Lexical ambiguity across the adult lifespan
Date
1987
Authors
Norton, Janice Marie
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Age differences in ambiguity detection were examined in the current study. Thirty (each) young (20-35), middle-aged (40-55), and older (60-73) adults were assessed on their ability to detect lexically ambiguous sentences that were presented in isolation (Experiment 1) or preceded by a contextual paragraph (Experiment 2).
In Experiment 1 the bias inherent in the ambiguous sentences was determined for each age group. Based on this information, sentences for Experiment 2 were selected and contextual paragraphs were designed. Paragraphs were congruent with either the dominant or subordinate meaning of the sentence. For one half of the context/sentence pairs, the relationship between the context and the ambiguous sentence was explicit, for the remaining context/sentence pairs, it was implicit.
Fewer ambiguities were detected when the context was congruent with the dominant than subordinate interpretation of the sentence. Similarly, fewer ambiguities were detected when the context/ambiguity relationship was implicit than when it was explicit. A context by implicitness interaction indicated than the effects of context were only manifest in the implicit condition. Detection latencies were longer in the dominant than the subordinate condition, and in the implicit versus the explicit condition.
Whereas there were no age differences in accuracy or latency of detecting isolated sentences, there were age differences in latencies for ambiguities preceded by context. Older subjects were slower to detect ambiguities than were young and middle-aged subjects, particularly in the implicit condition. The pattern of latencies varied by age. For young subjects, a dominant context led to the longest latencies and subordinate to the shortest latencies. Middle-aged subjects were relatively unaffected by context, and older subjects detected isolated ambiguities most rapidly.
Of the cognitive (verbal & perceptual) and demographic variables (age, sex, & education) assessed, crystallized ability best predicted both accuracy and latency of detection.
It was hypothesized that age differences did not occur when sentences were presented in isolation because access of multiple meanings is an automatic process unaffected by age. Age differences did occur when multiple meanings had to be held in working memory while inferences were drawn from contextual information.