The role of sublexical graphemic processing in reading

Dual-route models of reading postulate the existence of two separate mechanisms: The lexical route allows words to be recognized in their holistic form, and the sublexical route proceeds by converting the written sublexical entities of a word or a nonword into their corresponding phonological equivalents. Sublexical reading is assumed to require three stages of processing: graphemic parsing, graphophonemic conversion, and phoneme blending. This study provides evidence in favor of the existence of a graphemic parsing process which occurs prior to grapheme-phoneme conversion. A group of normal subjects read nonwords which contained multiletter graphemes significantly more slowly than graphemically simple nonwords. These results, best interpretable in the context of a recent dual-route model of reading, confirm previous data obtained in pathology which suggest the functional independence of this cognitive procedure.     

Graphemic complexity and multiple print-to-sound associations in visual word recognition

It has recently been reported that words containing a multiletter grapheme are processed slower than are words composed of single-letter graphemes (Rastle & Coltheart, 1998; Rey, Jacobs, Schmidt-Weigand, & Ziegler, 1998). In the present study, using a perceptual identification task, we found in Experiment 1 that this graphemic complexity effect can be observed while controlling for multiple print-to-sound associations, indexed by regularity or consistency. In Experiment 2, we obtained cumulative effects of graphemic complexity and regularity. These effects were replicated in Experiment 3 in a naming task. Overall, these results indicate that graphemic complexity and multiple print-to-sound associations effects are independent and should be accounted for in different ways by models of written word processing.     

The role of nasals in reading: a normative study in French

Dual-route models of reading assume that reading can be done in two ways. A most common lexical route, on the one hand, allows regular and irregular words to be read while a second sublexical route allows nonwords and novel words to be read. A graphemic processing stage in sublexical reading is assumed to assemble the individual letters of a word or a nonword into multiletter graphemes prior to grapheme-phoneme conversion. The purpose of this study was to determine whether vowel/nasal clusters required as much time to be processed asvowel/vowel and consonant/consonant clusters in sublexical nonword reading in French. Results indicate that nonwords that contain vowel/nasal clusters are read significantly faster than nonwords comprising vowel/vowel and consonant/consonant clusters. Furthermore, nonwords that contain single-letter graphemes are read significantly faster than nonwords comprising vowel/nasal clusters and nonwords comprising vowel/vowel and consonant/consonant clusters. These results taken as a whole support the idea that nasals act as diacritic marks rather than being processed by means of a graphemic parsing procedure.     

An alternative to grapheme-phoneme conversion rules?

When orthographic factors were tightly controlled in a lexical decision task, it was observed that orthographic similarity rather than homophony with a word led to increased reaction times to nonwords. This result suggested that the pseudohomophone effect is not a phonological effect. Instead, a conversion of the graphemes of a stimulus item into different graphemes via a set of grapheme-grapheme conversion rules was supported. When phonological factors were tightly controlled and orthographic similarity varied, evidence for the existence of grapheme-grapheme rules was provided in both a lexical decision task and a task in which subjects were required to say whether an item was pronounced in the same way as a word. Even in the latter task, in which the likelihood of phonological recoding was optimized, it appeared that grapheme-phoneme rules were rarely, if ever, used.     

Neighbourhood density effects in reading aloud: new insights from simulations with the DRC model.

A series of nine simulations with the Dual Route Cascaded (DRC) model (Coltheart, Rastle, Perry, Langdon, & Ziegler, 2001) investigated neighbourhood density (N) effects in nonword and word naming. Two main finding emerged from this work. First, when naming nonwords there are two loci for the effect of N in the model, contrary to Coltheart et al.'s single locus explanation of what the model is doing. The early N effect involves interactive activation between the orthographic lexicon and the letter units such that high N facilitates letter identification, which in turn affects the nonlexical route. The late N effect arises from activation in the orthographic lexicon that feeds forward to the phonological lexicon and primes phonemes in the phoneme system. Second, when naming words the presence/absence of an effect of N on the Letter Units through feedback from the lexical level depends on the parameter settings. Implications and suggestions for future directions are made.     

Whammies and double whammies: The effect of length on nonword reading

In the work presented here, the length effect in nonword reading aloud was investigated in order to assess whether that effect is driven by the number of letters in a string or by the number of graphemes in a string. Simulation work with the Dual-Route Cascaded (DRC) model (e.g., Coltheart, Curtis, Atkins, & Haller, 1993; Coltheart & Rastle, 1994) uncovered a surprising finding regarding the length effect; the same result was obtained in an experiment with human subjects. The results are discussed in terms of the DRC model, with particular reference to serial processing and interphoneme inhibition, two properties critical to understanding the effect reported here.     

Tracking the transition from sublexical to lexical processing: On the creation of orthographic and phonological lexical representations

Participants read aloud nonword letter strings, one at a time, which varied in the number of letters. The standard result is observed in two experiments; the time to begin reading aloud increases as letter length increases. This result is standardly understood as reflecting the operation of a serial, left-to-right translation of graphemes into phonemes. The novel result is that the effect of letter length is statistically eliminated by a small number of repetitions. This elimination suggests that these nonwords are no longer always being read aloud via a serial left-to-right sublexical process. Instead, the data are taken as evidence that new orthographic and phonological lexical entries have been created for these nonwords and are now read at least sometimes by recourse to the lexical route. Experiment 2 replicates the interaction between nonword letter length and repetition observed in Experiment 1 and also demonstrates that this interaction is not seen when participants merely classify the string as appearing in upper or lower case. Implications for existing dual-route models of reading aloud and Share's self-teaching hypothesis are discussed.     

Reading without phonology?

Responses to items such as brane are slower and/or more error prone than responses to items such as slint in lexical decision (is this string spelt like a real word?). The received view is that this “pseudohomophone” effect is attributable to phonological receding. Taft (1982) has challenged this view, offering instead a grapheme-grapheme account which assumes that graphemes that map onto a common phoneme develop the ability to activate each other without reference to phonological mediation.

Taft's grapheme-grapheme account is tested in two experiments. Experiment 1 shows that the presentation of a pseudohomophone facilitates the response to a subsequently presented word (e.g., groce-gross). Experiment 2 shows that nonword letter strings that are translatable into words by the application of putative grapheme-grapheme rules (e.g., gloce-gloss) produce no facilitation. These results are consistent with the notion of a phonological influence but inconsistent with the grapheme-grapheme account. Loci for this pseudohomophone priming effect are discussed.     

A meta-analysis of executive functioning in dyslexia with consideration of the impact of comorbid ADHD

Grapheme coding was examined in French Grade 6 and Grade 8 children and adults who learned English as a second language (L2). In Experiments 1 and 2, three conditions were compared in a letter detection task in L2: (1) simple grapheme (i.e., detect “a” in black); (2) complex language-shared grapheme (i.e., “a” in brain) and (3) complex L2-specific grapheme (i.e., “a” in beach). The data indicated that graphemes in L2 words were functional sub-lexical orthographic units for these L2 learners. Moreover, L2-specific graphemes took longer to process than language-shared complex graphemes. Using the same task, Experiment 3 examined phonological influences by manipulating the cross-language congruency of grapheme-to-phoneme mappings (detect “a” in have [congruent] vs. take [incongruent]). The outcome of this study offers preliminary evidence of graphemic coding during L2 word recognition both at the orthographic and the orthography-to-phonology mapping levels.     

A Computational and Empirical Investigation of Graphemes in Reading

It is often assumed that graphemes are a crucial level of orthographic representation above letters. Current connectionist models of reading, however, do not address how the mapping from letters to graphemes is learned. One major challenge for computational modeling is therefore developing a model that learns this mapping and can assign the graphemes to linguistically meaningful categories such as the onset, vowel, and coda of a syllable. Here, we present a model that learns to do this in English for strings of any letter length and any number of syllables. The model is evaluated on error rates and further validated on the results of a behavioral experiment designed to examine ambiguities in the processing of graphemes. The results show that the model (a) chooses graphemes from letter strings with a high level of accuracy, even when trained on only a small portion of the English lexicon; (b) chooses a similar set of graphemes as people do in situations where different graphemes can potentially be selected; (c) predicts orthographic effects on segmentation which are found in human data; and (d) can be readily integrated into a full‐blown model of multi‐syllabic reading aloud such as CDP++ (Perry, Ziegler, & Zorzi, 2010). Altogether, these results suggest that the model provides a plausible hypothesis for the kind of computations that underlie the use of graphemes in skilled reading.     

Surface dyslexia: A case report and some theoretical implications to reading models

This paper reports on the psycholinguistic investigation of a surface dyslexic aphasic patient's abilities to handle written material. The analysis of paralexic errors produced in reading aloud single words and nonwords classically suggested that the patient was using an analytical strategy parsing the letter string stimulus, from left to right, into graphemes, and assigning phonemic values to graphemes. The patient's results were found to be sensitive to irregularities in correspondence between graphemes and phonemes not only in reading aloud but in lexical decisions, writing on dictation, rhyming, and written-word comprehension. Moreover, the patient's linguistic behavior brought out the reverse pattern observed in deep-dyslexic performances within word/nonword and content/function word dimensions. It was found that some semantic information about written words could be retrieved from both phonological and nonphonological processes presumably operating concurrently and both providing converging or conflicting pieces of meaning to the understanding of written words. Some considerations derived from the observation of this pathological reading behavior are discussed, contributing to a psycholinguistic model of normal reading.     

Multiletter Word Units in Visual Recognition: Direct Activation by Supraletter Visual Features

Multiletter priming effects have been interpreted as evidence for the representation of separable multiletter units in the visual word recognition system (Whiteley & Walker, 1994). The reported experiments examine whether the activation of such units is pre- or post-lexical. Experiments 2 and 3 employed priming in an alphabetic decision task in which subjects made a discrimination response to test stimuli which could be classed as either targets or foils. Targets were single letters, or consonant bigrams, present or absent in an immediately preceding word, or (Experiment 3 and 4) they were whole words semantically associated or not to a preceding word. Foils were single non-alphanumeric characters, a character plus a letter, or a word with one letter replaced by a character. Experiment 1 was a preliminary to determine the parameters of a sequential presentation manipulation. Experiment 2 compared conditions of simultaneous and sequential presentation where letters of prime words were presented together, or one at a time in rapid succession. With simultaneous presentation, responses to bigram targets were facilitated when these appeared in the prime word, while responses to individual constituent letters of those bigrams were not facilitated. Additionally, responses to primed bigram targets were faster than responses to primed single letter targets. The sequential presentation of prime words resulted in a qualitative change in the response pattern indicative of the disruption of multiletter unit activation. That change was replicated in Experiment 3 where semantic priming confirmed that the prime words were being processed to a level of meaning. The observations challenge a post-lexical account of the multiletter priming effects. Finally, Experiment 4 addressed the question of whether bigram priming reflects the intentional use of prime information to predict following targets. Strategic interpretations are undermined and it is argued that multiletter units are activated automatically as part of normal visual word recognition.     

Processing Unpointed Hebrew: What Can We Learn from Determining the Identicalness of Monosyllabic and Bisyllabic Nouns

The aim of this study was to determine whether Hebrew readers reference phonological information for the silent processing of unpointed Hebrew nouns. A research paradigm in which participants were required to perform consecutive same/different judgments regarding the identicalness of members of stimulus pairs was used for answering this question. Twenty-eight students (mean grade 4.9) participated in the study. The nouns used in preparing the word stimulus pairs were comprised of various amounts of syllabic information (monosyllabic versus bisyllabic) and differed in the degree this information was represented by their letter graphemes. The main findings suggest that the processing of the identicalness of unpointed Hebrew words may not involve the referencing of their phonological information.     

Can the dual-route cascaded computational model of reading offer a valid account of the masked onset priming effect?

The masked onset priming effect (MOPE) refers to the empirical finding that target naming is faster when the target (SIB) is preceded by a briefly presented masked prime that starts with the same letter/phoneme (suf) than when it does not (mof; Kinoshita, 2000, Experiment 1). The dual-route cascaded (DRC) computational model of reading (Coltheart, Rastle, Perry, Langdon, & Ziegler, 2001) has offered an explanation for how the MOPE might occur in humans. However, there has been some empirical discrepancy regarding whether for nonword items the effect is limited to the first-letter/phoneme overlap between primes and targets or whether orthographic/phonological priming effects occur beyond the first letter/phoneme. Experiment 1 tested these two possibilities. The human results, which were successfully simulated by the DRC model, showed priming beyond the first letter/phoneme. Nevertheless, two recent versions of the DRC model made different predictions regarding the nature of these priming effects. Experiment 2 examined whether it is facilitatory, inhibitory, or both, in order to adjudicate between the two versions of the model. The human results showed that primes exert both facilitatory and inhibitory effects.     

Search reaction time for single targets in multiletter stimuli with brief visual displays

The same set of Ss was run In an LT (single late target letter followed an earher multiletter display) and an ET (single early target preceded a later multiletter display) condition. On one-half the trials, the multiletter set included the target. and on the other half, the target was absent. The task of S was to push the “yes” (“no”) button if the target was present (absent), and reaction times were recorded. The most plausible processing model assumed that LT comparisons took place in a verbal-acoustic store and that ET comparisons took place in a visual store. It further assumed that processing within these stores was self-terminating, with rates that differed on “same” and “different” comparisons and which changed as the multiletter set increased. Classes of serial and parallel models that are falsified or supported by the present and similar data are discussed.     

A comparative quantitative analysis of Greek orthographic transparency

Orthographic transparency refers to the systematicity in the mapping between orthographic letter sequences and phonological phoneme sequences in both directions, for reading and spelling. Measures of transparency previously used in the analysis of orthographies of other languages include regularity, consistency, and entropy. However, previous reports have typically been hampered by severe restrictions, such as using only monosyllables or only word-initial phonemes. Greek is sufficiently transparent to allow complete sequential alignment between graphemes and phonemes, therefore permitting full analyses at both letter and grapheme levels, using every word in its entirety. Here, we report multiple alternative measures of transparency, using both type and token counts, and compare these with estimates for other languages. We discuss the problems stemming from restricted analysis sets and the implications for psycholinguistic experimentation and computational modeling of reading and spelling.     

Sex differences in the reading process revealed by a letter-detection task

Four experiments used a task in which males and females had to work through a passage and circle instances of a target letter. Coltheart, Hull and Slater (1975) have previously used this task to show that females have greater difficulty than males in detecting target letters, especially letters in the and other silent letters. The four experiments failed to replicate Coltheart et al.'s findings and, in fact, frequently found significant sex differences in the opposite direction. This discrepancy in results could not be accounted for in terms of procedural differences between the two studies. It was concluded that Coltheart et al.'s (1975) results with the letter detection task may not be reliable.     

Processing complex graphemes in handwriting production

Recent studies on handwriting production and neuropsychological data have suggested that orthographic representations are multilevel structures that encode information on letter identity and order, but also on intermediate-grained processing units such as syllables and morphemes. This study on handwriting production examined whether orthographic representations also include a graphemic-processing level. French adults wrote words containing an embedded one-, two-, or three-letter grapheme (e.g., a in clavier, ai in prairie, ain in plainte) on a digitizer. The results for letter duration revealed that the timing of movement processing depends on grapheme length (e.g., the duration of a for one-letter graphemes was shorter than that for two-letter graphemes, which, in turn, was shorter than that for three-letter graphemes). Two- and three-letter graphemes start to be processed before we start to write them. The results therefore revealed that orthographic representations also encode information on grapheme complexity.