LINGUISTIC ASPECTS OF RESPIRATORY PHENOMENA

The respiratory system is the source of power for nearly all speech sounds. There are a few languages, such as Zulu and Xhosa, in which the acoustic energy in some segments is derived from movements of the tongue, and there is a slightly larger number of languages using upward or downward movements of the glottis to produce an increase or decrease of the pressure of the air in the vocal tract. Apart from these cases, speech sounds are produced by the movement of air that occurs with decreasing volume of the lungs. The respiratory muscles involved have been discussed in a series of papers summarized previously (Ladefoged, 1962). In general, we found that during the first part of an utterance beginning after a deep inspiration, the external intercostals remain in action, regulating the pressure of the air below the vocal cords by checking the descent of the rib cage, thus counteracting the relaxation pressure. As the lung volume decreases, the action of the external intercostals diminishes and eventually ceases altogether when the lung volume is slightly less than the volume after a normal inspiration, at which moment the relaxation Pressure is sufficient to provide the power required for a normal conversational utterance. From then on, expiratory activity is needed to maintain the pressure below the vocal cords, and accordingly the internal intercostals come into action with gradually increasing intensity. When the lung volume is a little below that at the end of a normal expiration, the action of the internal intercostals is supplemented by various other muscles, such as the external obliques. rectus abdominis, and latissimus dorsi. An example of the electromyographic data obtained in these investigations is reproduced in FIGURE 1 (Draper er al., 1958). This part of the record shows the changeover from the use of one set of muscles, the external intercostals, to the use of another set, the internal intercostals. Using adequate techniques to record subglottal pressure and lung volumes, it should be possible to derive the relaxation pressure curve from data of this kind, since at the moment when the external intercostal activity ceases and the internal intercostal activity begins, the subglottal pressure must be equd to the relaxation Pressure for that lung volume. The respiratory activity outlined above is of little linguistic interest in that it occurs in the pronunciation of all utterances irrespective of their linyistic components. There are, however, some variations in respiratory activity that have linguistic correlates. In particular, we should note that all stressed syltables are accompanied by an increase in subglottal pressure produced by the actlon of the respiratory muscles: many pitch changes are also accompanied by an Increase in subglottal pressure, and in some languages certain consonants are associated with particular patterns of respiratory activity. I have not had the facilities to investigate the muscular activtiy involved in all these phenomena; in the remainder of this paper I shall discuss them in terms of the subglottal Pressure variations that I have been able to observe. For the limited purpose of studying the subglottal pressure variations associated with the linguistic components of an utterance, we have found that variations in tracheal pressure can be simply correlated with suitable records of esoPha@ pressures. We are a t present obtaining esophageal pressure records from a small latex balloon, 1.5 cm in diameter and 2.0 cm long, sealed to the end of a Plythene