An interesting coincidence today. Nicholas Carr commented on Amazon's unseemly tags and the way in which people have exploited the explicit nature of tags To in effect provide a pass through from French Art Movies to porn (and yes there is a difference!) Carr is one of the great curmudgeons of our time, unafraid to challenge orthodoxy. He provides the stark waring: What they say about sexually transmitted diseases seems to apply equally well to data in the Web 2.0 age: You're not just sleeping with your partner; you're sleeping with your partner's partners. At the same time Dave Weinberger picks up on a post by Tom Hopkins on the relationship between categorisation and truth. Dave argues that we have moved away from the idea that there is some fundamental category, which is more important (in effect an essence) to multiple categories, which includes the use of tags in social computing.
I think this issue of tags and categories is one of the main issues facing knowledge management and social meaning at the moment. In effect we are continuing to use old ways. We like categories, so even when we are trying to express dynamic relationships we stick to them. We increase the number and give equal status, but we are still looking for truth linked to the validity of a tag. Socrates is human, Socrates is hungry etc. This dependence on categories and more specifically the assumption of common meaning in the use of tags remains one of the main challenges to progress in social computing. My own view is that it is limited, and leads to cluster or (as in the Carr case to easily exploitable a systems. It smacks a bit of the cow-chicken-grass difference.
Patrick's Lamb's book on Taxonomy has given me some more ideas on this, as has our own work on how do you signify meaning on narrative. I plan to reflect more on that in the future, but for the moment my comment to Weinberger's blog is reproduced below:
It might be easier if you abandoned the primacy of categories alltogether. Alloing multiple categories, none of which are essential is an improvement but it still fits within a static, or equilibrium based model. If we look at this through the lens of complex adaptive systems theory then we would talk about “coalesences”, “attractors” and “modulators” amoung other things. All of which are more closely related to definining things in terms of their relationship than categories.A now classic test on this is “Which is the odd one out? Cow, Chicken or Grass?” Anglo-saxon (NA, England) tends to be based on categories, grass is the odd one out as its a vegitable. Asia, African, Soouth Europe, Celts tend to eliminate chicken as the cow has a relationship with crass.
Now this has implications for social computing. By increasing tags (a form of category) we do enable richer meaning to emerge but it is limited by common use of language and sophistication of tags. If instead we look at the network relationships we get richer meaning faster.
Comments (2)
Interesting article.. i'm fascinated with the practical application of taxonomy structures...
cheers,
Arjun.
Posted by Arjun THomas | June 1, 2007 6:39 AM
Posted on June 1, 2007 06:39
reference your point about "defining things in terms of their relationship than categories." Here is a paragraph from the May 19th Science News (p. 314) that supports your observation. The article is about the role of bacterial communities in human health and disease:
"New studies are also showing that microbes within a community work together to influence health, a finding that may have a large impact on conventional views of disease. Instead of an illness being caused by the presence or absence of a single pathogen, "the real pathogenic agent is the collective," says David Relman, an infectious-disease investigator at Stanford University." (http://www.sciencenews.org/articles/20070519/bob9.asp -- subscription required)
Posted by christopher bellavita | June 2, 2007 5:16 PM
Posted on June 2, 2007 17:16