Roses Are Red, Socks Are Blue: Switching Dimensions Disrupts Young Children's Language Comprehension

Abstract

Language is used to identify objects in many different ways. An apple can be identified using its name, color, and other attributes. Skilled language comprehension requires listeners to flexibly shift between different dimensions. We asked whether this shifting would be difficult for 3-year-olds, who have relatively immature executive function skills and struggle to switch between dimensions in card sorting tasks. In the current experiment, children first heard a series of sentences identifying objects using a single dimension (either names or colors). In the second half of the experiment, the labeling dimension was switched. Children were significantly less accurate in fixating the correct object following the dimensional switch. This disruption, however, was temporary; recognition accuracy recovered with increased exposure to the new labeling dimension. These findings provide the first evidence that children's difficulty in shifting between dimensions impacts their ability to comprehend speech. This limitation may affect children's ability to form rich, multi-dimensional representations when learning new words.

Fig 1. Mean Accuracy by Block.

Proportion of time spent looking to the target object out of the total time spent looking at both objects during the critical window for trials before the dimensional change (Pre-Switch) and after the dimensional change (Post-Switch). Chance = 0.5. Data points represent the proportion for each subject averaged across trials. Error bars represent +/- 1 SE.

Fig 2. Mean Accuracy by Block Halves.

Proportion of time spent looking to the target object out of the total time spent looking at both objects during the critical window for trials in the first half and second half of the Pre-Switch and Post-Switch blocks. Chance = 0.5. Data points represent the proportion for each subject averaged across trials. Error bars represent +/- 1 SE.

Fig 3. Time Course of Fixations by Block.

Time course of fixations to the target object on Pre-Switch and Post-Switch trials. Fixations are plotted as the empirical log-odds averaged across participants. Chance = 0 log-odds. Data points are observed behavioral data and lines the growth curve fits (ribbons around the lines indicate +/- 1 SE).

Fig 4. Time Course of Fixations by Block Halves.

Time course of fixations to the target object on trials in the first half and second half of the Pre-Switch and Post-Switch blocks. Fixations are plotted as the empirical log-odds averaged across participants. Chance = 0 log-odds. Data points are observed behavioral data and lines the growth curve fits (ribbons around the lines indicate +/- 1 SE).

Fig 5. Mean Accuracy by Block by EF.

Proportion of time spent looking to the target object out of the total time spent looking at both objects during the critical window for Pre-Switch and Post-Switch trials for children with low EF (left panel) compared to high EF (right panel). Chance = 0.5. Data points represent the proportion for each subject averaged across trials. Error bars represent +/- 1 SE.

Fig 6. Time Course of Fixations by Block by EF.

Time course of fixations to the target object on Pre-Switch and Post-Switch trials for children with low EF (left panel) compared to high EF (right panel). Fixations are plotted as the empirical log-odds averaged across participants. Chance = 0 log-odds. Data points are observed behavioral data and lines the growth curve fits (ribbons around the lines indicate +/- 1 SE).

Fig 7. Mean Accuracy by Block Halves by EF.

Proportion of time spent looking to the target object out of the total time spent looking at both objects during the critical window for trials in the first and second half of the Pre-Switch and Post-Switch blocks. Children with low EF are plotted in the left panel and high EF in the right panel. Chance = 0.5. Data points represent the proportion for each subject averaged across trials. Error bars represent +/- 1 SE.

Fig 8. Time Course of Fixations by Block Halves by EF.

Time course of fixations to the target object on trials in the first and second half of the Pre-Switch and Post-Switch blocks. Children with low EF are plotted in the left panel and children with high EF in the right panel. Fixations are plotted as the empirical log-odds averaged across participants. Chance = 0 log-odds. Data points are observed behavioral data and lines the growth curve fits (ribbons around the lines indicate +/- 1 SE).