| Search engines, databases, and expert systems are classes of technological mechanisms designed to penetrate and extract information from archives of (primarily textual) information. In order to fulfill users' retrieval objectives, such systems provide (and impose) formal models for converting the mass of archived information into explicitly structured forms. This structure then enables users to extract aspects of knowledge relevant to very specific needs. However, all eventual success in retrieving precise and sufficient relevant information from such archives is always in proportion to users' ability to learn query strategies that must be expressed in terms of the formal languages within which such systems operate. That is, to be effective, most retrieval systems presuppose familiarity both with the internal dynamics of the textual archive, as well as with the formal interfaces, techniques, and languages they require for manipulating and navigating within them. These could be called text-level expertise and system-level expertise, but is a more informal type of conceptual matching possible?; The current work borrows concepts from information retrieval, machine learning, and communication theory in an attempt to implement the kind of process that results from the exchange of shared knowledge and shared experiences similar to what emerges from the dialogical encounter with an elder. There is no attempt to duplicate this human elder encounter in all respects, but acknowledging this impossibility, one can still attempt to ask what happens in the exchange of ideas with an elder simulated by software. More specifically, from the standpoint of execution, we may ask, "How does this cyclical dialogue become implemented?" and from the standpoint of the writer, we may ask, "What does the creative idea-exchanging dialogue comprise?", "How is it perceived?", and "How does it influence the writing process?" These are the questions underlying the dissertation. |
