’ ’
’
’
an exploration of Salish phonetics
Screen Shot 2018-02-27 at
10.23.41 PM
Hul’q’umi’num' Alphabet (red: sounds not in English)
’ a aa ch ch’ e ee h hw i ii k kw kw' l lh m n o oo ou p p’ q q’ qw qw’ s sh t t’ th tl’ ts ts’ tth tth’ u w x xw y
This research was supported by the Jamie Cassels Undergraduate Research Award | Supervised by Dr. Sonya Bird
Thanks to Dr. Sonya Bird for her guidance and support throughout this research project, to Dr. Donna Gerdts and Elder Ruby Peter for sharing their expertise, to Thomas Jones and Margaret Seymour for project assistance, and especially to Bernard David for sharing his Hul’q’umi’num’ story. I would like to acknowledge that I worked and lived as a visitor on the unceded territories of the WSÁNEĆ (Saanich), Lkwungen (Songhees), and Wyomilth (Esquimalt) Peoples of the Coast Salish First Nations.Hay ch q’a’
Hul’q’umi’num
’
• Dialect of Coast Salish Hul’q’umi’num’-Halq'eméylem-hən̓q̓əmin̓əm̓ language • Territory extends along the Salish Sea from Nanoose to Malahat on Vancouver Island and neighboring islands and is spoken by the Snaw-naw-us, Snuneymuxw, Stz’uminus, Penelakut, Lyackson, Halalt, Quw’utsun, Lake Cowichan, and Malahat Peoples • 40 fluent first-language speakers, 200 fluent second-language speakers, over 1,000 semi-fluent speakers and language learners, as well as several hundred fluent understanders (who understand but do not speak)CONTEXT
• With few Elders remaining, adult second-language learners are championing the language revitalization movement and they are concerned with speaking in a way that honours their Elders • Little research done on second language pronunciation of Coast Salish languagesGOALS & METHODS
PHONETICS & PHONOLOGY
GOAL
Support L2 development of oral proficiency in Hul’q’umi’num’ by: • Documenting segmental phonetics of an Elder’s speech to better understand production of Hul’q’umi’num’ • Comparing and contrasting written language to what is produced • Analyzing Hul’q’umi’num’ rhythm • Traditional rhythm: stressed timed, syllable-timed, and mora-timed classes • Classes are distinguished by infants: they can discriminate languages between rhythm classes but not within rhythm classes (Nazzi, Bertoncini, & Mehler, 1998)• Dauer (1983): two classes are on opposite ends of a continuum
RHYTHM
(Boersma & Weenink, 2017)
Boersma, P. & Weenink, David (2017). Praat: doing phonetics by computer [Computer program]. Version 6.0.36, retrieved 11 November 2017 from http://www.praat.org/ Dauer, R. M. (1983). Stress-timing and syllable-timing reanalyzed. Journal of phonetics. 11, 51-62. Dellwo, V. (2006). “Rhythm and speech rate: A variation coefficient for ΔC,” in Language and Language Processing, edited by P. Karnowski and I. Szigeti (Peter Lang, Frankfurt am Main), pp. 231–241. First Voices Hul’q’umi’num’ Community Portal (2018). Retrieved from http://www.firstvoices.com/en/HULQUMINUM Grabe, E. & Low, E. L. (2002). Acoustic correlates of rhythm class. Laboratory phonology, 7(515-546). Hul’q’umi’num’ Language Academy: ta’ulthun sqwal (2016). Retrieved from http://sqwal.hwulmuhwqun.ca/learn/learning-the-sounds/alphabet/ Nazzi, T., Bertoncini, J., & Mehler, J. (1998). Language discrimination by newborns: Toward an understanding of the role of rhythm. Journal of Experimental Psychology: Human Perception and Performance, 24(3), 756-766. Ramus, F., Nespor, M., & Mehler, J. (1999). Correlates of linguistic rhythm in the speech signal. Cognition, 73(3), 265-292. White, L., & Mattys, S. L. (2007). Calibrating rhythm: First language and second language studies. Journal of Phonetics, 35(4), 501-522.
REFERENCES
PHONETIC FINDINGS
CONTRIBUTIONS & FUTURE DIRECTIONS
• Collaborative research: jumping off point for collaborative work on Hul’q’umi’num’ pronunciation learning and teaching • Phonetic structures: baseline for rhythm of Salish languages; documentation of pronunciation patterns in continuous, rapid speechCONTRIBUTIONS
FUTURE DIRECTIONS
STRESS-TIMED - - - SYLLABLE TIMED
more syllable types, V reduction few syllable types, no V reduction • More work connecting phonetic findings to rhythm (clitics, prosodic parsing, and reduction) • L2 speech – do learners move through a syllable-timed inter-language?RHYTHM
SHORTCOMINGS
• Only used a single person’s speech sample – rhythm metrics may reflect individual differences • Cannot generalize this finding Hul’q’umi’num’ Spanish French English Dutch 30 40 50 60 70 35 37 39 41 43 45 47 49 Var co V %V Rhythm of Languages Around the World La ng ua ge d ata p oin ts ( En glis h, D utc h, Fre nc h Sp an ish ) sou rce d f rom W hit e a nd Ma tty s (2 00 7) *Standard error not includedSpanish French English Dutch Hul’q’umi’num'
%V 48 45 38 41 39 VarcoV 41 50 64 65 53 VarcoC 46 44 47 44 60 ΔV 32 44 49 49 56 ΔC 40 51 59 49 88 Figure 4. Rhythm classes by %V and VarcoV Table 2. Rhythm Metrics Across Languages • there is a considerably larger difference in VarcoC and ΔC of Hul’q’umi’num’ than other languages, meaning that consonantal length is very dispersed • Hul’q’umi’num’ fits in with the stress-timed class according to %V and VarcoV stress-timed Figure 1. Text grid of consonant cluster reduction and vowel alternation Figure 2. Text grid of glottal stop deletions and consonant cluster reduction • Metrics proposed for quantifying rhythm acoustically:
ΔC and ΔV: standard deviation of consonantal & vocalic interval duration
%V: proportion of vocalic material (Ramus et al. 1999)
VarcoC and VarcoV: same as ΔC and ΔV but normalized for speech rate
PVI: the mean of the differences between successive intervals (Grabe & Low, 2002) (First Voices, 2018)
METHODS
Collaboratively with Hul'q'umi'num' linguists and language experts: 1) Audio recording of story by Bernard David, Tl’isla, from Stz’uminus was transcribed by fluent speakers of Hul’q’umi’num’ 2) Praat used to phonetically segment & transcribe the same story 3) Consonants and vowels were segmented so that rhythm metrics could be calculated • <a7u> pronounced as long <aa> ([a:]) • Abundant loss of <7> ([ʔ]) • Consonant cluster reductions, vowel alternations, especially <u> ([ə]) Figure 3. Text grid showing a-glottal schwa pronounced as long <aa> ([a:]) VMackenzie Marshall | Department of Linguistics | 2018
• Consonants can be strung together in complex sequences, often in clitics e.g. kwakwəxwəw̓txwəɬtsθam̓š ‘knocking on the house for me’e.g. θəyt tst tseʔ ‘we will fix it’ à tst clitic expresses meaning of “we”
• Sounds and how they are strung together are difficult to learn, and when in longer segments, so is the rhythm • Phonetics (segment length), phonology (segments, clusters), and morphology (clitics) affect the rhythm of the language (First Voices, 2018) (Abercrombie, 1965) Table 1.
Written: xw kwunus hwu 7i7shul | Pronounced: xwnus hwi7i7shul7
Written: 7apun 7i7 kw7 lhq7… | Pronounced: 7apun iw lhq7…
La be lle d se gm en ts tr an sc rib ed wi th H ul ’q ’u m i’n um ' A lp ha be t no tIP A/ N AP A: [u ] r ep re se nt s [ə ] an d [7 ] r ep re se nt s [ʔ ]
Written: 7ula7ulh | Pronounced 7ulaalh
• Salish languages have very complex sound systems
• Very little phonetic documentation; none for Hul’q’umi’num’
• Hul’q’umi’num’ has 37 consonants; English has only 21 consonants
(First Voices, 2018)
