A pattern-masked arrow negatively biased the "free choice" between 2 manual responses or between 2 vocal responses. This apparently nonconscious influence occurred only when the free-choice trials were intermixed randomly with other trials that terminated in fully visible arrows, which directed a response of the same modality (manual vs. vocal) as that involved in the free-choice test trials. This indicates that recent conscious processing of the association between specific stimuli and specific responses is needed to activate the nonconscious influence of masked arrows on response selection. Because this influence occurred only when a concrete association was activated, it appears not to be based on deep comprehension of the stimuli and instead is attributable to simple stimulus-response bonds.