Application of Acoustical Analysis to Linguistic Problems

Abstract

An oscillographic record of a speech utterance can only be interpreted in the light of linguistic knowledge. The segregation of temporal portions of such a record into speech sounds is based upon phonetic observation. By careful auditory analysis the phonetician decides on the number of meaningful sound patterns which recur in essentially the same auditory form in a given language. These are called speech sounds or phonemes. After an objective record of an utterance has been divided into individual speech sounds, the physical characteristics of the sounds can be described. Much of this work has been done, but not as much as is desirable. Descriptions of certain consonantal sounds, especially, are inadequate. However, one further step of analysis is necessary. The linguist typically divides any utterance into two kinds of “sound-feature,” distinctive and non-distinctive. The distinctive features will be those structural features which give rise to the perception of the phonemes. All others, such as features of individual voice quality, are non-distinctive. We have some information on the structure of vowel phonemes, but practically none for consonantal phonemes.