Robin van Rijthoven

79 Role of semantics in a phonics through spelling intervention 4 Introduction The spelling of phonologically transparent words is rather straightforward. Starting from the phonemic segmentation of a word, the speller applies phoneme-grapheme correspondence rules to arrive at the correct spelling (Vanderswalmen et al., 2010). For words that are phonologically less transparent, the speller may use a strategy to search for analogy with similar spellings of known words, building on previously acquired phonological andorthographic representations (Allen, 1992). Research shows that children with dyslexia have severe problems in building up these representations (Lyon et al., 2003), which can be explained by their phonological deficit (Conrad, 2008; Göbel & Snowling, 2010). Consequently, children with dyslexia make more phonological errors in reading as well as in spelling compared to typically developing children (e.g., Bourassa et al., 2006). In addition, they also make more orthographic and morphological spelling errors. This can be explained from the fact that orthographic and morphological spelling builds upon the phonological base (Nunes et al., 1997). An overall delay in spelling development in children with dyslexia showing a phonological deficit that persists into adolescence can thus be expected (see, e.g., Bourassa & Treiman, 2003). Given that variation in semantic abilities has previously been shown to impact reading (Nation & Snowling, 2004; Torppa et al., 2010; Van Bergen et al., 2014) and spelling ability (Ouellette, 2010; Tainturier & Rapp, 2001), it might well be the case that children with dyslexia may compensate their low spelling outcomes through semantic abilities. However, semantic abilities has not yet been included as a predictor in responsiveness to intervention studies. Therefore, in the present study, the impact of semantic abilities on a response to phonics through spelling intervention in children with dyslexia was examined. Comparison of learning to spell for children with and without dyslexia In alphabetic writing systems, learning to spell starts with the segmentation of spoken words into singular sounds, followed by phoneme-to-grapheme mapping and word synthesis (Vanderswalmen et al., 2010). This so-called phonetic strategy helps to build up bi-directional connections between phonological and orthographic representations (Bosman & van Orden, 1997). Strong and bi-directional connections could facilitate a selfteaching mechanism (Share, 1995) to foster spelling development (Burt & Tate, 2002). Learning to spell, however, requires more than one-to-one phoneme-grapheme mapping. After mastering this phonetic strategy, children encounter words that are orthographically more complex. In this phase, word spelling inconsistencies need to be learned in order to become a skilled speller (Ehri, 2000; Treiman, 2018). Specific knowledge of morphological and orthographic patterns is then connected to already existing semantic representations, so that analogies of common words can be used in the spelling of uncommon words (Allen, 1992; Treiman et al., 1994). The accumulating