A group of native Mandarin Chinese speakers listened to tapes of six American females speaking in Chinese as their second language and evaluated each speaker on a set of 22 traits. Another group of native Chinese speakers listened to the same tapes and evaluated the six speakers on areas of Chinese language proficiency. The trait evaluations were then correlated with the language evaluations. Evaluations of 8 of the 22 traits showed significant correlations with evaluations in every language area. In contrast, seven other traits showed no significant correlations with any language area. Listening order significantly affected ratings on 12 of 22 traits. The results indicate two distinct groups of traits. One group is the traits that the trait judges somehow associated with language proficiency. In evaluating these traits, the judges depended on cues of language proficiency. The other group is the traits that apparently the trait judges evaluated independently of spoken language proficiency.