A perceptual study of the voicing feature of stop consonants after type II and III laser cordectomies

ABSTRACT

This article deals with the intelligibility of stop consonants in onset position (initial and intervocalic) as pronounced by 10 French speakers treated by a type II or III cordectomy. These surgeries cause a less efficient glottal closure and a vibratory asymmetry. The voicing contrast in consonants may be altered. The patients’ vocal samples were made up of non-words of the CVCVCVC type, where C = /p, t, k, b, d, g/ and V = /a, i, u/(n = 360), and were given to 22 French listeners for evaluation. Two identification tests were created and the listeners were asked to choose the consonants they thought they recognized (n = 7920). The results were presented in the form of a confusion matrix highlighting possible confusions of voicing and/or places of articulation. We showed high identification rates of consonants, which revealed an intact voicing contrast in this dysphonic population.

KEYWORDS:

• Identifications
• laser cordectomy
• stop consonants
• voicing contrast

Acknowledgements

I would particularly like to thank Jean-Marc Beltzung for his comments and unwavering support. I also thank Mark Tutton and Michael Kelly for their valuable corrections and the members of the HEGP’s unit ‘‘Voice’’, Daniel Brasnu, Stephane Hans and Lise Crevier Buchman, for welcoming us into the laryngology department, where we met the patients. I alone am responsible for any errors this article may contain.

Declaration of interest

The author reports no conflicts of interest.

Notes

¹ The relative intensity of the burst is calculated in order to normalize the differences between speakers as well as recording conditions. This index represents the difference between the value (average of the maximum peaks of amplitude) calculated in the release/burst of the consonant and the peak of intensity of the following vowel (Flack, 2007). Values near 0 indicate that the intensities of the burst and the vowel are close, that is to say, the intensity of the release of the consonant is high. In contrast, high and positive values indicate that the release of the stop is much less intense than the subsequent vowel. In other words, the higher a positive value is, the lower the intensity of the burst of the consonant is in relation to the following vowel.

² VOT or Voice Onset Time refers to laryngeal timing, and allows the contrast of voiced stops with their unvoiced counterparts. Laryngeal timing defines the relationship between the movement of the glottis and the establishment of the supra-laryngeal articulations. The measure of VOT corresponds to the “time interval between the release of the consonant and the onset of the first laryngeal vibrations” (Lisker & Abramson, 1964). In French, VOT is negative for voiced consonants and positive for unvoiced ones.

³ In a previous study, we showed an increase in the oral airflow, the intraoral pressure and the estimated transglottal pressure in patients treated by type II-III laser cordectomies compared to healthy subjects. We interpreted these results as a compensatory mechanism applied to fight against the laryngeal resistance and to facilitate the initiation of the voicing (Wallet 2015).

⁴ Even if the sample size needed is of 10 listeners with a confidence level at 95%, we doubled the sample size to be sure that their answers truly reflect the population.

⁵ Although the questionnaires can be considered as solid screening tools, we are aware that they do not replace hearing evaluations. The lack of such evaluations could be seen as a limitation of this perceptive study.