The creative reading theory can be viewed in a variety of ways. At one level, it represents a merging of a process theory and a knowledge theory. Because the two portions are treated equally in the theory, it is possible to determine where the ``power'' of the theory originates. Viewed from another perspective, the theory can be seen as being made up of two basic aspects. The first is the framework of the six supertasks. In addition to producing a more complete understanding of the reading process, this framework has also permits a large number of past research efforts to be related in a common environment. This is partially why my model of creative reading was possible; I was able to take advantage of a great deal of part research ideas which had never been put together before. The second aspect is the theory of creative understanding which supports the overall reading process. Another interpretation of the overall creative theory is to view the work in terms of the extensive evaluation performed on both the model and the theory. The level of evaluation is unique in that it treated the model as an opportunity to perform empirical evaluation using accepted tests. Although some subfields in artificial intelligence have been doing this for decades, the type of ``scruffy'' AI theory which this work is has generally been forced to avoid such empirical evaluation.
Ultimately, the evaluation demonstrates that ISAAC is a capable reader for these three stories based on independent evaluation. In fact, the ISAAC system is performing at a level indistinguishable from four human students, with respect to the higher-level aspects of comprehension. This is significant in light of other facts about the model--without creative understanding, ISAAC would not be able to comprehend the stories discussed in this dissertation; at least not without extensive knowledge re-engineering designed to put precisely the correct concepts into the basic ISAAC memory system. Thus, the ability to perform creative understanding gives the model the ability to be a more capable reader. As such, the model provides validation that the theory is sufficient to explain the behavior of creative reading. Using ISAAC as a testbed, I have been able to demonstrate that my theory can explain these abilities. ISAAC has shown itself able to creatively understand novel concepts; concepts which are essential to the proper comprehension of the stories considered.
The goal of this research has been deceptively simple--to demonstrate that understanding contains a creative component. Most of the creativity research done in artificial intelligence has concentrated on the idea that creativity, however it is explained, is an element of design or invention. There have been some implicit concessions given to the thought that understanding also needs to be creative, but these ideas were never developed to any degree of theoretical support. Instead, understanding has largely been thought of the process by which ``new'' concepts are matched to existing ones, thereby allowing the prediction, abduction, and explanation of the concepts to take place. My work has focused on the idea that this ``straightforward'' matching is actually only one aspect of the understanding process, that understanding involves a rich set of tasks in order to accomplish its goal.
I said this was a ``deceptively'' simple idea due to the level of complexity which arose once I began to study it. In the large, my work has impact on a number of areas. Understanding understanding is the first of these. Understanding reading is the second. Understanding creativity is the third. These are inseparable in my work--remove one, and the entire troika collapses. As I described in the future directions section, each of these can lead to more research, for each is a complex area with nuances I have yet to begin to tease apart. And, it eventually all goes back to that idea which has driven and focused my work--reading is fundamental.