My Version
description
Transcript of My Version
Historical Comparative Acoustics and Prosodic Reanalysis: The Case of Tibetan
Prosodic Typology: State of the Art and Future Prospects
February 24-26, 2010Berlin, German
Nancy J. CaplowUniversity of North Texas
Phonological classification
• Distinction is some dialects are tonal, some are not
• Tibetan dialects are traditionally divided into two categories
• Archaic dialects: non-tonal; rich in consonant clusters
• Innovative dialects: tonal; simplified onsets and codas
Dialect variation and classification
According to recent estimates (Tournadre 2005, 2008)
• More than 220 varieties of Tibetan (or “Tibetic”)
• Some are similar• Others are mutually incomprehensible
• 25 distinct groups
Comparing monosyllabic words
‘arrow’ mda′
‘horse’
rta
‘fire’ me / myi
‘medicine’ sman
Archaic Dialects
Innovative Dialects
Nda ɕta
Nda xtata0 ta
me sman
nyi ʰmɛnme0 me
What are archaic dialects?
• Considered to preserve older forms of the language
• Evidence comes in part from Written Tibetan (WT), developed in the 7th century
• Consonant clusters correspond with consonant clusters in WT
• Innovative dialects • Writing has apparently lost consonants and
developed contrastive tone over time
Correspondence with WT
• For all the dialects (non-tonal or tonal, written or never written)
• Sound patterns consistently correspond with WT.
• The non-tonal archaic dialects show segmental correspondence with WT consonants and consonant clusters.
• The tonal innovative dialects show suprasegmental correspondence with WT consonants and consonant clusters.
Correspondences with WT
boulder brag
brag ʈʰa0k
non-tonaltonalNote: tonal [ʈʰ] + L < WT gr, dr, and sbr
cold grang.mo [graχ.'mo] [ʈʰa0ŋ.mo]
Correspondences with WT
brain ’klad.pa
χlat.'pa lɛEt.tanon-tonal
tonalNote: tonal [l] + H < WT rl, gl, bl, sl
flute gling.bu [χlɛm.'bu] [lɪ Eŋ.bu]
Characteristics of Proto-Tibetan
Archaic dialects• Lacking in tonal contrasts• Rich in consonant clusters
Evidence• Correlations with WT• Geographic distribution
Geographic distribution is intriguing
• Non-tonal Archaic dialects • Spoken at the western and eastern edges of
the Tibetan language area.• Geographically peripheral
• Separated by the tonal Innovative dialects• Spoken across the vast expanse of the
Tibetan plateau.• geographically central
The similarity of the peripheral dialects
• Rule out borrowing• No contact
• Rule out coincidence• Similarities are systematic
• Similarity is due to inheritance of features from a common parent
• Proto-Tibetan was rich in consonant clusters and lacking in tone
Previous accounts of genesis of tone
• Focus on monosyllabic words• Explain tonogenesis in terms of consonantal
edge effects• Voiced onsets ↔ lower pitch register• Voiceless onsets ↔ higher pitch register• Laryngeal final consonants ↔ falling contour
in the preceding vowel• Tone patterns are attributed to correlations
between consonant types and glottal states
Universal physiological correlations
• Hombert, Ohala, and Ewan (1979)• A voiced onset consonant causes lowering of F0
in the vowel that follows.• A voiceless onset consonant causes raising of F0
in the vowel that follows.
• Why?• The voiced / voiceless distinction corresponds to
differences in larynx height, vocal fold tension, and transglottal airflow, all of which effect the rate of vibration of the vocal folds, and thus the pitch of the following vowel.
Implications
• Means that correlations between voicing and F0 are widespread cross-linguistically
• Pitch differences are unintentional, and occur even in languages that do not use tone contrastively
• The differences are of a discernible magnitude
• Available to be phonologized as tonal contrasts
Summarization
• Tone patterns observed on monosyllabic words• High and low pitch register• Level and falling pitch contour
• Some aspects of the tone patterns observed on polysyllabic words remains unaccounted
• Work provides a diachronic / phonetic explanation for tone patterns observed on disyllabic words• Focusing today on disyllabic non-verbs
Tone in Disyllabic Non-Verbs
willow tree lcang.ma
bridge zam.pa
brain klad.pa
Balti(West)
Amdo(East)
ɬtʃaχ.'ma
tʃaŋ.ma
Tokpe Gola(Central)
xtʃaŋ.'ma
zam.'pa sa0m.pa sam.'pa
xlat.'pa lɛEt.ta ᵊlai.'pa
nak.'po
sar.'pʰa
na0k.pu
sam.pa
nox.'χu
sʰo.'ma
black nag.po
new gsar.pa
σ2 always high in Tokpe Golaloose, floppy
flat
young, young person
fire
arrow
gun, rifle
leg, foot
arm, hand
limbs (legs and arms)
lhug.lhug
leb.leb
gzhon.gzhon
me
mda′
me. mda′
rkang.pa
lag.pa
rkang.lag
HH
LH
LH
L
L
LH
HH
LH
HH
Tone vs. Stress in Disyllabic Non-Verbs
willow tree lcang.ma
bridge zam.pa
brain klad.pa
black nag.po
new gsar.pa
Balti(West)
Amdo(East)
ɬtʃaχ.'ma
tʃaŋ.ma
Tokpe Gola(Central)
xtʃaŋ.'ma
zam.'pa sa0m.pa sam.'pa
xlat.'pa lɛEt.ta ᵊlai.'pa
nak.'po
sar.'pʰa
na0k.pu
sam.pa
nox.'χu
sʰo.'ma
Stress reported for Balti
Sprigg (1966, 2002); Bielmeier (1985, 1988)
• Second-syllable stress on disyllabic nouns – Except when that syllable is a non-stress-bearing
suffix• Higher pitch on that second syllable• Few noun-verb minimal pairs
My perceptions• First-syllable stress on disyllabic verbs
Stress reported for Amdo
de Roerich (1958): Rebkong Amdo• Non-tonal• No mention of stress
Sun (1986): Ndzorge Amdo• Stress on the last syllable of polysyllabic
words• “Stronger articulatory force, high falling
tune”
Stress reported for Amdo
Haller (2004): Themchen Amdo• Disyllabic words usually stressed on second
syllable• No evidence that stress is phonemically
contrastive
My perceptions: Rebkong Amdo• Second-syllable stress on nouns• Stronger articulatory force, sharp falling
pitch• First-syllable stress on verbs
Minimal pairs for stress in BaltiN
V
N
V
N
V
A
V
brain
to be tired
rope
to grind
bread
to carry
rare
to clothe
xlat.'pa
'xlat. pa
tʰak.'pa
'tʰak. pa
kʰur.'pa
'kʰur. pa
ʂkøn.'mo
'skøn.ma
Borrowed nouns in Baltihotel
teacher
school
film
hospital
ho.'ʈɨl
ʈi.'tʃɚ
su.'kul
ɸi.'lim
hɑs.pɨ.'ʈal
Disyllabic words in Amdowillow tree
star
shooting star
new
flat
N + verbalizer:
(to) snow
(to) answer
(to) shoot an arrow
lcang.ma
skar.ma
skar.zla
leb.leb
gangs.babs
lan.gyab
mda′.phen
Stress patterns in Proto-Tibetan
• Disyllabic non-verbs in Balti and Rebkong Amdo are stressed on σ2
• Can reconstruct a pattern of σ2 stress for disyllabic non-verbs in Proto-Tibetan
• Disyllabic verbs in Balti and Rebkong Amdo are stressed on σ1
• Can reconstruct a pattern of σ1 stress for disyllabic verbs in Proto-Tibetan
Taking this a step further…
• Reconstruct the acoustic correlates of stress for Proto-Tibetan, by comparing the acoustic correlates of stress in Balti and Rebkong Amdo
• Historical comparative acoustics
• Parameters measured• Pitch
• Pitch slope
• Intensity
• Vowel duration
• Vowel quality
Pitch findings – Balti Non-verbs
• For both speakers• Isolation forms and the sentence frame
• Pitch is significantly higher on σ2, the stressed syllable
Pitch findings – Rebkong Non-verbs
• For speaker AR_04• Isolation forms and the sentence frame
• Pitch is significantly higher on σ2, the stressed syllable
Pitch slope contrasts for speaker AR_05
• For speaker AR_05, it is pitch slope, rather than average pitch, which is relevant.
• Pitch slope is significantly “more downward” in σ2 than in σ1.
• It is this slope contrast that lends prominence to σ2, which we perceive as stress.
F0 findings (pitch & pitch slope)
• For Balti
• Pitch is a significant acoustic correlate of stress for non-verbs produced by both speakers
• For Rebkong Amdo
• Pitch is a significant acoustic correlate of stress for non-verbs produced by AR_04
• For speaker AR_05, it is pitch slope that is the significant acoustic correlate of stress
F0 as a correlate of stress in Proto-Tibetan
• Pitch and pitch slope are both reflexes of F0• F0 is a prominent correlate of σ2 stress in
both Balti and Rebkong Amdo • Prominent correlate of the σ2 stress pattern
reconstructed for non-verbs in Proto-Tibetan
Considering intensity…
For both Balti and Rebkong Amdo
• Intensity does not function as a correlate of stress
• Intrinsic correlations between intensity and vowel height govern intensity patterns
• Low vowels have higher intrinsic intensity
• High vowels have lower intrinsic intensity
• To compare intensity across syllables, it’s necessary to control for contrasts in vowel height
Intensity does not convey stress in PT
• Intensity does not consistently serve as a correlate of stress in non-verbs in Balti and Rebkong Amdo
• Cannot re-construct intensity as a correlate of the stress reconstructed for Proto-Tibetan
Acoustic correlates of stress
Non-verbs• Rebkong Amdo: pitch and pitch slope• Balti: pitchProto-Tibetan• Prominent fundamental frequency
Verbs• Rebkong Amdo: pitch and intensity• Balti: pitch and intensity
ReferencesCaplow, Nancy J. 2009. The role of stress in Tibetan tonogenesis: a study in historical comparative
acoustics. PhD dissertation. University of California Santa Barbara.Hombert, Jean-Marie, John J. Ohala, and William G. Ewan. 1979. Phonetic explanations for the development
of tones. Language 55(1): 37-58.Huang Bufan. 1995. Conditions for tonogenesis and tone split in Tibetan dialects. Linguistics of the Tibeto-
Burman area 18(1): 43-62.Matisoff, James A. 1970. Glottal dissimilation and the Lahu high-rising tone: a tonogenetic case-study.
Journal of the American Oriental Society 90(1): 13-44.Matisoff, James A. 1973. Tonogenesis in Southeast Asia. In L. Hyman (ed.), Consonant types and tone, pp.
71-96. Southern California Occasional Papers in Linguistics. Los Angeles: University of Southern California.
Mazaudon, Martine. 1977. Tibeto-Burman tonogenetics. Linguistics of the Tibeto-Burman Area, 3(2): 1-123.Sun, Jackson T.-S. 1997. The typology of tone in Tibetan. In Chinese Languages and Linguistics IV:
Typological studies of languages in China. Symposium Series of the Institute of History and Philology, Academia Sinica, Number 2. Taipei: Academia Sinica.
Sun, Jackson T.-S. 2001. Variegated tonal developments in Tibetan. Paper presented at the 34th International Conference on Sino-Tibetan languages and linguistics. Kunming, October 2001.
Tournadre, Nicolas. 2005. L’aire linguistique tibétaine et ses divers dialectes. Lalies n°25. Presse de l’école normale supérieure.
Tournadre, Nicolas. 2008. Arguments against the concept ‘Conjunct’ / ‘Disjunct’ in Tibetan. In Brigitte Huber, Marianne Volkart and Paul Widmer (eds.), Chomolangma, Demawend und Kasbek: Festschrift für Roland Bielmeier zu seinem 65. Geburtstag. Band I: Chomolangma, pp. 281-308. Halle (Saale): International Institute for Tibetan and Buddhist Studies (IITBS) GmbH.