ABSTRACT
This chapter describes a method of modeling human reading with a computer simulation that reads the text word by word, and whose processing time or effort on each word corresponds to the human gaze duration on that word. The simulation program, called READER, required the development of a system architecture that permitted the mapping between the humans’ and the simulation’s processing effort. Several factors constrained the modeling. The most obvious is that the simulation’s processing effort on each word of the text should have some correspondence to the human gaze duration. Some of the main empirical findings are reported in another chapter of this volume by Just and Carpenter. The chapter describes for choosing production systems as the formalism for the computer model. The term “production system” has a very unfortunate ambiguity, in that it refers both to an operating environment, like a TOPS operating system, and a particular program that runs within that environment.
