Poster D27, Friday, August 17, 4:45 – 6:30 pm, Room 2000AB

Can morphological structure compensate for missing phonological information during reading? Evidence from skilled and un-skilled readers

Tali Bitan1,2, Yael Weiss3, Tammar Truzman4, Laurice Haddad4, Bechor Barouch1, Tami Katzir5;1Psychology Department, IIPDM, University of Haifa, Israel, 2Department of Speech Pathology, University of Toronto, Canada, 3Psychology Department University of Texas at Austin, US, 4Department of Communication Sciences and Disorders, University of Haifa, Israel, 5Department of Learning Disabilities, The E.J. Safra Brain Research center, University of Haifa, Israel

Brain plasticity implies that the reading system could be tailored to demands of specific languages and orthographies. Theoretical models suggest that readers can rely on the word’s morphological structure when reading a phonologically opaque orthography, and that morphological segmentation develops in late stages of reading acquisition. We tested these hypotheses in Hebrew, a morphologically rich language with two levels of orthographic transparency, and in individuals with varying levels of reading skill. In Hebrew, most words are composed of a consonant root interleaved with a vowel template. It has two versions of script: the common un-pointed script is opaque, with partial representation of vowels; and the transparent, pointed script, which is mainly used in first and second grade children, fully represents all vowels. In a series of behavioral and fMRI experiments we tested adult typical and dyslexic readers as well as 2nd and 5th grade children. Participants read aloud 96 noun words in 2 levels of morphological complexity: 48 mono-morphemic words and 48 bi-morphemic words composed of a root and template. Words were presented with or without points. For children in both age groups, bi-morphemic words were identified more accurately than mono-morphemic words. In contrast to the predictions, this was only true for the transparent, script, while the reverse (higher accuracy for mono-morphemic words) was found for the un-pointed script, especially in 2nd grade. Namely, the morphological structure facilitated word recognition, but when phonological information was scarce the shared root increased competition and resulted in interference. Finally, correlations with a standardized measure of morphological awareness were found only for 2nd grade children, and only in bi-morphemic words. For adults, only dyslexic readers benefitted from the morphological structure showing faster responses to bi-morphemic compared to mono-morphemic words. Dyslexic reader also read unpointed words faster than pointed words, presumably because pointed words were less familiar, and they could not benefit from the high phonological transparency. Neuroimaging results in adults show that 1) in regions associated with orthographic processing (i.e. Occipito-temporal ctx.) only dyslexic readers showed overactivation for bi-morphemic compared to mono-morphemic words, when the words were presented with diacritics. Thus, when reading the less familiar pointed script dyslexic readers rely on the words’ morphological structure to access their orthographic representation. 2) Both typical and dyslexic adult readers showed increased activation for mono-morphemic compared to bi-morphemic words in left inferior & middle frontal gyri associated with phonological and morphological decomposition, across pointed and unpointed words. These results suggest that adult Hebrew readers engage in morphological decomposition of the root and template during the lexical-phonological access. Altogether these results suggest that readers of a morphologically rich language develop morphological segmentation skills already at an early age. Thus, unskilled readers, such as young children and adult dyslexic readers can rely on morphological segmentation during word recognition, to compensate for their immature orthographic lexicon or for impaired phonological abilities. However, we didn’t find an increase in morphological segmentation in the non-transparent script, suggesting that morphological and phonological segmentation enhance each other in the process of word recognition.

Topic Area: Grammar: Morphology

Back