JCPSLP VOL 15 No 1 March 2013

Computer-assisted assessment and intervention

Children’s naming as a function of neighbourhood density Skott E. Freedman

The majority of previous studies examining the role of neighbourhood density (ND) during children’s naming have limited analyses to the semantic level (i.e., whether or not a word was correctly retrieved). In order to investigate potential effects on articulation, the present investigation explored the influence of ND on preschool children’s naming accuracy using both semantic and phonological analyses. Thirty-seven typically developing children participated in a picture naming task consisting of 30 stimuli differing in ND (low, high). Results indicated that words with more semantic neighbours facilitated naming, such that lexical-semantic representations were more accurately retrieved than those with fewer neighbours. A similar facilitative effect was found at the phonological level; words with high phonological ND were articulated with greater levels of accuracy than those with low phonological ND. Findings are interpreted in the context of lexical facilitation. T he notion that the lexicon significantly affects phonological development was proposed over 35 years ago by Ferguson and Farwell (1975, p. 437), who argued that “a phonic core of remembered lexical items and the articulations that produced them is the foundation of an individual’s phonology.” Phonological neighbourhood density (ND) is a variable representing such an interaction. ND indexes the number of meaningful words in a given language (deemed neighbours ) that can be created by adding, deleting, or substituting a phoneme in any word position (Luce & Pisoni, 1998; Vitevitch & Luce, 1998). Consider the English word “back”. By substituting the first sound, the word “pack” would be considered a neighbour, as would the word “bat” by substituting the final sound. Words with many neighbours such as “back” have high phonological ND and presumably reside in dense neighbourhoods, while words with few neighbours such as “budge” have low ND and are said to reside in sparse neighbourhoods.

Studies investigating the relationship between ND and speech production have reported that words with high ND are better articulated and acquired than those with low ND. Munson and Solomon (2004) conducted an acoustic analysis of adults’ word productions varying in ND. Results indicated that adults articulated words with high ND to a greater degree (e.g., expanded vowel spaces) than words with low ND, suggesting a greater need for intelligibility when words have many similarly sounding items. Additionally, Storkel and colleagues have repeatedly found that both children and adults learn words with high ND more accurately than words with low ND (Storkel, Armbruster, & Hogan, 2006; Storkel & Rogers; 2000). Lexical acquisition has also been explored with respect to ND. In order to determine how ND might influence children’s first words, Storkel (2004) conducted an analysis of the earliest acquired words based on parental report. Findings revealed that early acquired words are overall higher in ND, suggesting a production advantage (at least at the semantic level) for words with many phonological neighbours. Although effects of ND on speech production have been investigated previously using picture-naming tasks, relevant to the present study, the extent to which production has been measured remains largely restricted to a semantic analysis. That is, productions have been scored as correct if the target lexical-semantic representation was retrieved irrespective of phonological accuracy (e.g., [tuf] for “tooth”). This is perhaps the appropriate analysis for use with adult participants, who demonstrate few production errors during naming. Children, in contrast, produce speech sound errors during language acquisition. Notably the influence of ND on children’s naming accuracy using a semantic analysis has been studied only in a small number of experiments; even fewer have examined productions at the phonological level. In order to understand how ND and other factors might influence children’s naming accuracy at the semantic level, Newman and German (2002) administered a naming task to typical school-aged children. Phonological ND was calculated for two word lists. Results showed that words with low ND were named more accurately (i.e., correct lexical-semantic representations were retrieved) than words with high ND, with the authors interpreting the results as stemming from lexical competition. On words with high ND, children may have not retrieved the appropriate target due to inhibiting competitors. However, in a follow-up study with children with word-finding impairment (German & Newman, 2004), the opposite result was found: words with high ND

This article has been peer- reviewed Keywords children naming neighbourhood density phonological development

Skott E. Freedman

19

JCPSLP Volume 15, Number 1 2013

www.speechpathologyaustralia.org.au