Quick note on taxonomic transparency

Notice that I am not using the word “ontology.” I’ll get into why later, but if you’ve read any Heidegger you can guess…

Hope Olson, Sandy Berman, and many others who have done work based on theirs, have shown how classification systems tend not to represent all users well. Hope Olson has described the problem in terms of its philosophic roots by deconstructing classification systems using methods that come from Derrida. I think it is clear and very interesting the way classification systems reflect positivistic assumptions about reality that lead to a presumed “view from nowhere” that reflects the dominant culture. I’m glad that there is a lot of work being done in this area right now, and especially the way people are tying these ideas to new topics like folksonomies and contemporary identity issues.

However, I can’t help thinking that an unjust “official” classification system is not a big problem as long as it is transparent to the user. Let’s say I’m deviant in some way, and I want to find information that is relevant to me using institutions that are connected to the Library of Congress through its standards. Am I really going to expect the Library of Congress to be hip to my understanding of things? No, I am not going to expect that. I am going to do what I am accustomed to doing as a deviant person, which is my choice anyway: I am going to translate the official language, in which I am fluent, into my own deviant street language. That kind of translation skill is part of daily life for anyone who negotiates between formal and informal contexts, streets and offices, subcultures and dominant cultures. The official classification system may reinforce a sense of alienation, but as long as it is transparent to the user, I would not agree that it presents a barrier to access. The cognitive skills involved in making use of an imperfect classification system are very interesting to me (whether we’re talking about reference librarians or end users).

The next plateau of information technology, which we are beginning to reach, is, as I see it, presenting a new problem concerning these issues. Simplified interfaces with artificial intelligence under the hood are “freeing” the user from having to understand either the mechanism of the search or the underlying organizational system, but one thing that this means is a loss of transparency. The user is expected to trust the virtual helper to understand his or her information needs, and it is likely that this virtual helper is going to make assumptions about the user and the context that stem from the assumptions that are embedded into the official classification structure. Without access to the official vocabulary or the mechanism of the search, the user is less able to adapt the tool to his or her needs. Without access to the original text, the user doesn’t have the opportunity to do a translation…