|
| Int'l Symposium - Recombinative Generalization in Spelling and Reading |
| Monday, May 31, 2004 |
| 9:00 AM–10:20 AM |
| Independence West |
| Area: VBC/EDC; Domain: Applied Research |
| Chair: Kathryn Saunders (University of Kansas) |
| Discussant: William V. Dube (E.K. Shriver Center, University of Massachusetts Medical School) |
| Abstract: . |
| |
| Recombinative Generalization in Constructed Spelling by Preschool Children |
| MONIKA M. SUCHOWIERSKA (University of Kansas), Katherine L. Stewart (University of Kansas), Kathryn Saunders (University of Kansas) |
| Abstract: Recombinative generalization is the demonstration of novel arrangements of previously established linguistic units. Two typically developing children were trained and tested on a computer with a touch-sensitive monitor. They spelled words by touching individual letters on the screen. A total of 60 consonant-vowel-consonant words were divided into two matrices containing five sets of six words each. Within sets, the words had overlapping onsets (initial consonant) (e.g., hop-hed) or rimes (vowel-consonant ending) (e.g., hop-bop). Across sets within a matrix, the words had overlapping rimes. Each set contained two training words and four generalization words. Tests for recombinative generalization determined whether children spelled correctly novel words that were constructed by recombining onsets and rimes (two words) or phonemes (two words) in ways that had not been directly trained. Only one matrix was trained. The results show that one child recombined onsets and rimes in the trained matrix, but made vowel errors on phoneme recombination tests. The other child initially did not show recombination in the trained matrix, but this performance emerged following training on auditory abstraction. Neither child spelled words correctly on the untrained matrix. These findings are a promising step in the development of a computerized technology for teaching rudimentary spelling. |
| |
| Adults with Mental Retardation Demonstrate Recombinative Generalization in a Constructed Spelling Task |
| KATHERINE L. STEWART (University of Kansas), Monika M. Suchowierska (University of Kansas), Lisa Chaney (University of Kansas), Kathryn Saunders (University of Kansas) |
| Abstract: Adults with mild mental retardation showed recombinative generalization in spelling of consonant-vowel-consonant (CVC) words after matrix training that ensured exposure to all the necessary within-word components. A total of 96 words were divided into four matrices containing six sets of four words each. Within each set, words had overlapping onsets (e.g., rag, red) and rimes (e.g., rag, lag). All words in a given matrix had one of two rimes (e.g., ag or ed) and 12 different onsets. Upon hearing the spoken word, participants constructed the words on a touch-sensitive computer monitor by touching individual letters in the correct order. After learning two words in a set (e.g., rag and led), participants demonstrated recombinative generalization by constructing CVC words that had not been directly trained (e.g., red and lag). In general, participants showed recombinative generalization within the first trained matrix and a related matrix that shared the same vowels (i.e., a and e) but showed overgeneralization of one or both vowels trained across matrices with a different set of vowels (i.e., u and i). One future direction will involve using a matrix that includes all four vowels to force a discrimination between the vowels. |
| |
| Recombinative Generalization in Reading and Spelling: Effects of Matching Dictated Syllables to Printed Syllables |
| DEISY DAS GRAÇAS DE SOUZA (Federal University of Sao Carlos, Brazil), Julio C. De Rose (Federal University of Sao Carlos, Brazil), Thais Cazati (Federal University of Sao Carlos, Brazil), Edson Huziawara (Federal University of Sao Carlos, Brazil), Daniela Toledo (Federal University of Sao Carlos, Brazil) |
| Abstract: In previous studies we showed recombinative generalization of reading after students learned to read whole words, to construct words with their syllables, and to match syllables to the corresponding sounds. This study presents effects of these procedures on spelling to dictation. Students first learned to match printed words to the corresponding dictated words. New printed words were presented together with familiar words and students could exclude the familiar words and learn to match the new words to the corresponding dictated words. Students also constructed the words with movable letters in the presence of the printed word. Matching printed to dictated syllables was introduced after the students demonstrated emergent reading of the trained words. The syllables were components of the same words that were being trained in the teaching unit. The whole program comprised 17 units. Tests at various stages and at the end of the program assessed reading and spelling to dictation of trained and new words. Children read most of the training and generalization words (a replication of results from previous studies). Of more interest for the present purposes, scores for spelling dictated words increased markedly throughout the study and closely approximated the scores for reading generalization words. |
|
| |