Skip to main content
. Author manuscript; available in PMC: 2011 Jan 1.
Published in final edited form as: Cognition. 2009 Sep 26;114(1):42–55. doi: 10.1016/j.cognition.2009.08.012

Figure 4.

Figure 4

Vowel spaces of the three talker groups based on the cardinal vowels /a/, /i/, and /u/ (as in “pot”, “heat”, and “booth”). Points indicate the mean locus of each vowel across all tokens in a group; error bars indicate one standard deviation of the mean. In addition to a consistently higher F1, Groups 2 and 3 exhibited significant fronting of the vowel /u/, characteristic of White dialects of American English, compared to Group 1. The vowel space of Group 3 talkers was virtually identical to Group 2 talkers, consistent with the pattern of listener-categorized racial identity (Fig. 2).