Verbal (word) and nonverbal (design) paired-associate tasks were administered to Huntington's disease (HD) patients and healthy control subjects. An AB-AC paradigm, in which the cue stimuli were paired with different responses on the learning (e.g., BED-REST) and test trials (e.g., BED-SHEET), was used. It was hypothesized that HD patients would continue to respond with AB associations on the AC trials. The results were contrary to expectations: Patients showed impaired learning of both verbal and nonverbal associations but did not display a perseverative response style, even when the associative strength between word pairs was manipulated to elicit perseverations. Patients made more nonperseverative than perseverative errors in all conditions, an error pattern similar to that of control subjects. HD patients did not demonstrate increased susceptibility to proactive interference on these associative learning tasks.