Butcher, K. R. & Kintsch, W. (2003). Text comprehension and discourse processing. In Handbook of psychology (Vol. 4). Hoboken, NJ: John Wiley & Sons, pp.575-595.
The following is a critical summary I produced as part of the Human Factors class I was in last term.
This chapter from Handbook of Psychology is an overview of key concepts in the psychological approach to text comprehension and discourse processing, including various types of memory, different aspects and models of comprehension, and models of knowledge representation (including Latent Semantic Analysis, a vector-based analysis of proximity and relationship to other words in order to gauge meaning).
Butcher & Kintsch point out that while some “discourse analysts view language essentially as a means for information transmission” (575), others also connect language to social interaction, where language is used “establish social roles, to regulate social interactions, to amuse, and to entertain” (575). Meanwhile, the connectionist models of language attempts to bridge the mathematical, formal models with the “disorderly part of language” (576), the irrationality of language.
The chapter often refers to the difference between short-term working memory and long-term working memory. The chapter references the reading-span test developed by Daneman and Carpenter “to measure the functional capacity of working memory for reading” (576), in which test subjects are asked to memorize the first and last words of a sentence as they are reading it aloud. This test has been used to show that “individuals with high working memory capacity are at a comprehension advantage” (577). While high working memory capacity is seen as a clear advantage, the chapter also points out the crucial role of domain knowledge in comprehension.
It seems that I am particularly susceptible to the power of the seductive detail, that tidbit of information that is not directly linked to the main subject matter, but is interesting on its own and as such tends to be remembered by the reader (585). The fact that the idea of the seductive detail was in and of itself a seductive detail in this article particularly appeals to me. In fact, someone should test whether a good seductive detail has the potential of affecting a reader’s overall impression of a text. Another similar seductive detail in this text (seeing as it is not discussed at length) is the idea that a reader has to have some base knowledge in order to learn anything from a new text (576). Although the chapter briefly refers to the importance of domain knowledge, this topic could easily have been discussed more fully.
In our class discussion, Dr. Blustein discussed the mathematics involved in the statistical analysis used in Latent Semantic Analysis, pointing out the difficulty of The Vocabulary Problem, and how this approach does not take into account the possibility of the same word being used in one text but not always with the same meaning (i.e. a text about the management of an apple orchard using Apple computers. He also suggested that this approach is not as good as the best Information Retrieval tools.