The phonetic variation of plosives in Estonian
Date
2024-07-15
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Kui me räägime, muutuvad häälikud ebamäärasemaks ja üksteisele sarnasemaks võrreldes sellega millised nad on eraldi hääldades. Selline sarnasemaks muutumine – koartikulatsioon ehk kaasahääldus – esineb kõigis maailma keeltes, igaühes pisut isemoodi, olenevalt sellest, millised häälikud keeles on. Oma doktoritöös tegelesin sulghäälikute koartikulatsiooni akustilise uurimisega. Sulghäälikud erinevad teistest häälikutest selle poolest, et kui teiste häälikute hääldamisel toimub kõnetraktis pidev õhuvool, siis sulghäälikute hääldamisel on eristatavad faasid: õhuvool katkestatakse, tekib sulg, heli kaob ja seejärel vallandub sulg väikse pahvatusega. Koartikulatsiooni tõttu võib nende faaside hääldus nõrgeneda ja vallandumisfaas võib ka kaduda. Väitekirjas uurisin, kuidas käituvad seotud kõnes eesti keele sulghäälikud /k p t/.
Töös uurisin seotud kõne korpuste põhjal sulghäälikute hääldusvariantide ehk allofoonide esinemist ja seda mõjutavaid tegureid nagu häälduskoht ja välde. Leidsin et eesti keele lühikesed sulghäälikud häälduvad seotud kõnes isoleeritud hääldusega võrreldes palju nõrgemalt. See on tingitud ilmselt lühikestest kestustest, mille jooksul ei jõua kõneleja kõiki hääldusliigutusi täielikult teha või teeb mitu tükki korraga. Nõrgenemine esines nii spontaanses kui loetud kõneregistris. Pikki sulghäälikuid hääldati pisut tugevamalt ja nende vallandumised ära ei kadunud, kuid päris helitud allofoonid olid vähemuses nii lühikeste kui pikkade häälikute puhul. Eri välteid eristas peamiselt hääliku kestus. Vallandumisfaasi kestus, mis mõnedes teistes keeltes eri väldetes erineb, oli eesti keeles kõigis pikkuskategooriates sama kestusega. Erinevused väldete vahel ilmnesid vallandumisfaasi spektris. Teises ja kolmandas vältes olevate sulghäälikute vallandumise spektris oli viiteid nende pisut tugevamale hääldusele. Samas vallandumise intensiivsus oli absoluutskaalal kõrgem hoopis lühikestes häälikutes.
Töö tulemused kinnitasid universaalsete tendentside kehtimist eesti keeles, kuid lisaks andsid uut infot näiteks varem peaaegu uurimata vallandumisfaasi akustika kohta ja andsid ideid edasisteks uurimusteks.
When we speak, the sounds of speech become more similar to each other compared to when they are pronounced separately. This phenomenon of sound segments becoming similar – coarticulation – occurs in all languages of the world, a little different in each. Plosives differ from other sounds in that, while there is a constant airflow in the speech tract when pronouncing other sounds, there are distinct phases when pronouncing plosives: the airflow is interrupted, a closure is created, the sound disappears, and then the closure opens with a slight burst. Due to coarticulation, the pronunciation of these phases may be reduced and the burst phase may be lost. In my dissertation, I studied the behaviour of Estonian plosives /k p t/ in connected speech. In the thesis I studied allophonic variation and the factors influencing it, like place of articulation and quantity. degree. The material was taken from large speech corpora. The work showed that the short plosives in Estonian language are pronounced much weaker in connected speech compared to the isolated pronunciation. This is probably due to short durations during which the speaker is unable to complete all the articulatory movements or makes several at a time. Weakening occurred in both spontaneous and read registers. Long plosives were a little more stable and their bursts did not disappear, but completely voiceless tokens were outnumbered in both short and long plosives. Different quantity degrees were mainly distinguished by the duration of the sound. The duration of the burst phase, which differs between quantity degrees in some other languages, was the same for all quantity degrees in Estonian. Differences were evident in the spectrum of the burst phase. There were hints to slightly stronger pronunciation in the burst spectrum in the second and third quantity degrees. The burst intensity, on the other hand, was higher in singletons. The results of the work confirmed the validity of universal tendencies in Estonian, but also provided new information about, for example, the acoustics of the previously unexplored burst phase and provided ideas for further studies.
When we speak, the sounds of speech become more similar to each other compared to when they are pronounced separately. This phenomenon of sound segments becoming similar – coarticulation – occurs in all languages of the world, a little different in each. Plosives differ from other sounds in that, while there is a constant airflow in the speech tract when pronouncing other sounds, there are distinct phases when pronouncing plosives: the airflow is interrupted, a closure is created, the sound disappears, and then the closure opens with a slight burst. Due to coarticulation, the pronunciation of these phases may be reduced and the burst phase may be lost. In my dissertation, I studied the behaviour of Estonian plosives /k p t/ in connected speech. In the thesis I studied allophonic variation and the factors influencing it, like place of articulation and quantity. degree. The material was taken from large speech corpora. The work showed that the short plosives in Estonian language are pronounced much weaker in connected speech compared to the isolated pronunciation. This is probably due to short durations during which the speaker is unable to complete all the articulatory movements or makes several at a time. Weakening occurred in both spontaneous and read registers. Long plosives were a little more stable and their bursts did not disappear, but completely voiceless tokens were outnumbered in both short and long plosives. Different quantity degrees were mainly distinguished by the duration of the sound. The duration of the burst phase, which differs between quantity degrees in some other languages, was the same for all quantity degrees in Estonian. Differences were evident in the spectrum of the burst phase. There were hints to slightly stronger pronunciation in the burst spectrum in the second and third quantity degrees. The burst intensity, on the other hand, was higher in singletons. The results of the work confirmed the validity of universal tendencies in Estonian, but also provided new information about, for example, the acoustics of the previously unexplored burst phase and provided ideas for further studies.
Description
Väitekirja elektrooniline versioon ei sisalda publikatsioone