Language exposure predicts children’s phonetic patterning: Evidence from language shift Supplementary Material

Related article: https://muse.jhu.edu/article/864633

[Download PDF file ] Supplementary 1

This PDF contains more detailed information about the construction and annotation of the corpus.

[Download PDF file ] Supplementary 2

This PDF describes the additional steps taken to clean and standardize the formant measures.

[Download PDF file ] Supplementary 3

This PDF gives additional details about clip annotation category counts and percentages, by maternal language profile and by child age, and includes figures representing the relationship between both speech production outcomes and the percentage of monolingual Quechua clips.

[Download PDF file ] Supplementary 4

This PDF provides individual vowel plots for the children in the study, grouped by age.

The relevant figures from the main article are repeated here in color, with their captions.

Figure 1a (p. 472).

Daylong recording collection materials: Example shirt used to house recorder during data collection.

Figure 1b (p. 472).

Daylong recording collection materials: The LENA recorder was worn inside of the cloth pocket for the duration of the recording.

Figure 3 (p. 478).

Example area plot of language proportions by number of clips annotated. Area plots were used to track progress toward language proportion stability during daylong recording annotation.

Figure 4 (p. 478).

Example plot of Spanish language proportion variance by number of clips annotated. Variance was computed over a moving window of sixty clips. This plot was used to track progress toward variance stability during daylong recording annotation.

Figure 5 (p. 481).

Proportion of language categories, by maternal language profile. Numbers on barplot reflect percentages of each category. Note: one family did not report maternal language profile.

Figure 6 (p. 482).

Proportion of language categories, by child age (years). Numbers on barplot reflect percentages of each category.

Figure 7 (p. 483).

Children’s vowel spaces by maternal language profile. Ellipses represent 95% CIs, or approximately 2 SDs of all data, assuming a normal t-distribution. Individual points represent a random subset of eight tokens per vowel category.

Figure 8 (p. 485).

Vowel category dispersion by percentage of Spanish clips containing target child. Each point represents one child. Ribbons represent 95% confidence intervals.

Figure 9 (p. 486).

Coarticulation difference by maternal language profile and biphone sequence: there was no reliable effect of profile on coarticulation. Circles and triangles represent each child’s coarticulatory difference across -man ‘allative’ and -pi ‘locative’ morpheme boundaries, respectively. Boxplot hinges represent the interquartile range. Points are jittered horizontally to avoid overlap.

Figure 10 (p. 487).

Coarticulation difference by percentage of Spanish clips containing the target child. Data points represent each child’s average coarticulatory difference by word environment for -man ‘allative’ (blue; dashed line) and -pi ‘locative’ (yellow; solid line). Ribbons represent 95% confidence intervals.

Figure 11 (p. 487).

Coarticulation difference by percentage of Spanish clips containing an adult or nontarget child. Data points represent each child’s average coarticulatory difference by word environment for -man ‘allative’ (blue; dashed line) and -pi ‘locative’ (yellow; solid line). Ribbons represent 95% confidence intervals.

Figure A1 (p. 501).

Clip annotation category counts for each child. Numbers on barplot reflect the number of clips from each category.