Visual identification of complex words: A masked priming study with Italian children

There is general agreement that, early in processing, morphological segmentation proceeds irrespectively of semantic transparency, so that DEALER is broken down into DEAL- and -ER just as CORNER is broken down into CORN- and -ER. It is much less clear, however, when this mechanism emerges in developing readers, with previous studies showing: (i) morpho-orthographic decomposition since 3rd grade (Quémart et al., 2011); (ii) no morpho-orthographic decomposition up to 5th grade (Beyersmann et al., 2012); (iii) morpho-orthographic decomposition maturing between 4th and 7th grade (Schiff et al., 2012). 128 typically-developing, Italian 3rd, 4th, and 5th graders participated in a masked priming, lexical decision task. 120 prime-target pairs were selected, 40 for each of the following conditions: a) transparent (farinoso-FARINA, mealy-MEAL); b) opaque (violenza-VIOLA, violence-VIOLET); c) orthographic (costume-COSTO, costume-COST). Each pair was contrasted with an unrelated baseline (timoroso-FARINA, fearful-MEAL). In a mixed-effects model carried out on RTs, CONDITION was found to interact with RELATEDNESS: morphological priming only emerged in the transparent condition (related: M = 1300 ms, SD = 444.4; unrelated: M = 1379, SD = 442.3), whereas no sign of facilitation was found in either the opaque (related: M = 1391, SD = 446.0; unrelated: 1379, SD = 412.1) or the orthographic condition (related: M = 1370, SD = 430.9; unrelated: M = 1399, SD = 431.4). Data are consistent with the hypothesis that in the course of reading acquisition form-meaning mapping is crucial to detect morphemic units. Accordingly, there is no evidence for morpho-orthographic parsing in primary school, not even in a language, i.e., Italian, characterised by a quasi-perfect one-to-one orthography-to-phonology mapping.