Phonological and Semantic Overlap Between Words: Effects on Recognition of Familiar and Recently Acquired Words in Early Childhood

Abstract

Classical research on word recognition describes how when adult participants are presented with pairs of words that share phonemes or belong to the same semantic category, the processing of the second word is facilitated provided the first word is related to the second. These observations have suggested that words in the mental lexicon are connected via phonological and semantic links. However, more recently, these association rules have been found to evolve in children during the early phases of language acquisition, at the time in which their language processing skills are being developed together with a fast growth of the child’s vocabulary. The present dissertation deals with the phenomena of facilitation and interference in word recognition due to phonological and semantic similarities during early childhood. It does so through three interrelated studies. In a longitudinal design, the first study investigated whether phonological and semantic priming would be better predicted either by the infants’ vocabulary size or by their age. The second study investigated the impact of joint phonological and semantic similarities on word recognition. While past studies had investigated both factors separately, Study II investigated the impact of both factors in combination. The third and last study dived into the question of how word recognition is affected by the similarities among the novel words to learn. Therefore, the mechanisms of facilitation/interference of word recognition due to phonological and semantic overlap was again investigated in young children, but under controlled learning experimental protocol; comparing the effects of word overlap on recently acquired words. Overall, the results of this dissertation show that (i) word recognition is primarily modulated by the number of words in the mental lexicon of children, (ii) phonological interference of word recognition can be alleviated by introducing a semantic similarity, and (iii) that children recognise better recently acquired words when the novel words shared either phonological or semantic similarities. Result (i) provides the first evidence of the effect of children current vocabulary on phonological interference effects. These results align with models of spoken word recognition that suggest that during speech processing, phonologically related words are activated and compete for recognition in line with the Cohort Model and the Neighbourhood Activation Model; thus, the more words a child knows more related associates are activated. In addition, result (ii) shows that adding a similarity in meaning reduces the interference in target recognition induced by the phonological similarity between words. These results resonate with the Distributed Cohort Model, which suggests that semantic aspects of words reduce the activation of phonologically related words to the target supporting the recognition of the intended word. Finally, together the results of this dissertation highlight the differential impact of words simple vs. compound overlap on familiar and recently acquired words. Specifically, it was found that while simple phonological or semantic similarities interfered with familiar word recognition, simple similarities between novel words supported novel word recognition. These results support the hypothesis of the LEX model of word learning that predicts better learning of novel words of similar form or similar meaning than of words with similar form and meaning.