Nested incremental modeling in the development of computational theories: the CDP+ model of reading aloud

At least 3 different types of computational model have been shown to account for various facets of both normal and impaired single word reading: (a) the connectionist triangle model, (b) the dual-route cascaded model, and (c) the connectionist dual process model. Major strengths and weaknesses of these models are identified. In the spirit of nested incremental modeling, a new connectionist dual process model (the CDP+ model) is presented. This model builds on the strengths of 2 of the previous models while eliminating their weaknesses. Contrary to the dual-route cascaded model, CDP+ is able to learn and produce graded consistency effects. Contrary to the triangle and the connectionist dual process models, CDP+ accounts for serial effects and has more accurate nonword reading performance. CDP+ also beats all previous models by an order of magnitude when predicting individual item-level variance on large databases. Thus, the authors show that building on existing theories by combining the best features of previous models--a nested modeling strategy that is commonly used in other areas of science but often neglected in psychology--results in better and more powerful computational models.     

Genetic and environmental influences on individual differences in printed word recognition

The genetic and environmental etiologies of individual differences in printed word recognition and related skills were explored in 440 identical and fraternal twin pairs between 8 and 18 years of age. A theoretically driven measurement model identified five latent variables: IQ, phoneme awareness, word recognition, phonological decoding, and orthographic coding. Cholesky decomposition models on these five latent constructs revealed the existence of both common and independent genetic effects, as well as non-shared environmental influences. There was evidence for moderate genetic influences common between IQ, phoneme awareness, and word-reading skills, and for stronger IQ-independent genetic influences that were common between phoneme awareness and word-reading skills, particularly phonological decoding. Phonological and orthographic coding skills in word recognition had both significant common and significant independent genetic influences, with implications for "dual-route" and "connectionist" reading models, subtypes of reading disabilities, and the remediation of reading disabilities.     

Better the DVL you know: acronyms reveal the contribution of familiarity to single-word reading

Current theories of reading are divided between dual-route accounts, which propose that separable processes subserve word recognition for orthographically regular and irregular strings, and connectionist models, which propose a single mechanism mapping form to meaning. These theories make distinct predictions about the processing of acronyms, which can be orthographically illegal and yet familiar, as compared with the processing of pseudowords, which are regular but unfamiliar. This study examined whether acronyms are processed like pseudowords and words. Event-related potentials (ERPs) were recorded as subjects viewed familiar and unfamiliar acronyms, words, pseudowords, illegal strings, and--as the targets of the substantive behavioral task--proper names. Familiar acronyms elicited repetition effects on the N400 component, a functionally specific index of semantic activation processes; repetition effects for familiar acronyms were similar in magnitude, timing, and scalp distribution to those for words and pseudowords. The similarity of the brain response to familiar--but--illegal and unfamiliar--but--legal classes of stimuli is inconsistent with predictions of dual-route models of reading.     

In search of subtypes of Chinese developmental dyslexia

The dual-route model offers a popular way to classify developmental dyslexia into phonological and surface subtypes. The current study examined whether this dual-route model could provide a framework for understanding the varieties of Chinese developmental dyslexia. Three groups of Chinese children (dyslexics, chronological-age controls, and reading-level controls) were tested on Chinese exception character reading, pseudocharacter reading (analogous to English nonword reading), novel word learning, and some phonological and orthographic skills. It was found that Chinese exception character reading and pseudocharacter reading were highly correlated and that orthographic skills was a better predictor of both Chinese exception character and pseudocharacter reading than was phonological skills. More than half (62%) of the children in the dyslexia sample were classified as belonging to the surface subtype, but no children were classified as belonging to the phonological subtype. These results suggested that the lexical and sublexical routes in Chinese are highly interdependent or that there may be only one route from print to speech as suggested by the connectionist models. Chinese dyslexic children generally are characterized as having delays in various phonological and orthographic skills, but some, such as those identified as surface dyslexics in the current study, are more severely impaired.     

Visual word recognition in bilinguals: phonological priming from the second to the first language

In this study, the authors show that cross-lingual phonological priming is possible not only from the 1st language (L1) to the 2nd language (L2), but also from L2 to L1. In addition, both priming effects were found to have the same magnitude and to not be related to differences in word naming latencies between L1 and L2. The findings are further evidence against language-selective access models of bilingual word processing and are more in line with strong phonological models of visual word recognition than with the traditional dual-route models.     

The Utility of Item-Level Analyses in Model Evaluation: A Reply to Seidenberg and Plaut

Seidenberg and Plaut (this issue) argue that the implications of our analyses (Spieler & Balota, 1997) for the two extant connectionist models of word naming are limited by two factors. First, variables outside the scope of these models influence naming performance, so it is not surprising that the models do not account for much of the variance at the item level. Second, there is error variance associated with large item-level data sets that obviously should not be captured by these models. We point out that there are a number of variables that have been incorporated within the targeted connectionist models that should provide these models an advantage over the simple predictor variables that we selected as a baseline to evaluate the efficacy of the models (e.g., log frequency, length in letters, and number of orthographic neighbors). We also point out that there is considerable consistency across four large-scale studies of item means. Finally, we provide evidence that even under conditions of a standard word-naming study (with a small set of items), simple word frequency, orthographic neighborhoods, and length accounted for more variance than the extant connectionist models. We conclude that item-level analyses provide an important source of evidence in the evaluation of current models and the development of future models of visual word recognition.     

On the Anglocentricities of current reading research and practice: the perils of overreliance on an "outlier" orthography

In this critique of current reading research and practice, the author contends that the extreme ambiguity of English spelling-sound correspondence has confined reading science to an insular, Anglocentric research agenda addressing theoretical and applied issues with limited relevance for a universal science of reading. The unique problems posed by this "outlier" orthography, the author argues, have focused disproportionate attention on oral reading accuracy at the expense of silent reading, meaning access, and fluency, and have significantly distorted theorizing with regard to many issues-including phonological awareness, early reading instruction, the architecture of stage models of reading development, the definition and remediation of reading disability, and the role of lexical-semantic and supralexical information in word recognition. The dominant theoretical paradigm in contemporary (word) reading research--the Coltheart/Baron dual-route model (see, e.g., J. Baron, 1977; M. Coltheart, 1978) and, in large measure, its connectionist rivals--arose largely in response to English spelling-sound obtuseness. The model accounts for a range of English-language findings, but it is ill-equipped to serve the interests of a universal science of reading chiefly because it overlooks a fundamental unfamiliar-to-familiar/novice-to-expert dualism applicable to all words and readers in all orthographies.     

Phonology in multisyllabic word recognition

The present study used two word-naming experiments to investigate whether a phonological representation is computed in the perception of multisyllabic words. In the first experiment, 100- or 250-msec beginning syllable previews were given, followed by the whole word. The results indicated that no phonological representation was computed at 100 msec, but a phonological representation was computed at 250 msec. In Experiment 2, the second syllable was given as a prime for either 100 or 250 msec. The results showed that no phonological representation was computed for the end of a multisyllabic word. Results are discussed in terms of current dual-route theory and connectionist models.This research was supported by a National Science Foundation grant BNS90110067. We thank Kim Clements, Cindy Connine, Fred Smith, Steve Straight, Sheldon Li, Steve Specht, Greg Boheimer, Deb Briihl, Rich Topoloski, Anna Chomiak and Chris Bruun for their support.     

Word recognition in developmental dyslexia: A connectionist interpretation

We present a study of the accuracy, consistency, and speed of word naming in a dyslexic boy, JM, who has severe impairments in the ability to use sub-lexical, phonological reading strategies. For words that he can recognise, JM's naming latencies do not differ from those of control subjects matched for reading age, and he is generally consistent from one occasion to the next. He can also match printed homophones with their definitions--a skill that requires access to well-specified orthographic representations. The data are interpreted as evidence for the creation of efficient recognition devices for words within JM's sight vocabulary. However, he shows a continuing inability to use phonological decoding strategies to deal with words that he cannot recognize by sight. Overall we argue our results pose problems for stage models of reading development, and that they may best be interpreted within a connectionist framework of the development of word recognition skills.     

German inflection: single route or dual route?

The German plural system has recently become a focal point for conflicting theories of language, both linguistic and cognitive. Marcus et al. (1995) highlight the German plural as support for the dual-route account of inflectional morphology first proposed by Pinker and colleagues (Pinker & Prince, 1988). On the dual-route account, inflectional morphology is universally subserved by a symbolic rule route which deals with regular inflection and an associative memory component which deals with irregular inflection. This contrasts with single-route connectionist systems. We seek to counter supposed evidence for the dual-route account through large-scale simulations as well as through experimental data. We argue that, in its current form, the dual-route account is incapable of generating experimental data provided by Marcus et al. (1995) as support. Finally, we provide direct quantitative comparisons between single-route and dual-route models of German plural inflection and find single-route performance superior on these tests.     

Cross-language computational investigation of the length effect in reading aloud

The authors examined whether 2 computational models of reading, the dual-route cascaded model (M. Coltheart, K. Rastle, C. Perry, R. Langdon, & J. C. Ziegler, 2001) and the connectionist 2-layer model (M. Zorzi, G. Houghton, & B. Butterworth, 1998), were able to predict the pattern that the length effect found in reading aloud is larger in German than in English (J. C. Ziegler, C. Perry, A. M. Jacobs, & M. Braun, 2001). The results showed that the dual-route cascaded model, which uses a serial mechanism for assembling phonology, successfully predicted this cross-language difference. In contrast, the connectionist model of Zorzi et al. (1998) predicted the opposite: a larger length effect in English than in German. Both the success of one model and the failure of the other highlight fundamental differences between 2 major classes of computational models.     

ROC and z-ROC parameters in word memory: experimental and pre-experimental effects

In this study, ROC and z-ROC parameters are applied in word recognition memory, with words classified according to frequency and class. In doing so, unbiased techniques can be used to assess hits and false alarms and to obtain standardized distributions for each experimental condition. Our results confirm the existence of differences in word recognition memory for experimental exposition time, type of words and other pre-experimental effects. These data are consistent with familiarity and recollection models when words' frequency is used, and with familiarity, recollection and dual-route models when words are organized according to class. Finally, we argue in favour of an interactive processing model (words organized according to frequency) versus a processing model nearer to modularity (words organized according to class).     

War and peace: morphemes and full forms in a noninteractive activation parallel dual-route model.

This article introduces a computational tool for modeling the process of morphological segmentation in visual and auditory word recognition in the framework of a parallel dual-route model.     

Prelexical facilitation and lexical interference in auditory word recognition.

Phonological priming effects were examined in an auditory single-word shadowing task. In 6 experiments, target items were preceded by auditorily or visually presented, phonologically similar, word or nonword primes. Results revealed facilitation in response time when a target was preceded by a word or nonword prime having the same initial phoneme when the prime was auditorily presented but not when it was visually presented. Second, modality-independent interference was observed when the phonological overlap between the prime and target increased from 1 to 3 phonemes for word primes but not for nonword primes. Taken together, these studies suggest that phonological information facilitates word recognition as a result of excitation at a prelexical level and increases response time as a result of competition at a lexical level. These processes are best characterized by connectionist models of word recognition.     

Challenging the widespread assumption that connectionism and distributed representations go hand-in-hand

One of the central claims associated with the parallel distributed processing approach popularized by D.E. Rumelhart, J.L. McClelland and the PDP Research Group is that knowledge is coded in a distributed fashion. Localist representations within this perspective are widely rejected. It is important to note, however, that connectionist networks can learn localist representations and many connectionist models depend on localist coding for their functioning. Accordingly, a commitment to distributed representations should be considered a specific theoretical claim regarding the structure of knowledge rather than a core principle, as often assumed. In this paper, it is argued that there are fundamental computational and empirical challenges that have not yet been addressed by distributed connectionist theories that are readily accommodated within localist approaches. This is highlighted in the context of modeling word and nonword naming, the domain in which some of the strongest claims have been made. It is shown that current PDP models provide a poor account of naming monosyllable items, and that distributed representations make it difficult for these models to scale up to more complex language phenomena. At the same time, models that learn localist representations are shown to hold promise in supporting many of the core reading and language functions on which PDP models fail. It is concluded that the common rejection of localist coding schemes within connectionist architectures is premature.     

BRINGING COMPUTATIONAL MODELS OF WORD NAMING DOWN TO THE ITEM LEVEL

Abstract— Early noncomputational models of word recognition have topically attempted to account for effects of categorical factors such as word frequency (high vs low) and spelling-to-sound regularity (regular vs irregular) More recent computational models that adhere to general connections: principles hold the promise of being sensitive to underlying item differences that are only approximated by these categorical factors In contrast to earlier models, these connectionist models provide predictions of performance for individual items In the present study, we used the item-level estimates from two connectionist models (Plaut, McClelland, Seidenberg, & Patterson, 1996, Seidenberg & McClelland, 1989) to predict naming latencies on the individual items on which the models were trained The results indicate that the models capture, at best, slightly more variance than simple log frequency and substantially less than the combined predictive power of log frequency, neighborhood density, and orthographic length. The discussion focuses on the importance of examining the item-level performance of word-naming models and possible approaches that may improve the models' sensitivity to such item differences     

Naming cAsE aLtErNaTeD words

A dual-route approach was used as an initial framework to examine the relation between presentation format and lexical processing in a naming task. In Experiments 1 and 3, words were presented in lowercase versus case-alternated format. Presentation format interacted with word frequency and regularity: For irregular words (e.g., pint), case alternation was additive with frequency, whereas for regular words (e.g., mint), case alternation and frequency interacted. Experiment 2 dissociated the locus of case-alternation effects from those of stimulus intensity. Stimulus intensity was additive with frequency and regularity, suggesting that whereas stimulus intensity affects encoding, case alternation affects lexical processing at a postencoding stage in the word recognition system. It was concluded that a dual-route approach provides a suggestive but incomplete account of how case alternation influences lexical processing. As an alternative to a dual-route approach, we show that the present results can be addressed and successfully simulated using an implemented version of Norris's (1994) multilevel model.     

Nonword reading: Comparing dual-route cascaded and connectionist dual-process models with human data

Two prominent dual-route computational models of reading aloud are the dual-route cascaded (DRC) model, and the connectionist dual-process plus (CDP+) model. While sharing similarly designed lexical routes, the two models differ greatly in their respective nonlexical route architecture, such that they often differ on nonword pronunciation. Neither model has been appropriately tested for nonword reading pronunciation accuracy to date. We argue that empirical data on the nonword reading pronunciation of people is the ideal benchmark for testing. Data were gathered from 45 Australian-English-speaking psychology undergraduates reading aloud 412 nonwords. To provide contrast between the models, the nonwords were chosen specifically because DRC and CDP+ disagree on their pronunciation. Both models failed to accurately match the experiment data, and both have deficiencies in nonword reading performance. However, the CDP+ model performed significantly worse than the DRC model. CDP++, the recent successor to CDP+, had improved performance over CDP+, but was also significantly worse than DRC. In addition to highlighting performance shortcomings in each model, the variety of nonword responses given by participants points to a need for models that can account for this variety. (PsycINFO Database Record (c) 2012 APA, all rights reserved).     

A CONNECTIONST MODELING APPROACH TO WORD RECOGNITION AND DYSLEXIA

Abstract— Word recognition plays an essential role in learning to read, skilled reading, and dyslexia. The goal of the research I describe is to develop a theory of word recognition that is realized as a connectionist simulation model. Experience with a first-generation model suggests that the approach can reveal general principles underlying word recognition and its impairments. Although computational modeling introduces new problems of method and interpretation, it contributes in an essential way to understanding reading and other aspects of cognition.     

Connectionist models of recognition memory: constraints imposed by learning and forgetting functions

Multilayer connectionist models of memory based on the encoder model using the backpropagation learning rule are evaluated. The models are applied to standard recognition memory procedures in which items are studied sequentially and then tested for retention. Sequential learning in these models leads to 2 major problems. First, well-learned information is forgotten rapidly as new information is learned. Second, discrimination between studied items and new items either decreases or is nonmonotonic as a function of learning. To address these problems, manipulations of the network within the multilayer model and several variants of the multilayer model were examined, including a model with prelearned memory and a context model, but none solved the problems. The problems discussed provide limitations on connectionist models applied to human memory and in tasks where information to be learned is not all available during learning.