Beck's influential cognitive account of anxiety has led to the prediction that individuals vulnerable to anxiety should favor threatening interpretations of ambiguity (e.g., Beck & Clark, 1988; Beck, Emery, & Greenberg, 1986). The current study introduces a novel adaptation of the RSVP technique, previously used in text comprehension research, to evaluate this hypothesis. Results suggest that a group of 24 high trait anxious students did indeed selectively impose threatening interpretations on unconstrained ambiguous sentences. In contrast, a matched group of 24 low trait anxious students appeared to selectively impose non-threatening interpretations on such ambiguous sentences. These findings are fully consistent with the predicted anxiety-linked interpretative bias. Specific testable hypotheses are developed concerning the types of interpretative idiosyncrasies that plausibly may contribute to pathological anxiety.