Phonological priming and orthographic analogies in reading

Automatic activation of phonology in silent reading is parallel: evidence from beginning and skilled readers

The picture-word interference paradigm was used to shed new light on the debate concerning slow serial versus fast parallel activation of phonology in silent reading. Prereaders, beginning readers (Grades 1-4), and adults named pictures that had words printed on them. Words and pictures shared phonology either at the beginnings of words (e.g., DOLL-DOG) or at the ends of words (e.g., FOG-DOG). The results showed that phonological overlap between primes and targets facilitated picture naming. This facilitatory effect was present even in beginning readers. More important, from Grade 1 onward, end-related facilitation always was as strong as beginning-related facilitation. This result suggests that, from the beginning of reading, the implicit and automatic activation of phonological codes during silent reading is not serial but rather parallel.

What do children do in the rime-analogy task? An examination of the skills and strategies used by early readers

Despite the intense debate surrounding the use of orthographic analogy in the clue word paradigm, little is known about the skills and strategies children actually use and how these compare with their everyday reading of single words. This study, with 4- and 5-year-olds (N=125), supports previous work which suggests children rely on phonological, rather than orthographic, priming in the clue word task since children most frequently produced rhyming words in response to the clue word. The extent to which phoneme and rhyme-based skills, along with letter-sound knowledge, predicted children's performance in the analogy task and in a test of single word reading was contrasted and compared. Our findings suggested that the balance of skills which children drew upon was determined by the demands of the task. The implications of these findings for the validity of the 'orthographic'-analogy task and for teaching beginning readers is discussed.

The limitations of orthographic analogy in early reading development: performance on the clue-word task depends on phonological priming and elementary decoding skill, not the use of orthographic analogy

Two experiments investigated the mechanisms underlying analogical transfer in the clue-word reading task developed by Goswami and her colleagues. Across both experiments, an equivalent number of "analogy" responses were made regardless of whether the clue word was seen or just heard. In addition, the number of "analogy" responses to words sharing both orthographic and phonological overlap with the clue words was no greater than that shown to words sharing only pronunciations. These results provide no evidence for the view that beginning readers make genuine orthographic-based analogies. Instead, the findings are interpreted within a framework in which phonological priming, in combination with the children's own partial decoding attempts based on limited orthographic knowledge, account for their performance on the clue-word task. It is concluded that the extent to which beginning readers make orthographic analogies is overestimated and as a consequence, theories that emphasize the importance of orthographic analogy as a mechanism driving the development of early reading skills need to be questioned.

Recombinative generalization of within-syllable units in prereading children

This study demonstrates recombinative generalization of within-syllable units in prereading children. Three kindergarten children learned to select printed consonant-vowel-consonant words upon hearing the corresponding spoken words. The words were taught in sets; there were six sets, presented consecutively. Within sets, the four words that were taught had overlapping letters, for example, sat, mat, sop, and sug. Tests for recombinative generalization determined whether the children selected novel words with the same components as the trained words (e.g., mop and mug). Two children demonstrated recombinative generalization after one training set, and the 3rd demonstrated it after two training sets. In contrast, 2 other children, who received tests but no training, showed low accuracy across six sets. The 3 experimental children then demonstrated highly accurate printed-word-to-picture matching, and named the majority of the printed words. These findings are a promising step in the development of a computerized instructional technology for reading.

Units of analysis in nonword reading: evidence from children and adults

Four experiments examined variations in children's (chronological age range: 5 years 7 months to 9 years 10 months) and adults' reading strategy as a function of task demands. Experiment 1 found that less skilled readers (mean reading age: 8 years 8 months), though able to make use of rime-based spelling-to-sound correspondences (reading "by analogy"), predominantly used simple grapheme-phoneme-level correspondences in reading isolated unfamiliar items. Skilled children (mean reading age: 11 years 6 months) were more likely to adopt an analogy strategy. Experiments 2 and 3 adopted versions of the "clue word" technique used by U. Goswami (1986, Journal of Experimental Child Psychology, 42, 73-83; 1988, Quarterly Journal of Experimental Psychology, 40A, 239-268) and found a much higher incidence of analogical responding by children of all ages, suggesting that reading strategy is task-dependent. Experiment 4 showed that adults' nonword-reading strategy is determined by list composition, in that grapheme-phoneme correspondences are used more when the list context contains nonwords. It is concluded that both adults and young children exhibit considerable flexibility and task-dependence in the levels of spelling-to-sound correspondence (analogies vs decoding) that they use and that grapheme-phoneme correspondences are preferred when maximum generalization to unfamiliar items is required.

Inferring sublexical correspondences from sight vocabulary: evidence from 6- and 7-year-olds

We report an experiment designed to investigate 6-to-7-year-old children's ability to acquire knowledge of sublexical correspondences between print and sound from their reading experience. A computer database containing the printed word vocabulary of children taking part in the experiment was compiled and used to devise stimuli controlled for grapheme-phoneme correspondence (GPC) frequency and rime neighbourhood consistency according to the children's reading experience. Knowledge of GPC rules and rime units was compared by asking children to read aloud three types of nonword varying in regularity of GPC and consistency of rime pronunciation. Results supported the view that children can acquire knowledge of both GPC rules and rime units from their reading experience. GPC rule strength affects the likelihood of a GPC response; rime consistency affects the likelihood of a rime response.

Not all reading disabilities are alike

In this article, reading disability is defined broadly to refer to below-average achievement in reading comprehension as assessed by a standardized test. With our research we tried to answer the question of whether all children with reading disability share a common etiology of deficient phonology, or constitute heterogeneous groups. The answer to this question was sought in four studies that examined reading disabilities from the perspective of componential skills of reading. In Part 1, the results of the first study are reported. A principal-components analysis of the performance of 139 children from Grades 3, 4, and 6 on reading-related tasks yielded two factors: decoding and comprehension. However, factor analyses conducted for each grade separately indicated that orthographic skill and processing speed could possibly constitute a third component. The orthography-speed factor emerged as a factor only in the 6th grade. Part 2 of this article reports the findings of three studies that analyzed the componential skills profiles of poor readers. It was found that the poor readers constituted heterogeneous groups and that four different types of poor readers could be identified with deficiency in any one of the following skills: (a) decoding only, (b) comprehension only, (c) a combination of decoding and comprehension, and (d) a combination of orthographic processing and reading speed. It was also found that the criteria used in selecting poor readers influenced the distribution of the ratio of the four types of poor readers within any given group.

Orthographic analogies and phonological priming: a comment on Bowey, Vaughan, and Hansen (1998)

J. A. Bowey, L. Vaughan, and J. Hansen (1998, Journal of Experimental Child Psychology, 68, 108-133) carried out two experiments on 6- and 7-year-old children's use of orthographic analogies in word reading. They reported that, following apparently stringent controls for phonological priming effects, beginning analogies (beak-bean) were more frequent in this age group than rime (beak-peak) analogies. From this, they concluded that beginning readers do not reliably use orthographic rimes in reading, even in the clue word task (p. 129). However, the clue word task was not used in this study. This comment highlights two problems with Bowey et al.'s paper. The first is a theoretical one, and the second is methodological. Firstly, Bowey et al. base their investigation on a misunderstanding of U. Goswami and P. E. Bryant's (1990, Phonological skills and learning to read, Hillsdale, NJ: Erlbaum) claims about the role of rhyme and analogy in beginning reading. Secondly, methodological weaknesses, in particular unintended intralist priming effects, seriously limit the conclusions that can be drawn from Bowey et al.'s booklet analogy task.

The limitations of orthographic rime analogies in beginners' word reading: a reply to Goswami (1999)

U. Goswami (1999, Journal of Experimental Child Psychology, 72, 210-219) argues that the findings of J. A. Bowey, L. Vaughan, and J. Hansen (1998, Journal of Experimental Child Psychology, 68, 108-133) are uninterpretable. This paper examines each of Goswami's criticisms of the methodology employed by Bowey et al. (1998). None can explain the differential analogy and phonological priming effects reported by Bowey et al. More fundamentally, none can explain the critical finding of Bowey et al. that, when phonological priming effects are controlled, the size of the end analogy effect is no greater than that of beginning and medial vowel analogy effects. Furthermore, some of Goswami's criticisms cast considerable doubt on the generalizability of findings from her version of the clue word task.

Sublexical Inferences in Beginning Reading: Medial Vowel Digraphs as Functional Units of Transfer

Two experiments evaluated young children's use of lexical inference. Experiment 1 compared transfer from shared rimes (e.g., "beak"-"peak"), or heads (e.g., "beak"-"bean"), under three conditions: (a) when both clue word pronunciation and orthography were present at transfer; (b) when only the pronunciation of the clue word was given; and (c) when the clue was pretaught. Equivalent transfer occurred in both conditions (a and b) where clue word pronunciations were provided at transfer, but no transfer was found when the clue word was pretaught (condition c). Experiment 2 investigated transfer from three pretaught clue words sharing rimes (e.g., "leak"-"peak"), or vowel digraphs (e.g., "leak"-"bean"). Children demonstrated lexical transfer under these conditions, but improvements were equivalent for vowel and rime analogous words. Results are interpreted in terms of models of vowel transfer. Copyright 1998 Academic Press.

Beginning readers' use of orthographic analogies in word reading

This research re-investigated the claim that beginning readers exploit information from the orthographic rime of clue words to help them to decode unfamiliar words. In Experiment 1, first-grade children were equally able to use orthographic information from the beginning, middle, and end of clue words to identify unfamiliar target words. Moreover, the improvement in reading end-(or orthographic rime-) same target words following clue word presentation reflected phonological priming. In second-grade children, with correction for retesting effects, improvement following clue word presentation for end-same and beginning-same target words was equivalent, although end-same target words improved more than middle-same target words. In Experiment 2, both first- and second-grade children were able to use orthographic information from the beginning, middle, and end of clue words to identify unfamiliar words. Clue word presentation enhanced the reading of beginning-same and end-same target words more than middle-same target words. Improvement was the same for beginning-same and end-same target words. Target word improvement following clue word presentation was greater than that for phonologically primed words only in children reading target words sharing the beginning sequence of the clue word.

Do children need concurrent prompts in order to use lexical analogies in reading?

In two experiments, children were taught a word and then given reminders of that word while reading unfamiliar analogous words. Reminders of the pronunciation of taught words facilitated reading of rime analogous words (e.g. beak-peak) over head analogous words (e.g. beak-bean) in Experiment 1, and onset analogous words (e.g. stilt-stem), over medial analogous words (e.g. stilt-milk) in Experiment 2. Importantly, however, privileged improvement for rime and onset analogous words was not evident in the absence of these concurrent prompts. The findings suggest that the analogy model developed by Goswami (1993) may have limited applicability in naturalistic settings.

Phonological recoding and self-teaching: sine qua non of reading acquisition

The self-teaching hypothesis proposes that phonological recoding functions as a self-teaching mechanism enabling the learner to independently acquire an autonomous orthographic lexicon. Successful decoding encounters with novel letter strings provide opportunities to learn word-specific print-to-meaning connections. Although it may not play a central role in skilled word recognition, phonological recoding, by virtue of its self-teaching function, is regarded as critical to successful reading acquisition. This paper elaborates the self-teaching hypothesis proposed by Jorm and Share (1983), and reviews relevant evidence. Key features of phonological recoding include an item-based rather than stage-based role in development, the progressive "lexicalization" of the process of recoding, and the importance of phonological awareness and contextual information in resolving decoding ambiguity. Although phonological skills have been shown to be primary in reading acquisition, orthographic processing appears to be an important but secondary source of individual differences. This implies an asymmetrical pattern of dissociations in both developmental and acquired reading disorders. Strong relationships between word recognition, basic phonological processing abilities and phonemic awareness are also consistent with the self-teaching notion. Finally, it is noted that current models of word recognition (both PDP and dual-route) fail to address the quintessential problem of reading acquisition-independent generation of target pronunciations for novel orthographic strings.