posted 30 Nov 2009 in Volume 13 Issue 3
Thought leader: Knowledge briefs - The social value of knowledge
First, let’s face it: there is *no way* to automate the creation of meaningful knowledge flows. Academic information scientists and AI proponents have been trying to crack the problem of automatic classification for more than 20 years, with little success, after wasting what must be hundreds of millions in dosh. People still mostly interact with sources using shot-in-the-dark word searches. I have been monitoring the failure of the automated approach (along with most colleagues who work in the practical side of information retrieval and knowledge management) for most of that time. Indeed there is a chapter on the subject in Taxonomies: Frameworks for Corporate Knowledge.
Partly, there is no automated solution because computers are not conscious (although they might appear to be, but that is fakery – see the Turing Test). Partly, the difficulty starts because the question behind it presupposes that meaning exists in some unchanging platonic world outside of the influence of a social context. Over the past 2000 years, science and philosophy have pretty much disproved the existence of Plato’s ‘Ideal Forms’. So it is pretty safe to say that fully automated classification is literally meaningless.
And, even if an automated solution were possible, it would have no social value. Who cares what a computer ‘thinks’ (especially when it can’t really think at all anyway.) The massive trick being missed by pursuing the automated approach is the opportunity to tap into people’s own native intelligence as a component of a greater social intelligence.
If you want to create knowledge communities the last thing you want to do is to automate tagging and categorisation. The acts of tagging and classifying are of key significance and create the opportunity for people to collaborate as never before. Each act of classification is an important contribution to the group, and adds to the value of the information streams being created.
The first thing you want to do is keep it simple for people and build up gradually, interest group by interest group. For example, start with a group of people interested in environmental or economics topics. Huge value would then be added by giving them the opportunity to tick, say, five boxes, plus a ‘don’t know’ option. Imagine if many people collaborated on, for instance, the environment, simply by ticking the most pertinent boxes – such as air, land, water, species or simply environment (if the item is about the whole shooting match). It’s easy, quick and very social, since the collection that is built by the users goes far beyond what any of them can do alone. Massive value is added from the start. Much, much more can be added by introducing facets – one by one. That is the Google beater!
Another advantage of this approach is that the statistics of what people are clipping become meaningful because they reflect a human choice, rather than a computer algorithm. Our Open Intelligence software shows these kinds of self signifying statistics changing in real time.
Jan Wyllie is a founding director of Open Intelligence and Trend Monitor. He is also author of one of Ark Group’s most successful reports, Taxonomies: Frameworks for Corporate Knowledge. He can be contacted at email@example.com.
To visit the latest self-signifying knowledge site, go to http://trendsmap.com