These studies are an examination of the role of phonological encoding in pronunciation tasks and in reading for meaningo The main purpose was to ascertain whether phonological encoding occurs prior to lexical access by the application of the rules of grapheme-phoneme correspondence, postlexically with the phonological representation of the word being retrieved from the internal lexicon and used for processing the meaning of the sentence, or not at all, and whether the nature of the experimental task affects encoding. In the first series of experiments young children, poor adolescent readers, skilled adult readers and deaf children were presented with different sentences, some of which were meaningful and some of which were meaningless. Several of the meaningless sentences became meaningful if they were encoded phonologically and most of the hearing subjects erred in judging these sentences. Pronunciation, usage and spelling of homophones was also tested to investigate whether these could have caused the homophone effect. In a further set of experiments children's knowledge of, and their ability to apply the rules of grapheme-phoneme correspondence was tested by pronunciation, usage and spelling tasks. Knowledge of the rules benefitted children in these tasks, but did not help them to decide on the meaning of sentences containin"g familiar and unfamiliar words which were either orthographically regular or not. The general conclusions were that pre-lexical phonological encoding is necessary for pronunciation tasks, but that reading for meaning can take place without the reader having first to encode the sentence phonologicallyo He may do so in certain situations, and phonological encoding is then likely to take place post-lexically for the purpose of processing the meaning of the sentence o Phonological encoding depends on specific factors such as the nature of the experimental task, the age and ability of the subjects, and the materials used for testing them.