Biomedical subjects
Publications and source records attributed to R Shillcock.
A computational analysis of a large British English database was performed and frequencies occurrence of grapheme-phoneme correspondences were obtained. A computer program was implemented, which used these frequencies to predict the probabilities of all possible pronunciations of any given string of graphemes. These results led to a proposal for a quantitative method of measurement of the orthographic depth of different languages.
A new theory of visual word recognition is based on the fact that the fovea is split in humans. When a reader fixates the center of a written word, the initial letters of the word that are to the left of fixation are projected first to the right cerebral hemisphere (RH) while the final letters are projected to the left cerebral hemisphere (LH). This paper explores the possibility that this has consequences for the early processing of the beginning and ends of centrally fixated words: specifically that lexical decision RTs are affected by the number of letters to the left of fixation but not by the number of letters to the right of fixation. For centrally presented five- and eight-letter words, we manipulated number of letters presented to the right or to the left of a fixation point (Experiment 1). We found that longer latencies to longer letter strings characterised the processing of the initial letters of words while LH word recognition features characterised the ends of words. Experiment 2 was a lateralized version of Experiment 1, and revealed the well established visual field and word length interaction. The results supported the split fovea theory.
We argue that the projection of the visual field to the cortex constrains and informs the modeling of visual word recognition. On the basis of anatomical and psychological evidence, we claim that the higher-level cognition involved in word recognition does not completely transcend initial foveal splitting. We present a schematic connectionist model of word recognition that instantiates the precise splitting of the visual field and the contralateral projection of the two hemifields. We explore the special nature of the exterior (i.e., first and last) letters of words in reading. The model produces the correct behavior spontaneously and robustly. We analyze this behavior of the model with respect to words and random patterns and conclude that the systematic division of the visual input has predictable, general informational consequences and is chiefly responsible for the exterior letters effect.
This paper presents an experiment that examined two related questions. First, it examined the effects of animacy on the production of different syntactic structures and word orders by Catalan-speaking children. Secondly, it explored the relationship between age and the production of different syntactic structures by these children. The results of a picture description experiment run with eighty Catalan-speaking children aged 4;11 to 11;11 show that participants tended to produce more object-dislocated descriptions when the patient was animate than when the patient was inanimate. The results also show that while the dislocation of object clauses is a construction already consolidated at 5;0, the passive structure is a construction still not fully acquired at 11;0. A comparison between the results obtained in the present experiment with existing results from similar experiments with English-speaking children shows that there is a cross-language difference in the age at which children start producing passive clauses. We argue that frequency of exposure to a particular syntactic structure is an important factor that contributes to the acquisition of that syntactic structure. We also suggest that the effects of animacy on the production of object-dislocated descriptions can be explained by means of conceptual and lexical accessibility.
Some of the implications of a model of visual word recognition in which processing is conditioned by the anatomical splitting of the visual field between the two hemispheres of the brain are explored. The authors investigate the optimal processing of visually presented words within such an architecture, and, for a realistically sized lexicon of English, characterize a computationally optimal fixation point in reading. They demonstrate that this approach motivates a range of behavior observed in reading isolated words and text, including the optimal viewing position and its relationship with the preferred viewing location, the failure to fixate smaller words, asymmetries in hemisphere-specific processing, and the priority given to the exterior letters of words. The authors also show that split architectures facilitate the uptake of all the letter-position information necessary for efficient word recognition and that this information may be less specific than is normally assumed. A split model of word recognition captures a range of behavior in reading that is greater than that covered by existing models of visual word recognition.
Nonfluent, Broca-type dysphasics are characterized by impaired syntactic processing. However, grammaticality judgements and certain on-line tasks have shown some preservation of this processing in such subjects. We report an experiment with nonfluent dysphasics in which they read aloud th-initial nonwords (e.g., thuz) in sentential contexts that predicted a function word or a content word. This paradigm was first used by Campbell and Besner (1981) to demonstrate syntactic effects on pronunciation: normal subjects pronounce word-initial th- as voiced in function word contexts and unvoiced in content word contexts, reflecting a regularity in the English lexicon. Poorer performance by the dysphasic subjects on this task is the default prediction of most "syntactic" accounts of agrammatism, including an account based on the impairment of functional projections, which we discuss. We replicate Campbell and Besner's effect in our normal control group and in the dysphasic group, with no significant difference between the two groups. We conclude that syntactic influences on pronunciation may be unimpaired in nonfluent dysphasia, and that the task used resembles the class of online tasks, in its capacity to elicit unimpaired processing. We argue that this result is compatible with the account of agrammatism discussed if the latter is grounded in a distributed, constraint-based processing device allowing graceful degradation of functioning.
The line-bisection task is the standard assessment of unilateral visual neglect. It supplies one effect, the crossover effect, that challenges models of neglect: in left neglect, the rightward displacement of the midpoint of the line becomes a left displacement for small lines. We review the various attempts to account for the cross-over effect, before describing a computational model of performance in the line-bisection task that produces a cross-over effect quite naturally in its damaged state. The model trades on aspects of several current theories of neglect, including independent attentional processing in the two hemispheres, each of which possesses an attentional gradient in which the contralateral field is accentuated. We assume a small residual noise, along the same gradient, in the damaged hemisphere. When lesioned to simulate right hemisphere damage, the model produces line bisections similar to human performance, in terms of the relationship with line length, a variable cross-over point for the smaller lines, and an amelioration of performance with leftside, but not rightside cueing.
Speech is continuous, and isolating meaningful chunks for lexical access is a nontrivial problem. In this paper we use neural network models and more conventional statistics to study the use of sequential phonological probabilities in the segmentation of an idealized phonological transcription of the London-Lund Corpus; these speech data are representative of genuine conversational English. We demonstrate, first, that the distribution of phonetic segments in English is an important cue to segmentation, and, second, that the distributional information is such that it might allow the infant, beginning with only a sensitivity to the statistics of subsegmental primitives, to bootstrap into a series of increasingly sophisticated segmentation competences, ending with an adult competence. We discuss the relation between the behavior of the models and existing psycholinguistic studies of speech segmentation. In particular, we confirm the utility of the Metrical Segmentation Strategy (Cutler & Norris, 1988) and demonstrate a route by which this utility might be recognized by the infant, without requiring the prior specification of categories like "syllable" or "strong syllable."