There’s been a bit of brou-ha-ha recently over and essay on folksonomies written by Elaine Peterson, Associate Professor and Information Resources Specialist at Montana State University. She talks about the disadvantages with using folksonomies – the posh word for data tagging – and ends with this bold statement:
Folksonomy is a scheme based on philosophical relativism, and therefore it will always include the failings of relativism. A traditional classification scheme will consistently provide better results to information seekers.
(Boldification mine.) I don’t believe this is completely true, however I don’t believe it’s completely false either. And the reason? Well, anyone can hit a nail in with a heavy screwdriver, but using a hammer is much better. That’s right, it’s horses for courses between taxonomies on the right and folksonomies on the left.
Elaine Peterson makes some claims about why folksonomies are a dodgy classification system, but I believe in a lot of instances they don’t matter. She says:
The issue is that adding enough of those individual interpretations [how each user views a particular piece of data] through tags can lead to inconsistencies within the classification scheme itself.
Of course, different people can have different views on particular data and that’s a great thing. For example, one of my favourite bands could be labelled in any music shop under “rock”, “folk” and “jazz”, depending on which album, or indeed which track, you listen too. How would a traditional, inflexible taxonomy handle that?
The fact is it wouldn’t, you get what you’re given whether you like it or not. I’m not saying there aren’t times when that kind of classification isn’t needed, as it is.
For example, part of what I do during the working day involves the set of codes used to categorise every ailment, illness and disease known to man. ‘H33’ would mean something very specific, and it would have a parent, and that would have a parent, all the way up to the major groups – cardio-vascular, muscular etc. In that case, a very specific and rigid classification methodology is not just useful, it’s vital.
However, when it comes to things that people interact with in a more random and rambling manner, tags are the way to go. What Elaine Peterson sees as the weaknesses of folksonomies, I see as it’s strengths. For instance:
- …user assigning a heading of “San Francisco”, while another uses “Frisco”…
- …spelling white horse as whit horse…
- …tagging White Horse when the image is of a white cat…
- …using an esoteric tag known to very few…
- Each Internet user is bringing to bear on the item a different linguistic and cultural background…
- There are no right or wrong classification terms in a folksonomic world
That, to me at least, is wonderful. Let’s use these apparent ‘wrong’ things to make sure we cover every base. I may want to search for ‘Frisco’ and not ‘San Fransisco’, I may discover a fantastic picture of a white cat when I’m looking for white animals. I might find some items relevant to me based on esoteric and culturally-specific search terms. And that is all good.
If you want something more specific, or you want to reward tags that are more ‘right’ then by all means have some weighting in there. So, if only one person has tagged my picture as ‘whit horse’ and a thousand have tagged it ‘white horse’, guess which one has the higher weighting? But don’t penalise someone just for having a different opinion to you about something.
When it comes to the ever-expanding world of internet user-driven applications (a la Flickr etc) tagging is not just a useful tool, it’s vital to the growth and reach of the system. Tagging allows people to search and find, discover, create connections they would never have thought of by themselves, follow unexpected paths. And maybe they’ll find something that is just perfect.
I don’t believe any inflexible, rigid categorisation structure will give you that: the joy of unexpected discovery.