Mechanisms that improve referential access

Abstract

Two mechanisms, suppression and enhancement, are proposed to improve referential access. Enhancement improves the accessibility of previously mentioned concepts by increasing or boosting their activation; suppression improves concepts' accessibility by decreasing or dampening the activation of other concepts. Presumably, these mechanisms are triggered by the informational content of anaphors. Six experiments investigated this proposal by manipulating whether an anaphoric reference was made with a very explicit, repeated name anaphor or a less explicit pronoun. Subjects read sentences that introduced two participants in their first clauses, for example, "Ann predicted that Pam would lose the track race," and the sentences referred to one of the two participants in their second clauses, "but Pam/she came in first very easily." While subjects read each sentence, the activation level of the two participants was measured by a probe verification task. The first two experiments demonstrated that explicit, repeated name anaphors immediately trigger the enhancement of their own antecedents and immediately trigger the suppression of other (nonantecedent) participants. The third experiment demonstrated that less explicit, pronoun anaphors also trigger the suppression of other nonantecedents, but they do so less quickly--even when, as in the fourth experiment, the semantic information to identify their antecedents occurs prior to the pronouns (e.g., "Ann predicted that Pam would lose the track race. But after winning the race, she..."). The fifth experiment demonstrated that more explicit pronouns--pronouns that match the gender of only one participant-trigger suppression more powerfully. A final experiment demonstrated that it is not only rementioned participants who improve their referential access by triggering the suppression of other participants; newly introduced participants do so too (e.g., "Ann predicted that Pam would lose the track race, but Kim..."). Thus, both suppression and enhancement improve referential access, and the contribution of these two mechanisms is a function of explicitness. The role of these two mechanisms in mediating other referential access phenomena is also discussed.

The role of recognition memory in anaphor identification

“…On the other hand, the Dopkins and Ngo recognition decrement reflects a memory process. Under this view, then, the Dopkins and Ngo phenomenon has no relevance for the Gernsbacher (1989) phenomenon and thus no relevance for anaphor comprehension.…”

“…As a consequence of the act of recognition, other words in the passage are subject to a recognition decrement, through a process such as Dopkins and Ngo proposed. In order to advance the forgoing interpretation, we must deal with two objections. According to the first objection, the Gernsbacher (1989) and Dopkins and Ngo (2002) phenomena do not reflect the same underlying process. On one hand, the Gernsbacher recognition decrement reflects an anaphor comprehension process.…”

“…In other words, the function of the process is to suppress words that might potentially be the antecedent but actually are not. As a consequence of the process, competitors are subject to a recognition decrement following the resolution of a repeated-noun anaphor (Gernsbacher, 1989(Gernsbacher, , 1990 Gernsbacher's interpretation of her results may warrant closer scrutiny in the light of some recently-reported results. Dopkins and Ngo (2002) observed something similar to Gernsbacher's recognition decrement using a procedure that followed Gernsbacher except that the order of the words in the stimulus sentences was scrambled.…”

“…

AbstractIn studies of anaphor comprehension, the capacity for recognizing a noun in a sentence decreases following the resolution of a repeated-noun anaphor (Gernsbacher, 1989). In studies of recognition memory, the capacity for recognizing a noun in a scrambled sentence decreases following the recognition that another noun has occurred before in the scrambled sentence (Dopkins & Ngo, 2002).

…”

“…The phenomenon came to light in a study by Gernsbacher (1989). The stimulus materials for the critical trials in the study were short, two-clause, sentences.…”