ABSTRACT
This chapter examines the nature and construction of such visual descriptions. The visual description used for word recognition could comprise holistic information about word shape or information derived in parallel concerning the identities of the letters in the words. This model assumes the existence of three levels of representation in visual word recognition: one specifying the visual features present in each letter; one specifying the letter at each position in the word; and one specifying the word. The notion that orthographic priming reflects some form of competitive interaction between primes and targets fits with recent interactive activation accounts of visual word recognition. The neutral baseline enables us to assess whether orthographic priming is facilitatory or inhibitory in nature, and it provides a within experiment control to test whether any intrusion errors are greater than chance. The data are consistent with a left-to-right identification process, and show no indication of a benefit for the report of the end-letters in the strings.
