Open-bigrams as orthographic processing units in Arabic: Evidence from the flanking-letters lexical-decision task

Abstract

Using the “flanking-letters lexical decision” task, Dare and Shillcock The Quarterly Journal of Experimental Psychology, 66, 487–504, (2013) and Grainger et al. Acta Psychologica, 146, 35-40, (2014) demonstrated that word is facilitated when the flanking bigrams are present in the target word (e.g., RO ROCK CK), regardless of their position (e.g., CK ROCK RO), compared to different flanking bigrams (e.g., DA ROCK SH). This finding aligns with the Open Bigram Model proposed by Grainger and Van Heuven, (2004), which posits that orthographic representations in the Latin script are encoded by an unordered set of ordered letter bigrams. Employing the same task and experimental design, we replicated this key finding in Arabic. We observed a facilitative bigram-relatedness effect in both the repeated and the switched conditions. These results suggest that bigram coding reflects a universal orthographic mechanism, with letter bigrams functioning as representational units in Arabic, similar to their role in Latin scripts. Our findings also suggest that letter-position coding in Arabic may be more flexible than previously thought for Semitic scripts. We evaluate these conclusions within the framework of the Open Bigram Model and contrast them with the PONG model, which assumes absolute position coding.