An online calculator to compute phonotactic probability and neighborhood density on the basis of child corpora of spoken American English

Abstract

An online calculator was developed (www.bncdnet.ku.edu/cml/info_ccc.vi) to compute phonotactic probability--the likelihood of occurrence of a sound sequence--and neighborhood density--the number of phonologically similar words--on the basis of child corpora of American English (Kolson, 1960; Moe, Hopkins, & Rush, 1982) and to compare its results to those of an adult calculator. Phonotactic probability and neighborhood density were computed for a set of 380 nouns (Fenson et al., 1993) using both the child and adult corpora. The child and adult raw values were significantly correlated. However, significant differences were detected. Specifically, child phonotactic probability was higher than adult phonotactic probability, especially for high-probability words, and child neighborhood density was lower than adult neighborhood density, especially for words with high-density neighborhoods. These differences were reduced or eliminated when relative measures (i.e., z scores) were used. Suggestions are offered regarding which values to use in future research.

Figure 1

Scatter plots of child versus adult positional segment average (top), biphone average (middle), and neighborhood density (bottom). Solid line indicates the linear regression fit line. Dashed line is a reference line indicating a perfect correlation.

Figure 2

Scatter plots of difference scores (adult – child) versus child positional segment average (top), biphone average (middle), and neighborhood density (bottom). Solid line indicates the linear regression fit line. Dashed line is a reference line indicating a difference score of zero (i.e., adult = child). Points falling below the line (i.e., a negative difference score) indicate that the child value is higher than the adult value. Points falling above the line (i.e., a positive difference score) indicate that the child value is lower than the adult value.

Figure 3

Scatter plots of child versus adult positional segment average z scores (top), biphone average z scores (middle), and neighborhood density z scores (bottom). Solid line indicates the linear regression fit line. Dashed line is a reference line indicating a perfect correlation.

Figure 4

Scatter plots of z score difference (adult – child) relative to the child positional segment average z score (top), the child biphone average z score (middle), and the child neighborhood density z score (bottom). Solid line indicates the linear regression fit line. Horizontal dashed line is a reference line indicating a difference score of zero (i.e., adult = child). Points falling below the line (i.e., a negative difference score) indicate that the child value is higher than the adult value. Points falling above the line (i.e., a positive difference score) indicate that the child value is lower than the adult value. Vertical dashed line indicates a z-score of 0.00, differentiating low probability or density (values below 0.00) from high (values greater than 0.00).