What does a very dense book written by a cognitive linguist have to do with Information Architecture, especially when there are some good texts already on this subject, such as the O’Reilly book, or the one written by Peter Van Dijck? More than you might think. So much, in fact, that this may turn into a several post series.

Women, Fire, and Dangerous Things

George Lakoff’s study “Women Fire and Dangerous Things” is the book in question. The book was a big deal when it came out in fields from linguistics to philosophy to computer science, due to some of the assertions the author makes about how human beings construct categories in order to make sense of the world. Information Architecture is about classifications - what if we could apply the key findings from experiments throughout the history of category theory in order to predict how users will label the content in a website, saving time and money in usability research?

Info Architect Donna Maeur summarizes many of the points in this document, including the important point Lakoff makes about prototype effects. Classical category theory is based on set theory, in which categories are envisioned as real-world ‘containers’, mutually exclusive. The categories we construct using language, however, don’t have one-to-one mapped real-world counterparts. Instead, they exhibit prototype effects, in which one member is more central, the ‘best example’ while others are peripheral at best. People have demonstrated in experiments that they are slower at categorizing these items.

Prototype effects imply not that one member is more (or less) of a member than others. Instead, the point is that there is some internal structure to our categories as well.

One of the coolest examples in the book of a result of prototype effects involves a story by Jean Luis Borges, in which a categorization scheme for animals is described. It includes a) those that belong to the Emperor, b) embalmed ones, c) those that are trained, d) suckling pigs, e) mermaids, f) fabulous ones … and so so. Lakoff writes of how a long-term study of an aboriginal tribe’s classification system revealed that such a scheme is not so far-fetched: their classification lumped together women, fire, and dangerous things in one group, while me, kangaroos, storms, and rainbows made of another, in addition to several more equally ludicrous ones.

Now for the take-away related to information architecture: the researcher found in the system that there was one category that could only be explained as “miscellaneous”, or “everything else”. Mythology connected men and kangaroos and rainbows, same for women, fire, and dangerous things. But this final example had no rational explination.

So, next time you are knocking your head against those last couple pages as you map out the categories in a navigation menu, consider the miscellaneous category. The hardest part, in my experience, is the name. While I can’t speak for the results of usability research with regard to such categories, I know that I tend to accept them, and even look for them on complicated sites.

Next time, I’ll talk more about the other “basic” application to info architecture, basic-level categories.