Speech Production Errors as Related to Speech Intelligibility in Deaf Children

Abstract

The mean intelligibility of the recorded speech of 40 deaf children was 18.7% to untrained listeners, with a range from zero to 76%. Segmental and suprasegmental errors in the speech were compared with the intelligibility scores. Errors of place of articulation and voicing remained in essentially the same proportion for all speakers, regardless of intelligibility. Errors of manner of articulation and combined place-and-manner showed a slight systematic decrease from the poorest to the best speakers. Omissions decreased sharply but not quite systematically. Vowel and diphthong errors showed the most marked and systematic decrease as intelligibility improved. The correlation of segmental errors with intelligibility was −0.80. The dispersion of the data was related in part to poor pitch control and excessive variability of intonation. Children of deaf parents generally showed a greater proportion of segmental errors, poorer control of intonation and rate of speech, and poorer speech intelligibility than children of hearing parents. [This research was partially supported by Grant No. 09252 from the National Institute of Neurological Diseases and Stroke.]