Abstract:We evaluated the Dual Route Cascaded (DRC) model of visual word recognition using Greek behavioral data on word and nonword naming and lexical decision, focusing on the effects of syllable and bigram frequency. DRC was modified to process polysyllabic Greek words and nonwords. The Greek DRC and native speakers of Greek were presented with the same sets of word and nonword stimuli, spanning a wide range on several psycholinguistic variables, and the sensitivity of the model to lexical and sublexical variables w… Show more
“…The 2001 version of the implemented DRC model consists of localist representations for single-syllable words (this model has also been applied to the processing of visible language in German, see Ziegler et al, 2000 and in Greek, Kapnoula et al, 2017). A single node (a lexical entry) corresponds to the correct spelling of each word known to the model in the orthographic input lexicon (the OIL) and a single node that corresponds to the phonological representation of each word it knows (the phonological output lexicon [the POL]).…”
The notion that some mental processes are "automatic" while others are "controlled" is a distinction that appears in virtually all cognition textbooks, as well as in thousands of papers and book chapters. Indeed, so entrenched is the automatic side of this distinction that various leading computational accounts make no mention of it, but instead assume it implicitly. These models, and the field more generally, assume that processing is stimulus triggered and does not need any form of attention or an intention as a preliminary. Further, the fundamental processing dynamics underlying such automatic processing is widely seen as consisting of interactive activation and autonomous in that it unfolds in the same way across contexts. I review a number of findings from my lab that lead me to a different conclusion. Visual word recognition requires a consideration and integrated understanding of automaticity, attention, intention, context, and cognitive processing. I present various findings that challenge the preeminent role ascribed to interactive activation as implemented in the dominant computational models. I conclude that, going forward, the time is due for computational models of visual word recognition (and researchers in the field more generally) to acknowledge that the findings reported here constitute benchmarks that constrain theory and present opportunities for making meaningful advances in our understanding of visual word recognition (and perhaps of cognition more generally). A few proposals for how we might think about some of these processes are offered.
Public Significance StatementI provide a review of 4 decades of work on visual word recognition as seen from the perspective of my lab. In particular, I consider the interrelations between several computational models, attention, intention, context, and the underlying processing dynamics.
“…The 2001 version of the implemented DRC model consists of localist representations for single-syllable words (this model has also been applied to the processing of visible language in German, see Ziegler et al, 2000 and in Greek, Kapnoula et al, 2017). A single node (a lexical entry) corresponds to the correct spelling of each word known to the model in the orthographic input lexicon (the OIL) and a single node that corresponds to the phonological representation of each word it knows (the phonological output lexicon [the POL]).…”
The notion that some mental processes are "automatic" while others are "controlled" is a distinction that appears in virtually all cognition textbooks, as well as in thousands of papers and book chapters. Indeed, so entrenched is the automatic side of this distinction that various leading computational accounts make no mention of it, but instead assume it implicitly. These models, and the field more generally, assume that processing is stimulus triggered and does not need any form of attention or an intention as a preliminary. Further, the fundamental processing dynamics underlying such automatic processing is widely seen as consisting of interactive activation and autonomous in that it unfolds in the same way across contexts. I review a number of findings from my lab that lead me to a different conclusion. Visual word recognition requires a consideration and integrated understanding of automaticity, attention, intention, context, and cognitive processing. I present various findings that challenge the preeminent role ascribed to interactive activation as implemented in the dominant computational models. I conclude that, going forward, the time is due for computational models of visual word recognition (and researchers in the field more generally) to acknowledge that the findings reported here constitute benchmarks that constrain theory and present opportunities for making meaningful advances in our understanding of visual word recognition (and perhaps of cognition more generally). A few proposals for how we might think about some of these processes are offered.
Public Significance StatementI provide a review of 4 decades of work on visual word recognition as seen from the perspective of my lab. In particular, I consider the interrelations between several computational models, attention, intention, context, and the underlying processing dynamics.
“…According to the “strong version” of ODH, in shallow/transparent orthographies with predictable letter-sound correspondences correct pronunciations for both words and nonwords could be computed via the sublexical pathway. In contrast to this theoretical assumptions, there is behavioral evidence ( Raman et al, 1996 ; Pagliuca et al, 2008 ; Marcolini et al, 2009 ; Difalcis et al, 2018 ; Ripamonti et al, 2018 ), neuropsychological evidence ( Ardila and Cuetos, 2016 ), modelling ( Seidenberg, 2011 ; Kapnoula et al, 2017 ), and neuroimaging ( Ischebeck et al, 2004 ; Danelli et al, 2015 ; Rueckl et al, 2015 ; Marinelli et al, 2016 ; Protopapas et al, 2016 ) suggesting that even in transparent orthographies the lexical-semantic processes are used for skilled reading of words. This evidence contradicts the “strong version” of ODH suggesting that there are distinct lexical-semantic and sublexical pathways in both shallow and deep orthographies which suggest that reading in all languages is supported by a neural network characterized by a universal dual-pathway architecture, and orthographic depth may influence the division of labor between the lexical-semantic and sublexical pathways ( Paulesu et al, 2000 ; Das et al, 2011 ; Cherodath and Singh, 2015 ; Mei et al, 2015 ; Oliver et al, 2017 ).…”
IntroductionAccording to the strong version of the orthographic depth hypothesis, in languages with transparent letter-sound mappings (shallow orthographies) the reading of both familiar words and unfamiliar nonwords may be accomplished by a sublexical pathway that relies on serial grapheme-to-phoneme conversion. However, in languages such as English characterized by inconsistent letter-sound relationships (deep orthographies), word reading is mediated by a lexical-semantic pathway that relies on mappings between word-specific orthographic, semantic, and phonological representations, whereas the sublexical pathway is used primarily to read nonwords.MethodsIn this study, we used functional magnetic resonance imaging to elucidate neural substrates of reading in Czech, a language characterized by a shallo worthography. Specifically, we contrasted patterns of brain activation and connectivity during word and nonword reading to determine whether similar or different neural mechanisms are involved. Neural correlates were measured as differences in simple whole-brain voxel-wise activation, and differences in visual word form area (VWFA) task-related connectivity were computed on the group level from data of 24 young subject. Trial-to-trial reading reaction times were used as a measure of task difficulty, and these effects were subtracted from the activation and connectivity effects in order to eliminate difference in cognitive effort which is naturally higher for nonwords and may mask the true lexicality effects.ResultsWe observed pattern of activity well described in the literature mostly derived from data of English speakers – nonword reading (as compared to word reading) activated the sublexical pathway to a greater extent whereas word reading was associated with greater activation of semantic networks. VWFA connectivity analysis also revealed stronger connectivity to a component of the sublexical pathway - left inferior frontal gyrus (IFG), for nonword compared to word reading.DiscussionThese converging results suggest that the brain mechanism of skilled reading in shallow orthography languages are similar to those engaged when reading in languages with a deep orthography and are supported by a universal dual-pathway neural architecture.
“…Although, in principle, dual-route models predict length effects only for words which are not represented in the orthographic lexicon, such as pseudowords, in reality many low frequency words may be novel to a reader. Therefore, it can be assumed that the dual-route view predicts WF × WL interaction (Balota et al, 2004;Kapnoula et al, 2017). Noteworthy, due to early divergence of the two routes immediately after the letter encoding stage (Perry et al, 2014), this interaction should start to emerge early (Fig.…”
Central questions in the study of visual word recognition and developmental dyslexia are whether early lexical activation precedes and supports decoding (a dual-stage view) or not (dual-route view), and the locus of deficits in dysfluent reading. The dual-route view predicts early word frequency and length interaction, whereas the dual-stage view predicts word frequency effect to precede the interaction effect. These predictions were tested on eye movements data collected from (n = 152) children aged 9–10 among whom reading dysfluency was overrepresented. In line with the dual-stage view, the results revealed an early word frequency effect in first fixation duration followed by robust word length effect in refixation probability and an interaction of word frequency and word length in summed refixation duration. This progression was advanced in fluent reading to be observable already in first fixation duration. Poor reading fluency was mostly explained by inflated first fixation durations, and to stronger word frequency and length effects in summed refixation duration. This pattern of results suggests deficits in early letter encoding and slowness in serial grapheme-phoneme conversion. In contrast to the widely held belief, the holistic orthographic processing of words seemed to be intact.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.