The authors report a series of experiments in which they use the masked congruence priming paradigm to investigate the processing of masked primes in the manual and verbal response modalities. In the manual response modality, they found that masked incongruent primes produced interference relative to both congruent and neutral primes. This finding, which replicates the standard finding in the masked congruence priming literature, is presumed to reflect the conflict that arises between two incompatible responses and, thus, to index the extent of processing of the masked prime. Somewhat surprisingly, when participants were asked to respond verbally in the same task, masked incongruent primes no longer produced interference, but masked congruent primes produced facilitation. These findings are surprising because they suggest that the processing of nonconsciously perceived primes extends to the response level in the manual, but not verbal, response modality. The authors propose that the modulation of the masked congruence priming effect by response modality is due to verbal, but not manual, responses being mediated by the lexical-phonological production system. Keywords: masked priming, stimulus-response compatibility, response priming, visual awareness, response modality, lexical access in speech production How extensively are subliminally presented stimuli processed? In a seminal paper in this area of research, Dehaene et al. (1998, p. 599) claimed that their results resolved the issue over the extent of processing of subliminally presented (masked) primes. In their experimental paradigm, participants categorized a visible target stimulus (e.g., EIGHT) as being either larger or smaller than five. Unbeknownst to the participants, the visible targets were preceded by a masked prime stimulus (e.g., three) that fell either on the same side of five as the target (i.e., congruent prime) or on the opposite side of five (i.e., incongruent prime). Participants responded by pressing a "smaller than five" button with one hand and a "larger than five" button with the other hand. Despite participants being unable to detect the presence of the prime stimuli, the primes produced clear behavioral and neurophysiological effects. Behaviorally, Dehaene et al. found that response times in the incongruent condition were significantly slower than in the congruent condition. This behavioral effect, referred to as the masked congruence effect (MCE), has been observed in a wide variety of paradigms with nonorthographic stimuli, including meta-contrast masking of arrows (Vorberg, Mattler, Heinecke, Schmidt, & Schwarzbach, 2003) and spatial locations (Leuthold & Kopp, 1998;Neumann & Klotz, 1994) and pattern masking of arrows (Eimer & Schlaghecken, 1998). Subsequently, the MCE has also been replicated several times with orthographic stimuli (Kunde, Kiesel, & Hoffmann, 2003;Naccache & Dehaene, 2001;Van Opstal, Reynvoet, & Verguts, 2005). It was the neurophysiological findings, though, that led Dehaene et al. to claim to have "resolved the d...