Roderick W. Barron

Word recognition in early reading: A review of the direct and indirect access hypotheses☆

Abstract Research on the development of word recognition, like that on fluent word recognition, has been strongly influenced by the dual-route model. One route is non-lexical and indirect because access to lexical meaning is mediated by pre-lexical phonological representations assembled through the application of grapheme-to-phoneme correspondence (GPC) rules. The other route, which is independent of the first, is regarded as lexical and direct because orthographic representations of whole words are used to retrieve lexical meanings or post-lexical phonological representations. Evidence is reviewed for two opposing developmental hypotheses based on the dual-route model: fluent readers use both direct and indirect access to lexical meaning, while beginning readers use (1) only indirect access or (2) only direct access. It was concluded that neither mode of access predominates in early reading. A review of other evidence suggests two reasons why the dual-route model fails to provide a satisfactory account of the development of early word recognition. First, it does not offer an adequate characterization of the orthographic units represented in early lexicons. Second, the independence of the two routes prevents lexical information from being acquired through the application of GPC rules. Finally, alternatives to the dual-route model are discussed. One of the most promising is a single process lexical model in which it is proposed that acquisition and performance in word recognition can be accounted for by interactions among orthographic and phonological units of various sizes in the lexicon.

Treatment, subtype, and word type effects in dyslexic children's response to remediation.

Children diagnosed as accuracy-disabled or rate-disabled readers (Lovett, 1984a, 1986, 1987) were randomly assigned to one of three treatment conditions providing training in word recognition and decoding skills (DS), oral and written language (OWLS), or classroom survival skills (an alternative treatment control). The experimental treatment programs exerted a positive treatment effect on the word recognition performances of both groups, but treatment-specific changes in contextual reading and oral language skill were not observed. Pre- and post-treatment comparisons on an experimental word recognition task revealed a post-test advantage for DS-trained children of both subgroups. Accuracy-disabled readers demonstrated treatment-specific gains for both orthographically regular words (e.g., wade) and for exception words (e.g., broad), with their gains greater on exception words. Rate-disabled children demonstrated treatment-specific gains only for exception words, but exhibited these gains following both the DS and the OWLS treatments. Although both experimental treatment programs were associated with a positive outcome for the rate disabled subgroup, DS training was associated with relatively greater treatment gain. These data suggest that the critical variables underlying the effectiveness of the DS treatment include the specific lexical knowledge these disabled readers acquired, their greater reliance on an orthographic pattern procedure in word recognition, and/or the fact that newly acquired items were practiced to a point approximating automatization.

The role of illustrations in children's inferential comprehension

Illustrations are a salient source of information in childrens books, yet their effect on childrens reading comprehension has been studied only through literal factual recall. The purpose of the current study was to determine the effect of illustrations on bridging inferences, an important aspect of meaning making in comprehension models. Identical short stories were presented under different illustration conditions with pictures that represented different parts of the story. Participants were 73 7- to 11-year-olds. Illustrations both facilitated and interfered with inferencing depending on the type of information depicted; however, this effect was reduced as grade increased. Additional findings were that the overall ability to make inferences increased with age and working memory was a significant predictor of this skill. Results are discussed in relation to cognitive and developmental models of comprehension.

The effect of orthographic structure and lexical meaning on “same-different” judgments

Pairs of high frequency English words, orthographically acceptable pseudo-words, and non-word letter strings were presented in a “same”-“different” task. The mean reaction times for “same” judgments were ordered; real words were faster than pseudo-words, and pseudo-words were faster than non-words. The RTs for the “different”, judgments showed no differences among the three types of words, except in the first two days of practice in a blocked presentation condition when the difference between the real words and non-words was marginally significant. These and other results suggest that “same” judgments are based upon a comparison process which efficiently uses higher order semantic and orthographic information in words, whereas “different” judgments are based upon comparison process which performs a self-terminating search of the graphemic information in words. The results were also discussed with reference to hierarchical models of word perception and reading.

Methylphenidate increases selectivity of visual scanning in children referred for hyperactivity.

Visual scanning patterns were investigated in 32 children referred for symptoms of hyperactivity in a doubleblind crossover comparison of methylphenidate and placebo treatments. Total errors, response latency, and visual fixations were recorded as the child scanned computer-generated visual matching-to-sample problems. Results indicated that the number of fixations on the standard stimulus in the matching task was significantly larger in the methylphenidate state. Drug treatment also resulted in a significant increase in the number of systematic comparisons between the standard and the variants in the task. However, the increased selectivity of attention to the standard stimulus was not accompanied by a reduction of total errors. It was suggested that the stimulant drug may increase attentional selectivity even when such a shift fails to produce improvement in task performance.

Teaching prereading skills with a talking computer

The phonological awareness skills of nonreaders were trained using an oddity task (e.g., which word in the series ‘sit’, ‘fit’, ‘cat’ has the odd sound in its middle position). As training progressed, the basis of the oddity decision was shifted from rhyming, to consonant onsets, to consonant and vowel phonemes. The words were spoken by a DECtalk speech synthesizer. One of the experimental groups was given printed as well as computer generated speech feedback while the other was given just computer speech feedback. The alternative training control group based their oddity decisions on meaning rather than sound and was also given just computer speech feedback. Only children with low letter-sound knowledge showed pre-test to post-test gains in performance on a rhyming task compared to the control group, and these gains were not influenced by print feedback. In contrast, only children with high letter-sound knowledge, who were given print feedback during learning, showed pre-test to post-test gains in performance on a phoneme deletion task compared to the control group. These results indicate that a combination of high letter-sound knowledge and print feedback facilitates awareness of phonemes among children who cannot yet read or spell, but awareness of rimes is not facilitated by either high letter-sound knowledge or print feedback. Although consistent with bi-directional, causal models of phonological awareness and literacy, these results indicate that the definition of literacy employed by such models may require expansion. This new definition should include proto-literacy — knowledge of letter-sound and other print-sound relationships that are learned before becoming literate and that may influence the acquisition of awareness of some sub-syllabic units of speech.

The Sound-To-Spelling Connection: Orthographic Activation in Auditory Word Recognition and Its Implications for the Acquisition of Phonological Awareness and Literacy Skills

Documented cases of individuals who are fluent in the acarne skill of talking backwards can be divided into two groups (Cowan, Braine & Leavitt, 1985; Cowan & Leavitt, 1982; Cowan, Leavitt, Massaro & Kent, 1982). One group appears to activate the orthographic representation of the word and then pronounce it by reversing the spelling (e.g., “bomb” is reversed as /bamab/) while a second group pronounces the word by reversing the phonemes (e.g., “bomb” is reversed as /mab/). The second group, however, is not completely immune from the effects of activating and using orthographic information during the process of reversing speech because the spelling of the word can influence choice of the units to be reversed. The two phonemes /k/ and /s/, for example, which are typically represented by the single letter x in words such as FOX, are treated as a single phoneme in about one third of their backward utterances. Like backwards talkers, I intend to take a backwards approach to orthographic knowledge because this chapter will be concerned with the phenomenon, interpretation, and significance of the activation of orthographic information during the process of auditory word recognition.

The effects of lexical and semantic information on same-different visual comparison of words.

Previous research has indicated that phonemic and orthographic factors cannot account for the fact that words (clear/clear) are responded to more rapidly than orthographically legal nonwords (creal/creal) in a same-different visual comparison task. However, the role of semantic and lexical factors is less certain. The effects of semantic similarity on both same and different judgments were evaluated in several experiments. In the first experiment, subjects were not any slower on semantically related (rang/rung) than on unrelated (rang/rank) different judgments even with a 3,000-msec interval between the first and second word. In Experiment 2, subjects based their judgments on whether or not the first letter of each word was visually identical. Same judgments were not any faster for semantically related than unrelated items even though other evidence indicated that subjects were processing the whole word and not just the first letter. Experiment 3 showed that the word/orthographically legal nonword difference could be replicated with the first-letter visual comparison task employed in Experiment 2. These and related results were discussed with reference to the idea that the word/orthographically legal nonword difference is due to the facilitating effects of a lexical entry upon the encoding, but not the comparison of an item.

Locus of the effect of a distinguishing feature in a memory search task

Subjects memorized one to four 8- or 16-sided random forms in a memory search task. The positive set forms (“yes” response) differed in number of sides from the negative set forms (“no” response) for the different set (DS) group, but this distinguishing feature was not available to the same set (SS) group. Mean reaction time increased as a linear function of log2 of the positive set size for both groups, but the increase was greater for the SS than the DS group, suggesting that memory search rather than an encoding stage of information processing was influenced by the availability of a distinguishing feature. In a transfer task which followed, new forms were introduced in which the positive and negative set forms differed in number of sides for both groups. In this task, the two groups did not differ in memory search, but in encoding. It was proposed that (a) availability of a distinguishing feature influences search time because the information specifying the number of sides of the set of memorized forms can be used to influence the speed with which individual forms are examined in memory; (b) previous experience with a distinguishing feature influences encoding because the DS group had learned to extract the information specifying the set of memorized forms (not the information specifying individual forms) more efficiently than the SS group.

Using talking computers to remediate reading and spelling disabilities: The critical role of the print-to-sound unit

Talking computers employing computer-generated speech feedback have been used to remediate the literacy skills of dyslexic readers. A computer program is described that employs DECtalk, a highlevel speech synthesizer, to narrate instruction involving intensive training in identifying whole words or in identifying and blending word segments corresponding to onsets, rimes, and phonemes. Procedures for developing individualized instruction are described as well as for constructing and editing the speech and graphics features of the program. Neurologically impaired dyslexic children trained with this program achieved greater acquisition and transfer of word recognition skill when their training involved segmented rather than whole word feedback.