Comprehension skill: Suppression versus knowledge activation

Suppressing Irrelevant Information: Knowledge Activation or Inhibition?

“…Experiment 3 demonstrated that participants with greater general knowledge, regardless of reading skill, more quickly suppressed the irrelevant meaning of the ambiguous words in general-topic sentences. As predicted by D. S. McNamara's (1997) knowledge-based account of suppression, ambiguity effects are influenced by greater activation of knowledge related to the intended meaning of the homograph.…”

“…As an alternative to Gernsbacher's proposal, McNamara (1997) proposed a “knowledge-based” account of the suppression results (Gernsbacher & Faust, 1991; Gernsbacher et al., 1990). The knowledge-based account was grounded in Kintsch's (1988, 1998) construction-integration (CI) model of comprehension.…”

“…McNamara's (1997) account was also inspired by a large body of reading research showing that skilled comprehenders more actively process information than less skilled readers. In particular, skilled readers tend to generate inferences that repair conceptual gaps, whereas less skilled readers tend to ignore gaps and fail to make the necessary inferences (e.g., Long, Oppy, & Seely, 1994; Magliano, Wiemer-Hastings, Millis, Muñoz, & McNamara, 2002; Oakhill & Yuill, 1996; Oakhill, Yuill, & Donaldson, 1990).…”

“…McNamara (1997) proposed that the results reported by Gernsbacher et al (1990) could be explained more parsimoniously in terms of enhancement of relevant information rather than inhibition of irrelevant information. According to Kintsch's (1988, 1998) CI model, incoming information and associated knowledge are represented as an associative network of nodes (concepts, ideas, or propositions) and links (relations or actions).…”

“…McNamara (1997) reasoned that if skilled comprehenders activated more knowledge associated with the context provided in the sentence, then more links to the relevant meaning of the ambiguous word would be created and the irrelevant meaning would quickly lose activation. In support of that hypothesis, she demonstrated within a CI simulation that the number of activated associations to the relevant meaning predicted the rate of activation loss for the irrelevant meaning.…”