ABSTRACT
In the current study, we examined the role of intelligence and executive functions in the resolution of temporary syntactic ambiguity using an individual differences approach. Data were collected from 174 adolescents and adults who completed a battery of cognitive tests as well as a sentence comprehension task. The critical items for the comprehension task consisted of object/subject garden paths (e.g., While Anna dressed the baby that was small and cute played in the crib), and participants answered a comprehension question (e.g., Did Anna dress the baby?) following each one. Previous studies have shown that garden-path misinterpretations tend to persist into final interpretations. Results showed that both intelligence and processing speed interacted with ambiguity. Individuals with higher intelligence and faster processing were more likely to answer the comprehension questions correctly and, specifically, following ambiguous as opposed to unambiguous sentences. Inhibition produced a marginal effect, but the variance in inhibition was largely shared with intelligence. Conclusions focus on the role of individual differences in cognitive ability and their impact on syntactic ambiguity resolution.
Acknowledgements
The authors would like to thank Laurie A. Carr and Elizabeth Davis for their help collecting the data. We would also like to thank Ioanna Markostamou for helpful comments.
Notes
1Recent work (Wurm & Fisicaro, Citation2014) has suggested several problems with this kind of procedure, specifically regarding the interpretation of the “residualized” variables. In order to be as transparent as possible we report a follow-up in the Discussion to ensure that results are not due to any artefact of residualizing our predictors.
2Example code for the first linear mixed effects analysis is presented in Section A of the supplemental material.
3At the suggestion of a reviewer, we have included an additional analysis of the Intelligence × Sentence Structure interaction in Section B of the Supplemental Material. There is some concern over the degrees of freedom with z-statistics and the fact that they are anti-conservative. However, the model comparison presented in the Supplemental Material confirms a significant improvement in model fit.
4The same may also be true of particularly weak ambiguities as well (e.g., coordination ambiguity).
5It should also be noted that vocabulary also showed some relationship with structure type, as there was a marginal interaction between vocabulary and syntactic structure. However, the bivariate correlations with vocabulary were highly similar to extracted intelligence variable (compare and ).