Number Magnitude Orients Attention, but Not against One's Will

Article excerpt

Recent evidence has shown that uninformative numbers can trigger attention shifts congruent with the spatial representation of number magnitude (Fischer, Castel, Dodd, & Pratt, 2003). In the present study, three spatial-cuing experiments whose aim was to qualify the automaticity of this number-mediated orienting are described. Experiment 1 replicated the phenomenon, showing that uninformative numbers can evoke orienting in a simple detection task. In Experiment 2, target location was random, but the participants were encouraged to shift attention to the left in response to large numbers and to the right in response to small numbers. No evidence for strong automaticity was observed, since the participants' performance was better when left-side targets were preceded by large numbers than when they were preceded by small numbers and vice versa. Experiment 3 corroborated this pattern by comparing gaze- and number-mediated cuing under conditions of real counterpredictiveness. The results indicate that unlike gaze-driven orienting, number-mediated orienting is not obligatory.

Visuospatial attention orienting has been investigated by presenting cues such as central informative arrows or peripheral uninformative abrupt onsets (e.g., Jonides, 1981). Through these signals, two distinct modes of attentional control have been identified: an exogenous and an endogenous mode. In recent years, the dichotomy between goal-directed orienting mediated via central informative symbolic cues and stimulus-driven orienting mediated by peripheral abrupt onsets has been challenged by studies highlighting that (1) a number of cues with high biological relevance, such as eye gaze and body shadows, generate attention shifts whose features do not conform to the dichotomy (e.g., Friesen & Kingstone, 1998; Galfano & Pavani, 2005), and (2) centrally presented symbolic cues can orient attention even when they are uninformative (e.g., Hommel, Pratt, Colzato, & Godijn, 2001).

Fischer, Castel, Dodd, and Pratt (2003) tested whether similar effects can be obtained using numbers as (uninformative) cues. This idea was inspired by the notion of a mental number line (MNL) representing numbers from left to right (Restle, 1970). Fischer et al. presented the numbers 1,2,8, and 9 at fixation. After different stimulus onset asynchronies (SOAs), a target requiring a simple detection response appeared in one of two peripheral boxes. The target had the same probability to appear in the left or the right box, irrespective of number magnitude, and the participants knew that the numbers were task irrelevant. If mere observation of numbers activated their position on the MNL, one might expect either a left or a right attentional shift, depending on number magnitude. The results showed that the participants were, indeed, faster when targets appearing on the left box were preceded by small numbers and when targets appearing on the right box were preceded by large numbers. Because numbers were uncorrelated with target location, Fischer et al. concluded that merely viewing numbers evokes automatic magnitude-dependent attentional shifts.

The present study assessed the degree of automaticity of number-mediated attentional shifts. According to the intentionality criterion (Jonides, 1981), orienting of attention is automatic if it resists suppression and occurs regardless of participants' expectations. In Fischer et al.'s (2003) study, numbers were task irrelevant; thus, it seems reasonable to conclude that they evoked an involuntary shift of attention, because the participants had no incentive for it to happen. However, to conclude that numbers elicited an obligatory shift of attention, it is necessary to show that this occurred even when participants intended for it not to happen.

In Experiment 1, we established a baseline condition by replicating Fischer et al.'s (2003) findings. Experiments 2 and 3 tested the criterion of intentionality for numbermediated attention shifts by manipulating expectancies. …