This paper reports two sentence-completion experiments exploring how nonnative speakers employ universal semantic and discourse information, which are implicit causality and consequentiality biases associated with psychological verbs, to solve pronouns. The results point out that intermediate-advanced and advanced Chinese-speaking English learners display weaker implicit causality and consequentiality biases than native English speakers in pronoun resolution. Instead, nonnative speakers show a general subject or first-mention bias. These findings propose that nonnative speakers do not employ semantic and discourse information in understanding as effectively as native speakers.