Both Heli Nurmi and MCMorgan have commented on the CritLit2010 week 4 reading from Clay Shirky Shirky: Ontology is Overrated — Categories, Links, and Tags.
I can’t help feeling that the idea that search based on content and tags will replace heirarchical categories is in one sense overstated, but in another sense doesn’t go far enough.
To use tags we need to know them, but if I have forgotten the name of an animal which is large and tusked, a heirarchical search may find it for me. Of course for this to work I may have to be flexible in how I approach the refinement and take account both of the logic and language choices made by whatever entity has constructed the category tree – and perhaps think of geographical locations such as Africa rather than bodily characteristics like tusks.
And if what I remember is the trunk instead of the tusks (or India insteadof Africa), then that should also work just fine as well, since there is no fundamental reason for requiring an item to be uniquely located in the heirarchy. In the case of Yahoo, as mentioned by Shirky, the restriction was just to avoid spamming, and in the case of an OS like Windows any distinction between “real” and “virtual” subdirectories is just a matter of catering to mental laziness (either of programmers or users).
The purpose of a heirarchy is not for location but for navigation and nothing prevents multiple heirarchies being used for the same data. Indeed, contrary to Shirky, computer memory is NOT organized heirarchically but rather more on the basis of frequency of access, and even a simple operating system like Windows provides more than one heirarchy (desktop/myDocuments= c:/docs&settings/me/docs ). This is true even in libraries where location is on a linear system of shelves and often “popular titles” may be located in a separate section closer to the checkout rather than deep within the main stacks. In both cases the navigational heirarchy is linked to the linearly organized actual locations by a table lookup system (library location maps or computer file access system).
The tree graph of a unique categorization system is popular because it is easy to navigate, and so such systems will continue to be important and ability to use them will continue to be a Critical Literacy for the 21st century.
But it may be increasingly important to understand and use more complex relationship patterns as well, and in particular to decide which approach is more appropriate to a particular research problem. So students will need to know not just that there are many possible arrangements of the “is a” relationship (often pretentiously referred to as “ontologies”) but also how to make use in their searching of other relationship types such as “uses” (for file types and other technologies) or “is a ‘friend’ of”, and so on (..not to mention the need for awareness of evolving interpretations of words like ‘friend’).
But this is not new. Using the index of a book is much like searching by keyword or content, and requires the same basic literacy or skill of being adaptable to the language and mindset of the author or cataloguer.
Perhaps it is ironic that in the past the keyword approach dominated in the small scale of a single document and now, as Shirky points out it does so more on the large scale with heirarchies working well in more limited domains, but in fact for both kinds of search the basic literacy required is the same.
Unfortunately, in my experience our educational systems do a poor job of developing this critical literacy. The lack of imagination and flexibility that most students display when trying to google something is just as disappointing as their need to ask an instructor for a simple definition that they could easily find by using the index of their textbook.
If we could find some way of developing this critical literacy in our students it would solve a problem that is not as new as many like to claim, but perhaps the good news is that it may be less necessary in future than before.
In fact, there are now sophisticated information mining technologies that really are new (with Google and the tools being used by people like Dave Snowden just being the tip of the rapidly growing iceberg), and once these are more widely deployed it may be unnecessary for an individual to think of ‘film’ as an alternative to ‘movies’ or for someone obsessed with the ‘homosexual agenda’ to look up the actual views of ‘gay rights’ proponents. If such concept mapping becomes fully automated then an ability to translate one’s queries into a variety of alternate linguistic or “ontological” frameworks may no more critical in future than penmanship is now.