Consonant bias in visual-word recognition: is it phonological in nature?