Acoustic and perceptual analysis of word-initial consonant clusters in children with normal and disordered phonology

Linda J. Louko, Syracuse University


The purpose of this study was to perceptually and acoustically assess and compare consonant clusters (such as /spr/, /str/) attempted by children with and without disordered phonology. Subjects were sixteen 4 1/2 to 5 year old children (8 girls and 8 boys) in two groups of 8 each. All subjects were monolingual English-speaking children with normal hearing and developmental histories, and age-appropriate language skills. Eight exhibited normal phonological development (NP), while eight had disordered phonology (DP), exhibiting many speech sound production errors. In addition, all of the DP subjects reduced consonant clusters quite consistently.

All subjects were given a consonant cluster production task consisting of 19 real words representing three common three-element clusters and all of the two element clusters and singletons that make up those clusters, paired with the /i/ vowel. All productions were phonetically transcribed from audio-videotapes and were acoustically analyzed using the Computerized Speech Lab (CSL). Acoustic measurements included /s/ duration, stop gap, frication duration, aspiration duration, and voice onset time (VOT).

Few significant differences were found between the stops produced by the DP subjects in their attempts at clusters vs. the stops evident in the cluster productions of the NP Group. Significant differences between the groups did occur for the stop gap measure in /str/, /skr/ and /tr/ clusters. Overall, it did not appear that the DP children were attempting to retain acoustic features of the deleted elements of their attempted clusters. Similarly, there was little statistical evidence of acoustic distinctions between singleton consonants produced by the DP Group in place of clusters and those produced by the NP Group for correct singletons.

There was no clear evidence that any of the children in the DP Group were systematically differentiating between intended singletons and the singletons they were using for clusters, in spite of the fact that they could auditorily discriminate between clusters and singletons. This seems to indicate that their errors were relatively "stable," and also seems to show a lack of productive "phonological knowledge." In addition, each of the DP subjects reduced clusters in a unique way. This suggests that speech-language pathologists need to consider individual patterns of cluster reduction when treating children with disordered phonology.

Future research should study larger numbers of children who exhibit the same pattern of cluster reduction in order to investigate the issue of group differences more thoroughly.