next up previous index
Next: Evaluation contributions Up: Contributions Previous: Creative understanding task contributions

Reading contributions

  In order to produce a testbed for my ideas involving creative understanding, it was necessary to place the process within a larger reasoning domain--I chose reading for this purpose. Reading fit my needs since it is a way in which humans learn new things; therefore, there had to be a mechanism which would permit these novel concepts to be understood. Because of this, I believed there would be a close relationship between reading and creative understanding. This belief was supported by the research, as seen in the last eight chapters of this presentation.

My first contribution to the area of reading research is to present a set of tasks which are sufficient to permit the process to occur in a reasoner--as my claim stated,   THE SPECIFIC TASKS REQUIRED TO ALLOW READING TO OCCUR INCLUDE SENTENCE PROCESSING, AGENT IDENTIFICATION, MEMORY RETRIEVAL, UNDERSTANDING OF CONCEPTS, AND CONTROL OF RESOURCES. The identification of these tasks resulted from a combination of studying past reading research from the three fields which supported my work and experimenting with various combinations of tasks in the computational model, allowing the results to refine the theory. There were a number of tasks which were proposed and discarded as it was discovered that they did not support the reading process to the degree necessary.

Next, my research has shown that   ORGANIZING THE TASK SET INTO SUPERTASKS ALLOWS TASKS WITH SIMILAR PURPOSES TO COORDINATE THEIR EFFORTS. If the supertask organization is not permitted, the individual tasks have a more difficult time coordinating their efforts. These two contributions are in support of the claim that   THE REQUIRED TASKS WORK LARGELY INDEPENDENTLY BUT ARE ORGANIZED WITH A HIGHER-LEVEL STRUCTURE.

Since the tasks are working in a more-or-less autonomous fashion while at the same time needing to coordinate their activity, it is possible to conclude that     COMMUNICATION IS REQUIRED BETWEEN THE SUPERTASKS. Specifically, I have demonstrated that two types of communication are sufficient to allow coordination to occur. First,   INFORMATION DISCOVERED BY ONE SUPERTASK NEEDS TO BE AVAILABLE TO OTHER SUPERTASKS. This is an implicit sort of   communication;   in the ISAAC model, this is accomplished via a set of shared memory representations. Once any supertask adds information to the global representations, the other supertasks may access it. In addition to this implicit communication, my work has contributed the fact that     EXPLICIT COMMUNICATION REQUESTS PASS BETWEEN THE SUPERTASKS IN ORDER TO FACILITATE COORDINATION OF PROCESSING.         In the ISAAC system, these explicit communication take the form of the suggestions, anticipations, expectations, and requests.

All of these contributions, taken together, lead to the higher-level claim that   THE BEHAVIOR KNOWN AS READING IS MADE UP OF A COLLECTION OF SMALL TASKS. It is possible to take another step higher and arrive at the top-most claim of this research, by combining the reading contributions, the creative understanding contributions, and the knowledge contributions. As stated in the opening chapter,   A READER NEEDS TO BE CREATIVE TO COMPREHEND REAL-WORLD TEXTS, WHICH WILL CONTAIN NOVEL CONCEPTS THAT MUST BE UNDERSTOOD. However, this set of claim-related contributions are not the only ones my research makes. In particular, I have a set of contributions which relate to the evaluation area of my work--this is the focus of the next section.  


next up previous index
Next: Evaluation contributions Up: Contributions Previous: Creative understanding task contributions
Kenneth Moorman
11/4/1997