On the role of graphemes, syllables and morphemes in reading: evidence from different tasks and languages

<p dir="ltr">The research presented in this thesis centres on the crucial question of how individuals recognise written words when reading, and what role do various linguistic units play in this process. Specifically, this thesis investigates whether readers primarily access whole words or rely on smaller sublexical units, like graphemes, syllables, and morphemes. These units, otherwise called reading units, have been found to have a facilitatory role in retrieving meaning and pronunciation (Carreiras & Grainger, 2004). But first and foremost, why are readers relying on larger units than letters during reading? By undertaking cross-linguistic research and investigating this issue using a variety of experimental procedures (pseudoword naming, eyetracking, and lexical decision) this thesis contends that three main aspects, which are a source of linguistic variety, contribute to cross-linguistic differences in sublexical processing: orthographic depth, morphological complexity, and syllabic complexity. In the study presented in the first experimental chapter (Chapter 2), I explore how orthographic depth and the knowledge of grapheme-phoneme correspondences influence pseudoword reading by introducing the use of Entropy as a mean to calculate pronunciation variability. The study investigates four European languages (English, German, French, Italian) and examines different age groups (adults, children in grades 2, 3, and 4) as well as linguistic backgrounds (monolingual and bilingual children). The results reveal that English, with its deep orthography, exhibits greater pronunciation variability, that orthographic depth is a multidimensional construct and that cross-linguistic differences in pseudoword pronunciation only arise when the material employed is representative of the language under investigation. In the second experimental chapter (Chapter 3), currently under review, delves into morphological processing in greater depth, and examines how it is affected by morphological complexity and orthographic depth. This was achieved by contrasting two languages that differ on both aspects: English and Italian. The findings underscore that orthographic depth significantly influences morphological processing, more so than morphological complexity. In the third experimental chapter (Chapter 4), the thesis will revolve on the interplay between syllabic and morphological processing when reading multi-syllabic, multi-morphemic words in German. Using eye-tracking, this study finds that skilled native German speakers primarily employ syllables as their preferred units of analysis during silent reading for comprehension. In summary, the findings of these three experimental chapters show that orthographic depth is the main cause of cross-linguistic differences in reading behaviour. Reliance on units larger than letters in reading aloud and silent reading is mostly driven by the specific orthography demands thus providing further evidence for the Orthographic Depth Hypothesis (Katz & Frost, 1992) and the Psycholinguistic Grain Size Theory (Ziegler & Goswami, 2005). Results are also discussed in terms of the Flexible-unit-size Hypothesis of Brown and Deavers (1999), and implications for theoretical and computational modelling are considered.</p>