Reduction and frequency analyses of vowels and consonants in the Buckeye speech corpus

Abstract

In a casual conversation American speakers tend to talk fast and to reduce or change sounds of phonetic symbols defined in an English dictionary which we would find in citation speech style. This study examined how much reduction of pronunciation Americans make from the dictionary prescribed symbols to the real speech ones and how frequently Americans use vowels and consonants in the Buckeye speech corpus. The corpus was recorded by 40 American male and female subjects for an hour per each subject. Results were as follows: First, the Americans produced a reduced number of vowels and consonants in daily conversation. The reduction rate from the dictionary transcriptions to the real transcriptions was around 38.2%. There was not much difference between the vowels and consonants in the reduction. Second, the Americans used more front high and back low vowels while 78.7% of the consonants accounted for stops, fricatives, and nasals. This indicates that the segmental inventory has nonlinear distribution in the speech corpus. From those results we conclude that there is a substantial reduction in the real speech from the dictionary symbols and suggest that English educators consider pronunciation education reflecting the real speech data.