The Role of Variations in the Degree of Regularity During Novel Word Reading: Evidence from a Novel Word Training Study with Skilled Readers

It is well established that oral vocabulary contributes to the process of learning to read novel words, but questions remain about the nature of the cognitive mechanism(s) that support this association. When an individual attempts to read a new word for the first time, they are thought to apply their knowledge of phonics rules. Regular words, such as drab, follow phonics rules and can be fully decoded using letter-sound mappings. However, irregular words, such as swab, contain at least one letter-sound that is not the most common mapping. This means that irregular words can only be partially decoded; decoding the word swab would result in a pronunciation rhyming with the word scab. Therefore, to read irregular words correctly, an individual must be able to match their decoding attempt with the correct pronunciation of a word in oral vocabulary. One mechanism, termed mispronunciation correction, is thought to be particularly helpful when learning to read new words with irregular mappings between letters and sounds. According to this view, a reader attempts to read a word (e.g. busy) using the most common grapheme-to-phoneme correspondences (“bussy”), but when that word is not recognised orally, they are prompted to search their oral vocabulary for a similar sounding word that makes sense in the context, and then adjust their pronunciation accordingly (“bizzy”). Research interest on mispronunciation correction has accrued over recent years and suggests that the ability to correct mispronunciations might play an important role within reading acquisition. However, little work has directly addressed how this mechanism operates during the reading of orally known but visually novel words, and many questions remain.

Chapter 1 is a literature review that will begin by defining oral vocabulary and word regularity. Two theories will be introduced, namely the lexical quality hypothesis (Perfetti, 1992, 2007) and the self-teaching hypothesis (Share, 1995, 1999), and how those theories view the role of oral vocabulary in learning to read new words will be presented. The review will then outline evidence for the association between oral vocabulary and word reading. The mispronunciation correction mechanism will be described with reference to how it has been defined and operationalised. Prior work investigating mispronunciation correction will be reviewed and discussed separately for studies taking an individual differences approach, an intervention approach (aimed at teaching mispronunciation correction as a strategy), and a training study approach (aimed at understanding whether this mechanism plays a causal role in learning to read new words). A case will be made that we need to better understand: (a) whether v   mispronunciation correction operates across the lifespan as, for instance, skilled adult readers encounter novel written words that are orally familiar; (b) whether the extent to which a word is irregular influences the likelihood that readers will be able to arrive at the correct pronunciation.

Chapter 2 reports an empirical training study that asked: (1) when reading orally known but visually novel words for the first time, do skilled readers use mispronunciation correction to resolve mismatches between decoded and known pronunciations?; (2) does the probability of success differ for words with varying degrees of irregularity of vowel graphemes? A two-day training study was conducted, where participants learnt 18 novel nonwords on day one. On day two, participants completed a reading aloud task in which the trained words were presented in contextual sentences, along with another six untrained words. The regularity of trained words was manipulated in order to create three conditions; regular, somewhat irregular, and very irregular. Reading accuracy was the dependent variable. Results supported the operation of the mispronunciation correction mechanism within this sample of skilled readers, but no evidence was observed that the degree of irregularity (somewhat versus very irregular vowel pronunciations) influenced the probability of reading success. Implications of these findings and directions for future research are discussed.