The Phonetics and Phonology of Nasal Gestures Patrice Speeter Beddor University of Michigan Supported by NSF.

Slides:



Advertisements
Similar presentations
Tone perception and production by Cantonese-speaking and English- speaking L2 learners of Mandarin Chinese Yen-Chen Hao Indiana University.
Advertisements

Sounds that “move” Diphthongs, glides and liquids.
SPPA 403 Speech Science1 Unit 3 outline The Vocal Tract (VT) Source-Filter Theory of Speech Production Capturing Speech Dynamics The Vowels The Diphthongs.
Basic Spectrogram & Clinical Application: Consonants
Acoustic Characteristics of Consonants
Plasticity, exemplars, and the perceptual equivalence of ‘defective’ and non-defective /r/ realisations Rachael-Anne Knight & Mark J. Jones.
Glides (/w/, /j/) & Liquids (/l/, /r/) Degree of Constriction Greater than vowels – P oral slightly greater than P atmos Less than fricatives – P oral.
Human Speech Recognition Julia Hirschberg CS4706 (thanks to John-Paul Hosum for some slides)
Acoustic Characteristics of Vowels
Darkness in /l/ as a gradual phonetic property. Evidence from three Catalan dialects Daniel Recasens Universitat Autònoma de Barcelona & Institut d’Estudis.
The sound patterns of language
Phonetic variability of the Greek rhotic sound Mary Baltazani University of Ioannina, Greece  Rhotics exhibit considerable phonetic variety cross-linguistically.
The Sound Patterns of Language: Phonology
Infant sensitivity to distributional information can affect phonetic discrimination Jessica Maye, Janet F. Werker, LouAnn Gerken A brief article from Cognition.
Speech perception 2 Perceptual organization of speech.
Speech Science XII Speech Perception (acoustic cues) Version
Syllables and Stress, part II October 22, 2012 Potentialities There are homeworks to hand back! Production Exercise #2 is due at 5 pm today! First off:
Nuclear Accent Shape and the Perception of Prominence Rachael-Anne Knight Prosody and Pragmatics 15 th November 2003.
Evidence of a Production Basis for Front/Back Vowel Harmony Jennifer Cole, Gary Dell, Alina Khasanova University of Illinois at Urbana-Champaign Is there.
Niebuhr, D‘Imperio, Gili Fivela, Cangemi 1 Are there “Shapers” and “Aligners” ? Individual differences in signalling pitch accent category.
Narrow phonetic transcription
Speech Perception Overview of Questions Can computers perceive speech as well as humans? Does each word that we hear have a unique pattern associated.
Part Two Distinctive features and Natural classes Phonology: The study of the sound system - i.e. how sounds relate to and interact with each other in.
Development of coarticulatory patterns in spontaneous speech Melinda Fricke Keith Johnson University of California, Berkeley.
TEMPLATE DESIGN © Listener’s variation in phoneme category boundary as a source of sound change: a case of /u/-fronting.
Chapter 6 Features PHONOLOGY (Lane 335).
1 Perceptual explanations of articulatory variability in the realisation of the nasal feature for the consonants. J. Vaissière.
Research on teaching and learning pronunciation
Chapter three Phonology
Stop Place Contrasts before Liquids Edward Flemming MIT.
Consonants and vowel January Review where we’ve been We’ve listened to the sounds of “our” English, and assigned a set of symbols to them. We.
Fricatives + Voice Onset Time March 31, 2014 In the Year 2000 Today: we’ll wrap up fricatives… and then move on to stops. This Friday, there will be.
Interarticulator programming in VCV sequences: Effects of closure duration on lip and tongue coordination Anders Löfqvist Haskins Laboratories New Haven,
Present Experiment Introduction Coarticulatory Timing and Lexical Effects on Vowel Nasalization in English: an Aerodynamic Study Jason Bishop University.
Phonetics and Phonology
Diphthongs Five most frequent diphthongs in Māori are /ai ae au ou ao/. Mergers between /ai~ae/ and /au~ou/ [3] (Figure 2). Only one of these mergers is.
Segmental factors in language proficiency: Velarization degree as a signature of pronunciation talent Henrike Baumotte and Grzegorz Dogil {henrike.baumotte,
An Introduction to Linguistics
Speech Perception 4/6/00 Acoustic-Perceptual Invariance in Speech Perceptual Constancy or Perceptual Invariance: –Perpetual constancy is necessary, however,
Perceived prominence and nuclear accent shape Rachael-Anne Knight LAGB 5 th September 2003.
Nasal endings of Taiwan Mandarin: Production, perception, and linguistic change Student : Shu-Ping Huang ID No. : NA3C0004 Professor : Dr. Chung Chienjer.
Infant Speech Perception & Language Processing. Languages of the World Similar and Different on many features Similarities –Arbitrary mapping of sound.
Psych 156A/ Ling 150: Psychology of Language Learning Lecture 5 Sounds III.
1 Speech Perception 3/30/00. 2 Speech Perception How do we perceive speech? –Multifaceted process –Not fully understood –Models & theories attempt to.
Jiwon Hwang Department of Linguistics, Stony Brook University Factors inducing cross-linguistic perception of illusory vowels BACKGROUND.
Phonology The sound patterns of language Nuha Alwadaani March, 2014.
Speech Science Fall 2009 Oct 28, Outline Acoustical characteristics of Nasal Speech Sounds Stop Consonants Fricatives Affricates.
5aSC5. The Correlation between Perceiving and Producing English Obstruents across Korean Learners Kenneth de Jong & Yen-chen Hao Department of Linguistics.
Acoustic Cues to Laryngeal Contrasts in Hindi Susan Jackson and Stephen Winters University of Calgary Acoustics Week in Canada October 14,
Speech Science IX How is articulation organized? Version WS
Evaluating prosody prediction in synthesis with respect to Modern Greek prenuclear accents Elisabeth Chorianopoulou MSc in Speech and Language Processing.
Epenthetic vowels in Japanese: a perceptual illusion? Emmanual Dupoux, et al (1999) By Carl O’Toole.
Sensation & Perception
Phonetic Context Effects Major Theories of Speech Perception Motor Theory: Specialized module (later version) represents speech sounds in terms of intended.
The Effect of Pitch Span on Intonational Plateaux Rachael-Anne Knight University of Cambridge Speech Prosody 2002.
The long-term retention of fine- grained phonetic details: evidence from a second language voice identification training task Steve Winters CAA Presentation.
Lecture 2 Phonology Sounds: Basic Principles. Definition Phonology is the component of linguistic knowledge concerned with rules, representations, and.
1 Cross-language evidence for three factors in speech perception Sandra Anacleto uOttawa.
Parsing acoustic variability as a mechanism for feature abstraction Jennifer Cole Bob McMurray Gary Linebaugh Cheyenne Munson University of Illinois University.
Tongue movement kinematics in speech: Task specific control of movement speed Anders Löfqvist Haskins Laboratories New Haven, CT.
Stop + Approximant Acoustics
Nuclear Accent Shape and the Perception of Syllable Pitch Rachael-Anne Knight LAGB 16 April 2003.
A STUDY ON PERCEPTUAL COMPENSATION FOR / /- FRONTING IN A MERICAN E NGLISH Reiko Kataoka February 14, 2009 BLS 35.
How We Organize the Sounds of Speech 김종천 김완제 위이.
Elaine R. Hitchcocka, Ph.D., Laura L. Koenigb,c, Ph.D.
Why sonority and intra-oral pressure?
Jessica McKee Speech, Language and Hearing Sciences
Speech Perception (acoustic cues)
A Japanese trilogy: Segment duration, articulatory kinematics, and interarticulator programming Anders Löfqvist Haskins Laboratories New Haven, CT.
Presentation transcript:

The Phonetics and Phonology of Nasal Gestures Patrice Speeter Beddor University of Michigan Supported by NSF

Anthony Brasher Rose Letsholo Susan Lin Chandan Narayan Chutamanee Onsuwan Samantha Sefton In collaboration with:

Background The temporal and spatial extent of coarticulatory vowel nasalization is known to be highly variable in ways that depend on phonetic context and prosodic structure, including, among many other factors: obstruent voicing in NC clusters (Malécot, 1960) frication in NC clusters (Ohala & Busà, 1995; Busà, 2007) vowel length (Whalen & Beddor, 1989) syllable structure (Krakow, 1989; Cohn, 1990; Solé, 1995) stress (Vaissière, 1988; Krakow, 1993)

Central hypothesis Hypothesis: The temporal and spatial extent of the nasal gesture is relatively constant across a range of contextual conditions, although the temporal alignment of nasal and oral gestures can differ across contexts in predictable ways. Three-pronged approach: production, perception, phonology For several years, cross-language research on coarticulatory variation in our lab has been motivated by the expectation that the nasal (lowered velum) gesture is more stable than might appear from the vowel nasalization literature.

Central hypotheses Production: In tautosyllabic VN sequences, duration of coarticulatory V nasalization and duration of N co-vary, such that velum lowering in the syllable rhyme is relatively stable across phonetic contexts. Investigated for 2 VN contexts: following C voicing preceding V length Phonology: The co-variation between V and N in production and perception is consistent with widely attested phonological patterns. ~ Perception: Listeners treat N and V as perceptually equivalent and are more sensitive to total nasalization in the rhyme than to the precise alignment of oral and nasal gestures. ~

Variation in temporal alignment Oral closure Vowel Lowered velum Same-sized velum gesture initiated earlier Longer post-N oral constriction C N Shorter N murmur VN C More extensive V nasalization V

This pattern is precisely what we find in voiceless (relative to voiced) contexts in English. Variation in temporal alignment: effects of voicing English Stimuli: /C E nC/ words where coda C = /s t d z/ (e.g., bent, bend, sent, send, sense, dense, dens). Speakers: 5 native speakers of American English. Acoustic measures: N duration, post-N oral closure duration, and duration of V nasalization. In measuring V nasalization, FFT spectra were inspected in 10 ms increments across the V. Nasalization onset = first spectrum with an identifiable low-frequency FN and/or a broadening of F1 BW and lowering of F1 amplitude.

Measuring V Nasalization (a) (c) (d)(e) (f) (b) Relative Amplitude Frequency (in Hz)

N Duration (ms) V Duration (ms) ~ Variation in temporal alignment: effects of voicing AMERICAN ENGLISH Scatterplot of VNC voiceless (filled) and VNC voiced (unfilled) tokens from 5 speakers. Data show inverse relation between N and V durations. (R 2 statistics for trend lines range from.27 to.45.)

Variation in temporal alignment: effects of voicing Speaker 1 Speaker 2 Voiceless Post-N Closure (ms) More extensive V nasalization V Same-sized velum gesture initiated earlier Longer post-N oral constriction C N Shorter N murmur

Later onset of the velum gesture in voiced contexts is likely due to aerodynamic factors, with nasal leakage before full velum closure and the continued raising of the velum after closure (which expands the oral cavity) facilitating voicing (Hayes & Stivers, 2000; Solé, 2007). Ikalanga VNC: As in other Bantu languages, NC sequences are traditionally analyzed as prenasalized N C (Mathangwane, 1999). Most NCs in Ikalanga are voiced, although NC [voiceless] occur in borrowed–including some common–words. Stimuli: Extensive word list with voiced VNCV (e.g., [da n da]) and voiceless VNCV (e.g., [ke n ta]) sequences Speakers: 6 native speakers of Botswanan Ikalanga Variation in temporal alignment: effects of voicing However, not all languages exhibit N shortening in pre-voiceless contexts. Our prediction is that these languages should also not show more extensive vowel nasalization in these contexts.

Variation in temporal alignment: effects of voicing Duration of oral and nasalized V, N, and C in VNCV sequences. IKALANGA ~ No significant effect of voicing on duration of N or coarticulatory nasalization; total duration of nasalization (V and N) is constant across contexts. Duration (ms)

Variation in temporal alignment: effects of vowel length A second context known to give rise to variation in N duration is length of a preceding V. Short V + Long N Shorter oral constriction Long V + Short N V N Oral closure Vowel V N Same-sized velum gesture Lowered velum More extensive V nasalization

Variation in temporal alignment: effects of vowel length Thai (Onsuwan, 2005) Stimuli: Extensive set of CVN and CV:N words produced by 4 speakers American English Stimuli: CVN words with (long) tense and (short) lax Vs (e.g., seen, sin, pain, pen) produced by 6 speakers This pattern is upheld by data from Thai and English.

Variation in temporal alignment: effects of vowel length Duration of N and oral and nasalized V in Thai and English CVN sequences. (Thai data from Onsuwan, 2005) The longer N is, the less extensive is the anticipatory nasalization on V.

Variation in temporal alignment Voicing VNT (shorter N, greater V NAS) vs. VND: English (our lab; Cohn) Vowel length V:N (shorter N, greater V NAS) vs. VN: English, Thai (our lab) Frication VNS (shorter N, greater V NAS) vs. VNT: Italian (Busà 2007) Japanese (Hattori et al. 1958) The inverse relation between duration of N and duration of V nasalization is found for other contexts and in other languages as well: Production summary: Coarticulatory vowel nasalization shows considerable temporal variation across contexts, yet nasal gestures in codas have temporal stability: contexts with shorter N—including contexts that would appear to induce N shortening for substantially different phonetic reasons—have more anticipatory vowel nasalization.

Variation in temporal alignment: Perceptual equivalence between V and N ~In view of the co-variation in production between V and N, for the past few years we have tested the perceptual hypothesis that listeners are more sensitive to total nasalization in the rhyme than to the precise alignment of oral and nasal gestures. ~ ~ Our method uses a variant of the trading relations paradigm in which we orthogonally vary duration of N and temporal extent of V nasalization. We have tested this approach with speakers of different languages, and with various types of stimuli, and I report here one set of findings. Stimuli: original [gaba] and [gamba] Using cross-splicing techniques, we co-varied N duration with duration of vowel nasalization, yielding: (a) 9-step [b-mb] series which incrementally replaced portions of oral pulsing for the stop with nasal murmur (b) 2 degrees of vowel nasalization: slight (20%) and heavier (52%) ~

 should be hardest for listeners  should be easiest for listeners If heard as perceptually equivalent, then: Listeners: 28 American English speakers Perceptual equivalence between V and N ~ Stimuli were paired in 3 types of discrimination pairings to determine whether listeners treat nasality on V and N as perceptually equivalent.  N-only pairs: shorter Ns paired with longer Ns (V NAS held constant)  "Different nasality" pairs: slight V NAS + short N duration paired with heavier V NAS + long N duration  "Similar nasality" pairs: slight V NAS + long N duration paired with heavier V NAS + short N duration ~

Different nasality pair ga S m S ba - ga L m L ba Similar nasality pair ga L m S ba - ga S m L ba Perceptual equivalence between V and N ~ ~ g a a m b a ~ ~ ~

Pooled responses of 28 listeners ~ Prediction: Different Nasality > N-only > Similar Nasality (Slight VNAS+ short N (Heavy VNAS + short N paired with heavier V NAS paired with Slight VNAS + short N) + long N) Perceptual equivalence between V and N ~

Prediction: Different Nasality > N-only > Similar Nasality Perceptual equivalence between V and N ~ ~

Production and perception summary Summary Nasal gestures in codas have temporal stability: contexts with shorter N have more anticipatory vowel nasalization. The perceptual counterpart to (relative) temporal stability of nasalization is that listeners are sensitive to total nasalization across the syllable rhyme, responding to vocalic and consonantal nasality as though perceptually equivalent.

Phonological implications The phonetic findings are consistent with widely attested phonological patterns: Phonologically, nasal codas are robust. Some languages allow only nasals as codas, and languages that have lost coda consonants and for which we have a detailed chronology (e.g., Romance, Chinese; Chen & Wang, 1975 ) show that N loss follows a slower trajectory than does oral stop loss. This robustness is in keeping with temporal stability of the nasal gesture in the syllable rhyme. ~

Phonological implications ~ ~ At the same time, the tight link between N and coarticulatory V nasalization in production and perception should mean that N and V are closely related in phonology, as found, for example, in the historical change VN > V. The phonological evolution of this change is influenced by the same factors that trigger the phonetic co-variation reported here: voiceless obstruents, fricatives, preceding long vowels ( Hajek, 1997; Sampson, 1999 ) are also contexts that are especially likely to lead to phonological VN > V. ~

Phonological implications VN VVVVVVV VVV(VN) Data from Hajek for 9 Northern Italian dialects At the same time, the tight link between N and coarticulatory V nasalization in production and perception should mean that N and V are closely related in phonology, as found, for example, in the historical change VN > V. The phonological evolution of this change is influenced by the same factors that trigger the phonetic co-variation reported here: voiceless obstruents, fricatives, preceding long vowels ( Hajek, 1997; Sampson, 1999 ) are also contexts that are especially likely to lead to phonological VN > V. ~ ~ ~ Dialect VNT VND ~ ~~~ ~~ ~ ~~ ~

Phonological implications There is evidence of this proposed "later stage" in additional experiments that we have conducted with bet-bent stimuli. (Recall that English N is often extremely short in a pre-voiceless context, and especially short before [t].) It's tempting to speculate that co-variation/equivalence may be a phonetic pre- condition to such changes. At a later stage, listener insensitivity to the precise segmental alignment of the target property may lead to a point at which the original source—which is only variably present in the input—is no longer informative. I hypothesize that co-variation in production and equivalence in perception between coarticulatory source (N) and effect (V) serve as a phonetic path to sound changes in which the coarticulatory source is lost over time but its coarticulatory effects remain. ~

Additional perceptual testing The approach was identical to that for /gaba/-gamba/: V nasalization and N duration (0-85 ms in 9 steps) were orthogonally varied and the stimuli were paired in 3 types of discrimination pairings.  N-only pairs  "Different nasality" pairs  "Similar nasality" pairs We speculated that, because of the extreme shortness of N in this context in natural productions, listeners might be especially sensitive to vowel nasalization in these pairs. If so, then Different and Similar Nasality bet-bent pairs might be discriminated equally well—and better discriminated than N-only pairs. /b E t - b E) nt/

Additional perceptual testing / bEt - bE)nt / As expected, listeners attended more to vowel nasalization than to total nasalization across the syllable rhyme in the Similar Nasality pairings (the Similar Nas —Different Nas difference was not significant). However, this outcome is not representative of the responses of all listeners.

Additional perceptual testing / bEt - bE)nt / "Perceptual equivalence" pattern (as for /gaba-gamba/) V NAS dominant pattern

N Duration (stimulus number) Additional perceptual testing: / bEt - bE)nt / V0 V33 V66 %V NAS "Perceptual equivalence" listener Parallel differences are seen in the identification data for these listeners. "V dominant" listener

Phonological implications Taking together the findings for /gaba-gamba/ and /b  t-b  nt/, we find—as expected—that (some) listeners are more sensitive to vowel nasalization in contexts that trigger especially short N and extensive coarticulatory nasalization. Moreover, we find that not all listeners arrive at the same generalization about what constitutes a distinction between VC and VNC. Some listeners are more sensitive to differences in total nasalization while others are particularly sensitive to differences in V nasalization. ~~ Such an account of phonologization differs from the often-cited view of the listener as misperceiving a speaker's intended signal (Ohala, 1981, 1993; Blevins, 2007). Regarding phonologization of coarticulation (here, VN > V), we suggest that listener insensitivity to the precise segmental timing of the target property (nasalization) can—and does—lead to a point at which the original source (N) is less informative than the coarticulatory effect (V). ~ ~

Conclusion Ohala: listeners factor out the acoustic "distortion" due to coarticulation, as long as they detect the source of the distortion: SpeakerListener /  n/ /  n/ distorted by vocal tract into reconstructed as [ E ) n]heard as [ E) n] because of knowledge of coarticulatory effects of nasals

Conclusion Ohala: But if the source of the "distortion" goes undetected, the coarticulatory effects will not be parsed properly: SpeakerListener /  n/ / E ) / distorted by vocal tract into reconstructed as [ E ) (n)]heard as [ E ) ] [ E) ] later produced as In the current approach: the perceptual consequence of co-variation in production is that listeners formulate equivalence classes based on the (lawfully) variable input signal (e.g., [  n], [  n ], [  ]) and they arrive at generalizations (e.g., presence vs. absence of nasality in the syllable rhyme) fully consistent with the input. ~~~ If copied by other speakers, this misperception could become a systematic change in a language.

Conclusion Production: Data from different languages and for different phonetic contexts show that, in tautosyllabic VN sequences, duration of coarticulatory V nasalization and duration of N co-vary, such that velum lowering in the syllable rhyme is relatively stable across phonetic contexts. Phonology: Co-variation in production and equivalence in perception between coarticulatory source and effects may be phonetic precursors to the coarticulatory effect becoming a distinctive property in a language. Perception: Listeners treat N and V as perceptually equivalent and are more sensitive to total nasalization in the rhyme than to the precise alignment of oral and nasal gestures, although in contexts in which N is often absent in production, listeners attend more to V. ~ ~

References Busà, M. G. (2007) Coarticulatory nasalization and phonological developments, in M. J. Solé, P. S. Beddor, and M. Ohala (eds.), Experimental Approaches to Phonology, pp Oxford: OUP. Blevins, J. (2007) Interpreting misperception, in M. J. Solé, P. S. Beddor, and M. Ohala (eds.), Chen, M. Y., Wang, W. S-Y. (1975) Sound change: actuation and implementation. Language 51, Cohn, A. (1990) Phonetic and phonological rules of nasalization. UCLA WPP 76, Hajek, J. (1997) Universals of Sound Change in Nasalization. Boston, MA: Blackwell. Krakow, R. A. (1993) Nonsegmental influences on velum movement patterns: syllables, sentences, stress, and speaking rate. In M. K. Huffman and R. A. Krakow (eds.), Nasals, Nasalization, and the Velum, pp New York: Academic Press. Malécot, A. (1960) Vowel nasality as a distinctive feature in American English. Language 36, Ohala, J. J. and Busà, M. G. (1995) Nasal loss before voiceless fricatives: a perceptually-based sound change. Rivista di Linguistica 7, Mathangwane, J. T. (1999) Ikalanga Phonetics and Phonology: A Synchronic and Diachronic Study. Stanford, CA: CSLI Publications. Onsuwan, C. (2005) Temporal relations between consonants and vowels in Thai syllables. University of Michigan doctoral dissertation. Sampson, R. (1999) Nasal Vowel Evolution in Romance. Oxford: OUP. Solé, M.-J. (1995) Spatio-temporal patterns of velopharyngeal action in phonetic and phonological nasalization. Language and Speech 38, Solé, M.-J. (2007) Compatibility of features and phonetic content: the case of nasalization, in Proceedings of the 16th International Congress of Phonetic Sciences, Saarbrücken, Germany. Vaissière, J. (1988) Prediction of velum movement from phonological specifications. Phonetica 45, Whalen, D. H., Beddor, P. S. (1989) Connections between nasality and vowel duration and height: elucidation of the Eastern Algonquian intrusive nasal. Language 65,

Prediction: Different Nasality > N-only > Similar Nasality (Slight VNAS+ short N (Heavy VNAS + short N paired with heavier V NAS paired with Slight VNAS + short N) + long N) Perceptual equivalence between V and N ~ V S N S -V L N L ~ ~ N only V L N S -V S N L ~ ~ English Ikalanga Both groups of listeners showed the predicted pattern although, not surprisingly, English listeners had greater difficulty discriminating stimuli with similar nasality.

More extensive V nasalization V Variation in temporal alignment Oral closure Vowel Lowered velum Same-sized velum gesture initiated earlier Longer post-N oral constriction C N Shorter N murmur VN C