The phonetics of prosody in Yoloxóchitl Mixtec
Christian DiCanio1 cdicanio@buffalo.edu Joshua Benn1 jbenn@buffalo.edu Rey Castillo García2 [email protected]
Departament of Linguistics University at Buffalo1 Secretaria de Educación Pública (Guerrero)2
11/17/17
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 1 / 59 Introduction Motivation
t˜a3 ti1k˜ı14 yaa14 k˜a4 ndi4 yaa14 sa4-ndu3ta3=ndu2 t˜a4 sa3k˜a4 ndi4
How do both instances of /yaa14/ ‘ash’ differ?
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 2 / 59 2 Tone production in pre-pausal/connected speech contexts.
3 Tone production and utterance-level intonational patterns.
Introduction Research questions
Prosody can influence the production of tones in tonal languages.
1 The influence of information structure on tone.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 3 / 59 3 Tone production and utterance-level intonational patterns.
Introduction Research questions
Prosody can influence the production of tones in tonal languages.
1 The influence of information structure on tone.
2 Tone production in pre-pausal/connected speech contexts.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 3 / 59 Introduction Research questions
Prosody can influence the production of tones in tonal languages.
1 The influence of information structure on tone.
2 Tone production in pre-pausal/connected speech contexts.
3 Tone production and utterance-level intonational patterns.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 3 / 59 Background I: Focus and tone How is prosodic prominence marked in tone languages?
1 Phonological Intonational pitch accents or boundary tones can influence tonal contour shapes. e.g. Kipare (Herman, 1996), Shekgalagari (Hyman and Monaka, 2011), Thai (Luksaneeyanawin, 1998)
2 Phonetic Prosodic prominence is marked via phonetic lengthening, register shift, or pitch range expansion. e.g. Mandarin (Xu, 1999), Akan (Kügler and Genzel, 2011)
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 4 / 59 Background I: Focus and tone Register shift
High tones in Mandarin undergo raising and F0 range expansion when in focus (Xu, 1999).
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 5 / 59 Background I: Focus and tone Phonetic marking of domains
Prominent positions in the prosodic hierarchy undergo processes of phonetic enhancement domain-initial strengthening (Fougeron and Keating, 1997; Keating et al., 2000)
focal F0 range expansion (Xu, 1999) and gestural hyperarticulation (Mücke and Grice, 2014) stress-related hyperarticulation (Byrd and Choi, 2010; de Jong, 1995; Krivokapić and Byrd, 2012)
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 6 / 59 Background I: Focus and tone Prosodic lengthening and strengthening
Stressed syllables undergo greater prosodic lengthening under focus than unstressed syllables do. English (Turk and Sawusch, 1997; Turk and White, 1999), Dutch (Cambier-Langeveld and Turk, 1999), Swedish (Heldner and Strangert, 2001)
Intonational pitch accents are aligned with stressed syllables in non-tonal languages (Gussenhoven, 1983). Is focus aligned with stress in tone languages?
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 7 / 59 4 Cho (in press) Laboratory Phonology 8
equation parameter manipulations. Figure 2 shows schematized movement trajectories that correspond to changes in four dynamical parameters (stiffness, target, intergestural timing, and shrinking).Background (Although I: Focus andshrinking tone is not a dynamical parameter but a scaling of two parameters as described below, it will be called ‘parameter’ for the sake of Dynamicalsimplicity.) parameters Figure 3 also visualizes (Cho,idealized kinematic 2006) manifestations of different dynamical specifications in four dynamical parameters by relating some kinematic measures to each other. (Figures 3a-f are adapted from Beckman, et al. (1992).)
(a) Change in Stiffness (b) Change in Target larger Target Target
gesture 2 gesture 2 gesture 1 gesture 1
smaller Target Onset displacement Onset more stiff less stiff
(c) Change in Intergestural Timing (d) Change by shrinking
Target larger Target
gesture 2 gesture 2 gesture 1 gesture 1
displacement Onset smaller Target Onset
truncated earlier phase later phase less stiff more stiff
Time Time
Figure 2. Hypothetical movement trajectories that correspond to a change in each parameter. (a) DiCanio et al (UB/SEP) Prosody in YM 11/17/17 8 / 59 show change in stiffness; (b) change in target; (c) change in intergestural timing,; and (d) change by shrinking. Empty circles indicate the timepoint of the peak velocity attainment.
(1) Stiffness. Variation in articulatory movement duration is thought to be controlled by the stiffness parameter: the stiffer the spring (the articulator), the faster the movement. An idealized pattern in a pure change in stiffness is visualized in Figure 2a, and its corresponding kinematic relationships in Figures 3a-b. If stiffness is the only parameter underlying kinematic differences, there should be a change in peak velocity (i.e., the maximum velocity that the articulator attains during the gesture), but not in displacement (i.e., the amount of spatial distance that the articulator travels), therefore showing vertical distribution of the datapoints (Figure 3a). In addition, (with a change only in stiffness) Background I: Focus and tone
1. How do tones change in prosodically weak/strong environments? at boundaries?
2. Is this phonological or phonetic?
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 9 / 59 Background I: Focus and tone Language background Yoloxóchitl Mixtec (YM)
Otomanguean, spoken in Guerrero, Mexico (∼4000 speakers). Phonological/phonetic fieldwork (Castillo García (2007), DiCanio et al. (2014), DiCanio (submitted a, b), Palancar et al. (2016)).
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 10 / 59 Background I: Focus and tone Language background Yoloxóchitl, Guerrero
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 11 / 59 Background I: Focus and tone Language background
All roots are minimally composed of bimoraic feet, consisting of either monosyllabic stems with long vowels (CVV) or disyllabic stems with shorter vowels (CVCV) (Castillo García, 2007).
No codas.
Glottalization is contrastive: /yaP4a1/‘grey’, /saP3ma4/‘cloth to wrap tortillas’
Final syllables are prominent. Nasal vowels only occur on stem-final syllables. Restricted vowel contrasts on non-final syllables. 9 tones on stem-final syllables, but only 5 on non-final syllables. Final syllable lengthening
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 12 / 59 Background I: Focus and tone Language background Tone is lexical and morphological
Twenty-six tonal melodies are possible on a disyllabic word.
Melody Word Gloss Melody Word Gloss 1.1 ta1ma1 without appetite 4.13 na4ma13 is changing 1.3 na1ma3 to change (intr) 4.14 nda4ta14 is splitting up 1.4 na1ma4 soap 4.24 ya4ma24 Amuzgo person 1.32 na1ma32 I will change myself 4.42 na4ma42 I often pile rocks 1.42 na1ma42 my soap 13.2 hi13ni2 has seen 3.2 na3ma2 wall 13.3 na13na3 has photographed (self) 3.3 na3ma3 to change (tr) 13.4 na13ma4 has piled rocks 3.4 na3ma4 sprout 14.2 na14ma2 I will not change 3.42 na3ma42 I will pile rocks 14.3 na14ma3 to not change 4.1 ka4nda1 is moving (intr) 14.4 na14ma4 to not pile rocks 4.2 na4ma2 I am changing 14.13 na14ma13 to not change oneself 4.3 na4ma3 it is changing 14.14 nda14ta14 to not split up 4.4 na4ma4 is piling rocks 14.42 na14ma42 I will not pile rocks
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 13 / 59 Table 3: San Juan Quiahije Chatino tone sandhi, data from Cruz (2011).
Word Gloss Word Gloss Phrase Gloss knaH ‘snake’ ˜ı ML 3S knaH ˜ı ML ‘his/her snake’ ktaL ‘tobacco’ ˜ı ML 3S ktaL ˜ı ML ‘his/her tobacco’ snaH ‘apple’ ˜ı ML 3S snaH ˜ı 0 ‘his/her apple’ skw˜aL ‘I threw’ ˜ı ML 3S skw˜aL ˜ı 0 ‘I threw him/her’
87 these languages may be much larger than previously assumed, i.e. a high tone with no floating tone 88 is phonologically distinct from one with a floating super-high tone (Cruz & Woodbury 2014). 89 Tone is not merely lexical within Oto-Manguean languages, but often serves a morphological 90 role within verbal and personalBackground morphological I: Focus and tone alternations.Language background Yoloxóchitl Mixtec (YM) illustrates 91 the high functional load of tone in a morphological system, as shown in Table 4. YM has 9 tones, 92 /4, 3, 2,Morphological 1, 13, 14, 24, 42, 32/ (“4” tone is high and “1” is low).
Table 4: Yoloxóchitl Mixtec tonal morphology. Data from Palancar et al. (2016).
Morphology ‘to break’ (tr) ‘hang’ (tr) ‘to change’ (intr) ‘to peel’ (tr) ‘to get wet’ Stem ta3PBi4 tSi3k˜u2 na1ma3 kwi1i4 tSi3i3 NEG ta14PBi4 tSi14k˜u2 na14ma3 kwi14i14 tSi14i3 COMP ta13PBi4 tSi13k˜u2 na13ma3 kwi1i4 tSi13i3 INCOMP ta4PBi4 tSi4k˜u2 na4ma13 kwi4i14 tSi4i4 1S ta3PBi42 tSi3k˜u2=ju1 na1ma32 kwi1i42 tSi3i2
93 Tonal changes in the initial syllable of the YM verb root indicate negation, completive (per- 94 fective) aspect, or incompletive aspect. On polysyllabic words, the penultimate syllable’s tone is 95 replaced by the morphological tone. In monosyllabic words, the morphological tone is simply ap- 96 pended to the left edge of the syllable, creating complex tonal contours. The 1st person singular is 97 marked by tone /2/ at the right edge of the root unless the root contains a final tone /2/ or /1/. In 1 98 this environment, the allomorph is /=ju /. It is possible to combine several tonal morphemes on a 14 (3)2 99 single stem in YM, e.g. /tSi i / ‘I will not get wet.’
100 2.2. Stress
101 Stress is usually fixed in Oto-Manguean languages and universally occurs at the right or left DiCanio et al (UB/SEP) Prosody in YM 11/17/17 14 / 59 102 edge of morphological roots (a xes never receive stress). Most roots/stems are maximally di- 103 syllabic and, as a result, languages are described as having stem-initial or stem-final stress. The 104 presence of stress in Oto-Manguean phonological systems has been motivated on the basis of dis- 105 tributional asymmetries in the word (DiCanio 2008, Hernández Mendoza 2017, Hollenbach 1984). 106 That is, more segmental and tonal contrasts surface on a stressed syllable than on an unstressed 107 syllable. Stress interacts with tone in this way. In some languages, like Mazahua (Knapp Ring 108 2008), tone is only contrastive on the stressed, initial syllable of the word. Of the 94 languages 109 surveyed in §2.1, some description of stress was found for 70, summarized in Table 5. 110 Oto-Manguean languages are almost evenly split in terms of stress alignment. Of the 58 111 languages without monosyllabic root structure, 25/58 (43%) have stem-final stress and 21/58 112 (36%) have stem-initial stress. Stem-penultimate stress is also described for certain Zapotec lan-
4 Background I: Focus and tone Methodological constraints
How do we elicit information structure differences in YM?
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 15 / 59 Background I: Focus and tone Methodological constraints Methodological issues I
Most work on information structure involves reading. e.g. in Mandarin (Chen and Gussenhoven, 2008; Xu, 1999), Guaraní (Clopper and Tonhauser, 2013), Arabic (de Jong and Zawaydeh, 2002), German (Mücke and Grice, 2014), Dutch (Peters et al., 2014), etc.
There is no native literacy in YM and many speakers are not functionally literate in Spanish.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 16 / 59 Background I: Focus and tone Methodological constraints Solution I
Corpus linguist/Syntactian’s solution: Just mine a corpus for natural examples!
Issue: You are not controlling for tone or word structure.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 17 / 59 Background I: Focus and tone Methodological constraints Methodological issues II
A Q&A paradigm following a short story can naturally elicit NPs of different types (narrow, broad, contrastive focus).
e.g. in Akan (Kügler and Genzel, 2011), Guaraní (Clopper and Tonhauser, 2013)
This works well for contexts of narrow or contrastive focus, but not so well for broad focus.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 18 / 59 Background I: Focus and tone Methodological constraints Why?
1. YM (and other Mixtecan languages) use pronominal clitics for animate entities that have been backgrounded.
2. Mixtecan languages are object-dropping.
3. “Describe what happened.” is an odd demand after listening to a text. Speakers attempt to answer it by speculating about the actors’ intents in the text.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 19 / 59 Background I: Focus and tone Methodological constraints Solution II
Use a repetition task?
Issue: Speakers might mimic the prompt and this is less natural than a Q&A paradigm.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 20 / 59 Experiment: Focus and stress Stimuli elicitation for focus - a mixed design
Argument focus (after story) Rey: Who arrived? Speaker: John arrived.
Contrastive focus (after story) Rey: Did Marcus arrive? Speaker: John arrived.
Sentential focus (repetition) Rey: John arrived. Speaker: John arrived.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 21 / 59 Experiment: Focus and stress Focus in Yoloxóchitl Mixtec
1 3 3 3 4 4 4 2 3 4 4 (1) ni -ta Si yu Ba =˜o kwa yu nda Pa =˜o Sentential focus PERF-give father=2S horse hand=2S ‘Your father gave you a horse.’
(2) yu3Ba4=˜o4 ni1-ta3Si3=Ri4 kwa4yu2 nda3Pa4=˜o4 Argument focus father=2S PERF-give=3S horse hand=2S ‘Your father gave you a horse.’
3 4 4 1 3 3 4 4 2 3 4 4 (3) yu Ba =˜o ni -ta Si =Ri kwa yu nda Pa =˜o Corrective focus father=2S PERF-give=3S horse hand=2S ‘Your father gave you a horse.’
(4) ni1-ta3Si3 yu3Ba4=˜o4 kwa4yu2 nda3Pa4=˜o4 PERF-dar padre=2S caballo mano=2S ‘Tu padre te dio un caballo.’ DiCanio(5) yu et al3Ba (UB/SEP)4=˜o4 ni1-ta3Si3=Ri4 kwaProsody4yu in2 nda YM 3Pa4=˜o4 11/17/17 22 / 59 padre=2S PERF-dar=3S caballo mano=2S ‘Te dio un caballo tu padre.’ (6) yu3Ba4=˜o4 ni1-ta3Si3=Ri4 kwa4yu2 nda3Pa4=˜o4 padre=2S PERF-dar=3S caballo mano=2S ‘Fue tu padre que te dio un caballo.’
1 Experiment: Focus and stress Methods
Each answer/response was repeated six times by each respondent across two separate recording sessions (3 reps/session). Recording took place in San Luis Acatlán, a town near Yoloxóchitl. Each condition contained the same 28 target words which possessed nine tonal melodies: 1.1, 1.3, 1.4, 1.42, 3.2, 3.3, 3.4, 4.2, 4.4. Ten native speakers participated; a total of 5,040 utterances were analyzed (504/speaker). Target words segmented and analyzed using a script written in Praat (Boersma and Weenink, 2016).
Normalized F0 trajectories extracted over 5 time points and converted to log-normal values. Onset and vowel duration also extracted. Results analyzed using LMMs with lmertest (Kuznetsova et al., 2013). All reported results are significant.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 23 / 59 Experiment: Focus and stress Results
Since we have a mixed design, we will present the contrastive-argument focus comparison first and then compare them both to the sentential focus condition.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 24 / 59 Experiment: Focus and stress Results Results: Duration I
Onset duration by stress and focus type Vowel duration by stress and focus type
300 300
200 Word position 200 Word position unstressed unstressed
stressed stressed Vowel duration (ms) duration Vowel
100 100 Onset consonant duration (ms) duration Onset consonant
argument contrastive argument contrastive Focus Focus
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 25 / 59 Experiment: Focus and stress Results Results: Duration - comparative
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 26 / 59 Experiment: Focus and stress Results
Table: Durational patterns across focus types. Except for ratios and percentages, all numbers are in milliseconds.
C1 V1 C2 V2 σ1 σ2 σ-ratio Sentential focus 70 77 95 90 141 185 1:1.31 Contrastive focus 77 92 120 99 169 219 1:1.30 Argument focus 76 94 136 107 170 242 1:1.42
Maximum Lengthening 10% 22% 43% 19% 21% 31% under focus
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 27 / 59 Experiment: Focus and stress Results Results: Level and rising melodies
Globally, contrastive focus undergoes raising relative to argument focus.
Effect of focus type on level tonal melodies Effect of focus type on rising tonal melodies
non-final final non-final final
1.0
1 Tonal melody Focus 4.4 0.5 argument
3.3 contrastive
1.1 0.0
0 Tonal melody 3.4 Focus -0.5 argument 1.4
z-score normalized log(F0) normalized z-score contrastive log(F0) normalized z-score 1.3 -1.0 -1
-1.5
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 Time (normalized) Time (normalized)
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 28 / 59 Experiment: Focus and stress Results Results: Falling and complex melodies
Effect of focus type on falling tonal melodies Effect of focus type on complex tonal melody /1.42/
non-final final non-final final
1 1 Focus argument contrastive Focus
0 0 argument Tonal melody contrastive 4.2
3.2 z-score normalized log(F0) normalized z-score -1 log(F0) normalized z-score -1
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 Time (normalized) Time (normalized)
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 29 / 59 Experiment: Focus and stress Results Results: Tone - comparative
For most melodies, tone is lower in sentential focus than argument focus.
Effect of focus type on level tonal melodies Effect of focus type on rising tonal melodies
non-final final non-final final
Tonal melody Focus 1 4.4 1 argument
3.3 contrastive
1.1 sentential
0 0 Focus Tonal melody argument 1.3
contrastive 1.4
z-score normalized log(F0) normalized z-score -1 log(F0) normalized z-score -1 sentential 3.4
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 Time (normalized) Time (normalized)
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 30 / 59 Experiment: Focus and stress Discussion Discussion: duration and focus
Focus lengthens the stressed syllable in YM, but more on the onset than on the vowel. Why?
In Swedish onsets are lengthened when a syllable contains a phonologically short vowel (Heldner and Strangert, 2001).
Given that all vowels in the CVCV disyllables were short, vowel length may have influenced the domain of prosodic lengthening in YM.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 31 / 59 Experiment: Focus and stress Discussion Discussion: tone and focus
Contrastive focus is distinguished from narrow focus by increased F0 range and raising of tonal melodies.
Tones in fronted, focal NPs undergo F0 range expansion and raising relative to tones in sentential focus.
Tonal hyperarticulation, F0 raising, and range expansion mark focus.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 32 / 59 Experiment: Focus and stress Discussion Positional effects
Effect Tones Strong focus x position interaction T1.3, T1.4, T1.42 Weak focus x position interaction T1.1, T4.4, T4.2, T3.2 No focus x position interaction T3.3, T3.4
Why initial tone /1/?
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 33 / 59 Experiment: Focus and stress Discussion
Maintaining the level of /1/ enhances the syntagmatic contrast between it and the following tone.
The distance between tones in a /1.4/ melody is 2.5x as large under contrastive focus as under sentential focus.
Focus induces processes of tonal hyperarticulation that enhance syntagmatic contrast on the word.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 34 / 59 Experiment: Focus and stress Discussion
Focus No focus
Asymmetrical expansion occurs because low tones are near the F0 floor. High vowel displacement functions in a parallel way under different focus conditions (Cho, 2006; Mücke and Grice, 2014).
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 35 / 59 Part II: Phrase-final phenomena Part II: Phrase-final phenomena
Phrase-final position is a domain of articulatory strengthening (Barnes, 2006; Cho, 2006) and where articulatory gestures may reduce their velocity (Krivokapić and Byrd, 2012).
Declination is a universal phonetic process (Gussenhoven, 2004) but phrase-final tonal alternations may be phonological.
How does one separate domain-final effects from global effects in speech production?
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 36 / 59 Part II: Phrase-final phenomena Three possibilities
Impressionistic descriptions of Mixtecan languages mention phrase-final tonal alternations (Pike and Small, 1974; Pike and Wistrand, 1974), but such descriptions do not specify the origin of such effects.
H Declination H#
H Final lowering H#
Final allotony H H#
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 37 / 59 Part II: Phrase-final phenomena Positional effects
Tone production is sensitive to word and phrasal position.
High > Falling in Diuxi Mixtec (Pike and Oram, 1976) Low > Low falling in Ayutla Mixtec (Pankratz and Pike, 1967)
In a more complex tonal system, like YM, one anticipates less sensitivity of tone to phrasal position (Connell, 2017).
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 38 / 59 Part II: Phrase-final phenomena Declination and final lowering
Declination is a universal process in declarative utterances (Gussenhoven, 2004), but there are exceptions in tone languages: 1. It does not occur in a sequence of high tones, e.g. Mandarin (Xu, 1999), Taiwanese (Peng, 1997). 2. It only occurs in a sequence of low tones, e.g. Mambila (Connell, 2017), Yoruba (Laniran and Clements, 2003). 3. It does not occur, e.g. Choguita Rarámuri (Garellek et al., 2015), Embosi (Rialland and Embanga Aborobongui, 2017).
Final lowering occurs in tone languages: 1. It occurs for all tones, e.g. Kipare (Herman, 1996), Moro (Chung et al., 2016), Embosi (Rialland and Embanga Aborobongui, 2017) 2. It only occurs with low and falling tones, e.g. Mambila (Connell, 2017), Taiwanese (Peng, 1997), Akan (Kügler, 2017).
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 39 / 59 Experiment 2: positional effects Methods: positional effects on tone
20 tonal melodies were analyzed (1.1, 1.3, 1.42...) in disyllabic words in non-final contexts (before a PP/Adv) and utterance-final contexts.
Sa4Si24=Ra2 ndi3Si4 ‘He is eating corn.’ Sa4Si24=Ra2 ndi3Si4 Bi3t˜ı3 ‘He is eating corn now.’
The post-target word always had tone /3/. 288 repetitions for each speaker (36 words x 2 conditions x 4 repetitions); 9 speakers. Initial transcription in ELAN and segmentation in Praat. We used a script to analyze F0 dynamics and duration.
F0 was normalized and all data was analyzed using the same methods as experiment 1.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 40 / 59 Experiment 2: positional effects Results II: duration
Onset duration by stress and utterance position Vowel duration by stress and utterance position
non-final final non-final final
400 400
300 300
200 200 Vowel duration (ms) duration Vowel Onset consonant duration (ms) duration Onset consonant 100 100
0 0
penult ultima penult ultima penult ultima penult ultima Syllable position in word Syllable position in word
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 41 / 59 Experiment 2: positional effects Results II: level tone melodies
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 42 / 59 Experiment 2: positional effects Results II: falling and rising melodies
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 43 / 59 Experiment 2: positional effects Results II: melodies with final contours
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 44 / 59 Experiment 2: positional effects Discussion - experiment 2
1. Tone /4/ undergoes final raising.
2. Tone /3/ does not change.
3. Final lowering occurs only for lower register tones. Tones /2, 1/ and falling tones /42, 32/ lower in utterance-final position.
4. Rising tones ((/13, 24/) have distinct allotones in non-utterance-final position.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 45 / 59 Experiment 3: declination and tone Methods: Declination
We analyzed sentences that consisted of only level tone sequences. 10 sentences between 4-7 syllables in length; 2 with tone /4/, 4 with tone /3/, 4 with tone /1/. 10 sentences x 4 repetitions; 9 speakers. Initial transcription in ELAN and segmentation in Praat. We used a script to analyze F0 dynamics and duration.
F0 was normalized and all data was analyzed using the same methods as experiment 1. Two statistical methods to disambiguate declination from final lowering: trajectory modelling with and without utterance-final syllable.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 46 / 59 Experiment 3: declination and tone Results - declination
Occurs in sequences of tone /1/ and /3/, but not with tone /4/.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 47 / 59 Experiment 3: declination and tone Resultados - declination modelling
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 48 / 59 Discussion Declination, final lowering, or allotony?
All of them occur in YM!
Final raising occurs with tone /4/, final lowering occurs with lower register tones; declination occurs with tones /3/ and /1/; and positional allotony occurs with rising tones /13, 24/.
Declination is a distinct phenomenon from final lowering.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 49 / 59 Discussion Are there boundary tones?
No. If they were to exist, we would have to stipulate that they be extensions of the same preceding tones, i.e. H% only after /4/.
Non-final position Final position
Tonal changes in utterance-final position result from tonal hyperarticulation which expands the tonal range (Krivokapić and Byrd, 2012).
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 50 / 59 Discussion Final allotony
F0 rises require more time than level or falling trajectories, thus we might expect that they be limited to contexts with longer phonetic duration, e.g. phrase-final position (Sundberg, 1979; Zhang, 2004).
Allotony results from durationally-induced F0 levelling. Levelling is induced via articulatory undershoot (Parrell, 2014; Mücke and Grice, 2014).
high rising mid non-final position
final position
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 51 / 59 Discussion Variation in the production of rising tones
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 52 / 59 Conclusions Conclusions I
Both narrow focus (at utterance-initial position) and utterance-final position involve tonal range expansion and hyperarticulation.
The type of range adjustment differs with respect to low tones. Final lowering occurs with low or falling tones in utterance-final position, but no active lowering occurs for low tones under focus.
Phrase-final effects occur in addition to range expansion.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 53 / 59 Conclusions Conclusions II
Prosody in YM is marked primarily by adjustments to F0 range and hyper/hypoarticulation (de Jong, 1995; de Jong and Zawaydeh, 2002).
Change in range = postural target adjustment?
H
H M M L L
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 54 / 59 Conclusions Future plans
1 Parallel research on Itunyoso Triqui (IT) prosody.
2 Tone production in the YM and IT corpora.
3 EMA research in the UB Phonlab on the supralaryngeal articulation of information structure in English and Korean.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 55 / 59 Conclusions Acknowledgements
Support from NSF DEL/RI grant 1603323, Understanding Prosody and Tone Interactions through Documentation of Two Endangered Languages
Team Mixtec: Rey Castillo García (SEP, México), Jonathan Amith (Gettysburg College), and Joshua Benn (University at Buffalo)
Commentary from audiences at CILLA VIII and UC Santa Cruz.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 56 / 59 Appendix Results - declination by speaker
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 57 / 59 Appendix Prosodic marking
Accentual marking of heads/edges – intonational pitch accents are attracted to prominent positions in the prosodic hierarchy or on constituents with narrow focus (Gussenhoven, 2004; Pierrehumbert and Beckman, 1988). Non-accentual phonological marking of domains – prominent positions in the prosodic hierarchy license the application of specific phonological processes, e.g. tone spreading domains (Hsu and Jun, 1996; Hyman, 1990; Hyman and Monaka, 2011; Lee, 2014), positional neutralization (Barnes, 2006). Phonetic marking of domains – prominent positions in the prosodic hierarchy undergo processes of phonetic enhancement, e.g. domain-initial strengthening (Fougeron and Keating, 1997; Keating et al., 2000), focal F0 range expansion (Xu, 1999), stress-related hyperarticulation (Byrd and Choi, 2010; de Jong, 1995; Krivokapić and Byrd, 2012).
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 58 / 59 References Barnes, J. (2006). Strength and Weakness at the Interface: Positional Neutralization in Phonetics and Phonology. Berlin, New York: Mouton de Gruyter. Boersma, P. and Weenink, D. (2016). Praat: doing phonetics by computer [computer program]. www.praat.org. Byrd, D. and Choi, S. (2010). At the juncture of prosody, phonology, and phonetics – the interaction of phrasal and syllable structure in shaping the timing of consonant gestures. In Fougeron, C., Kuehnert, B., Imperio, M., and Vallee, N., editors, Laboratory Phonology 10, pages 31–59. Mouton de Gruyter. Cambier-Langeveld, T. and Turk, A. E. (1999). A cross-linguistic study of accentual lengthening: Dutch vs. English. Journal of Phonetics, 27:255–280. Castillo García, R. (2007). Descripción fonológica, segmental, y tonal del Mixteco de Yoloxóchitl, Guerrero. Master’s thesis, Centro de Investigaciones y Estudios Superiores en Antropología Social (CIESAS), México, D.F. Chen, Y. and Gussenhoven, C. (2008). Emphasis and tonal implementation in Standard Chinese. Journal of Phonetics, 36(4):724–746. Cho, T. (2006). Manifestation of prosodic structure in articulatory variation: Evidence from lip kinematics in English. In Goldstein, L. M., Whalen, D. H., and Best, C. T., editors, Laboratory Phonology 8: Varieties of Phonological Competence. Berlin, New York: Mouton de Gruyter. Chung, Y., Piccinini, P. E., and Rose, S. (2016). The interaction of polar question and declarative intonation with lexical tone in Moro. In Proceedings of Speech Prosody 8. Clopper, C. G. and Tonhauser, J. (2013). The prosody of focus in Paraguayan Guaraní. International Journal of American Linguistics, 79(2):219–251. DiCanio et al (UB/SEP) Prosody in YM 11/17/17 58 / 59 References Connell, B. (2017). Tone and Intonation in Mambila. In Downing, L. J. and Rialland, A., editors, Intonation in African Tone Languages, pages 132–166. Berlin/Boston: De Gruyter. de Jong, K. and Zawaydeh, B. (2002). Comparing stress, lexical focus, and segmental focus: patterns of variation in Arabic vowel duration. Journal of Phonetics, 30:53–75. de Jong, K. J. (1995). The supraglottal articulation of prominence in English: Linguistic stress as localized hyperarticulation. Journal of the Acoustical Society of America, 97(1):491–504. DiCanio, C., Amith, J. D., and Castillo García, R. (2014). The phonetics of moraic alignment in Yoloxóchitl Mixtec. In Proceedings of the 4th Tonal Aspects of Language Symposium. Nijmegen, the Netherlands. DiCanio, C., Benn, J., and Castillo García, R. (submitteda). The phonetics of information structure in Yoloxóchitl Mixtec. DiCanio, C., Zhang, C., Whalen, D. H., Amith, J. D., and Castillo García, R. (submittedb). Phonetic structure in Yoloxóchitl Mixtec consonants. Fougeron, C. and Keating, P. A. (1997). Articulatory strengthening at edges of prosodic domains. Journal of the Acoustical Society of America, 101(6):3728–3740. Garellek, M., Aguilar, A., Caballero, G., and Carroll, L. (2015). Lexical and post-lexical tone in Choguita Rarámuri. In Proceedings of the 18th International Congress of the Phonetic Sciences, pages 254–258. University of Edinburgh. Gussenhoven, C. (1983). Focus, Mode, and the Nucleus. Journal of Linguistics, 19(2):377–417. Gussenhoven, C. (2004). The Phonology of Tone and Intonation. Research Surveys in Linguistics. Cambridge University Press. Heldner, M. and Strangert, E. (2001). Temporal effects of focus in Swedish. Journal of Phonetics, 29(329–361). DiCanio et al (UB/SEP) Prosody in YM 11/17/17 58 / 59 References Herman, R. (1996). Final lowering in Kipare. Phonology, 13:171–196. Hsu, C.-S. and Jun, S.-A. (1996). Is tone sandhi group part of the prosodic hierarchy in Taiwanese? Journal of the Acoustical Society of America, 100:2824. Hyman, L. M. (1990). Boundary tonology and the prosodic hierarchy. In Inkelas, S. and Zec, D., editors, The phonology-syntax connection, pages 109–125. Chicago: University of Chicago Press. Hyman, L. M. and Monaka, K. C. (2011). Tonal and Non-tonal Intonation in Shekgalagari. In Frota, S., Elordieta, G., and Prieto, P., editors, Prosodic categories: Production, Perception, and Comprehension, pages 267–290. Springer Verlag. Keating, P., Cho, T., Fougeron, C., and Hsu, C.-S. (2000). Domain-initial articulatory strengthening in four languages. In Local, J., Ogden, R., and Temple, R., editors, Papers in laboratory phonology 6, chapter 10. Cambridge University Press. Krivokapić, J. and Byrd, D. (2012). Prosodic boundary strength: an articulatory and perceptual study. Journal of Phonetics, 40(3):430–442. Kügler, F. (2017). Tone and intonation in Akan. In Downing, L. J. and Rialland, A., editors, Intonation in African Tone Languages, pages 89–129. Berlin/Boston: De Gruyter. Kügler, F. and Genzel, S. (2011). On the prosodic expression of pragmatic prominence: The Case of Pitch Register Lowering in Akan. Language and Speech, 55(3):331–359. Kuznetsova, A., Brockhoff, P. B., and Christensen, R. H. B. (2013). lmerTest (R package). Laniran, Y. O. and Clements, G. N. (2003). Downstep and high raising: interacting factors in yoruba tone production. Journal of Phonetics, 31:203–250. Lee, S. J. (2014). Domains of H tone spreading and the noun class prefix in Xitsonga. Southern African Linguistics and Applied Language Studies, 32(1):21–34. DiCanio et al (UB/SEP) Prosody in YM 11/17/17 58 / 59 References Luksaneeyanawin, S. (1998). Intonation in Thai. In Hirst, D. and Di Cristo, A., editors, Intonation Systems: A survey of twenty languages, chapter 21, pages 376–394. Cambridge University Press. Mücke, D. and Grice, M. (2014). The effect of focus marking on supralaryngeal articulation - Is it mediated by accentuation? Journal of Phonetics, 44:47–61. Palancar, E. L., Amith, J. D., and Castillo García, R. (2016). Verbal inflection in Yoloxóchitl Mixtec. In Palancar, E. L. and Léonard, J.-L., editors, Tone and Inflection: New Facts and New Perspectives, chapter 12, pages 295–336. Mouton de Gruyter. Pankratz, L. and Pike, E. V. (1967). Phonology and Morphotonemics of Ayutla Mixtec. International Journal of American Linguistics, 33(4):287–299. Parrell, B. (2014). Dynamics of consonant reduction. PhD thesis, University of Southern California. Peng, S.-h. (1997). Production and perception of Taiwanese tones in different tonal and prosodic contexts. Journal of Phonetics, 25:371–400. Peters, J., Hanssen, J., and Gussenhoven, C. (2014). The phonetic realization of focus in West Frisian, Low Saxon, High German, and three varieties of Dutch. Journal of Phonetics, 46:185–209. Pierrehumbert, J. B. and Beckman, M. E. (1988). Japanese tone structure. Cambridge: MIT Press. Pike, E. V. and Oram, J. (1976). Stress and tone in the phonology of Diuxi Mixtec. Phonetica, 33:321–333. Pike, E. V. and Small, P. (1974). Downstepping terrace tone in Coatzospan Mixtec. In Brend, R. M., editor, Advances in Tagmemics, North Holland Linguistic Series, Number 9, pages 105–134. Amsterdam: North Holland. DiCanio et al (UB/SEP) Prosody in YM 11/17/17 58 / 59 Appendix Pike, E. V. and Wistrand, K. (1974). Step-up terrace tone in Acatlán Mixtec. Advances in tagmemics, pages 81–104. Rialland, A. and Embanga Aborobongui, M. (2017). How intonations interact with tones in Embosi (Bantu C25), a two-tone language without downdrift. In Downing, L. J. and Rialland, A., editors, Intonation in African Tone Languages, pages 195–222. Berlin/Boston: De Gruyter. Sundberg, J. (1979). Maximum speed of pitch changes in singers and untrained subjects. Journal of Phonetics, 7:71–79. Turk, A. and White, L. (1999). Structural influences on accentual lengthening in english. Journal of Phonetics, 27(2):171–206. Turk, A. E. and Sawusch, J. R. (1997). The domain of accentual lengthening in American English. Journal of Phonetics, 25:25–41. Xu, Y. (1999). Effects of tone and focus on the formation and alignment of F0 contours. Journal of Phonetics, 27:55–105. Zhang, J. (2004). The role of contrast-specific and language-specific phonetics in contour tone distribution. In Hayes, B., Kirchner, R., Kirchner, R. M., and Steriade, D., editors, Phonetically-based Phonology. Cambridge University Press.
DiCanio et al (UB/SEP) Prosody in YM 11/17/17 58 / 59