Phonological neighbourhood clustering effects on verbal short-term memory


Phonological similarity between words as measured by neighbourhood density counts the number of words differing from the target word by a single phoneme. The word cat has a denser neighbourhood (e.g. hat, cut, at, scat) than the word wag (e.g. bag, wan). However, density does not capture between-neighbour relationships (hat and at are also neighbours of each other, but bag and wan are not). A recent index called the clustering coefficient measures the proportion of neighbours of a word that are also neighbours of each other. In an immediate serial recall task of 6-word lists using non-repeated sampling, lists of words with high clustering coefficients (i.e. many neighbours were neighbours of each other) were better remembered than those with low clustering coefficients (i.e. few neighbours were neighbours of each other). The findings suggest that a network of overlapping similarities among a to-be-remembered word’s phonological neighbourhood enhances recall in short-term memory.

Back to Table of Contents