96 Whitney, C. (2001). How the brain encodes the order of letters in a printed word: The SERIOL model and selective literature review. Psychonomic Bulletin and Review, 8, 221–243. doi: 10.3758/bf03196158.
97 Yarkoni, T., Balota, D., & Yap, M. (2008). Moving beyond Coltheart’s N: A new measure of orthographic similarity. Psychonomic Bulletin & Review, 15, 971–979. doi: 10.3758/PBR.15.5.971.
98 Zorzi, M., Barbiero, C., Facoetti, A., Lonciari, I., Carrozzi, M., Montico, M., … & Ziegler, J. C. (2012). Extra‐large letter spacing improves reading in dyslexia. Proceedings of the National Academy of Sciences, 109, 11455–1459. doi: 10.1073/pnas.1205566109.
Notes
1 1 The reason for this focus is that the vast majority of skilled reading involves silent reading for meaning, while reading aloud is clearly a key step in the process of learning to read (Castles & Nation, this volume).
2 2 Figure 3.4 illustrates an extension of the Grainger and van Heuven model to the case of orthographic processing spanning multiple letter strings, and therefore provides the link with the section on multiple word processing.
3 3 In further extensions of this model, additional flexibility has been achieved via a set of position‐independent letter identities (a bag‐of‐letters; see Grainger et al., 2014, for a discussion of one possible implementation).
CHAPTER FOUR Word Recognition II : Phonological Coding in Reading
Marc Brysbaert
In his classic book on the psychology of reading, Huey (1908) was motivated by the question of whether silent reading was possible without inner pronunciation. This was a natural question for him to ask, given that historical accounts tell us that silent reading was rare before the late nineteenth century and, instead, people tended to read aloud, or mumble when they wanted to spare their neighbors (Manguel, 1996; Pearson & Goodin, 2010). Even nowadays, young children require months of practice before they can read silently (Wright et al., 2004) and most people report experiencing some kind of inner speech as they read (Alderson‐Day, & Fernyhough, 2015; Vilhauer, 2016).
Bakhtin (1981) argued that the inner voice need not be our own voice. When we know the writer well, we sometimes hear their voice as we read silently. Zhou and Christianson (2016) reported that text comprehension may even improve if we imagine someone else reading the text aloud. There is evidence for stable individual differences in the degree to which people experience inner speech (Roebuck & Lupyan, 2020; Vilhauer, 2017). For some individuals, it feels a powerful phenomenon, whereas others are surprised when told that some experience an inner voice in silent reading.
This chapter considers key questions and issues concerning the role of phonology in silent reading. I first review classic findings that point to phonological effects in reading. The nature of these phonological effects are discussed, and their implications for computational and neural models of word reading are considered. It will become clear that visual language processing cannot be understood properly without taking word phonology into account.
Evidence that Phonology Is Involved in Silent Reading
Our intuitive feeling of inner speech is not the only evidence pointing to a role of phonology in silent reading. Table 4.1 includes five other key findings.
Although there is good evidence that connections between orthography (the written code) and meaning are at the core of skilled word reading (see Grainger, Rastle, this volume), the findings in Table 4.1 make clear that phonology is involved too.
Phonology refers to the spoken forms of words. Phonemes are the smallest units of sound that distinguish between one word and another. The sounds /r/ and /l/ are different phonemes in English, because the spoken word /lap/ does not have the same meaning as /rap/. In alphabetic languages, phonemes are represented in written form by letters or combinations of letters, called graphemes. For instance, in English the grapheme ph refers to the phoneme /f/.
It is important to distinguish between phonology and articulation. Articulation refers to the muscle movements required for speech sounds to be made. It is possible to investigate the role of articulation in silent reading by asking people to read text while simultaneously repeating aloud “the the the.” This interferes with articulation, but does not interfere with silent reading (Besner, 1987). This finding tells us that the inner speech used in silent reading is not the same as the speech code for reading aloud. Hence the use of the term phonology refers to abstract sounds rather than articulation and the physical mechanism required to make speech sounds.
Table 4.1 Findings indicating that silent reading is affected by phonology
| Finding | Example | Reference |
|---|---|---|
| People need more time to read silently sentences that are difficult to pronounce (tongue‐twister effect) | It takes longer to decide that “Twenty toys were in the trunk” is a meaningful sentence than “Several games were in the chest.” | McCutchen & Perfetti (1982) |
| Letter detection is easier in stressed syllables than in unstressed syllables | Readers are more likely to notice the letter e in introspect than in apartment | Drewnowski & Healy (1982) |
| Homophone errors in semantic decision | People more often misclassify the written word rows as a flower than robs | van Orden (1987) |
| Phonological errors in proofreading | Proofreaders more often miss the error in “they would meat for” than in “they would moat for” | Daneman & Stainton (1991) |
| Parafoveal homophone priming in reading | People read rains more quickly following parafoveal preview of reins than ruins | Pollatsek et al. (1992) |
Reliance on phonology instead of articulation is further in line with the observation that silent reading is faster than reading aloud. It operates at 240–260 words per minute, depending on the length of the words (Brysbaert, 2019), whereas oral reading happens at the slower rate of 175–195 words per minute. At the same time, the speed of silent reading is not too fast for phonology to be involved, as listening comprehension remains good up to 270–290 words per minute (Carver, 1982; Kuperman et al., 2021).
Reading without phonology?
To further assess the importance of phonology for silent reading, we can look at groups of individuals who have suboptimal access to phonological information, such as people