THE DEVELOPMENT OF SPELLING-SOUND RELATIONSHIPS IN A MODEL OF PHONOLOGICAL READING

Developmental aspects of the spelling to sound mapping for English mon osyllabic words are investigated with a simple two-layer network model using a simple, general learning rule. The model is trained on both r egularly and irregularly spelled words, but extracts the regular spell ing to sound relationships which it can apply to new words, and which cause it to regularise irregular words. These relationships are shown to include single letter to phoneme mappings as well as mappings invol ving larger units such as multiletter graphemes and onset-rime structu res. The development of these mappings as a function of training is an alysed and compared with relevant developmental data. We also show tha t the two-layer model can generalise after very little training, in co mparison to a three-layer network. This ability relies on the fact tha t orthography and phonology can make direct contact with each other, a nd its importance for self-teaching is emphasised.