems that if the Method by which Experts are Knighted is authentic and honest, there cannot be much said disparagingly against its practice. Likewise, if you can facilitate a functioning Plurality of Voices to constitute the Group Average Decision, you have a winner as well. However, it is more than useful to note that the Real World can be seen as doing its utmost to disallow such cut-and-dry solutions from being realized.

12/26/2004 05:33:00 AM  
Blogger MelbournePhilosopher said...

Wikipedia has to be about the best example of group-think that I've ever seen. It just works so well.

On the other hand, I think organisations with leaders are common for a reason - i.e. they are effective.

Partly, that's because not all people are doers - some are passive, some are introverted, etc etc.

I like your point - plurality is frightening. I think this goes very deep in our culture, and I think it's a bad thing.

Great comments. Thanks for stopping by.

I might do a follow-up post integrating some of those ideas at some later date...

12/27/2004 11:56:00 PM  
Blogger Genius said...

It sounds like the person is trying to imply somthing mystical about the ability of groups to determine answers to such problems.
I imagine for starters most people use a maximum minimum sort of logic where they know hte weight of some cows lets say and they "know" this one weighs more than one and less than another. the average of guesses within bands should be reasonably good. extreme guesses may well define maximums and minimum plausible weights.
But this only works in certain environments dealing with certain problems.
For example in environments where people are heavily influence by each other one gets a heard effect
In other environments one gets a stagnation effect (because the majority almost always is uncertain about a particular action).Usually some sort of structure must becreated in order for the right environment to be created. and usually the creatin of thet environment will require some sort of a heirachy.
A person with a vision at the top people to manage it below him etc...

12/31/2004 08:36:00 AM  
Blogger Genius said...

the article itself is an example of a potential problem. he uses examples of people having sucess with the method - obviously ignoring failures with the method (which must also exist) a sampling bias. For example I could prove any theory by example if I had enough examples to search through.
Also I wonder -
Did the crowds agree with his book before he wrote it? Do they agree with it now?
who are the wisdom of crowds experts?

12/31/2004 08:44:00 AM  

Post a Comment

Links to this post:

Create a Link

<< Home