B.1.6.1. Relational vector vs feature-based accounts

Becker (2008, 2010) proposed a "relational vector account" of attentional guidance, opposed to "feature-based" selection hypotheses. She argued, on the basis of interesting experiments, that attention could be oriented according to relational properties. That is, a relational template, qualifying the target's position relative to distractors in the featural space, would guide attention. So, attention might be oriented toward "larger items" or towards "smaller items", without determining a specific target feature. Tentatively, this author even claimed that: "In sum, the relational account holds promise for synthesizing a wide range of results that were previously attributed to a feature-based or a saliency-based selection mechanism" (Becker, 2010, p. 261). The adequacy of this account should be evaluated regarding both exogenous and endogenous influences in the present experiments. It must be noted that this account was adapted for priming and for attentional set, although there is no evidence that these two phenomena are based on common mechanisms.