The activation of multiletter units in visual word recognition

Abstract

Influential accounts of visual word recognition argue that lexical access is entirely mediated via preliminary letter identification processcs(McClelland and Rumelhart, 1981; Seidenberg, 1987). Others advocate a role for visual features characterising the whole word (Haber and Haber, 1981). A third view holds that lexical access is, at least in part, mediated by sublexical units which are recovered by the action of a parsing mechanism operating on identified letter codes through the application of specific rules (e.g., Spoehr and Smith, 1973). Recent work by Alice Healy and her colleagues supports a multi-level view of word recognition incorporating representations not only at the letter and whole word levels, but also at an intermediate level where multiletter units can be activated directly by supraletter features (e.g., Healy and Drewnowski, 1983). The concept of an intermediate level of representation in visual word recognition remains controversial. A priming paradigm was employed to investigate the existence of directly activated multiletter units. Subjects were required to make a discrimination response to test stimuli which could be either targets or foils.
Targets were either single-letters or consonant-bigrarns which were present or absent in an immediately preceding word, and foils were either single keyboard characters or a character plus a letter. Experiment I verified an earlier finding that responses to consonant-bigram targets are facilitated when these appear in a prime word, while responses to the constituent letters of those bigrams are not facilitated (Greenberg and Vellutino, 1988). In addition, responses to primed bigrain targets were faster than responses to primed single letter targets.
Experiments 2 to 4 revealed that the bigram priming effect occurs only when both primes and targets appear in lower-case type. These observations provide evidence for the existence of directly activated multiletter units. Subsequent experiments supplied converging evidence for the importance of supraletter visual features in the activation of multiletter units and strengthened the view that the multiletter unit effects observed arise at a pre-lexical stage of processing when words are being processed to a level of meaning. Implications for theories of reading are discussed, and related developmental issues are considered.