Abstract
The definition of orthographic neighbour (Coltheart et al., 1977) was analysed in two experiments using a variety of the three-field technique (Humphreys et al., 1988). With this technique, a clearly visible prime (in lower-case letters) is followed by a briefly presented upper-case word which is immediately masked. Pairs of five-letter neighbouring words were selected. Only orthographically related pairs that differed from the prime by the third letter (womenWOVEN) or the fourth letter (frost-FRONT) showed (inhibitory) relatedness effects compared with an unrelated word condition. The results suggest that models of visual word recognition should be modified to address the fact that some letter positions are more important than others.

This publication has 22 references indexed in Scilit: