A developmental analysis of similarity neighborhoods in European Portuguese

Abstract

We present a developmental analysis of the structural organization of young children's and adults' lexicons for European Portuguese. The production lexicons of 3-,4-, and 5-year-olds, a receptive lexicon for 12- to 19-month-olds, and an adult lexicon were compared using the similarity neighborhoods (e.g., Charles-Luce & Luce, 1990). For each lexicon, similarity neighborhoods were computed for words with 3 to 8 phonemes, and phonological neighborhood sizes were compared. A phonological neighbor was defined as any word in one of the lexicons that differed from a given target by one phoneme substitution, deletion, or addition. Results showed structural differences between shorter (3-, 4-, and 5-phoneme) and longer (6- to 8-phoneme) words. There was no age effect for longer words, of which ca. 92% had no neighbors. Shorter words, in contrast, had more neighbors: in the children's lexicons, ca. 58% of shorter words had one to four neighbors, and 8% had five to seven neighbors; only ca. 36% had no neighbors. An age effect was found, whereby similarity neighborhoods had become increasingly dense over the course of childhod. The results are discussed in light of previous finds for English-speaking children and adults, and their implication for the development of spoken word recogniton by Portuguese listeners are considered.