Is word recognition automatic? A cognitive-anatomical approach

What are the implications of anatomical localization of component mental operations for cognitive models? In this paper we use the anatomical localizations of visual and auditory word processing that were previously reported from PET studies (Petersen, Fox, Posner, Mintun & Raichle, 1988. We hypothesize that two operations performed simultaneously by the same or heavily interconnected anatomical areas will produce specific interference. One task is repeating back (shadowing) auditory words as quickly as possible. This task is shown to interfere with shifts of visual attention in the direction of peripheral cues. Both tasks are known to require common attentional operations localized to the medial frontal lobe. The shadowing task does not interfere with operations involving priming of a visual word form. This kind of priming involves areas of the ventral occipital lobe not used during shadowing. Finally, both shadowing and semantic priming involve anterior semantic and intentional areas. Accordingly, they can interfere. The conditions under which they produce interference suggest that the interference involves operations performed by the anterior attention system. These experiments support the idea that words automatically activate visual word forms, but involve shared attentional systems for higher level processes.