Can phonological awareness predict concurrent reading outcomes in a deep orthography?

Phonological awareness can predict reading skills in typical readers (Bradly & Bryant, 1983; Stahl & Murray, 1994) and can distinguish between typical reading and reading deficit in alphabetic languages (Snowling, 1981; Stanovich & Siegel, 1994; Bryant et al., 1990). Yet the nature of phonological awareness and the causal link between phonological awareness and reading skill are subject to debate (Harm & Seidenberg, 1999; Castles & Coltheart, 2004; Blomert & Willems, 2010). Phonological awareness is often defined as sensitivity to the phonological structure of language and the ability to segment, isolate and manipulate the sounds of a specific language. We report the results of a study whose aim was to determine the sensitivity of phonological awareness tests in distinguishing between typical readers and deficit readers in Urdu. Urdu has a deep orthography, which however, presents readers with challenges different from the type offered by a language, such as English (Farukh & Vulchanova, 2014). The tasks included in the battery were typical phonological awareness tasks, such as phoneme manipulation, rhyme oddity, rhyming, and syllabification. Our results show that none of the phonological awareness tasks successfully predict concurrent poor reading skills. Most notably, one task, syllabification, predicts reading accuracy in the Control group only, but not in the Reading Deficit group. We discuss these results in light of the grain-size hypothesis and the orthographic depth hypothesis of reading, and from the point of view of stages in literacy acquisition.