Education

  Reading issues have always been a concern of educators; in the United States, there have been several distinct phases of the treatment of the process. I am most interested in the modern era, which began with the publication of   Huey's influential text, The Psychology and Pedagogy of Reading (read:huey). The period surrounding the publication of this book marked the transition from viewing reading solely as a social, cultural, and religious activity to viewing it as a general cognitive skill. It was at this point that a rigorous scientific study of reading and the teaching of reading began. In fact, Huey himself identified many of the core concepts still held to be true concerning reading today. After Huey's publication, reading research expanded at an incredible rate. It is reported in Smith   (read-ed:smith1) that prior to 1910, only 34 studies on reading had been published in the English language (no information is given for international studies). From 1910 to 1924, however, 436 new studies had been published. This trend has continued to modern day, with hundreds of reading studies being published every year.

Spurred on by this increasing interest in the scientific study of reading, a number of new researchers joined the field. Many of these were solely concerned with the proper methods for teaching the skill of reading and failed to recognize any higher level reasoning as being necessary. However, a number of researchers began with the belief that reading is more than simple decoding of symbols. Gray (stated in many places and summarized in [#!read-ed:mason1!#]), for example, believed that reading involved the recognition of words and facts, but also the critical examination of the concepts, their interrelations, and their significance. Although Gray's theory was never specified to a point where it can be critically examined, the elements which he discussed bear a resemblance to the reasoning supertask in the breakdown I have proposed. The recognition of words and facts can be seen as the sentence processing supertask which is responsible for converting ``raw'' text into internally understandable concepts. The critical examination of concepts is captured by my reasoning supertask, although memory would necessarily play a role. What Gray did not comment on was the need for both a story structure comprehender and a scenario comprehender, as well as some higher means of controlling the process   (my control supertask). The omitting of these are likely the result of the period in which the work was done--more emphasis was placed on the word decoding level in the early periods of educational research into reading. It was assumed that students had the skills necessary to understand material presented in other forms, for example, orally presented stories. The problem of reading was simply to convert the written word into a form which could then be comprehended.

Taking this idea of looking beyond word decoding as the driving force behind reading was   Russell (read-ed:russell1). He identified six types of ``thinking'' which needed to be taught for reading to be successful. These were perceptual thinking, associate thinking, concept formation, problem solving, critical thinking, and creative thinking. Russell's divisions focus on elements which are incorporated in the reasoning supertask of my own work; most importantly, note the acknowledgment of creativity playing a role in the reading process. Further, notice that Russell greatly lowered the effect which word decoding has--it is still there as ``perceptual thinking'' but it no longer is the primary focus of reading. Again, however, no acknowledgment of the need for either story structure or scenario comprehension is given; nor is control considered a valid area of exploration. These were considered outside the scope of reading and not included in theories of the behavior.

For a computationalist, none of these early theorists provided an exact enough account of their ideas for a complete theory to be implemented as a model. None of these early theories attempted to explain how the various steps of reading were accomplished, only what those steps were hypothesized to be. Fortunately, a new breed of researcher began to publish outlines of both the knowledge and processes which must exist for reading to occur, at a level of description that began to make modeling possible.   Graesser (read:graesser), for example, described six basic knowledge sources involved with textual comprehension: linguistic, rhetorical, causal, intentional, spatial, and roles, personalities, and objects. As Chapter 4 will describe, my theory also emphasizes multiple knowledge sources within the reading process. The ones described here correspond to the knowledge which is needed to understand the sentences (again, the decoding stage of reading) and to comprehend the scenarios being described by a text. Knowledge relating to story structure was not presented in Graesser's work; similarly, no representations of a meta-level are discussed. The control issue was not considered part of the reading skill.

Rather than focus on knowledge, another class of researchers chose to concentrate on the processing which goes into reading, developing a model of the reading task which was largely an information processing one.   The best known theory of this type is that of Gough (read:gough) who came up with a model of ``one second of reading,'' which was designed to describe all of the processes making up one second of oral reading. The model was strictly an information processing account and explained the functions of the reading system and how each manipulate the information coming into the system. The model was a bottom-up account of reading; this fact, and the fact that it was intended to account for only the first second of reading, means that the model is insufficient at explaining many of the skills needed in reading at much higher than the phrase level. In that respect, Gough's theory most closely resembles the sentence processing supertask of my own work. In fact, it goes beyond the level of detail which my supertask is currently described at--as I stated in Chapter 2, my work has focused on the aspects of reading which intersect with conceptual understanding; while the Gough model describes the more perceptual elements of reading, such as eye movement.

Other researchers were less concerned with detailed information processing models and concentrated on more vague ideas, less well-defined than process and knowledge.   Davis (read:davis), for example, focused on skills, which are sort of hybrid process and knowledge packets. He identified five skills necessary for reading to occur. These were: memory for word meanings; drawing inferences from the content; following the passage's structure; understanding and/or recognizing the author's purpose, attitude, tone, and mood; and being able to answer questions, either explicitly or in paraphrase. Notice that these skills are sort of an intermediate level between knowledge and process and include elements from both. The theory which results has the level of description which the ISAAC theory attempts to achieve. For example, Davis recognizes the importance of memory, one of the first to explicitly do so. Other work considered memory outside the scope of reading; Davis places it into his theory in order to make specific claims about the memory needed for reading, although it is limited to word meanings alone. Davis is also one of the earliest researchers to acknowledge that the ability to follow the structure of a text is an important aspect of reading--a fact seen in my work as the story structure comprehension supertask. Control is still not being explicitly discussed, and, as mentioned, memory is still limited. But, the awareness of reading as more than decoding is continuing to develop in the historical research.

The final study on reading from the education field to discuss is the work of     Hidi and Baird (read:hidi1). They argue that the concept of interestingness has been overlooked by other researchers studying reading. Most researchers argue that the importance of a sentence or phrase to the text's overall meaning is the primary influence on how it is processed and how well it is remembered later. Hidi and Baird, on the other hand, argue that the reader's interest in the text as a whole, as well as in particular sentences, also has a great effect on both processing and later retrieval. They go on to identify two types of interest--knowledge-triggered interest, which arises from the information in the text structure; and value-triggered interest, which is dependent on the reader's own values and desires at the time of reading. The result of this is the ability to produce a reading theory which models a more dynamic reading behavior than previously possible in the scope of the theories. A human reader does not read the same text twice in the exact same way. If the only factors influencing this were text-based, then an identical reading should occur. However, the ISAAC theory acknowledges that the reader shapes the meaning which is constructed from the text--it is the combination of the author's intended meaning and the reader's current state which determine the comprehension which is achieved. Reading the same material a second time will, therefore, not produce the same comprehension as the reader of the ``second'' text is not the same as the reader of the ``first'' one, if we could examine their internal state.

In conclusion, my work has focused on the ideas driving educational research; this has strengthened my results by enabling me to produce a theory which takes into account the reader as a student, not simply the reader as an experimental subject in the laboratory. Remember, reading is best researched as a situated activity (see Section 2.3); exploring theories which view reading in a non-laboratory setting is important. On the other hand, the above theories represent the ``best'' of over a hundred years of concentrated research by reading educators; yet, no complete theory was produced. The educators did not have the cognitive tools needed to fully explore and understand the reading behavior. In particular, the educational researchers did not (usually) attempt to produce models of the behavior, only descriptive theories. For the influence of a line of research which does build models, I need to turn to the second discipline, psychology.