From Pierre Levy’s book Collective Intelligence: mankind’s emerging world in cyberspace:
Groups learn even more slowly than individuals.
The flipside to this is that sometimes groups learn things that individuals can’t or won’t.
8 comments
Comments are closed.
Groups learn more slowly because “learning” at this scale is usually a product of consensus. It’s something that makes me wonder/worry about Wikipedia — the downside to encyclopedias is the requirement that knowledge be agreed upon. This doesn’t leave much room for “edge” knowledge, let alone interesting speculation/innovation. At the very least, theoretical camps have to be acknowledge. What linguists would agree on regarding language is pretty shallow, and mostly has to do with physical elements as in phonetics or syllable structure. But how language operates would look remarkably different when viewed in a Chomskian frame as opposed to a functional frame.
The idea that groups can discover things the individual cannot is an interesting one, but likely tied to the size of the group. Or, perhaps, that what emerges from the group is some final picture with new information when the group process itself is over — when all the threads are viewed as an object for analysis. A map can be generated with “core” knowledge, and levels more peripheral created from there.
Can you please provide a simple example of something a group can learn that an individual cannot?
I think the first time I heard someone make a comment like this was in 1997, when I had dinner with an Intel engineer working on one of their microprocessor design teams. He told me that it was sometimes rather dispiriting, because no one person understood the whole chip in detail, only the group as a whole did. Another good example is modern particle accelerators: if the Higgs is found, there is likely to be no one person who understands the entire chain of evidence.
There’s a good discussion of this in Baetjer’s “Software as Capital” (http://tinyurl.com/2bdwfal) that refers back to Austrian School economists and the car industry. Of course, the line, “Nobody in Detroit actually knows how to make a car,” is more poignant now than it was a decade ago 🙂
There is very good evidence for groups learning more slowly than individuals if you look at Search Engines. Every year the average search string (in health given the data we have but probably true more generally) increases by one word. So if the modal search in 2007 was 3 terms, in 2009, it was 5 terms. Individuals vary widely in their sophistication with search but the whole herd lumbers slowly to learn that more words in the box yields better search results.
David Deutsch states in his book “The Fabric of Reality” that we are not moving away from a state where one person could understand everything, but toward it as our explanations (theories) become broader and deeper.
In this vein, if the Higgs is found, I believe that there most likely will be individuals who *understand* the chain of evidence; there will be no one who has specialized factual knowledge of every aspect of the chain of evidence.
I think we’re talking about the same concept.
“David Deutsch states in his book “The Fabric of Reality†that we are not moving away from a state where one person could understand everything, but toward it as our explanations (theories) become broader and deeper.”
This may be the case in some areas, but in others new results are quite often the product of piecing together work from seemingly disparate areas.
As the knowledge base becomes broader, discovering new implications/results from what is already known can often become harder for individuals to do, so in that respect I think group research may be becoming more important.
One of the main characteristic of group is broad perspective so this can be taken for consideration when we are talking about collective intelligence. I`m more curious can we apply term “collective intelligence” on social networks?