Better to lose the anchor than the whole ship.

In a recent TICS article, Ahissar [1] proposed an interesting new theory for understanding dyslexia, the ‘perceptual anchor theory’. The idea is straightforward. Most psychophysical tasks used to investigate perceptual deficits in dyslexia repeat stimuli, such as tones in frequency discrimination or objects in rapid automatized naming (RAN). Theoretically, this massed repetition allows unimpaired participants to form a perceptual anchor. Indeed, Ahissar and colleagues [2] showed that dyslexics failed to form a perceptual anchor when the same reference stimulus (anchor) was repeated across trials or when a small set of stimuli was used repeatedly. These results were taken as evidence against the phonological deficit theory of dyslexia [3]. Here, I present three pieces of evidence that are inconsistent with the strong claims of the perceptual anchor theory. The first comes from RAN. According to Ahissar [1], perceptual anchoring accounts for RAN deficits because a small set of repeated items is used in RAN. Thus, the theory predicts that anchoring deficits should not exist in the first trials of RAN because anchoring needs time to build up. To assess this prediction, I re-analyzed the dyslexics’ naming performance