Children read affixed words and non‐words

Children aged seven, eight and nine, subdivided according to reading ability were asked to read affixed words and non‐words ending in ‐ed and ‐er morphemes. There was evidence that the performance of these young children was comparable with Taft's (1985) account of skilled reading which suggests that morphologically complex items are decomposed and this facilitates reading in children. Both better and poorer readers read affixed words more accurately than pseudo‐affixed words, and affixed words significantly better than non‐words with a word stem; all readers were better at reading non‐words with a word as opposed to a non‐word stem. The most notable finding was that words and non‐words ending in ‐er were read more accurately than those ending in ‐ed and adult word reading latencies showed no difference between the two types of ending. This is not straightforwardly explained by decomposition in Taft's (1985) terms, age of acquisition, affix frequency, or by any simplistic recourse to surface and/or base frequency. It is suggested that the explanation lies in the inconsistent pronunciation of the ‐ed ending, and that the syntactic rules for generating affixes need to be augmented by phonological rules in order to explain naming data in children. There was evidence that these children were in a transitional stage between a lexicon which is decomposed and one which is not, and that the lexicon becomes progressively decomposed as suggested by Taft (1985) and Seymour (1987).