Top-down contingencies of nonconscious priming revealed by dual-task interference.

According to the direct parameter specification (DPS) account, reaction time effects of invisible primes depend on top-down control settings directed to targets (Neumann & Klotz, 1994). If this hypothesis holds, effects of invisible primes should decrease in dual-task as compared with single-task conditions: Prior to the primes control settings for the alternative task should be activated, which renders a match between target-directed control settings and primes less likely. In Experiments 1-2 and 4-5, a second task interfered with the validity effect of invisible primes. Control conditions ruled out several alternative explanations. Interference was not due to higher spatial memory loads (Experiments 1 vs. 2), increased numbers of stimuli or responses (Experiments 3-5), or increased response latencies (Experiment 3). If predictable, alternative tasks did not interfere (Experiment 3). The results are in line with the DPS account and less so with some classical definitions of automatic processing (e.g., Posner & Snyder, 1975).