Robert, C. 1 , Zagar, D. 2 & Mathey, S. 1
1 Université Victor Segalen Bordeaux 2
2 Université de Bourgogne
Readers can generally recognize that a sequence of letters is a real word and have intuitions about how typical it looks as a word. In fact, previous research has reported that pseudoword processing was influenced by orthographic neighbourhood and by the frequency with which a letter sequence occurs in one given language. Other studies have found that wordlikeness ratings of pseudowords varied as a function of lexical neighbourhood and bigram frequency. Even though there is evidence that the similarity with words influences pseudoword processing, it is yet unclear what makes a sequence of letters more or less typical, more or less wordlike. The aim of the present study was to disentangle the role of wordlikeness judgments from various objective lexical variables (e.g., orthographic neighbourhood, bigram frequency, syllable frequency) on pseudoword lexical-decisions. Twenty-two participants performed a lexical decision task on 200 pseudowords. Then, they were asked to rate on a seven-point scale the degree to which each pseudoword looks like a real word. Results indicated that wordlikeness judgment was the strongest significant predictor of lexical decision latencies and errors, which suggests that a subjective variable of wordlikeness is a good predictor of pseudoword processing in the lexical decision task.