Do fixation cues ensure fixation accuracy in split-fovea studies of word recognition?