Learned helplessness theory explains the impaired performance that follows exposure to uncontrollable outcomes by assuming learned expectation of response-outcome independence that is transferred between tasks. Recent evidence has shown that introducing a second neutral stimulus, contingent on the offset of the uncontrollable stimulus, removes the subsequent interference. This finding has been claimed to support the view that the interference is a result of conditioned inattention rather than of the expectation of response-outcome independence. These conflicting explanations were examined in a series of four experiments that varied induction procedures (passive exposure or inescapability) and stimulus quality (aversive or nonaversive). All four experiments found the predicted interference, but only one, in which passive exposure was combined with an aversive stimulus, obtained results supporting the conditioned inattention hypothesis. We conclude that learned helplessness probably involves more than a single mechanism and that the passive exposure procedure may not be appropriate for demonstrating genuine helplessness deficits.