Post-oralized and devoiced nasals in PanaráMyriam LapierreSymposium on Amazonian Language IIApril 8, 2017University of California, Berkeley

1

2

Goal1. To present a case of post-nasal devoicing as

evidence of a phonetically unnatural sound change in Panará (Jê)

2. To discuss the implications of this sound change for the internal classification of Jê languages

Sound changes are phonetically grounded• Understanding mechanisms of speech production and perception can help us

understand recurring sound patterns in unrelated languages (Ohala 1993, Ohala & Ohala 1993)� Most likely arise from language universal factors, i.e. physiological and psychological

factors common to all humans

• A series of phonetically natural sound changes may collectively result in a system that is not phonetically natural (Hayes 1999; Hyman 2001)

• Optimality Theory: Incorporates two main principles� Markedness: militates toward phonetic well-formedness� Faithfulness: militates for contrast preservation� Sound change: reordering of constraints results in (phonetic) well-formedness in

some other aspect of the system

• In summary: Sound changes should be phonetically natural.

3

Post-nasal voicing (ND)

• The phenomenon of post-nasal voicing is of particular interest in the debate on whether sound changes should be phonetically natural or not.

• Cross-linguistic surveys claim that the most common process is for the postnasal stop to become voiced (Herbert 1986; Rosenthal 1989, Steriade 1993, and Hyman 2001)

• A number of authors have proposed a universal bias against post-nasal devoicing (*NT) (Pater 1996; Hayes 1999; Hayes & Stivers 2000) .� Articulatorily grounded in aerodynamic principles

4

… but sometimes NT• Evidence of post-nasal devoicing in a number of languages

• Mostly from Bantu languages� Tswana (Coetzee et al. 2007, Coetzee & Pretorius 2010)

� 4/12 speakers of Tswana show clear and consistent post-nasal devoicing� an additional 3/12 show infrequent post-nasal devoicing

� Shekgalagari (Solé et al. 2010)� impressionistic description

• Beguš (forthcoming) surveys eight cases of post-nasal devoicing� Argues that no case derives from a single atypical sound change� Rather, from a series of 2-3 natural and well-formed sound changes

5

NT in Panará• Post-nasal devoicing is categorical

� Productive for all speakers� Sound change is completely phonologized� If voicing is defined as 50% of the duration is voiced, then no stop in the corpus

can be classified as voiced.� Following Westbury & Keating 1986, Hayes 1999, Coetzee et al 2007

• Contributes to the body of literature providing evidence of unnatural sound changes

• Specifically, this sound change does not seem to result in an improvement of some other aspect of the grammar� Or to arise from a series of phonetically-grounded sound changes

6

Typological data on NT

• Out of a sample of 454 languages from UPSID (Maddieson & Ladefoged 1993)� 55 languages (12%) have prenasalized stops� 8 languages have devoiced and prenasalized stops (<2%)

• Why are these segments so rare?

7

alveolar closure

vocal folds vibrating

velum closed

lungs 8

alveolar closure

vocal folds closed

velum closed

lungs

Aerodynamics of [d]

velum open

alveolar closure

vocal folds vibrating

lungs 9

alveolar closure

vocal folds vibrating

velum closed

lungs

Aerodynamics of [nd]velum closed

alveolar closure

vocal folds vibrating

lungs

Outline• Part 1: Evidence of NT in Panará

� Description of the phenomenon� Data collection, methodology and results

• Part 2: Evidence that NT in Panará is a sound change from ND� 2 proposals for the internal classification of the family� Data from other Jê languages

10

• Jê language, ~600 speakers

• Spoken in central Brazil

Panará

11

PhonemesBilabial Alveolar Palatal Velar

Stop p t ts k

Nasal m n ɲ ŋ

Fricative s

Approximant w ɾ j

Short oral (9) Short nasal (6)

i ɯ u i ɯ ũ

e ɤ o ẽ ã õ

ɛ a ɔ

Long oral (9) Long nasal (5)

i: ɯ: u: i: ũ:

e: ɤ: o: ẽ: ã: õ:

ɛ: a: ɔ:

Consonants (12)

Vowels (29)

12

The phonological process

� In Panará, nasal consonants are only fully nasalized before nasal vowels

/m,  n,  ɲ,  ŋ/  à  [m,  n,  ɲ,  ŋ]  /  __  Ṽ  

� NCs appear as allophones of nasal stops� Process of post-oralization and devoicing of nasal consonants (Lapierre et al., 2016)

/m,  n,  ɲ,  ŋ/  à  [mp,  nt,  ns,  ŋk]  /  __  V

13

Data collection• Acoustic data was collected from 12 native speakers of Panará

� only analyzed two so far

• 50 target words� mono- or disyllabic� target syllable types included:

� NṼ, NCV, CV, CṼ

• Target words were presented inside a carrier phrase� “Ĩkje hẽ ka tõ syrĩ [X]”, (I say the word [X].)

• Words and carrier phrases presented on a laptop screen alongside a picture depicting the target word� 5 semi-randomized blocks

• Total of 250 target words produced per speaker

• Calculated percentage of voicing duration for phones in target syllables 14

Results

• One-way ANOVA (DV: voicing duration, IV: phone type)� F = 9055, p<.001

• Tukey HSD post-hoc analysis� All pairwise comparisons are significantly different (p<.001), � except (V, N), (C, NC-C)

• Panará exhibits a clear case of categorical post-nasal devoicing

15

V N C NC-N NC-C

voicing duration 95% 98% 7% 91% 4%

Internal classification of Jê• At the moment, controversial

� Especially with respect to the position of Panará within the structure

• Two recent proposals� Nikulin 2015� Lapierre et al. 2016

• I’m making a phonological argument here for the proposition by Lapierre et al. 2016� See Bardagil-Mas (forthcoming) for additional evidence on ergative case-

marking in favour this tree

16

Internal classification of Jê

17

Tree A (Lapierre et al. 2016)

Tree B (Nikulin 2015)

Internal classification of Jê

18

Tree A (Lapierre et al. 2016)

Tree B (Nikulin 2015)

Data from other Jê languages

• Realization of /m/ / V__V in Jê languages

• Which form should be reconstructed for Proto-Jê?� [mb] seem to be the most suitable option� It occurs most frequently

19

PNR TMB MBG APN KNS TPY XAV XER KNG XKL

[mp] [mp] [m] [mb] [mb] [mb] ~ [m] [b] [b] [bmb] [mb]

• Realization of /m/ / V__V in Jê languages

•

• Kaingang: /m/ à [bmb] / V __ V à [mb] / Ṽ __ V

• If a language has [bmb], then it also has [mb] (Stanton, 2015)� [mb] >> [bmb]

20

PNR TMB MBG APN KNS TPY XAV XER KNG XKL

[mp] [mp] [m] [mb] [mb] [mb] ~ [m] [b] [b] [bmb] [mb]

Data from other Jê languages

• Realization of /m/ / V__V in Jê languages

•

• Kaingang: /m/ à [bmb] / V __ V à [mb] / Ṽ __ V

• If a language has [bmb], then it also has [mb] (Stanton, 2015)� [mb] >> [bmb]

21

PNR TMB MBG APN KNS TPY XAV XER KNG XKL

[mp] [mp] [m] [mb] [mb] [mb] ~ [m] [b] [b] [bmb] [mb]

Data from other Jê languages

• Realization of /m/ / V__V in Jê languages

•

• It is uncontroversial that Tapayuna is most closely related to Kĩsêdjê

• In the 1960s, the Tapayuna were massacred by rubber barons. (Camargo 2010)� Only 31 speakers survived. � They were incorporated into Mebêngôkre & Kĩsêdjê speaking villages.� Language is critically endangered. 22

PNR TMB MBG APN KNS TPY XAV XER KNG XKL

[mp] [mp] [m] [mb] [mb] [mb] ~ [m] [b] [b] [bmb] [mb]

Data from other Jê languages

• Realization of /m/ / V__V in Jê languages

•

• Tapayuna [m] is due to recent heavy contact with Mebêngôkre.� Not an independent innovation.� Nearly all speakers are bilingual in Mebêngôkre.

� Many young speakers are L1 MBG and L2 TPY

23

PNR TMB MBG APN KNS TPY XAV XER KNG XKL

[mp] [mp] [m] [mb] [mb] [mb] ~ [m] [b] [b] [bmb] [mb]

Data from other Jê languages

• Realization of /m/ / V__V in Jê languages

•

• Panará and Timbira� oral release is devoiced

24

PNR TMB MBG APN KNS TPY XAV XER KNG XKL

[mp] [mp] [m] [mb] [mb] [mb] ~ [m] [b] [b] [bmb] [mb]

Data from other Jê languages

• Realization of /m/ / V__V in Jê languages

• Which form should be reconstructed for Proto-Jê?� [mb]� post-oralization occurs in 7/10 languages� oralization occurs in 9/10 languages

25

PNR TMB MBG APN KNS TPY XAV XER KNG XKL

[mp] [mp] [m] [mb] [mb] [mb] ~ [m] [b] [b] [bmb] [mb]

Data from other Jê languages

Internal classification of Jê

26

Tree A (Lapierre et al. 2016)

[mb]

[mb]

[mp]

[mb]

[mp]

[mp]

[mb]

[mb ~ m][m]

[mb]

[b]

[b]

[bmb]

[mb]

[b]

[m]

[bmb][mb]

Internal classification of Jê

27

Tree A (Lapierre et al. 2016)

[mb]

[mb]

[mp]

[mb]

[mp]

[mp]

[mb]

[mb ~ m][m]

[mb]

[b]

[b]

[bmb]

[mb]

[b]

[m]

[bmb]

4 innovations

[mb]

Internal classification of Jê

28

Tree B (Nikulin 2015)

[mb][bmb]

[mb]

[b]

[mp]

[m]

[mb]

[mb]

[mb ~ m]

[mp]

[b][b]

[bmb]

[mb]

[mb]

[mp]

[mp]

[m]

[mb]

[mb]

Internal classification of Jê

29

Tree B (Nikulin 2015)

[mb][bmb]

[mb]

[b]

[mp]

[m]

[mb]

[mb]

[mb ~ m]

[mp]

[b][b]

[bmb]

[mb]

[mb]

[mp]

[mp]

[m]

[mb]

5 innovations2 x [mb] à [mp]

[mb]

Internal classification of Jê

30

Tree B (Nikulin 2015)

[mb][bmb]

[mb]

[b]

[mp]

[m]

[mb]

[mb]

[mb ~ m]

[mp]

[b][b]

[bmb]

[mb]

[mp] [m][mb]

[mb][mp]

[mb]

Internal classification of Jê

31

Tree B (Nikulin 2015)

[mb][bmb]

[mb]

[b]

[mp]

[m]

[mb]

[mb]

[mb ~ m]

[mp]

[b][b]

[bmb]

[mb]

[mp] [m][mb]

[mb]

6 innovations2 x [mb] à [mp] à [mb]

[mp]

[mb]

Interim summary• I’ve shown evidence in favour of tree A over Tree B

� Tree A requires only 4 innovations (1x mb > mp)� Tree B is less parsimonious

� 5 innovations: 2x mb > mp� 6 innovations: 2x mp > mp > mb (back-mutation)

• Even if the internal classification of Tree A is incorrect, ND à NT still occurred (at least!) once.

32

Conclusion

• The change of ND à NT in Panará seems to be an instance of a phonological change for the worse� No synchronic evidence of an improvement in some other area of the

grammar resulting from this diachronic change� Unlike other documented cases of post-nasal devoicing

• Although we should avoid positing unnatural sound changes when possible, changes for the worse do occur in natural language� as evidenced by Panará ND à NT� no clear evidence that this occurred as a series of phonetically natural

sound changes

33

References• Alves. F. de C. (2007). Sistema Fonológico do Timbira Apãniekrá (Fonemas, Sílaba e Acento). Proceedings of the Macro-Jê meeting. Brasilia, Brazil.

• Bardagil-Mas, B. (forthcoming). Case and Agreement in Panará. Doctoral Dissertation, University of Groningen, The Netherlands.

• Begus, G. (forthcoming). Post-Nasal Devoicing and a Probabilistic Model of Phonological Typology, ms.

• Camargo, N. da S. (2010). Língua Tapayúna: Aspectos Sociolinguísticos e uma Análise Fonológica Preliminar. Master's thesis, Universidade Estadual de Campinas, Brazil.

• Coetzee, A. W., Lin, S., & Pretorius, R. (2007). Post-nasal devoicing in Tswana. In J. Trouvain, & B. William (Eds.), ICPhS XVI: Proceedings of the 16th international congress of phonetic sciences (pp. 861–864). Saarbrücken.

• Coetzee, A. W. & Pretorius, R. (2010). Phonetically grounded phonology and sound change: The case of Tswana labial plosives. Journal of Phonetics (38): 404–421.

• Gakran, N. (2005). Aspectos Morfossintáticos da Língua Laklãnõ (Xokleng) “Jê”. Master’s thesis, Universidade Estadual de Campinas, Brazil.

• Hayes, B. (1999). Phonetically Driven Phonology: The Role of Optimality Theory and Inductive Grounding. In M. Darnell, E. Moravscik, M. Noonan, F. J. Newmeyer, & K. Wheatly (Eds.), Functionalism and formalism in linguistics, Vol. I: General papers (pp. 243–285). Amsterdam: JohnBenjamins.

• Herbert, R (1986). Language Universals, Markedness Theory, and Natural Phonetic Processes. Berlin: Walter de Gruyter & Co.

• Hyman, L. M. (2001). On the limits of phonetic determinism in phonology: *NC revisited. In E. Hume, & K. Johnson (Eds.), The role of speech perception phenomena in phonology (pp. 141–185). NewYork: Academic Press.

• Lapierre, M., Bardagil-Mas, B., Salanova, A. P. (2016). The Nasal Consonants of Panará. Talk presented at the the 21st Workshop on the Structure and Constituency in Languages of the Americas (WSCLA 2016), Université du Québec à Montréal, Canada

• Lapierre, M., Bardagil-Mas, B., Salanova, A. P. (2016). A Reconstruction of Proto-Northern Jê Phonemics. Talk presented at Amazônicas VI, Universidad Nacional de Colombia, Leticia, Colombia

• Maddieson, I. & Ladefoged, P. (1993). Phonetics of Partially Nasal Consonants. In M. K. Huffman, & R. A. Krakow, Nasals, nasalization, and the velum (Vol. 5 of Phonetics and phonology, pp. 251-301). London: Academic Press.

• De Mattos, R. (1973) Fonêmica Xerente. Série Lingüística 1. SIL: Brasília.

• Nikulin, A. (2015). On the genetic unity of Jê-Tupí-Karib. Diploma thesis, Lomonosov Moscow State University, Russia.

• Nonato, R. (2014). Clause Chaining, Switch Reference and Coordination. Doctoral dissertation, MIT, Cambridge.

• De Oliveira, C. C. (2005). The Language of the Apinajé People of Central Brazil. Doctoral dissertation, University of Oregon, Eugene.

• Pater, Joe (1996) *NC. Proceedings of the North East Linguistic Society 26, Graduate Linguistic Student Association, University of Massachusetts, Amherst, pp. 227-239.

• Rosenthall, S. (1989). The phonologyofnasal-obstruentsequences. MA thesis, Linguistics, McGill University, Montréal.

• Salanova, A. P. (2001). A nasalidade em Mebengokre e Apinayé: o limite do vozeamento soante. Master's thesis, Universidade Estadual de Campinas, Brazil.

• Solé, M.-J.; Hyman, L, M.; & Monaka, K. C. (2010). More on Post-Nasal Devoicing: The Case of Shekgalagari. Journal of Phonetics 38 (4): 299–319.

• Quintino, W. P. (2000). Aspectos da Fonología Xavante. Master’s thesis, Universidade Estadual de Campinas, Brazil.

• Westbury, J. R., & Keating, P. A. (1986). On the naturalness of stop consonant voicing. Journal of Linguistics, 22, 145–166.

• Wiesemann, U. (1972). Die phonologische und grammatische Struktur der Kaingáng-Sprache. The Hague: Mouton. 34