Are the early stages of orthographic processing universal? Insights from masked priming with Semitic words

Two views contend to account for the processes at play during the early stages of visual word recognition. The first holds that these stages are not modulated by the idiosyncratic properties of different languages. The second maintains that the structural properties of the language determine the weighting of the different domains of linguistic knowledge (e.g., orthographic and morphological domains may be differentially weighted across languages). To explore this question, we focused on orthographic priming in Arabic. In this Semitic language, lexical representations are claimed to be based on morphological similarity, with little or no role for orthographic similarity. We conducted two masked priming experiments using the yes-no and go/no-go versions of the lexical decision task to determine if Arabic target words (e.g., (sic) 'mudyr', director) are facilitated by nonword primes that are orthographically but not morphologically related (i.e., pairs share neither a root nor a word pattern; e.g., (sic) 'maAyr') relative to unrelated primes. Results showed faster responses for the orthographically related target words than for the unrelated target words in the two experiments. These findings favor the view that the early phases of visual word processing in Semitic and Indo-European languages are fundamentally the same.