Unfamiliar word comprehension in adulthood: The influence of inference generation, cue selection, memory, and age

Studies of meaning derivation in adulthood reveal that, compared to young adults, older adults favor generalized interpretations. The primary goal of this study was to examine two explanations for over-abstraction in older adults: inferential processes and semantic cue sensitivity. Sixty-six adults (30 aged 19--37; 36 aged 64--87) read passages containing unfamiliar words. For each passage, participants were asked to think-aloud while reading; to indicate which passage cue was the most essential; and to rate four definition choices on a seven-point scale. Young adults rated generalized definitions (generalizations of precise definitions) higher than older adults, whereas older adults rated thematic definitions (story themes) and irrelevant definitions (off-target information) higher. To examine inferential processes, inferences in the think-aloud protocols were categorized as supporting overall comprehension (overall), supporting meaning attainment (supportive), or as embellishing the story (elaborative). Each inference was also categorized as specific or generalized. Older adults produced more elaborative and more generalized elaborative inferences whereas young adults produced more generalized supportive inferences. These results reveal that, when comprehension is challenging, generalization errors can arise regardless of age, but that there are qualitative age differences in generalization tendencies and errors. Young adults were more likely to generalize inferences that support meaning attainment and, consequently, they favored unfamiliar word meanings that generalized precise definitions; whereas older adults were more likely to generalize inferences that were thematic elaborations, and, consequently, favored meanings that reflected thematic interpretations. Hence, the hypothesis that age deficits in inferential processes contribute to the age-related tendency to produce generalized interpretations was supported. Age differences in semantic cue selection were not obtained, suggesting that age-related abstraction deficits are not due to problems selecting meaning-relevant cues. In addition, the data revealed that age differences in definition accuracy and inferential processes were due to the diminished performance of people 75 years of age and older: results that are consistent with longitudinal research revealing a phase of cognitive change characterized by a loss of crystallized or product abilities occurring around 80 years of age.