I've never billed myself as an information architect. It actually isn't clear to me that labour should be divided such that there is a discrete role of information architect. The abstract skill set of information architecture crosscuts just about every domain of conventional expertise.
But then, I'm probably not normal: I design information infrastructure end to end, which ranges from sculpting organizational policy to writing the code that implements it.
design of shared information spaces, or something to that effect. I submit that since information architecture uses information as raw material, it is virtually devoid of any content of its own. As such it is best demonstrated through copious examples.
Recently I have been exploring the clustering or partitioning of directed graphs, i.e., the fundamental structure of websites. This produces a natural
hierarchical decomposition along lines which are not likely to be obvious to a conscious designer.
I've also been interested in forensics on existing sites. When a site is already up, we can state ipso facto that the author of a given page thought whatever is on it was important enough to publish. Likewise, the paths people take through a site indicate clusters of information that could be rearranged for quicker uptake.
One of the most significant properties of hypertext is that it is uniquely amenable to the composition of conceptual structures which are considerably more complex than what one can achieve with a book. This capacity for complexity can be applied in ways as diverse as exploring an N-dimensional data object or getting a subtle point across. That said, there is a time and a place for traditional media.