comprehension, and knowledge systems, with a central role for the lexicon."/>
Figure 1.1 The Reading Systems Framework
(modified from Perfetti & Stafura, 2014)
consisting of word‐identification, comprehension, and knowledge systems, with a central role for the lexicon.
Reading and Reading Science in Historical Context
Humans have been reading for around 3,500 years. Or at least writing has been around for about that long, which is all we have to go on. Reading science is much younger. Although reports of patients with acquired reading disorders appeared earlier (Berlin, 1887; Kussmaul, 1878), Cattell’s (1886) experiments on the time it takes to read words and letter strings mark the beginning of experimental reading research. The broader research findings published by E. B. Huey (1908), who acknowledged contemporary research by Erdmann and Dodge, are the most substantial landmark for a beginning of reading science. Indeed, most of Huey’s observations in the Psychology and Pedagogy of Reading remain foundational for reading science: word perception, the “inner voice” in silent reading, meaning, and “interpretation,” the evolution of writing and the alphabetic principle. Notably omitted was dyslexia, a slight that was repaid by Orton (1925) when he ignored Huey’s book and its research in his classic work “Word Blindness.”
Much of the progress since has been enabled by tools that reveal the intricate and interleaved processes and knowledge interactions that occur rapidly in reading: Eye tracking, Event Related Potentials (ERPs) and chronometric behavioral measures can detect the processes that constitute the rapid stream of reading. The products of these processes – the slower stream of reading – are exposed by behavioral output measures, and by imaging tools that identify brain areas associated with these processes. Beyond laboratory tools, the development of computational modeling has added precision to theoretical accounts and large language corpora provide statistical tools for the modeling of reading processes.
Advance 1: The Word‐identification System in Skilled Alphabetic Reading
Visual processing and models of eye movements
We begin with the lower left portion of the Reading Systems Framework, the visual input that initiates the identification of a printed word. Pre‐dating modern‐day observations that the brain was not designed for reading (e.g., Dehaene, 2009), Huey (1908) pointed out that reading is “intensely artificial.” “The human eye and the human mind, the most delicate products of evolution, were evolved in adaptation to conditions quite other than those of reading” (p. 8).
The core visual constraint is that the acuity needed to identify a specific letter within a word is limited to one to two degrees of visual angle at normal viewing distance. Within this narrow window, only a single word or two (with the help of parafoveal viewing) can be identified during an eye fixation, although less precise visual information is available peripherally.
As detailed by Liversedge et al. (this volume), readers adapt to this limitation by making frequent eye‐fixations, directly fixating on between 60% and 80% of content words (Rayner et al., 2005). They also adjust their fixation rates (and the number of regressions) in response to text difficulty and reading goals, one of the key regulatory strategies in reading. Word fixations vary in duration, generally allowing three to five words to be fixated within a second of reading (Rayner et al., 2004, 2005). With assistance from word properties, context, and parafoveal viewing, a reader may approximate the reading rate implied by Gough’s (1972) one second of reading. The familiarity of a word, its predictability from context (Rayner et al., 2004), and the structure of the sentence (Clifton & Staub, 2011) all exert an effect on eye movement measures. Some measures reflect the more passive, automatized aspects of word identification (e.g., fixation durations), whereas others also reflect regulatory processes that help the reader make sense of the text (e.g., regressions). Together, eye‐tracking measures reflect how context and the linguistic properties of words affect how easily they are read and understood.
Skilled readers control their eye movements to accommodate the perceptual constraints on word identification while maintaining reading efficiency. How this is accomplished is the target of eye‐movement control models. Serial processing models assume that only a single word is in visual attention, for example, the EZ Reader model (Reichle et al., 1998, 2003). To accomplish rapid reading rates with serial processing, the brain must signal an eye movement before the word has been identified completely because the movement lags behind the brain’s launch signal. Thus, EZ Reader assumes a signal that word identification is imminent (not complete) is what prompts an eye movement. This signal comes earlier for a familiar word or one predictable from context. An alternative solution to perceptual constraints is to allow parallel processing on adjacent words (SWIFT model, Engbert et al., 2005). A more recent model allows for parallel processing of words and provides specific word identification mechanisms (Snell et al., 2018). The question of parallel versus serial processing of words remains a point of contention (see Grainger, and Liversedge et al., this volume).
Orthographic processing and models of word identification
The word‐identification system codes visual input as familiar orthographic units. The skilled reader has acquired an inventory of orthographic units – graphs, to use a neutral term – and connected them to language units (the word‐identification system in Figure 1.1)—allowing words to be identified.
From word superiority to interactive activation.
One of the most intriguing problems in reading science is how the reader’s knowledge of orthographic units is used in skilled reading (Grainger, this volume). The long‐standing answer is that readers come to recognize a word as a whole unit rather than a string of letters. J.M. Cattell’s famous experiments (1886; reviewed in Huey 1908) were intended to demonstrate this. After viewing a briefly exposed string of letters, Cattell attempted to report all the letters in the string. When the letters spelled a word, he could report more letters than when he viewed a random letter string.
In fact, Cattell’s experiments could not distinguish perception of the whole word from memory for some of its letters. Remembering enough letters would prompt retrieval of a word that contains them, making the report of the letter string a mix of perception, memory, and a bias to respond with words. Nevertheless, Cattell’s explanation (and Huey’s) stood unchallenged until the independent publications of experiments by Reicher (1969) and Wheeler (1970).
Reicher (1969) and Wheeler (1970) controlled for response bias by asking participants which of two letters had been briefly presented (and masked) in a particular position. For example, given the string lake, probing whether k or t had appeared in the third position would not produce a word bias, because either letter completes a word. The publication of these experiments stimulated a generation of research on the “word superiority effect,” eventually leading to a modified conclusion: Letters within nonword pronounceable strings (pseudowords) are also perceived better than random strings of letters. Letters in real words are perceived a little better than letters in these pseudowords, but the largest difference seems to concern the internal structure of the letter string, its word‐like orthography and phonology.
McClelland and Rumelhart (1981) explained both the word superiority effect and the pseudoword superiority effect in a new approach, a model that connected three hierarchical levels – words, letters, and letter segments – with bi‐directional activation between adjacent levels of the hierarchy. Activation
