The dependence of root extraction in a non-concatenated morphology on the word-specific orthographic context

In Hebrew, as in other Semitic languages, most words are formed in a non-concatenated way, with a root morpheme embedded in a word-pattern morpheme consisting of only vowels or vowels plus consonants. Previous research on visual word recognition in Hebrew has revealed a robust morphological root-priming effect, with word recognition facilitated by the prior sub-perceptual presentation of the root morpheme, along with a less stable and more fragile word-pattern priming effect, particularly in the nominal system. These findings support the theory that morphological principles govern lexical access, with the root morpheme as a main organizational unit of the mental lexicon. However, less research has been done to delineate the algorithm underlying decomposition. The current study explores the importance of the natural lexical orthographic context of a complex root + pattern word structure for root extraction, using on-line measures based on tracking eye-movements in sentence reading. A series of 4 experiments using a fast-priming paradigm demonstrated that detaching the root morpheme from its lexical orthographic structure hinders the root-priming effect. Presenting the root in a non-word or a pseudo-word, that is, a non-existent combination of a real root + a real pattern did not make any difference. These results suggest that mapping the orthographic root onto its morphological mental representation depends on the orthographic context in which its letters appear. This finding constrains the role of the root in visual word-recognition, highlighting the crucial conditions for extracting it in the natural setting of reading.