Priming and interference effects can be dissociated in the Stroop task: New evidence in favor of the automaticity of word recognition

Recently, Derek Besner and his colleagues (Bauer & Besner, 1997; Besner & Stolz, 1999a, 1999b; Besner, Stolz, & Boutilier, 1997; Stolz & Besner, 1999) have reported some experiments that apparently challenge the “automatic” character of word recognition in the Stroop task. The basic underlying idea in their studies is that if the Stroop effect is reduced or even eliminated under certain experimental conditions, this outcome would reasonably question the automatic nature of word processing, because one of the critical features of this kind of processing is that it cannot be prevented even when it is not relevant to the goal task. In their experiments, Besner et al. (1997) nicely show a reduction and even a total elimination of the Stroop effect when color was applied to a single letter instead of to all letters of a word, as is usually the case in the standard version of the task. The authors extended this result when different versions of the Stroop task were used (e.g., Besner & Stolz, 1999a, 1999b). They account for these results by assuming that subjects used domain-specific processing algorithms. Their argument is that if subjects apply a determined algorithm (mental set in their terminology) to process the target (e.g., semantic processing), this algorithm is also used to process the distractor, because it is not possible to activate two different algorithms simultaneously. Now, if both target and distractor belong to the same domain (words), Stroop interference is expected. But if subjects are told to respond to the color or a single letter, it is the mental set (letter processing) that is activated and therefore no semantic processing of the irrelevant word is accomplished, eliminating any Stroop-interference effect. Elimination of the Stroop effect has also been observed when other variables have been manipulated in Stroop-like tasks. For instance, Fuentes and Ortells (1993) showed that incongruent color words displayed in black and presented in close proximity to a colored central patch produced interference effects, as compared with two baseline conditions (noncolor words and no distractor at all). However, incongruent color words presented far from the central patch did not produce any effect at all. These results are in accord with previous studies that have shown that interference reduces or even disappears as target and distractors become more physically distinct either by increasing eccentricity (e.g., Eriksen & Eriksen, 1974; Gatti & Egeth, 1978) or by presenting them in different colors (e.g., Francolini & Egeth, 1980). At first sight, these results do not contradict at all the Besner et al. (1997) idea that word recognition is not automatic; we could simply assume that