Tracking Memes in the Wild, Part II

In my last post I talked about a meme tracking experiment and bemoaned the fact that it provided no way to track arbitrary memes in the wild. Luckily, an e-Literate reader put me on the track to a workable idea in his comment on a previous post.

Martin Terre Blanche points us to a post on his own blog which describes an a trick performed by Stephen Downes (jeez, does this guy ever sleep?) to aggregate blog posts from different blogs about a conference. The trick was that all the posts would use a particular text string or “shibboleth” phrase to identify them as being about the same topic. as Martin correctly points out, any unique (or relatively rare) text string could be used by Google or any other search engine to aggregate posts on a topic, as long as those posts were properly tagged.

There are several things that are interesting about this approach:

• It can be applied to any arbitrary content.
• The participants would decide what counts as the “meme.” So you could really find out what people perceive to be the common idea independent of the particular text content in individual posts. For example, a person writing about “idea viruses” who never uses the word “meme” in a post could still label that post as belonging to the “meme”…er…meme and we could aggregate it along with this post.
• It would be pretty simple to implement this as a web service with existing technology.

This last point really has me thinking. In my next post (after I walk the doggies), I will describe my own vision of how a meme tracking service could work.