The Acoustic Characteristics of Aspiration Merger in Korean Mi Jang The University of Texas at Austin Dec 06 2006 1 A three way contrast in Korean stops Lenis Aspirated and Tense stops p t k ph th kh p t k Minimal contrasts for Korean lenis and aspirated stops Lenis pa room Aspirated ph a bang tal moon k ta to fold up thal mask kh ta to dig up 2 Aspiration Merger in Korean The heteromorphemic sequences of lenis stop and h become an aspirated stop in the wordmedial position Kim Renaud 1986 p p law hak learning study p phak jurisprudence p phak t the first ha a clause tha the first clause tha 3 The phonetic correlates of the three way contrast VOT tense stop lenis stop aspirated stop Fo lenis stop aspirated stop tense stop The duration of the stop closure lenis stop aspirated stop tense stop Concomitant tongue and larynx movements and glottal opening found in the stroboscopic cine MRI experiment lenis stop aspirated stop 4 The distribution of the aspirated stop and stop plus h Within the Prosodic Word aspirated stop kipho air bubble iphak admission stop plus h ip hak i phak iph ak 5 The distribution of the aspirated stop and stop plus h Beyond the Prosodic Word a k ikap hanth nun sir hay That purse AP Hanthae Top dislike do AP Hantae dislikes the purse oa hay b Sony ka phado l l girl Nom AP waves ACC like do AP The girl likes waves 6 Previous Analysis Kim 1976 Coalescence analysis which merges the two segments h plus stop or stop plus h into a single aspirated stop directly Iverson Kim Renaud 1998 Aspiration Adjustment In a heterosyllabic cluster containing h and a lenis consonant spread marked features from the coda to the onset a stop h sip ten ho a number sipho ten number b h lenis stop nat to give birth ta DECL natha give birth DECL cf h exists in underlying representation because the Post Obstruent Tensing does not apply to this output i e nat a 7 Problems in the previous analysis Iverson Kim Renaud 1998 Variation depending on the different speech styles They cannot be explained in terms of coalescence extra careful careful casual ipphak iphak iphak admission sokkhi sokhi sokhi fast ADVERBIAL nattha natha give birth DECL Johnson Oh 1995 Both aspirated and tense consonants are lengthened intervocalically and there is little difference in emphatic lengthening between a single laryngeal consonant and two underlying consonants intervocalically To explain the stop gemination Iverson Kim Renaud assumed intermediate processes such as coda neutralization 8 geminate reduction Problems in the previous analysis aspirated stop lenis stop no rightward extension of spread glottis into following lenis stop kip deep kiph deep DECL kip ta kipt a deep DECL kiptha There is no unified explanation about the Aspiration Merger and no production study to provide phonetic evidence for the described pattern of the Aspiration Merger in Korean 9 The similarity between ph and p h phado waves ikap Hanth purse Hantae PART 10 Research Question Is there physical similarity between the two categories If so does the difference between the two categories lie in the timing of the gestures Another question is whether the Aspiration Merger is phonetically complete neutralization or incomplete neutralization Port O Dell 1985 the word final devoicing in German incomplete neutralization 11 The Accentual Phrase in Korean Beckman Pirrehumbert 1986 Pirrehumbert Beckman 1988 The prosodic units larger than a word based on the surface phonetic form of an utterance by looking at suprasegmental features such as intonation and final lengthening Jun 1993 1998 The prosodic hierarchy in Korean Prosodic Word Accentual Phrase Intonational Phrase An AP can have one or more Words and is marked by a tonal pattern phrase initial rise and phrase final rise i e LH LH or HH LH The AP initial strengthening Lenis stop voicing within AP 12 Goals To compare the underlying aspirated stop to stop plus h with and without the Aspiration Merger To vary the prosodic domains and to place the target segments across the Accentual Phrase AAP and within the Accentual Phrase APM 13 Speakers Three males and three females from Seoul Korea Their ages ranged from 25 45 They were recruited from the graduate students at the University of Texas at Austin 14 Speech Material 1 p h a AAP k munkap AP Hanth nun AP sir hay AP S NP That stationery case S NP Hanth Top VP dislike do Hanthae dislikes the stationery case b APM Pyu kap haks kwa AP pap l m k AP S PP Pyungkap name student with VP NP a meal V eat ending With Pyungkap I have a meal 15 p h in AAP 1a 300 280 260 240 220 200 180 160 140 120 100 Hz LHL ms 300 H 600 HL 900 1200 1500 16 p h in APM 1b 300 280 260 240 220 200 180 160 140 120 100 Hz ms 300 600 900 1200 1500 17 Speech Material 2 ph a AAP K ny ka AP phado l l oa hay AP S NP she Nom VP NP waves ACC V like do She likes waves b APM Pada phal un AP a u kil s AP S NP Pada name arms Top VP Adv very VP long past ending Pada s arms were very long 18 ph in AAP 2a 300 280 260 240 220 200 180 160 140 120 100 Hz ms 300 600 900 1200 1500 19 ph in APM 2b 300 280 260 240 220 200 180 160 140 120 100 Hz ms 300 600 900 1200 1500 20 Procedure Subjects were asked to read materials written in Korean orthography at a natural speaking rate The test sentences were presented in random order on separate slides in a timed PowerPoint presentation Speakers were recorded in a sound proofed booth using solid state recorder Maranz PMD 670 in the UT at Austin Phonetics Laboratory 21 Measurements Physically p h and ph have the same oral and glottal gesture but the difference lies in the timing of the gestures In ph the glottal spreading gesture peaks around the release while in p h it would be expected to lie in the h The timing difference in the gestures would be reflected in the duration of noise and the change of intensity at the noise onset 22 The Noise Duration The vocal fold spreading gesture is earlier in ph than in p h and the vocal fold closing phase of the gesture in ph is expected to come sooner as well 23 The Change of Intensity Since spreading the glottis occurs during the closure and it maximizes airflow up into the supralaryngeal vocal tract ph is expected to have greater pressure build up during closure than p h 24 The Preceding Vowel Duration p h with the Aspiration Merger is expected to have longer preceding vowel duration than p h without the Aspiration Merger 25 Hypotheses a In the AAP the duration of noise at the noise onset in p h is longer than that of noise in
View Full Document
Unlocking...