Aspects of Literary Comprehension

Researchers in the field of text comprehension have traditionally distinguished between two types of text: narratives, which focus on plot, story lines, characters, etc.; and expositions, which direct the reader’s attention to the factual aspects of a text, rather than the story per SE (see Britton and Black, 1985; Weaver and Kintsch, 1991, for a more thorough discussion of the two text types). Over the past two decades, cognitive science has made great strides towards understanding the comprehension process of these two text types (e. g., van Dijk and Kintsch, 1983). However, cognitive science has had relatively little to say on the topic of literary text comprehension. This volume is one of the first major research efforts investigating comprehension of literary texts. At the risk of “giving away the ending” of this story, let me say that Zwaan’s work represents an outstanding first effort and is likely to serve as the inspiration for much additional research into literary comprehension. Aspects OfLitcr-ary Coniprehension is a monograph apparently based on the author’s dissertation, carried out in both the Netherlands and the United States. Zwaan reports a series of experiments in which he applies the Kintsch and van Dijk (1978; van Dijk and Kintsch, 1983) model to comprehension of literary texts. In its earliest and simplest formulation, the Kintsch and van Dijk model (heretofore referred to as KvD) was a cyclical, recursive model, constrained by Working Memory resources, resulting in representation at two levels: a swface level, which encodes the verbatim trace of the passage, and the propositional textbase, which encodes the meaning of the text, but not necessarily the exact wording. As such it can be considered a type of “paraphrase”. Numerous studies have demonstrated that typically surface traces are fleeting, while textbase representations are better remembered over longer periods of time. The 1983 modification to the model recognized that for some texts, readers construct an additional level of representation, what KvD call a “situation model”.’ The situation model represents what the text was about, rather than what was actually said. While textbase representations are relatively static (since they are determined primarily by the text structure), situation models are quite flexible. They can change as a function of a reader’s goals or background knowledge, among other things, and may be represented in non-verbal form. For example, while reading a passage about the layout of a small town a reader might construct a mental map of the town (i. e., Perrig and Kintsch, 1985; Taylor andTversky, 1992; Weaver and Kintsch, 1987). Or, when reading aprogrammer’s manual,