West Coast Conference on Formal Linguistics (WCCFL) (38th : 2020)

A Hypothesis on the Origin of Preglottalized Sonorants in Kra-Dai Norquest, Peter 2020-03-08

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
73804-Norquest_P_Preglottalized_sonorants__WCCFL38_2020_handout.pdf [ 259.15kB ]
73804-Norquest_P_Preglottalized_sonorants__WCCFL38_2020_slides.pdf [ 254.27kB ]
Metadata
JSON: 73804-1.0389866.json
JSON-LD: 73804-1.0389866-ld.json
RDF/XML (Pretty): 73804-1.0389866-rdf.xml
RDF/JSON: 73804-1.0389866-rdf.json
Turtle: 73804-1.0389866-turtle.txt
N-Triples: 73804-1.0389866-rdf-ntriples.txt
Original Record: 73804-1.0389866-source.json
Full Text
73804-1.0389866-fulltext.txt
Citation
73804-1.0389866.ris

Full Text

A Hypothesis on the Origin of Preglottalized Sonorants in Kra-Dai  Peter Norquest                  West Coast Conference in Formal Linguistics 38 University of Arizona              The University of British Columbia norquesp@u.arizona.edu          March 8, 2020  0. Introduction  • A distinction between plain and preglottalized initial voiced stops is reconstructible within all primary branches of the Kra-Dai phylum at three places of articulation.  • This distinction may be hypothesized to be a secondary development within Kra-Dai, where the diachronic trajectories of what were originally medial voiced stops depended on whether or not the preceding vowel was schwa on analogy with a similar development in Proto-North Sarawak (Blust 2006, Norquest 2016).  • While this plain versus glottalized opposition in the voiced obstruent series is quite robust within Kra-Dai, the same is not true for the sonorants, which in many languages show only a two-way contrast between voiced and voiceless.  • However, the Kam-Sui branch of Kra-Dai is particularly conservative in this regard, and a four-way phonation opposition (preaspirated, voiceless, voiced and preglottalized) can be reconstructed for all sonorants with the exception of the lateral.  • The Hlai branch of Kra-Dai can now also be postulated as having retained evidence for a series of preglottalized sonorants. • While it has generally been assumed that these phonation differences reflect the different glottal states of the initial consonants of sesquisyllabic words before the presyllable was lost, the suggestion is put forward here that the preglottalized series of sonorants is not the result of conditioning by the presyllable initial.  • Rather, on analogy with the voiced stops, it occurs in an environment following schwa in which gemination occurred, followed by debuccalization of the first half of the geminate.   Figure 1: The distribution of the Kra-Dai phylum. The Kra-Dai family is composed of four major branches: (1) Kra, (2) Kam-Sui, (3) Hlai, and (4) Tai (3 branches); the smaller Be group is also shown. Lakkja and Biao are not shown, but are located to the east of the Tai group towards the Pearl River Delta.    Kra-Dai         Eastern Kra-Dai          Lakkja-Kam-Tai           Kam-Tai        Western Kam-Tai                   Be-Tai                  Kra            Biao  Lakkja  Kam-Sui             Ong-Be     Tai      Hlai  Figure 2: The Kra-Dai Phylogenetic Tree (Norquest 2015)    1. Background: Initial Voiced Stops in Kra-Dai  A distinction between plain and preglottalized initial voiced stops is reconstructible within the Kra-Dai phylum at three places of articulation, with a fourth (retroflex) occurring in the latter series:  Table 1: The plain ~ preglottalized contrast in Kra-Dai initial voiced stops Plain Preglottalized *b, *d, *ɟ *ʔb, *ʔd, *ʔɖ, *ʔɟ  However, words with plain voiced obstruent initials appear to be largely if not completely secondary – many of them can be shown to be loanwords from either Middle Chinese or one of the Austroasiatic families:  Table 2: Reflexes of plain initial voiced stops after non-schwa vowels in Lakkja-Kam-Tai *Kra-Dai ? *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai *b *b *b *v *b *ɓ *d *d *d *ɦ *d *ɗ *ɟ *d *zj *ʒ *ɟ *tɕ  (1) M. Chinese1  *Lakkja *Kam-Sui *Ong-Be *Tai  *Hlai  ‘goose’  ---  ---  ---  *vuːn X *bɯnh  *ɓunɦ ‘copper’ 銅 *dəwŋ  *doŋ  *dɔŋ  *ɦoːŋ  *doːŋ  *ɗuːŋ ‘money’ 錢 *dzjen  *diːn  *zjen  *ʒeːn  *ɟeːn  *tɕiːn  Norquest (2013) demonstrated, based on the evidence of an older layer of Austronesian/Kra-Dai cognates, that the original Kra-Dai voiced stops in initial position had devoiced before register splits occurred in the individual Kra-Dai families, allowing the voiced stops to merge with the original voiceless stops:  Table 3: Reflexes of plain initial voiced stops after non-schwa vowels in Kra-Dai Pre-Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai *b *p *p *p *p *pʰ *d *t *t *t *t *tʰ *ɟ *ts *c (*ts) *c *tɕʰ  (2) *RAn  *Lakkja *Kam-Sui *Ong-Be *Tai  *Hlai  ‘grandparent’ *apu  ---  ---  ---  *puː  *phuːʔ ‘mouth’ *baqbaq ---  *paːk  *paːk  *paːk  ---  ‘to fall’  *-tuq  ---  *tɔk  *tɔk  *tok  *thok ‘chest, liver’ *dəbdəb *tap  *tap  *t[a]p  *tap  ---  ‘sharp’  *ʈaɟəm2  ---  ---  ---  ---  *tɕhəːm  2. Background: Medial Voiced Stops in Kra-Dai  Norquest (2016) argues that initials which were formerly medials of sesquisyllabic forms were conditioned depending on whether the preceding vowel had been schwa (Table 5) or not (Table 4). This was true particularly in Proto-Tai and Proto-Hlai, where preglottalized and implosive voiced stops were conditioned by a preceding schwa:  Table 4: Reflexes of medial voiced stops after non-schwa vowels in Kra-Dai *Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai *C-b *w̥ *C-b *ʔb *C-b *ʋ *C-d *l ̥ *C-ʔɖ *r ̥ *C-d *ɾ *C-ɖ *j ̊(< *r)̥ *C-ʔɖ *r *C-ɖ *ɾ *C-ɟ *l ̥ *ʔj *j ̊ *ʔj *hj         																																																						1 Middle Chinese data are taken from Baxter & Sagart (2014). 2 It is assumed that the first syllable of this form was lost (*ʈaɟəm > *ɟəm) before intervocalic lenition could occur. (3) *Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai  *Hlai  ‘thin’  *C-báːŋ  *w̥aːŋ  *C-baːŋ  *ʔbjaŋ  *C-baːŋ  --- ‘bone’  *Cudə:́k ---  *C-ʔɖaːk *rɯ̥k  *C-dwoːk       *Cuɾɯːk ‘boat’  *Cuɖá:  *jẘaː  *C-ʔɖrwaː *rwaː  *C-ɖwaː *Cuɾaː ‘borrow’ *C-ɟáːm *lḁːm  *ʔjaːm  ---  *ʔjɯːm  ---  Table 5: Reflexes of medial voiced stops after schwa in Kra-Dai *Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai *Cəʔb *ʔb *ʔb *ʔb *ʔb *ɓ *Cəʔd *l ̥ *ʔd *r ̥ *ʔd *ɗ *Cəʔɖ *l ̥ *ʔɖ *r ̥ *ʔd *ɗ *Cəʔɟ *j ̊ *ʔɟ *j ̊ *ʔɟ *tɕ   (4) *Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai  *Hlai   ‘village’ *Cəʔbáːnʔ *ʔbaːnʔ  *ʔbaːnʔ  ---  *ʔbaːnʔ  --- ‘winnow basket’ *Cəʔdóŋʔ *lo̥ŋʔ  *ʔdɔŋʔ  *ro̥ːŋ X  *ʔdoŋʔ  *ɗoŋʔ ‘to stand’ *Cəʔɟún *jůːn  *ʔɟun  *jůn  *ʔɟɯn  *tɕu:n  3. Proto-North Sarawak Voiced Stop Fortition after Schwa  • Blust (2006) reconstructs a series of voiced aspirates in Proto-North Sarawak (PNS), based on a distinction in the daughter languages between plain/lenited voiced stops and phonetically ‘complex’ reflexes such as voiced aspirates, implosives and fricates.  • However, only Kelabit evinces actual voiced aspirates, whereas other languages show other reflexes including implosives, voiceless stops, and voiceless fricatives.  • Blust (p.c.) suggests that the voiced aspirates have arisen secondarily from consonant gemination resulting from lengthening of consonants after either (a) word-internal or (b) epenthetic initial schwa (which satisfies the requirement for a bisyllabic template).  To illustrate the phenomenon of post-schwa lengthening, examples of PNS plain medial voiced stops are shown in (5a), and geminate voiced stops following schwa in (5b). All forms are derived from Proto-Malayo-Polynesian (PMP), the Austronesian parent of PNS:  (5)  (a) Gloss  PMP3  PNS ash  *qabu  *abuh 3pl  *(si-)ida *idah rain  *quɟan  *uɟan digging stick *tuɢal  *tugal   (b) Gloss  PMP  PNS sugarcane *təbuh  *təbːuh faint  *mədan  *mədːan blink  *kəɟəp  *kəɟːəp sleep  ---  *məgːəl 																																																						3 Please note that traditional PMP phonemes are interpreted in the following way: *j = [ɖ], *z = [ɟ], *g = [ɢ], *R = [ʀ], *y = [j], and *e = [ə]. The reflexes of voiced geminates in languages of the four branches of PNS (Blust 2006: 321) are given in Table 6:  Table 6: Reflexes of Proto-North Sarawak voiced geminate stops PNS *bː *dː *ɟː *gː Bintulu ɓ ɗ ɟ g  Kenyah  Kenyah (Long San) ɓ ɗ ʄ ɠ Kenyah (Long Dunin) b/ɓ d/ɗ s ɠ Kenyah (Long Wat) b d ɟ g Kenyah (Long Anap) p t c k  Kelabitic  Kelabit (Bario) bh dh dh gh Kelabit (Long Napir) f s s k Kelabit (Pa’ Mada) p t t k Kelabit (Tring) p c c k      Berawan-Lower Baram   Berawan (Long Terawan) p c c k Berawan (Long Jegan) p c c k Narum f t c k Kiput s s c k Miri f s s k  It is apparent that at least some of these changes were areal in nature, and occurred after the break-up of PNS, since the same kinds of changes happen in languages from different branches. The languages above are regrouped below in table 2 according to the broad direction in which these changes took place:  Table 7: Reflexes of PNS geminate voiced stops by type of change PNS *bː *dː *ɟː *gː  Shortening  Kenyah (Long Wat) b d ɟ g  Implosion  Kenyah (Long San) ɓ ɗ ʄ ɠ Kenyah (Long Dunin) ɓ/b ɗ/d s ɠ Bintulu ɓ ɗ ɟ g  Aspiration  Kelabit (Bario) bh dh dh gh  Devoicing  Kelabit (Pa’ Mada) p t t k Kenyah (Long Anap) p t c k Kelabit (Tring) p c c k Berawan (Long Terawan) p c c k Berawan (Long Jegan) p c c k  Devoicing with Frication  Narum f t c k Kiput s s c k Kelabit (Long Napir) f s s k Miri f s s k  I agree with Blust in reconstructing original gemination as the earliest stage of these medial voiced stops in PNS, which had multiple outcomes in the daughter languages (including both voiced aspirates and implosives) on the basis that (1) if voiced aspirates were to be reconstructed, the devoicing of i.e. bh would be predicted to result in an aspirated voiceless stop ph, not in a plain voiceless stop p, and (2) it seems strange that a voiced aspirate would become an implosive, since this would require a reversal of glottal aperture from lax to constricted. I therefore predict the possible trajectories of change for an intervocalic voiced geminate (using the bilabial place of articulation as an example) to be the following:  (6)  (a) -b:- > -ʔb- > -ɓ- (b)  -b:- > -bp- > -p- (c)  -b:- > -bh- ( > -ph- > -f-)  Prentice (1974) drew attention to the bilabial split in North Sarawak and a similar distinction in two Idahan languages of Sabah: the Kadazan dialect of Coastal Dusun and the Timugon dialect of Lowland Murut. In Sabahan, there is a similar contrast in fortis and lenis reflexes of voiced stops. Focusing on Kadazan, examples of the lenis/fortis split in medial position are given below4, in which it can be seen that the conditions of the split are the same as those for PNS. Intervocalic lenition occurs after non-schwa vowels in (7a), and gemination is hypothesized to have occurred in the forms in (7b), preventing lenition:  (7)  (a) Gloss  PMP  Kadazan   cloud  *ʀabun  gavun  (*b > v)   housepost *hadiʀi  to-igi  (*d > Æ)   paddy  *paɖaj  paaj  (*ɖ > Æ)   indicate  *tuɟuq  tuuʔ  (*ɟ > Æ)     (b) Gloss  PMP  Kadazan   stab  *təbək  tobok  (< *təbːək)   hiccough *sədu  sodu  (< *sədːu)   sting, smart *hapəɖəs podos  (< *pədːəs)   pinch  *kəɟut  kodut  (< *kədːut)   																																																						4 Kadazan data have been drawn from Prentice and various publications by Blust, and supplemented by Antonissen (1958). 4. Proto-North Sarawak Sonorants after Schwa  In contrast with the voiced stops, PNS voiceless obstruents don’t undergo lengthening after schwa:   (8)   Gloss  PMP  PNS     (a) fathom  *dəpa  *dəpa four  *əpat  *əpat   blowpipe *səput  *səput   (b) design, tattoo *bətik  *bətik   bamboo sp. *bətuŋ  *bətuŋ   fart  *qətut  *ətut  (c) stick, adhere to *dəkət  *dəkət   bracelet  *ləku  *ləkuʔ   (d) full, satiated *bəsuʀ  *bəsuʀ   flesh, muscle *həsi  *əsi   rice mortar *əsuŋ  *əsuŋ  Neither do sonorants:  (9)   Gloss  PMP  PNS   (a) fat, grease *ləmu  *ləmu   weak  *ləmaq  *ləmaʔ   sweet  *əmis  *əmis   (b) correct  *kəna  *kəna   six  *ənəm  *ənəm   full (container) *pənuq  *pənuʔ   (c) sea turtle *pəɲu  *pəɲu   to swallow *təɲəl  *təɲəl fat, grease *məɲak  *məɲak   (d) ankle  *bəŋil  *bəŋil   deaf  *dəŋəl  *dəŋəl   (e) vulva, vagina *təli  *təli   three  *təlu  *təlu   buy  *bəli  *bəlih  (f)  husked rice *bəʀas  *bəʀas   core of tree *təʀas  *təʀas   k.o. tree *təʀəp  *təʀəp  • Question: is it possible to extend the PNS intervocalic voiced stop preglottalization paradigm to other phoneme classes?  5. Preglottalized Sonorants in Kra-Dai  While the plain versus preglottalized opposition in the voiced obstruent series is quite robust within Kra-Dai, the same is not true for the sonorants. However, the Kam-Sui branch of Kra-Dai is particularly conservative in this regard, and a four-way phonation opposition can be reconstructed for all sonorants, with the single exception of a preglottalized lateral (which has merged with the voiceless lateral):  Table 8: The four-way phonation distinction in Kra-Dai sonorants Preaspirated Voiceless Plain Preglottalized *hm, *hn, *hȵ, *hŋ *m̥, *n,̥ *ȵ̥, *ŋ ̊ *m, *n, *ȵ, *ŋ *ʔm, *ʔn, *ʔȵ, *ʔŋ *hl, *hr, *hʀ *l,̥ *r,̥ *ʀ̥ *l, *r, *ʀ *ʔl, *ʔr, *ʔʀ *hw, *hj *w̥, *j ̊ *w, *j *ʔw, *ʔj  The modern Sui phonetic reflexes have been examined in some detail (Edmondson et al 2004). These are given below using the bilabial nasal series as an example. All Kam-Sui languages are tonal and distinguish between high (H) and low (L) registers, the former correlating with original voiceless initials and the latter with original voiced initials, as shown below using the bilabial series again as an example:  Table 9: Phonation and register Preaspirated Voiceless Plain Preglottalized *hm    >   m̥H *m̥    >    mH *m    >    mL *ʔm    >    ʔmH  It has generally been assumed that these phonation differences reflect the different glottal states of the initial consonants of sesquisyllabic words before the presyllable was lost. For example, an initial aspirated stop or fricative (with spread glottis) would result in a preaspirated sonorant, an initial voiceless initial would result in a voiceless sonorant, and an initial voiced initial would result in a voiced sonorant. However, this fails to explain the origin of the preglottalized series of sonorants, as no word-initial consonants with glottal constriction can be reconstructed:  (10)  (a) *tʰ-m  > *hm  (b) *t-m > *m̥  (c) *d-m >  *m  (d) ??? >  *ʔm  The hypothesis is put forward here that this final preglottalized series of sonorants, in contrast with the other three series, is not the result of conditioning by the presyllable initial. Rather, on analogy with the voiced stops, it occurs in an environment following schwa in which gemination occurred, followed by debuccalization of the first half of the geminate. Under this hypothesis, (10d) above may be revised as the following:  (11)  *Cəm-   >  *Cəm̆ː-  >  *Cəʔ̆m-  > *ʔm-  Examples of all four phonation types are given below:  (12)  *Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai  *Hlai  ‘dog’   *kʰ[u]máː *kʰ-mwaː *k-hmaː *m̥aː  *m̥aː  *hmaː ‘ditch’  *[t]-m̥ˠáːŋ ---  *T-m̥jaːŋ *m̥aŋ  *m̥ɯəŋ  --- ‘ant’  *r-móȶ  *mot  *r-mət  *muːʔ  *moc  *hmuȶ ‘bear’  *kəʔmˠúj *k-Nuːj  *ʔmjeː  ---  *m̥wiː  *ʔmuj Since other branches of Kra-Dai generally fail to preserve preglottalized sonorants, and since the Kam-Sui preglottalized sonorants usually correspond with voiceless sonorants in those branches, it can be posited that Proto-Kam-Sui is archaic in this aspect of its phonological inventory. This distinction was lost permanently in some branches (Lakkja, Ong-Be, Tai) as the original preglottalized sonorants series merged with the voiceless sonorant series. It was, on the other hand, preserved indirectly in the Hlai branch, where the voiceless series of sonorants merged with the preaspirated series, leaving the preglottalized series to fill the vacancy that was created in what became a chain shift:  (13)  (a) Kra-Dai *hm, *m̥, *m  >  Proto-Hlai *hm  (b) Kra-Dai *ʔm   >  Proto-Hlai *m̥  Table 10: Reflexes of medial sonorants after schwa in Kra-Dai *Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai *Cəm *m̥ *ʔm *m̥ *m̥ *ʔm *Cən *n ̥ *ʔn *n ̥ *n ̥ *ʔn *Cəȵ *ȵ̥ *ʔȵ *ȵ̥ *ȵ̥ *ʔȵ *Cəŋ *ŋ ̊ *ʔŋ *ŋ ̊ *ŋ ̊ *ʔŋ *Cəl *l ̥ *l ̥ *l ̥ *l ̥ *ʔl *Cər *j ̊(< *r)̥ *ʔr *ʃ *Cr *hr *Cəʀ (*j)̊ *ʔʀ (*ʃ) (*ʀ̥) (*hr) *Cəw *w̥ *ʔw *w̥ *ʔw (*ʔw) *Cəj *j ̊ *ʔj *j ̊ *ʔj (*ʔj)  Some examples are given below:  (14)  *Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai  *Hlai  ‘thick’  *tsəʔnáː *ts-Naː  *ʔɳaː  *nḁː  *nḁː  *ʔna: ‘cold’  *kəʔȵít  *k-Ni:t  *ʔȵit  *ni̥ːt  *ni̥t  --- ‘stupid’  *Cəʔŋáːŋh *ʔŋaːŋh  *ʔŋaːŋh  *ŋə̊ːŋ X  ---  --- ‘gills’  *Cəʔŋˠáːk ---  *ʔȵaːk  *ŋaːk  *ŋɯ̊ək  *ʔŋaːk ‘taro’  *pəʔrˠáːk *jåːk (<*rḁːk) *ʔrjaːk  *ʃaːk  *prɯək  *hraːk ‘moan’  *gəʔráːŋ *jåːŋ (<*rḁːŋ) *ʔraːŋ  ---  *graːŋ  --- ‘hungry’ *məʔjáːk ---  *m-ʔjaːk *jåk  *ʔjaːk  --- ‘stupid’  *Cəʔwáːʔ ---  *ʔwaːʔ  ---  *ʔwaːʔ  ---  6. Conclusion  There may be good phonetic motivations for the development of post-schwa geminate consonants (including sonorants) and their subsequent debuccalization in Pre-Kra-Dai:  • Iambic prosody: encouraged lengthening and fortition in post-schwa medial environment • Reduction of schwa duration: favored the ultimate deletion of schwa itself • Glottal stop acted as a final syllabic nucleus place-holder before ultimate deletion of the presyllable • Glottal stop was therefore a phonetic default which then became phonologized:   (15) *tsənáː  >  *tsən̆ːáː  >  *tsəʔ̆náː > *tsʔn̩áː  >   *ʔnáː   References  Antonissen, A. 1958. Kadazan-English and English-Kadazan Dictionary. Canberra: Government Printing Office.  Baxter, William H., and Laurent Sagart. 2014. Old Chinese: a new reconstruction. New York: Oxford University Press. 	Blust, Robert. 1995. Notes on Berawan Consonant Gemination. Oceanic Linguistics 34 (1): 123-38.  ---. 1998. The Position of the Languages of Sabah. In Pagtanáw: Essays on Language in Honor of Teodoro A. Llamzon, Ma. Lourdes S. Bautista (ed.). 29-52. Manila: Linguistic Society of the Philippines.  ---. 2001. Language, Dialect, and Riotous Sound Change: The Case of Sa’ban. In Papers from the Ninth Annual Meeting of the Southeast Asian Linguistics Society 1999, Graham Thurgood (ed.). 249-359. Arizona State University Program for Southeast Asian Studies, Monograph Series Press, Tempe: Arizona State University.  ---. 2002. Kiput Historical Phonology. Oceanic Linguistics 41 (2): 384-438.  ---. 2006. The Origin of the Kelabit Voiced Aspirates: A Historical Hypothesis Revisited. Oceanic Linguistics 45 (2): 311-38.  ---. 2007. Òma Lóngh Historical Phonology. Oceanic Linguistics 46 (1): 1-53.  ---. 2010. The Greater North Borneo Hypothesis. Oceanic Linguistics 49 (1): 44-118.  Edmondson, Jerold, John Esling, Jimmy Harris & James Wei. 2004. A phonetic study of Sui consonants and tones. Mon–Khmer Studies 34:47–66. Norquest, Peter. 2013. A Revised Inventory of Proto Austronesian Consonants: Kra-Dai and Austroasiatic Evidence. Mon-Khmer Studies 42: 102-126. ----. 2015. A Phonological Reconstruction of Proto-Hlai. Brill Academic Publishers.  ----. 2016. Revisiting the question of Austronesian implosives. Mother Tongue XX. Norquest, Peter & Sean Downey. 2013. ‘Expanding the PAn Consonant Inventory.’ Journal of the Southeast Asian Linguistics Society (JSEALS) 6: 99-145. Prentice, D.J. 1974. Yet Another PAN Phoneme? Oceanic Linguistics 13: 33-75.   A Hypothesis on the Origin of Preglottalized Sonorants in Kra-Dai  Peter Norquest                  West Coast Conference in Formal Linguistics 38 University of Arizona              The University of British Columbia norquesp@u.arizona.edu          March 8, 2020  0. Introduction  • A distinction between plain and preglottalized initial voiced stops is reconstructible within all primary branches of the Kra-Dai phylum at three places of articulation.  • This distinction may be hypothesized to be a secondary development within Kra-Dai, where the diachronic trajectories of what were originally medial voiced stops depended on whether or not the preceding vowel was schwa on analogy with a similar development in Proto-North Sarawak (Blust 2006, Norquest 2016).  • While this plain versus glottalized opposition in the voiced obstruent series is quite robust within Kra-Dai, the same is not true for the sonorants, which in many languages show only a two-way contrast between voiced and voiceless.  • However, the Kam-Sui branch of Kra-Dai is particularly conservative in this regard, and a four-way phonation opposition (preaspirated, voiceless, voiced and preglottalized) can be reconstructed for all sonorants with the exception of the lateral.  • The Hlai branch of Kra-Dai can now also be postulated as having retained evidence for a series of preglottalized sonorants. • While it has generally been assumed that these phonation differences reflect the different glottal states of the initial consonants of sesquisyllabic words before the presyllable was lost, the suggestion is put forward here that the preglottalized series of sonorants is not the result of conditioning by the presyllable initial.  • Rather, on analogy with the voiced stops, it occurs in an environment following schwa in which gemination occurred, followed by debuccalization of the first half of the geminate.   Figure 1: The distribution of the Kra-Dai phylum. The Kra-Dai family is composed of four major branches: (1) Kra, (2) Kam-Sui, (3) Hlai, and (4) Tai (3 branches); the smaller Be group is also shown. Lakkja and Biao are not shown, but are located to the east of the Tai group towards the Pearl River Delta.    Kra-Dai         Eastern Kra-Dai          Lakkja-Kam-Tai           Kam-Tai        Western Kam-Tai                   Be-Tai                  Kra            Biao  Lakkja  Kam-Sui             Ong-Be     Tai      Hlai  Figure 2: The Kra-Dai Phylogenetic Tree (Norquest 2015)    1. Background: Initial Voiced Stops in Kra-Dai  A distinction between plain and preglottalized initial voiced stops is reconstructible within the Kra-Dai phylum at three places of articulation, with a fourth (retroflex) occurring in the latter series:  Table 1: The plain ~ preglottalized contrast in Kra-Dai initial voiced stops Plain Preglottalized *b, *d, *ɟ *ʔb, *ʔd, *ʔɖ, *ʔɟ  However, words with plain voiced obstruent initials appear to be largely if not completely secondary – many of them can be shown to be loanwords from either Middle Chinese or one of the Austroasiatic families:  Table 2: Reflexes of plain initial voiced stops after non-schwa vowels in Lakkja-Kam-Tai *Kra-Dai ? *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai *b *b *b *v *b *ɓ *d *d *d *ɦ *d *ɗ *ɟ *d *zj *ʒ *ɟ *tɕ  (1) M. Chinese1  *Lakkja *Kam-Sui *Ong-Be *Tai  *Hlai  ‘goose’  ---  ---  ---  *vuːn X *bɯnh  *ɓunɦ ‘copper’ 銅 *dəwŋ  *doŋ  *dɔŋ  *ɦoːŋ  *doːŋ  *ɗuːŋ ‘money’ 錢 *dzjen  *diːn  *zjen  *ʒeːn  *ɟeːn  *tɕiːn  Norquest (2013) demonstrated, based on the evidence of an older layer of Austronesian/Kra-Dai cognates, that the original Kra-Dai voiced stops in initial position had devoiced before register splits occurred in the individual Kra-Dai families, allowing the voiced stops to merge with the original voiceless stops:  Table 3: Reflexes of plain initial voiced stops after non-schwa vowels in Kra-Dai Pre-Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai *b *p *p *p *p *pʰ *d *t *t *t *t *tʰ *ɟ *ts *c (*ts) *c *tɕʰ  (2) *RAn  *Lakkja *Kam-Sui *Ong-Be *Tai  *Hlai  ‘grandparent’ *apu  ---  ---  ---  *puː  *phuːʔ ‘mouth’ *baqbaq ---  *paːk  *paːk  *paːk  ---  ‘to fall’  *-tuq  ---  *tɔk  *tɔk  *tok  *thok ‘chest, liver’ *dəbdəb *tap  *tap  *t[a]p  *tap  ---  ‘sharp’  *ʈaɟəm2  ---  ---  ---  ---  *tɕhəːm  2. Background: Medial Voiced Stops in Kra-Dai  Norquest (2016) argues that initials which were formerly medials of sesquisyllabic forms were conditioned depending on whether the preceding vowel had been schwa (Table 5) or not (Table 4). This was true particularly in Proto-Tai and Proto-Hlai, where preglottalized and implosive voiced stops were conditioned by a preceding schwa:  Table 4: Reflexes of medial voiced stops after non-schwa vowels in Kra-Dai *Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai *C-b *w̥ *C-b *ʔb *C-b *ʋ *C-d *l ̥ *C-ʔɖ *r ̥ *C-d *ɾ *C-ɖ *j ̊(< *r)̥ *C-ʔɖ *r *C-ɖ *ɾ *C-ɟ *l ̥ *ʔj *j ̊ *ʔj *hj         																																																						1 Middle Chinese data are taken from Baxter & Sagart (2014). 2 It is assumed that the first syllable of this form was lost (*ʈaɟəm > *ɟəm) before intervocalic lenition could occur. (3) *Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai  *Hlai  ‘thin’  *C-báːŋ  *w̥aːŋ  *C-baːŋ  *ʔbjaŋ  *C-baːŋ  --- ‘bone’  *Cudə:́k ---  *C-ʔɖaːk *rɯ̥k  *C-dwoːk       *Cuɾɯːk ‘boat’  *Cuɖá:  *jẘaː  *C-ʔɖrwaː *rwaː  *C-ɖwaː *Cuɾaː ‘borrow’ *C-ɟáːm *lḁːm  *ʔjaːm  ---  *ʔjɯːm  ---  Table 5: Reflexes of medial voiced stops after schwa in Kra-Dai *Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai *Cəʔb *ʔb *ʔb *ʔb *ʔb *ɓ *Cəʔd *l ̥ *ʔd *r ̥ *ʔd *ɗ *Cəʔɖ *l ̥ *ʔɖ *r ̥ *ʔd *ɗ *Cəʔɟ *j ̊ *ʔɟ *j ̊ *ʔɟ *tɕ   (4) *Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai  *Hlai   ‘village’ *Cəʔbáːnʔ *ʔbaːnʔ  *ʔbaːnʔ  ---  *ʔbaːnʔ  --- ‘winnow basket’ *Cəʔdóŋʔ *lo̥ŋʔ  *ʔdɔŋʔ  *ro̥ːŋ X  *ʔdoŋʔ  *ɗoŋʔ ‘to stand’ *Cəʔɟún *jůːn  *ʔɟun  *jůn  *ʔɟɯn  *tɕu:n  3. Proto-North Sarawak Voiced Stop Fortition after Schwa  • Blust (2006) reconstructs a series of voiced aspirates in Proto-North Sarawak (PNS), based on a distinction in the daughter languages between plain/lenited voiced stops and phonetically ‘complex’ reflexes such as voiced aspirates, implosives and fricates.  • However, only Kelabit evinces actual voiced aspirates, whereas other languages show other reflexes including implosives, voiceless stops, and voiceless fricatives.  • Blust (p.c.) suggests that the voiced aspirates have arisen secondarily from consonant gemination resulting from lengthening of consonants after either (a) word-internal or (b) epenthetic initial schwa (which satisfies the requirement for a bisyllabic template).  To illustrate the phenomenon of post-schwa lengthening, examples of PNS plain medial voiced stops are shown in (5a), and geminate voiced stops following schwa in (5b). All forms are derived from Proto-Malayo-Polynesian (PMP), the Austronesian parent of PNS:  (5)  (a) Gloss  PMP3  PNS ash  *qabu  *abuh 3pl  *(si-)ida *idah rain  *quɟan  *uɟan digging stick *tuɢal  *tugal   (b) Gloss  PMP  PNS sugarcane *təbuh  *təbːuh faint  *mədan  *mədːan blink  *kəɟəp  *kəɟːəp sleep  ---  *məgːəl 																																																						3 Please note that traditional PMP phonemes are interpreted in the following way: *j = [ɖ], *z = [ɟ], *g = [ɢ], *R = [ʀ], *y = [j], and *e = [ə]. The reflexes of voiced geminates in languages of the four branches of PNS (Blust 2006: 321) are given in Table 6:  Table 6: Reflexes of Proto-North Sarawak voiced geminate stops PNS *bː *dː *ɟː *gː Bintulu ɓ ɗ ɟ g  Kenyah  Kenyah (Long San) ɓ ɗ ʄ ɠ Kenyah (Long Dunin) b/ɓ d/ɗ s ɠ Kenyah (Long Wat) b d ɟ g Kenyah (Long Anap) p t c k  Kelabitic  Kelabit (Bario) bh dh dh gh Kelabit (Long Napir) f s s k Kelabit (Pa’ Mada) p t t k Kelabit (Tring) p c c k      Berawan-Lower Baram   Berawan (Long Terawan) p c c k Berawan (Long Jegan) p c c k Narum f t c k Kiput s s c k Miri f s s k  It is apparent that at least some of these changes were areal in nature, and occurred after the break-up of PNS, since the same kinds of changes happen in languages from different branches. The languages above are regrouped below in table 2 according to the broad direction in which these changes took place:  Table 7: Reflexes of PNS geminate voiced stops by type of change PNS *bː *dː *ɟː *gː  Shortening  Kenyah (Long Wat) b d ɟ g  Implosion  Kenyah (Long San) ɓ ɗ ʄ ɠ Kenyah (Long Dunin) ɓ/b ɗ/d s ɠ Bintulu ɓ ɗ ɟ g  Aspiration  Kelabit (Bario) bh dh dh gh  Devoicing  Kelabit (Pa’ Mada) p t t k Kenyah (Long Anap) p t c k Kelabit (Tring) p c c k Berawan (Long Terawan) p c c k Berawan (Long Jegan) p c c k  Devoicing with Frication  Narum f t c k Kiput s s c k Kelabit (Long Napir) f s s k Miri f s s k  I agree with Blust in reconstructing original gemination as the earliest stage of these medial voiced stops in PNS, which had multiple outcomes in the daughter languages (including both voiced aspirates and implosives) on the basis that (1) if voiced aspirates were to be reconstructed, the devoicing of i.e. bh would be predicted to result in an aspirated voiceless stop ph, not in a plain voiceless stop p, and (2) it seems strange that a voiced aspirate would become an implosive, since this would require a reversal of glottal aperture from lax to constricted. I therefore predict the possible trajectories of change for an intervocalic voiced geminate (using the bilabial place of articulation as an example) to be the following:  (6)  (a) -b:- > -ʔb- > -ɓ- (b)  -b:- > -bp- > -p- (c)  -b:- > -bh- ( > -ph- > -f-)  Prentice (1974) drew attention to the bilabial split in North Sarawak and a similar distinction in two Idahan languages of Sabah: the Kadazan dialect of Coastal Dusun and the Timugon dialect of Lowland Murut. In Sabahan, there is a similar contrast in fortis and lenis reflexes of voiced stops. Focusing on Kadazan, examples of the lenis/fortis split in medial position are given below4, in which it can be seen that the conditions of the split are the same as those for PNS. Intervocalic lenition occurs after non-schwa vowels in (7a), and gemination is hypothesized to have occurred in the forms in (7b), preventing lenition:  (7)  (a) Gloss  PMP  Kadazan   cloud  *ʀabun  gavun  (*b > v)   housepost *hadiʀi  to-igi  (*d > Æ)   paddy  *paɖaj  paaj  (*ɖ > Æ)   indicate  *tuɟuq  tuuʔ  (*ɟ > Æ)     (b) Gloss  PMP  Kadazan   stab  *təbək  tobok  (< *təbːək)   hiccough *sədu  sodu  (< *sədːu)   sting, smart *hapəɖəs podos  (< *pədːəs)   pinch  *kəɟut  kodut  (< *kədːut)   																																																						4 Kadazan data have been drawn from Prentice and various publications by Blust, and supplemented by Antonissen (1958). 4. Proto-North Sarawak Sonorants after Schwa  In contrast with the voiced stops, PNS voiceless obstruents don’t undergo lengthening after schwa:   (8)   Gloss  PMP  PNS     (a) fathom  *dəpa  *dəpa four  *əpat  *əpat   blowpipe *səput  *səput   (b) design, tattoo *bətik  *bətik   bamboo sp. *bətuŋ  *bətuŋ   fart  *qətut  *ətut  (c) stick, adhere to *dəkət  *dəkət   bracelet  *ləku  *ləkuʔ   (d) full, satiated *bəsuʀ  *bəsuʀ   flesh, muscle *həsi  *əsi   rice mortar *əsuŋ  *əsuŋ  Neither do sonorants:  (9)   Gloss  PMP  PNS   (a) fat, grease *ləmu  *ləmu   weak  *ləmaq  *ləmaʔ   sweet  *əmis  *əmis   (b) correct  *kəna  *kəna   six  *ənəm  *ənəm   full (container) *pənuq  *pənuʔ   (c) sea turtle *pəɲu  *pəɲu   to swallow *təɲəl  *təɲəl fat, grease *məɲak  *məɲak   (d) ankle  *bəŋil  *bəŋil   deaf  *dəŋəl  *dəŋəl   (e) vulva, vagina *təli  *təli   three  *təlu  *təlu   buy  *bəli  *bəlih  (f)  husked rice *bəʀas  *bəʀas   core of tree *təʀas  *təʀas   k.o. tree *təʀəp  *təʀəp  • Question: is it possible to extend the PNS intervocalic voiced stop preglottalization paradigm to other phoneme classes?  5. Preglottalized Sonorants in Kra-Dai  While the plain versus preglottalized opposition in the voiced obstruent series is quite robust within Kra-Dai, the same is not true for the sonorants. However, the Kam-Sui branch of Kra-Dai is particularly conservative in this regard, and a four-way phonation opposition can be reconstructed for all sonorants, with the single exception of a preglottalized lateral (which has merged with the voiceless lateral):  Table 8: The four-way phonation distinction in Kra-Dai sonorants Preaspirated Voiceless Plain Preglottalized *hm, *hn, *hȵ, *hŋ *m̥, *n,̥ *ȵ̥, *ŋ ̊ *m, *n, *ȵ, *ŋ *ʔm, *ʔn, *ʔȵ, *ʔŋ *hl, *hr, *hʀ *l,̥ *r,̥ *ʀ̥ *l, *r, *ʀ *ʔl, *ʔr, *ʔʀ *hw, *hj *w̥, *j ̊ *w, *j *ʔw, *ʔj  The modern Sui phonetic reflexes have been examined in some detail (Edmondson et al 2004). These are given below using the bilabial nasal series as an example. All Kam-Sui languages are tonal and distinguish between high (H) and low (L) registers, the former correlating with original voiceless initials and the latter with original voiced initials, as shown below using the bilabial series again as an example:  Table 9: Phonation and register Preaspirated Voiceless Plain Preglottalized *hm    >   m̥H *m̥    >    mH *m    >    mL *ʔm    >    ʔmH  It has generally been assumed that these phonation differences reflect the different glottal states of the initial consonants of sesquisyllabic words before the presyllable was lost. For example, an initial aspirated stop or fricative (with spread glottis) would result in a preaspirated sonorant, an initial voiceless initial would result in a voiceless sonorant, and an initial voiced initial would result in a voiced sonorant. However, this fails to explain the origin of the preglottalized series of sonorants, as no word-initial consonants with glottal constriction can be reconstructed:  (10)  (a) *tʰ-m  > *hm  (b) *t-m > *m̥  (c) *d-m >  *m  (d) ??? >  *ʔm  The hypothesis is put forward here that this final preglottalized series of sonorants, in contrast with the other three series, is not the result of conditioning by the presyllable initial. Rather, on analogy with the voiced stops, it occurs in an environment following schwa in which gemination occurred, followed by debuccalization of the first half of the geminate. Under this hypothesis, (10d) above may be revised as the following:  (11)  *Cəm-   >  *Cəm̆ː-  >  *Cəʔ̆m-  > *ʔm-  Examples of all four phonation types are given below:  (12)  *Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai  *Hlai  ‘dog’   *kʰ[u]máː *kʰ-mwaː *k-hmaː *m̥aː  *m̥aː  *hmaː ‘ditch’  *[t]-m̥ˠáːŋ ---  *T-m̥jaːŋ *m̥aŋ  *m̥ɯəŋ  --- ‘ant’  *r-móȶ  *mot  *r-mət  *muːʔ  *moc  *hmuȶ ‘bear’  *kəʔmˠúj *k-Nuːj  *ʔmjeː  ---  *m̥wiː  *ʔmuj Since other branches of Kra-Dai generally fail to preserve preglottalized sonorants, and since the Kam-Sui preglottalized sonorants usually correspond with voiceless sonorants in those branches, it can be posited that Proto-Kam-Sui is archaic in this aspect of its phonological inventory. This distinction was lost permanently in some branches (Lakkja, Ong-Be, Tai) as the original preglottalized sonorants series merged with the voiceless sonorant series. It was, on the other hand, preserved indirectly in the Hlai branch, where the voiceless series of sonorants merged with the preaspirated series, leaving the preglottalized series to fill the vacancy that was created in what became a chain shift:  (13)  (a) Kra-Dai *hm, *m̥, *m  >  Proto-Hlai *hm  (b) Kra-Dai *ʔm   >  Proto-Hlai *m̥  Table 10: Reflexes of medial sonorants after schwa in Kra-Dai *Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai *Cəm *m̥ *ʔm *m̥ *m̥ *ʔm *Cən *n ̥ *ʔn *n ̥ *n ̥ *ʔn *Cəȵ *ȵ̥ *ʔȵ *ȵ̥ *ȵ̥ *ʔȵ *Cəŋ *ŋ ̊ *ʔŋ *ŋ ̊ *ŋ ̊ *ʔŋ *Cəl *l ̥ *l ̥ *l ̥ *l ̥ *ʔl *Cər *j ̊(< *r)̥ *ʔr *ʃ *Cr *hr *Cəʀ (*j)̊ *ʔʀ (*ʃ) (*ʀ̥) (*hr) *Cəw *w̥ *ʔw *w̥ *ʔw (*ʔw) *Cəj *j ̊ *ʔj *j ̊ *ʔj (*ʔj)  Some examples are given below:  (14)  *Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai  *Hlai  ‘thick’  *tsəʔnáː *ts-Naː  *ʔɳaː  *nḁː  *nḁː  *ʔna: ‘cold’  *kəʔȵít  *k-Ni:t  *ʔȵit  *ni̥ːt  *ni̥t  --- ‘stupid’  *Cəʔŋáːŋh *ʔŋaːŋh  *ʔŋaːŋh  *ŋə̊ːŋ X  ---  --- ‘gills’  *Cəʔŋˠáːk ---  *ʔȵaːk  *ŋaːk  *ŋɯ̊ək  *ʔŋaːk ‘taro’  *pəʔrˠáːk *jåːk (<*rḁːk) *ʔrjaːk  *ʃaːk  *prɯək  *hraːk ‘moan’  *gəʔráːŋ *jåːŋ (<*rḁːŋ) *ʔraːŋ  ---  *graːŋ  --- ‘hungry’ *məʔjáːk ---  *m-ʔjaːk *jåk  *ʔjaːk  --- ‘stupid’  *Cəʔwáːʔ ---  *ʔwaːʔ  ---  *ʔwaːʔ  ---  6. Conclusion  There may be good phonetic motivations for the development of post-schwa geminate consonants (including sonorants) and their subsequent debuccalization in Pre-Kra-Dai:  • Iambic prosody: encouraged lengthening and fortition in post-schwa medial environment • Reduction of schwa duration: favored the ultimate deletion of schwa itself • Glottal stop acted as a final syllabic nucleus place-holder before ultimate deletion of the presyllable • Glottal stop was therefore a phonetic default which then became phonologized:   (15) *tsənáː  >  *tsən̆ːáː  >  *tsəʔ̆náː > *tsʔn̩áː  >   *ʔnáː   References  Antonissen, A. 1958. Kadazan-English and English-Kadazan Dictionary. Canberra: Government Printing Office.  Baxter, William H., and Laurent Sagart. 2014. Old Chinese: a new reconstruction. New York: Oxford University Press. 	Blust, Robert. 1995. Notes on Berawan Consonant Gemination. Oceanic Linguistics 34 (1): 123-38.  ---. 1998. The Position of the Languages of Sabah. In Pagtanáw: Essays on Language in Honor of Teodoro A. Llamzon, Ma. Lourdes S. Bautista (ed.). 29-52. Manila: Linguistic Society of the Philippines.  ---. 2001. Language, Dialect, and Riotous Sound Change: The Case of Sa’ban. In Papers from the Ninth Annual Meeting of the Southeast Asian Linguistics Society 1999, Graham Thurgood (ed.). 249-359. Arizona State University Program for Southeast Asian Studies, Monograph Series Press, Tempe: Arizona State University.  ---. 2002. Kiput Historical Phonology. Oceanic Linguistics 41 (2): 384-438.  ---. 2006. The Origin of the Kelabit Voiced Aspirates: A Historical Hypothesis Revisited. Oceanic Linguistics 45 (2): 311-38.  ---. 2007. Òma Lóngh Historical Phonology. Oceanic Linguistics 46 (1): 1-53.  ---. 2010. The Greater North Borneo Hypothesis. Oceanic Linguistics 49 (1): 44-118.  Edmondson, Jerold, John Esling, Jimmy Harris & James Wei. 2004. A phonetic study of Sui consonants and tones. Mon–Khmer Studies 34:47–66. Norquest, Peter. 2013. A Revised Inventory of Proto Austronesian Consonants: Kra-Dai and Austroasiatic Evidence. Mon-Khmer Studies 42: 102-126. ----. 2015. A Phonological Reconstruction of Proto-Hlai. Brill Academic Publishers.  ----. 2016. Revisiting the question of Austronesian implosives. Mother Tongue XX. Norquest, Peter & Sean Downey. 2013. ‘Expanding the PAn Consonant Inventory.’ Journal of the Southeast Asian Linguistics Society (JSEALS) 6: 99-145. Prentice, D.J. 1974. Yet Another PAN Phoneme? Oceanic Linguistics 13: 33-75.   A Hypothesis on the Origin of Preglottalized Sonorants in Kra-DaiPeter Norquest West Coast Conference in Formal Linguistics 38University of Arizona The University of British Columbianorquesp@u.arizona.edu March 8, 2020Introduction• A distinction between plain and preglottalized initial voiced stops is reconstructible within all primary branches of the Kra-Dai phylum at three places of articulation. • This distinction may be hypothesized to be a secondary development within Kra-Dai, where the diachronic trajectories of what were originally medial voiced stops depended on whether or not the preceding vowel was schwa on analogy with a similar development in Proto-North Sarawak (Blust 2006, Norquest 2016). Introduction• While this plain versus glottalized opposition in the voiced obstruent series is quite robust within Kra-Dai, the same is not true for the sonorants, which in many languages show only a two-way contrast between voiced and voiceless. • However, the Kam-Sui branch of Kra-Dai is particularly conservative in this regard, and a four-way phonation opposition (preaspirated, voiceless, voiced and preglottalized) can be reconstructed for all sonorants with the exception of the lateral. • The Hlai branch of Kra-Dai can now also be postulated as having retained evidence for a series of preglottalized sonorants.Introduction• While it has generally been assumed that these phonation differences reflect the different glottal states of the initial consonants of sesquisyllabic words before the presyllable was lost, the suggestion is put forward here that the preglottalized series of sonorants is not the result of conditioning by the presyllable initial. • Rather, on analogy with the voiced stops, it occurs in an environment following schwa in which gemination occurred, followed by debuccalization of the first half of the geminate.The distribution of the Kra-Dai phylum• The Kra-Dai family is composed of four major branches: (1) Kra, (2) Kam-Sui, (3) Hlai, and (4) Tai (3 branches); the smaller Be group is also shown. • Lakkja and Biao are not shown, but are located to the east of the Tai group towards the Pearl River Delta.BiaoLakkjaThe Kra-Dai Phylogenetic Tree (Norquest 2015)• Although the hypothesis presented here is valid for the entire Kra-Dai phylum, the scope of the data provided in this presentation includes everything under the scope of Lakkja-Kam-Tai.• We have very little data on the Biao languages – one full wordlist for the standard language, but only a handful of forms for the other (at least three major) languages, which hinders comparative work.• I haven’t done sufficient work on the Krabranch to have good control of the data –this is a future goal.The plain ~ preglottalized contrast in Kra-Dai initial voiced stopsPlain Preglottalized*b, *d, *ɟ *ʔb, *ʔd, *ʔɖ, *ʔɟA distinction between plain and preglottalized initial voiced stops is reconstructible within the Kra-Dai phylum at three places of articulation, with a fourth (retroflex) occurring in the latter series:Reflexes of plain initial voiced stops after non-schwa vowels in Lakkja-Kam-Tai*Kra-Dai ? *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai*b *b *b *v *b *ɓ*d *d *d *ɦ *d *ɗ*ɟ *d *zj *ʒ *ɟ *tɕM. Chinese *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai‘goose’ --- --- --- *vuːn X *bɯnh *ɓunɦ‘copper’ 銅 *dəwŋ *doŋ *dɔŋ *ɦoːŋ *doːŋ *ɗuːŋ‘money’ 錢 *dzjen *diːn *zjen *ʒeːn *ɟeːn *tɕiːn*	Middle Chinese data are taken from Baxter & Sagart (2014).Reflexes of plain initial voiced stops after non-schwa vowels in Kra-DaiPre-Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai*b *p *p *p *p *pʰ*d *t *t *t *t *tʰ*ɟ *ts *c (*ts) *c *tɕʰ*RAn *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai‘grandparent’ *apu --- --- --- *puː *phuːʔ‘mouth’ *baqbaq --- *paːk *paːk *paːk ---‘to fall’ *-tuq --- *tɔk *tɔk *tok *thok‘chest, liver’ *dəbdəb *tap *tap *t[a]p *tap ---‘sharp’ *ʈaɟəm --- --- --- --- *tɕhəːmReflexes of medial voiced stops after non-schwa vowels in Kra-Dai*Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai*C-b *w ̥ *C-b *ʔb *C-b *ʋ*C-d *l̥ *C-ʔɖ *r ̥ *C-d *ɾ*C-ɖ *j̊ (< *r)̥ *C-ʔɖ *r *C-ɖ *ɾ*C-ɟ *l̥ *ʔj *j̊ *ʔj *hj*Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai‘thin’ *C-báːŋ *wḁːŋ *C-baːŋ *ʔbjaŋ *C-baːŋ ---‘bone’ *Cudə́:k --- *C-ʔɖaːk *rɯ̥k *C-dwoːk      *Cuɾɯːk‘boat’ *Cuɖá: *j̊waː *C-ʔɖrwaː *rwaː *C-ɖwaː *Cuɾaː‘borrow’ *C-ɟáːm *l̥aːm *ʔjaːm --- *ʔjɯːm ---	Reflexes of medial voiced stops after schwa in Kra-Dai*Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai*Cəʔb *ʔb *ʔb *ʔb *ʔb *ɓ*Cəʔd *l̥ *ʔd *r ̥ *ʔd *ɗ*Cəʔɖ *l̥ *ʔɖ *r ̥ *ʔd *ɗ*Cəʔɟ *j̊ *ʔɟ *j̊ *ʔɟ *tɕ*Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai	‘village’ *Cəʔbáːnʔ *ʔbaːnʔ *ʔbaːnʔ --- *ʔbaːnʔ ---‘winnow basket’ *Cəʔdóŋʔ *l̥oŋʔ *ʔdɔŋʔ *ro̥ːŋ X *ʔdoŋʔ *ɗoŋʔ‘to stand’ *Cəʔɟún *j̊uːn *ʔɟun *j̊un *ʔɟɯn *tɕu:nProto-North Sarawak Voiced Stop Fortition after Schwa(a) Gloss PMP PNSash *qabu *abuh3pl *(si-)ida *idahrain *quɟan *uɟandigging stick *tuɢal *tugal(b) Gloss PMP PNSsugarcane *təbuh *təbːuhfaint *mədan *mədːanblink *kəɟəp *kəɟːəpsleep --- *məgːəl*Please note that traditional PMP phonemes are interpreted in the following way: *j = [ɖ], *z = [ɟ], *g = [ɢ], *R = [ʀ], *y = [j], and *e = [ə].Reflexes of Proto-North Sarawak voiced geminate stopsPNS *bː *dː *ɟː *gːBintulu ɓ ɗ ɟ gKenyahKenyah (Long San) ɓ ɗ ʄ ɠKenyah (Long Dunin) b/ɓ d/ɗ s ɠKenyah (Long Wat) b d ɟ gKenyah (Long Anap) p t c kKelabiticKelabit (Bario) bh dh dh ghKelabit (Long Napir) f s s kKelabit (Pa’ Mada) p t t kKelabit (Tring) p c c kBerawan-Lower BaramBerawan (Long Terawan) p c c kBerawan (Long Jegan) p c c kReflexes of PNS geminate voiced stops by type of changePNS *bː *dː *ɟː *gːShorteningKenyah (Long Wat) b d ɟ gImplosionKenyah (Long San) ɓ ɗ ʄ ɠKenyah (Long Dunin) ɓ/b ɗ/d s ɠBintulu ɓ ɗ ɟ gAspirationKelabit (Bario) bh dh dh ghDevoicingKelabit (Pa’ Mada) p t t kKenyah (Long Anap) p t c kKelabit (Tring) p c c kBerawan (Long Terawan) p c c kBerawan (Long Jegan) p c c kPossible trajectories of change for an intervocalic voiced geminatePreglottalization > Implosion-b:- > -ʔb- > -ɓ-Partial devoicing > Full devoicing-b:- > -bp- > -p-Aspiration > (Devoicing > Frication)-b:- > -bh- ( > -ph- > -f-)SabahanVoiced Stop Fortition after Schwa(a) Gloss PMP Kadazancloud *ʀabun gavun (*b > v)housepost *hadiʀi to-igi (*d > Æ)paddy *paɖaj paaj (*ɖ > Æ)indicate *tuɟuq tuuʔ (*ɟ > Æ)(b) Gloss PMP Kadazanstab *təbək tobok (< *təbːək)hiccough *sədu sodu (< *sədːu)sting, smart *hapəɖəs podos (< *pədːəs)pinch *kəɟut kodut (< *kədːut)Proto-North Sarawak Voiceless Obstruents after SchwaGloss PMP PNS Gloss PMP PNS(a) fathom *dəpa *dəpa (c) stick, adhere *dəkət *dəkətfour *əpat *əpat bracelet *ləku *ləkuʔblowpipe *səput *səput(b) design, tattoo *bətik *bətik (d) full, satiated *bəsuʀ*bəsuʀbamboo sp. *bətuŋ *bətuŋ flesh, muscle *həsi *əsifart *qətut *ətut rice mortar *əsuŋ *əsuŋProto-North Sarawak Sonorants after SchwaGloss PMP PNS Gloss PMP PNS(a) fat, grease *ləmu *ləmu (d) ankle *bəŋil *bəŋilweak *ləmaq *ləmaʔ deaf *dəŋəl *dəŋəlsweet *əmis *əmis(b) correct *kəna *kəna (e) vulva, vagina *təli *təlisix *ənəm *ənəm three *təlu *təlufull (container) *pənuq *pənuʔ buy *bəli *bəlih(c) sea turtle *pəɲu *pəɲu (f) husked rice *bəʀas *bəʀasto swallow *təɲəl *təɲəl core of tree *təʀas *təʀasfat, grease *məɲak *məɲak k.o. tree *təʀəp *təʀəpThe four-way phonation distinction in Kra-Dai sonorantsPreaspirated Voiceless Plain Preglottalized*hm, *hn, *hȵ, *hŋ *m,̥ *n̥, *ȵ̥, *ŋ̊ *m, *n, *ȵ, *ŋ *ʔm, *ʔn, *ʔȵ, *ʔŋ*hl, *hr, *hʀ *l̥, *r,̥ *ʀ̥ *l, *r, *ʀ *ʔl, *ʔr, *ʔʀ*hw, *hj *w,̥ *j̊ *w, *j *ʔw, *ʔjPhonation, register and the origin of the four-way contrastPreaspirated Voiceless Plain Preglottalized*hm    >   mH̥ *m ̥   >    mH *m    >    mL *ʔm >    ʔmH(a) *tʰ-m > *hm(b) *t-m > *m ̥(c) *d-m > *m(d) ??? > *ʔmPhonation and the origin of the four-way contrast(a) *tʰ-m > *hm(b) *t-m > *m̥(c) *d-m > *m(d) *Cəm-	 > *ʔmExamples of the four-way contrast*Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai       *Hlai‘dog’ *kʰ[u]máː *kʰ-mwaː *k-hmaː *mḁː *mḁː      *hmaː‘ditch’ *[t]-m ̥ˠ áːŋ --- *T-mj̥aːŋ *mḁŋ *mɯ̥əŋ        ---‘ant’ *r-móȶ *mot *r-mət *muːʔ *moc     *hmuȶ‘bear’ *kəʔmˠúj *k-Nuːj *ʔmjeː --- *mw̥iː     *ʔmujPreservation of preglottalized sonorants in Proto-HlaiProto-Kra-Dai  Proto-Hlai(a) *hm > *hm*m̥ > *hm*m >   *m̥ > *hm(b) *ʔm > *m̥ Reflexes of medial sonorants after schwa in Kra-Dai*Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai *Hlai*Cəm *m ̥ *ʔm *m ̥ *m ̥ *ʔm*Cən *n̥ *ʔn *n̥ *n̥ *ʔn*Cəȵ *ȵ̥ *ʔȵ *ȵ̥ *ȵ̥ *ʔȵ*Cəŋ *ŋ̊ *ʔŋ *ŋ̊ *ŋ̊ *ʔŋ*Cəl *l̥ *l̥ *l̥ *l̥ *ʔl*Cər *j̊ *ʔr *ʃ *Cr *hr*Cəʀ (*j̊) *ʔʀ (*ʃ) (*ʀ̥) (*hr)*Cəw *w ̥ *ʔw *w ̥ *ʔw (*ʔw)*Cəj *j̊ *ʔj *j̊ *ʔj (*ʔj)Examples of medial sonorants after schwa in Kra-Dai*Kra-Dai *Lakkja *Kam-Sui *Ong-Be *Tai          *Hlai‘thick’ *tsəʔnáː *ts-Naː *ʔɳaː *n̥aː *n̥aː         *ʔna:‘cold’ *kəʔȵít *k-Ni:t *ʔȵit *n̥iːt *n̥it           ---‘stupid’ *Cəʔŋáːŋh *ʔŋaːŋh *ʔŋaːŋh *ŋ̊əːŋ X ---           ---‘gills’ *Cəʔŋˠáːk --- *ʔȵaːk *ŋaːk *ŋ̊ɯək        *ʔŋaːk‘taro’ *pəʔrˠáːk *jåːk *ʔrjaːk *ʃaːk *prɯək        *hraːk‘moan’ *gəʔráːŋ *jåːŋ *ʔraːŋ --- *graːŋ            ---‘hungry’*məʔjáːk --- *m-ʔjaːk *jåk *ʔjaːk            ---‘stupid’ *Cəʔwáːʔ --- *ʔwaːʔ --- *ʔwaːʔ            ---Conclusion: Phonetic motivations • Iambic prosody: encouraged lengthening and fortition in post-schwa medial environment• Reduction of schwa duration: favored the ultimate deletion of schwa itself• Glottal stop acted as a final syllabic nucleus place-holder before ultimate deletion of the presyllable• Glottal stop was therefore a phonetic default which then became phonologized:		*tsənáː  >  *tsə̆nːáː   >  *tsə̆ʔnáː  >  *tsʔ̩náː  >  *ʔnáː 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.73804.1-0389866/manifest

Comment

Related Items