Todd and I had a great discussion yesterday on What The Hell I Mean by my research interests. I started by asking him whether he agreed with the basic idea I've been toying with, that information systems decontextualize knowledge. He basically agreed, and eventually we wound our way around to the fact that even language is an information system, i.e., a system for encoding and transferring knowledge. The problem that I've been driving at is that every encoding step -- from experience into language, from language into database, from database into report -- is a translation, with all of the inherent problems of linguistic translation. In other words, the categories from one "language" to another don't necessarily match directly. When you translate from English to French to English, you make tradeoffs at each step and the end product is not exactly what you started with. When you move from narrative to "traditional" information system back to narrative, the same thing happens.
This is getting a little convoluted. Let's start with some premises:
* Knowledge is contestible.
* Humans negotiate the meaning of knowledge through shared linguistic encodings.
* Information systems decontextualize knowledge.
* No information system can produce a perfect replica of knowledge in another system.
It may be that by knowledge I really mean experience (what happens in the real world), and that by information system I mean encoding system. But the Big Win from this discussion was me realizing that I can bring some linguistic theory to bear on this, specifically stuff about translation, semantics, categorization, and metaphor.
But what does this all mean? I'd like to take an ends-means approach to this. Given that we use information systems for all kinds of decisions that affect the human condition (policy making), what are the costs of decoupling the knowledge in those systems from the context in which it was produced? We assume that "data" can be zapped around, shared, transferred with no consequences. I'm not sure...
Gah. This was so much clearer when I started the entry. I think I need to start a lit review of representation, translation, standardization, and ethics. Or something.
Posted on January 12, 2003 at 10:58 AM
Post a comment