Development of phonological constancy: 19-month-olds, but not 15-month-olds, identify words in a non-native regional accent

Abstract

By 12 months, children grasp that a phonetic change to a word can change its identity (phonological distinctiveness). However, they must also grasp that some phonetic changes do not (phonological constancy). To test development of phonological constancy, sixteen 15-month-olds and sixteen 19-month-olds completed an eye-tracking task that tracked their gaze to named versus unnamed images for familiar words spoken in their native (Australian) and an unfamiliar non-native (Jamaican) regional accent of English. Both groups looked longer at named than unnamed images for Australian pronunciations, but only 19-month-olds did so for Jamaican pronunciations, indicating that phonological constancy emerges by 19 months. Vocabulary size predicted 15-month-olds' identifications for the Jamaican pronunciations, suggesting vocabulary growth is a viable predictor for phonological constancy development.

Figure 1

Placement of images and areas of interest. Left and right images were centered vertically, and situated horizontally at 20% and 80% of the screen's width from the left border for the left and right images, respectively. Areas of interest (in dark gray) were placed around the left and right images. Fixations within those areas of interest indexed looking at either the left or right image.

Figure 2

Trial procedure used in this experiment. Each trial comprised a silent familiarization followed by a test phase, followed by a reward-sentence phase.

Figure 3

Mean percentage of fixation time to the target image. This was calculated by dividing total fixation time on the target image by the sum of fixation time on the target and distractor images. Both 15- and 19-month-olds fixated to the target image more than chance (50%) when target words were spoken in the native regional accent (Australian English), but only 19-month-olds fixated to the target image more than chance also when target words were spoken in the non-native regional accent (Jamaican Mesolect English). Error bars represent standard error of the mean.

Figure 4

Mean percentage of fixation time to the target image versus log-transformed expressive vocabulary size at 15 months. Although 15-month-olds did not fixate on the target image more than chance in the non-native regional accent (Jamaican Mesolect English), log-transformed expressive vocabulary size is positively correlated with total fixation time on the target image in the non-native regional accent. Log-transformed expressive vocabulary size did not predict their performance in the native regional accent (Australian English).

Figure 5

Mean percentage of fixation time the target image versus log-transformed expressive vocabulary size at 19 months. Nineteen-month-olds fixated on the target image more than chance in both the native (Australian English) and non-native (Jamaican Mesolect English) regional accent, but expressive vocabulary size did not predict their performance in either regional accent.

Figure 6

Horizontal location of raw gaze for 15- and 19-month-olds in the native regional accent condition from onset of the first repetition of the target word. Onset of the second repetition of the target word occurred at 2000 ms, and shaded areas indicate the presentation of auditory stimuli. Data is normalized so that gaze above 400 reflects looking to the target image side of the screen, and gaze below 400 indicates looking to the distractor image side of the screen. While analysis did not reveal any differences in looking patterns across ages, analysis of total fixation time on the target image revealed both 15- and 19-month-olds identified target images in the native regional accent (Figure 3).

Figure 7

Horizontal location of raw gaze for 15- and 19-month-olds in the non-native regional accent condition from onset of the first repetition of the target word. Onset of the second repetition of the target word occurred at 2000 ms, and shaded areas indicate the presentation of auditory stimuli. Data is normalized so that gaze above 400 reflects looking to the target image side of the screen, and gaze below 400 indicates looking to the distractor image side of the screen. While analysis did not reveal any differences in looking patterns across ages, analysis of total fixation time on the target image revealed that 19-, but not 15-month-olds, identified target images in the non-native regional accent (Figure 3).