SENRI ETHNOLOGICAL STUDIES 75: 199–221 ©2009

Issues in Tibeto-Burman Historical Linguistics

Edited by Yasuhiko Nagano

Is There a Himalayan Tone Typology?

Jonathan P. Evans
Academia Sinica

1. Introduction
2. Documentation of Typological Features
2.1 Binarity of Tone
2.2 Culminativity and Restricted Locations
2.3 Right Edge Effects
2.4 Morpho-Tonemics
2.5 The Tone-Laryngeal Interface
3. Discussion
4. Conclusions
Appendix

1. Introduction
It has long been noted that Tibeto-Burman tone systems differ widely in structure from each
other. This paper considers the tone systems of Western Tibeto-Burman languages, especially
those of the Qiangic, Bodic, Tani, Bodo-Garo and Kuki-Chin-Naga branches, to see what
commonalities, if any, emerge. While there does not appear to be an iron-clad typology, cer-
tain traits emerge from a cross-linguistic comparison of tone systems:

1) Prevalence of binary oppositions, with one member of the pair more marked than the other.
2) No more than one of each type of tonal distinction per word (culminativity).
3) Restricted location of tone distinctions (e.g., only stem-penultimate in Caodeng rGyalrong,
word-initial elsewhere).
4) Right edge effect of (historically) checked syllables, which may be either synchronically
predictable or distinctive (e.g., Tibetan dialects, Caodeng rGyalrong).
5) Morpho-tonemics:
a. Affixes may be extrametrical (toneless: Caodeng rGyalrong, Lhasa Tibetan, Muka
Qiang, Bodic, Bodo-Garo). However, this can lead to a reversal, where the suffixes
can become the carrier of the tone (Tamangic).
b. Tonal polarity under certain morphological conditions (Mianchi Qiang, Zhuokeji
rGyalrong).
199
200 Jonathan P. Evans

c. Morphologically determined tone deletions (Caodeng rGyalrong). Note that this does
not include deletions due to culminativity restrictions.
d. Morphosyntactically assigned tones (Caodeng rGyalrong).
6) Tone-laryngeal interface:
a. Breathiness (Tamangic, Kham)
b. Voicing/aspiration of Ci (Tibetan dialects, Tamangic).

This paper has two aims. The first is to assemble typologically relevant features of tonal
systems of languages in this area. The second aim is to thereby alert Himalayan linguists to
a list of features that may be present in tone systems of interest, which it is hoped will lead to
a greater number of detailed descriptions.
To date there have been various overviews of characteristics of Tibeto-Burman tone
systems (e.g., Matisoff 1999; Weidert 1987; Yip 2002). Sources note that languages in this
family range from having many tonal contrasts to none, and from displaying emerging tonal
contrasts to disappearing ones. With such variety, it is not possible to establish one set of
typological characteristics shared by all Tibeto-Burman languages.
The aim of this paper is to compare tonal properties of western TB languages, starting
from the Qiangic languages of Sichuan and Yunnan, spreading westward across the Tibetan
plateau, through Bhutan and Northern India, and into Nepal. In selecting sources of data, I
have focused on those papers that include discussions of word-level phenomena.
It should be noted at the outset that most, but not all languages in the Sino-Tibetan family
are tonal. Lexical pitch distinctions are not found in some varieties of rGyalrong (‘Jiarong’,
Qiangic; Sichuan Province, China) (Nagano 1984, 2003), in most of Northern Qiang (H. Sun
1981; Liu 1998; cf. Evans 2006a; Huang and Zhou 2006), Newari (Genetti 2003; Hargreaves
2003), Garo (Burling and Joseph 2007), some Tani languages (J. Sun 2003), most of Kiranti
(Ebert 2003), and various Tibetan dialects.
The extinct Tangut language had two tones (Nathan Hill, p.c., Gong 2003). In some
cases, changes in Tangut tone indicate derivational changes: ljii1 ‘trousers’, ljii2 ‘put on trou-
sers’; ŋewr1 ‘to count’, ŋewr2 ‘a number’. However, tonal changes do not always indicate
changes in part of speech: nər1, nər2 ‘yellow.’ At this time, nothing is known of word-level
prosody in Tangut.
The Sino-Tibetan language family is comprised of approximately 250 to 300 languages
(Matisoff 2003), few of which have been described in sufficient detail to include in this
study. In addition, this study does not examine languages with “omnisyllabic” tone systems
(Matisoff 1999), in which each syllable has its own tone assignment which is relatively unaf-
fected by neighboring syllables and word-level prosody. Such languages may be found in
Lolo-Burmese, as well as among the ‘dialects’ of Chinese (Chen 2000). The languages in
this study whose systems are closest to omnisyllabic are the Kuki-Chin-Naga languages Ao
and Kuki-Thaadow.1)
The following discussion follows the order of the topics given in the Introduction.
Phonological characteristics of tone are summarized by language in the appendix.
Is There a Himalayan Tone Typology? 201

2. Documentation of Typological Features

2.1 Binarity of Tone
Throughout the Himalayan area, there is a prevalence of binary tone oppositions, in which
one tone may be more marked (restricted) than the other. A surface distinction between two
tones, say /H/ and /L/, may be phonologically structured in multiple ways. The opposition
may be between two specified tones (phonological binarity), or between toned and toneless
syllables (privativity, unarity). Some binary systems also contain toneless syllables (Mianchi
Qiang, Evans 2008). See Hyman (2001b), Evans (n.d.) for further discussions of types of
tonal privativity. Many sources do not present a sufficiently detailed phonological analysis
to determine whether a surface binary opposition corresponds to phonological binarity or
privativity, so this paper does not split these into two typological categories.
The presence of /H/, /L/, and toneless syllables in Mianchi Qiang may be demonstrated
by numeral compounds with the generic classifier /qo (H)/, whose floating tone links to the
leftmost toneless syllable. When paired with a /L/-toned numeral, a classifier in this group
surfaces with /H/; if the numeral is toneless, (H) links to the numeral:

(1) Mianchi NUM-CL and linking of (H). (Evans 2008)

Low numerals: Toneless numerals:

/L-(H)/ /Ø-(H)/

/a-qo/ /si-qo/
[à-qó] [sı ́-qò]
one-CL three-CL
‘one (thing)‘ ‘three (things)‘

[nə̀-qó] ‘two (things)’ [tʂoú-qò] ‘six (things)’

[zə̀-qó] ‘four (things)’ [nə́-qò] ‘seven (things)’
[ʁuà-qó] ‘five (things)’ [gú-qò] ‘nine (things)’
[tʂʰɛ̀-qó] ‘eight (things)’ [ɦà.diú-qò] ‘ten (things)’

In some cases, complex surface tone patterns reduce to binarity on closer inspection. This
has been shown to be the case in Niuwozi Pumi (Ding 2006). In Ding’s analysis, the apparent
complex of word tones (more on this later) actually consists of just /H/ and /L/ pitches, and
/H/ can be specified to spread one syllable rightward (“High” tone) or not (“Falling”):

(2) Niuwozi Pumi mono- and disyllable pairs

High Falling Rising Form

(a) ʃiH ‘hundred’ ʃiF ‘louse’ ʃiR ‘new’ (citation)
ʃiH ɡeH ʃiF ɡeL ʃiL ɡeH (speech)

The presence of monosyllabic contours that then spread out on disyllables shows that
both /H/ and /L/ are phonologically active. Similar patterns may be observed in Kuki-
202 Jonathan P. Evans

Thaadow (Chin; Assam, Myanmar. Hyman 2007). In other cases, a surface contrast between
two pitches can be shown to consist of one phonologically specified tone (privative), with the
remaining pitches filled in by default rules. Muka (Southern) Qiang demonstrates “simple”
privativity, in which /H/ contrasts with Ø (surface /L/): ‘sickle’

(3) Muka Qiang tone patterns (Evans 2006b)

Surface σ σσ σσσ σσσσ tones

pattern:
a. all /H/ 'ŋo ‘ŋo mje ‘ŋo lo kwe ‘ra ko ɕi də T#
[ŋó] [ŋó mjé] [ŋó ló kwé] [rá kó ɕı ́ də́]
‘bovine’ ‘cow’ ‘old bovine’ ‘sickle’
b. /H/ on ʐu* ʐu* mje ʐu* lo kwe me* gu mi dʑi *
1st σ [ʐú] [ʐú mjè] [ʐú lò kwè] [mé gù mı ̀ dʑɪ ̀]
‘horse’ ‘mare’ ‘old horse’ ‘thunder’
c. /H/ on ksə zə* ksə zə* mje ksə zə* se ȵ̥i *
2nd σ [ksə̀ zə́] [ksə̀ zə́ mjè] [ksə̀ zə́ sè ȵ̥ı ̀]
‘musk deer’ ‘musk doe’ ‘musk deer liver’
d. /H/ on 'lu ɹa* 'lu ɹa* pə 'lu ɹa* ʐwe pə T#-*
1st two [lú ɹá] [lú ɹá pə̀] [lú ɹá ʐwè pə̀]
σ ‘Small ‘Small Heishui- ‘Small Heishui-
Heishui’ LOC’ field-LOC’
(place)
e. /H/ on last i i dʐu i lo kwe ba lu ba se Ø
σ [ı ́] [ı ̀ dʐú] [ı ̀ lò kwé] [bà lù bà sé]
‘chicken’ ‘pheasant’ ‘old chicken’ ‘thing’

The Zhuokeji dialect of Situ rGyalrong has a more complex type of privativity, in which
there are layers of tone assignment, and each layer has one tone to contribute (or not):2)
Is There a Himalayan Tone Typology? 203

(4) Zhuokeji tone assigment in tri- and quadrisyllabic phrase-non-final words (Lin forthcoming;
Evans n.d.)

trisyllabic quadrisyllabic
HL HL
lexicon: tamərdam samarpak kasənapri kɐnəsaksə
foot parsing: HL HL
final σ of (tamə)(rdam) (sama)(rpak) (kasə)(na)pri (kɐnə)(sa)ksə
quadrisyllables
is extrametrical
metrical tone: H H HL H H HL
/H/ on right (tamə)(rdam) (sama)(rpak) (kasə)(na)pri (kɐnə)(sa)ksə
syllable of first
foot
tone spread: H H HL H H HL
(tamə)(rdam) (sama)(rpak) (kasə)(na)pri (kɐnə)(sa)ksə
default tone: L H L H HL LH L LH HL
assign /L/ to all (tamə)(rdam) (sama)(rpak) (kasə)(na)pri (kɐnə)(sa)ksə
syllables still ‘flail‘ ‘shoulder ‘to cause to ‘to have lunch‘
toneless pole‘ dine‘

Systems of two tones are rampant in this area, including Lhasa Tibetan, if one analyses
the possibility of final contour as being conditioned by right-edge glottality (see J. Sun 1997,
for discussion of competing analyses of Lhasa tone). However, the Tamangic (or TGTM)
group is typified by four tones. Historically, these derived from two tones via a split. The
table below (Mazaudon 2005; Noonan 2003) shows that breathy phonation (in gray) is speci-
fied for some tones.

(5) Tamangic tones

Tamang Thakali Gurung Manang

tone Ris. Sahu Taglung Tukche Marpha Syang Ghachok Ngawal Nar-Phu
1. 54 44 55/44 54 43 43 33 33 53
2. 44 54 43 44/33 45 45 54 45 44
3. 33/22 11 33/22 11 33/22 11 11 54 12
4. 211 32 51 121 51 33/22 12 31 21/31

Several of these languages (although not all) could be analyzed as having a binary pitch
contrast coupled with a phonation contrast. Similarly, Kham (Kiranti; Nepal) has a four-way
opposition that has been described as opposition between two tones, combined with clear/
breathy phonation (D. Watters 2003).
The Tani/Mirish branch appears to include both binary tone systems and those with
204 Jonathan P. Evans

more. Apatani (Weidert 1987) has /L/ and /H/, which like Niuwozi Pumi can be specified to
spread rightward one syllable. Gallong (Adi-Galo) has three tones, which following Weidert’s
(1987) description, consist of /H/, /HM/, /HL/. Eastern Tani languages (e.g., Bokar) lack lexi-
cal tone (J. Sun 2003).
Of the languages and groups examined for this paper, only TGTM, Gallong, and the
languages of Nagaland are not typified by binarity.

2.2 Culminativity and Restricted Locations

Many prosodic systems in this area do not permit more than one tone specification per multi-
syllabic prosodic unit. It will be assumed that this unit is the prosodic word. This restriction
on tonality has gone by many names in the literature: restricted tone, (pitch) accent, word
tone. The optimality theory label OCP! (Myers 1997) has also been applied to this phenom-
enon. Following Hyman (2001a, n.d.), it will simply be called ‘tone’ in this paper.
In Mianchi Qiang, the culminativity restriction applies to /H/ tones, where only the first
/H/ in a prosodic word may surface (Evans 2008):

(6) Culminative restriction on /H/ in Mianchi Qiang

/L + L.H/ -> L-L.H /mɛ̀ + ɕı̀.peı́/ -> [mɛ̀-ɕı̀.peı́] ‘human body’

/LH + L.H/ -> LH-L.L /bʐě + ɕı̀.peı́/ -> [bʐě-ɕı̀.peı̀] ‘snake body’

In most other languages in this area, tone is only specified in one place, usually identified
in reference to the left or right edge of the word. In Zhuokeji, the only existing lexical tone
(/HL/) can only link to the last syllable in the prosodic word (4). In the Caodeng dialect of
rGyalrong, the marked tone falls on the stem-penultimate syllable. The counting requirement
is so strict that the tone falls on the prefix of a monosyllabic verb stem:

(7) Caodeng rGyalrong penultimate tone location (J. Sun 2008)

kɐ́d-ⁿdʒev ‘to roll’ kɐ-qɐ́se ‘to look for’

kɐ-sə́-ⁿdʒev ‘to cause to roll’ kɐ-qɐsə́se ‘to look for each other’

In Bodic, Mirish, and Niuwozi Pumi, only the tone specified for the first syllable/mor-
pheme is pronounced, although its contour may spread:
Is There a Himalayan Tone Typology? 205

(8) Niuwozi Pumi first syllable tone spread (Ding 2006)3

Monosyllable Disyllable Trisyllable Quadrisyllable

bɨHL bɨH ɡeL bɨH bʴõL bʴõ L bɨH bʴõL bʴõL ɡeL
‘honey’ ‘as for honey’ ‘roasted flour ‘as for roasted flour
with honey’ with honey’
bɨH bɨH ɡeH bɨH ɬiH ɹuL bɨH ɬiH pɜL tsɨL
‘sun’ ‘as for sun’ ‘sunflower stem’ ‘sunflower’
tʃʼɨLH tʃʼɨL mɐ̃H tʃʼɨL n̥ıH̃ dʒjɛ̃L tʃʼɨL n̥ıH̃ ʤjɛ̃LɹəL
‘dog’ ‘dog hair’ ‘dog-nose group’ ‘dog-nose groups’
tõL puH tõL puH kʼʉH tõL pu H mɜH ɬeL
‘donkey’ ‘donkey head’ ‘donkey tail’
dʒjõL dʒɨLH dʒjõ L dʒɨL kʼʉH dʒjõ L dʒɨL mɜH ɬeL
‘buffalo’ ‘buffalo head’ ‘buffalo tail’
ɹəL tʃʼɨLH ɹəL tʃʼɨL ʃõH ɹəL tʃʼɨL ʃõH ɡeH
‘liquor’ ‘clean liquor’ ‘as for clean liquor’
dəL ɹ̥əL ɹ̥ı̃r dəL ɹ̥əL ɹ̥ı̃L siH
‘concentrate’ ‘concentrated’

In Lavrung, the location of the pronounced tone is determined by a set of complex inter-
actions among input tones:

(9) Lavrung tone realization rules (J. Sun 2008)

(i) Accent the leftmost high (falling) tone.

(ii) If there are no high tones, then
a) accent the rightmost low tone if the accentual domain contains no toneless
syllables.
b) otherwise accent the leftmost low tone.

The following data exemplify the Lavrung tone association rules:

(10) Lavrung tone realizations

Input Output Type

spô’ ‘meadow’ + sʌsə̂ ‘wild berry’ → spôsʌsə ‘strawberry’ i
və̌ɣ ‘butter’ + dɀî ‘eat’ → vəɣ dɀî i
snə̌ɯ ‘broad bean’ + cʰʌ̌v ‘pod’ → snəɯcʰʌ̌v ‘broad-bean pod’ ii.a
və̌ɣ ‘butter’ + u-dɀı ̌ ‘eat [pfv] ’ → və̌ɣ u-dɀi ii.b

Because these tone rules (like the much simpler /H/ rule in Mianchi Qiang) rely on the
input tones to determine the output, rather than restricting tone to a certain location in the
word, these languages are not considered to have restrictions on where tone may surface.
206 Jonathan P. Evans

In Boro (Bodo-Garo; Burling and Joseph 2007), tones are culminative. Like the above
cases, tone is specified once per word; however, the location need not be specified, as the tone
category and morphological structure determine the pitch pattern for the entire word (see
Burling and Joseph 2007, for conventions on the location of tone markings):

(11) Boro tone examples

High (rising) Low level Falling

bái ‘break’ bài ‘buy’
gɯ-bá ‘thin’ gɯ-bà ‘embrace’ gɯ-bâ ‘vomit’
gɯ-bá-gɯn ‘will be thin’ gɯ-bà-gɯn ‘will embrace’ gɯ-bâ-gɯn ‘will vomit’

Boro syllables ending in /-p/ or /-t/ are more restricted in their tonal possibilities; e.g., the
main consultant did not draw tonal distinctions on syllables ending in /-p/.
In the closely related Bodo-Garo language Tiwa (Joseph and Burling 2001), one tone
per word is specified, and the specification can fall on any syllable. Words longer than three
syllables are not common:

(12) Tiwa tonal possibilities

Mono- Disyllables Trisyllables

1st σ 2nd σ 1st σ 2nd σ 3rd σ
H H-H M-H H-H-H M-H-H M-M-H
ná khú-jur kojá khú-jur-o yaŋ-gúl-o chor-ri-á
‘you’ ‘lip’ ‘red’ ‘on the lip’ ‘at the back’ ‘lime’
F H-L M-F H-M-L M-H-L M-M-F
nâ khân-jur paŋ-sı ̂ khân-jur-o paŋ-sı ̂-na che-la-râu
‘come out’ ‘ear’ ‘flute’ ‘on the ear’ ‘for the back’ ‘y. sis. Hu’

2.3 Right Edge Effects

As has been well documented, tonal Tibetan dialects have a possible right edge pitch contour,
conditioned by final glottal stop (at least historically). For those dialects which have lost the
glottal conditioning, this contour is distinctive. The wide range of tonal analyses of Lhasa
may be due to different degrees of preservation of this glottal feature among speakers. A
predictable pitch contour effect is also found in Caodeng rGyalrong.
Similarly, Garo syllables with final glottal stop correspond to /H/ toned syllables in other
Bodo-Garo languages, often with redundant glottal stop. In Boro, that glottal stop often ‘trav-
els’ with the /H/ tone to the final syllable of the word.
In Kuki-Thaadow, contour tones only appear in word-final position.
Is There a Himalayan Tone Typology? 207

2.4 Morpho-Tonemics
There are numerous interactions between tones and morphology that are possible. Many of
the sources do not mention these; those writings that focus on phonology often do not con-
sider morphological complexities, and sources related to morphosyntax often do not discuss
morphophonemics. However, there are a few patterns that have been noted, which are dis-
cussed presently.

2.4.1 Extrametrical affixes

Affixes can be extrametrical in some of these languages. Extrametricality does not have the
same implications in each case. Lhasa clitics, such as

“the perfective marker –pə- are extrametrical in that the host syllables they are attached to
are characterized by domain-final contours, as if the enclitics do not count as part of the tonal
domain (Qu 1981; Mazaudon 1977; Durand 1990).” (J. Sun 1997)

In Muka Qiang, lexically toneless words receive a postlexical /H/ on the last syllable;
for plurals, this tone falls on the last syllable before the plural marker, which shows that the
plural marker is extrametrical:

(13) Tonal pattern of lexically toneless words in Muka Qiang

a. /i/ [ı ́] ‘chicken’
b. /i-mje/ [ı ̀ mjé] ‘hen’
c. /i-lo.kue/ [ı ̀ lò kué] ‘old chicken’
d. /i-lo.kue= ŋa/ [ı ̀ lò kué ŋà] ‘old chickens’

In Caodeng rGyalrong, tone surfaces on the stem penult, regardless of the presence or
absence of suffixes. In Tamangic, “grammatical suffixes are reported to be devoid of distinc-
tive tones.” (Mazaudon 2005) As a result, they end up carrying some of the tonal melody, as
in the following examples from Risiangku Tamang (ibid.):
208 Jonathan P. Evans

(14) Risiangku tonal melody on verbs and suffixes

Figure 1 Affixes are also toneless in other languages of this area, such as Meithei
(Chelliah, 2003)

2.4.2 Tonal polarity

By definition, polarity can only occur in a binary tonal system. In Mianchi Qiang and in
Zhuokeji rGyalrong there are morphological processes that reverse a tone value. Because
Zhuokeji has a privative tonal system, this is equivalent to alternating the presence/absence
of tone.
In Zhuokeji, “almost all verbs resort to tone to achieve stem alternation. This morpho-
logical process involves tone polarity. That is, if a verb is on /HL/ in its Stem1 form, then it
switches to /Ø/ to achieve Stem2 form, and vice versa.” (Lin, forthcoming)
In Mianchi, negation of an existential verb causes the tone on that verb to switch between
high and low:

(15) Mianchi tonal polarity

Root NEG+root Gloss

ŋá mı̀-ŋà ‘have (wealth), there is (a matter)’
ʑı̀ mı̀-ʑı́ ‘there is (person)’
wɛ̀ mı̀-wɛ́ ‘there is (something fixed in place, as a tree)’
lɛ̀ mı̀-lɛ́ ‘there is (something contained)’
tɨ̀ mı̀-tɨ́ ‘there is (something on a surface)’

Kuki-Thaadow pronominal possessive proclitics have /L/ tone before /HL, H/, but /HL/
before /L/ (the rising tone on /zóoŋ/ ‘monkey’ comes from spreading of the proclitic’s /L/
Is There a Himalayan Tone Typology? 209

tone; /HL/ simplifies to /H/ in non-final position):

(16) Kuki-Thaadow tonal polarity

‘my’ ‘your’ ‘his/her’ ‘our (dual incl.)’
/L/: kà ûy nà ûy à ûy ı ̀ ûy ‘dog’ /ûy/
/L/: kà zǒoŋ nà zǒoŋ à zǒoŋ ı ̀ zǒoŋ ‘monkey’ /zóoŋ/
/HL/: ká kèel ná kèel á kèel ı ́ kèel ‘goat’ /kèel/

2.4.3 Tone deletions

In Zhuokeji rGyalrong, “the observational verb forms are all realized with the phrase-final
melody of /Ø/,” as seen in the following examples.

(17) Zhuokeji rGyalrong tone deletion (Lin, p.c.)

Present Imperfective Observational

a. ‘s/he is standing’ ŋa-rjap (/Ø/) na-rjap (/Ø/)
[H-L] [H-L]
b. ‘s/he is releasing (it)’ ŋɐ-lɐ̂t (/HL/) nɐ-lɐt (/Ø/)
[L-HL] [H-L]
c. ‘s/he is detouring’ ŋa-nə-pjol (/Ø/) na-nə-pjol (/Ø/)
[L-H-L] [L-H-L]
d. ‘s/he is resting’ ŋa-nə-nâ (/HL/) na-nə-na (/Ø/)
[L-H-HL] [L-H-L]
e. ‘s/he is putting (it) ŋa-sa-tak-tak (/Ø/) na-sa-tak-tak (/Ø/)
on top of (sth.)’ [L-H-H-L] [L-H-H-L]
f. ‘s/he is reading (it)’ ŋɐ-nə-mcɐ-rɐ̂w (/HL/) nɐ-nə-mcɐ-rɐw (/Ø/)
[L-H-H-HL] [L-H-H-L]

In Caodeng rGyalrong,
“Unprefixed verb forms are distinguished from prefixed ones by removal of inherent accent.
This quite idiocyncratic accent loss can be demonstrated by the following prefixed (18a) and
unprefixed (18b) forms of the accented verb kɐ-rɐ́ski ‘to pull’” (J. Sun 2008):

(18) Caodeng tone loss

a. kə-rɐ́ski ‘one who pulls’
tə-rɐ́skɛ ‘yousg will pull’ or ‘(Yousg) pull upward!’
jə-rɐ́skɛ ‘s/he will pull’
ʃə-rɐ́ski-jə kə ‘Let’s go pull!’
b. rɐskɛ ‘s/he will pull’
rɐskɛ-aŋ ‘I will pull’
rɐski-tsə ‘wedl will pull’
rɐski-jə kə ‘Let’s pull!’
210 Jonathan P. Evans

2.4.4 Tone insertions

Caodeng rGyalrong and Mianchi Qiang have morphosyntactic processes that insert tones. In
Caodeng, “The intensive forms of stative verbs are formed via partial reduplication” with the
addition of tone, if the verb is toneless:

(19) Caodeng tone insertion

Citation form Intensive

kə-xtɛʔ kə-xtə́xtɛ ‘to be big’
kə-wɐrneʔ kə-wɐrnə́rne ‘to be red’
kə-qɐrquʔ kə-qɐrqə́rqu ‘to be freezing’
kə-mɐ́rtsev kə-mɐrtsə́rtsev ‘to be spicy’

In Mianchi Qiang, if the verb root has a directional prefix, but is non-suffixed, then it
surfaces with a /LH/ tone, as in the third person perfective forms:

(20) Mianchi Morphological /LH/ tone

ʐɪ ́ ‘dry’ tə̀-ʐɪ ̌ ‘s/he dried (it).’

bʐà ‘big’ dà-bʐǎ ‘It got bigger.’

Similarly, causativized perfective verbs in which the causative suffix is the last syllable
of its clause appear with a /HL/ tone on the causative suffix:

(21) Mianchi Morphological /HL/ tone

Causative is the last syllable of its clause:
ɦà-ŋà bùʐà ʂè-qʰàlà-zə̂.
3-pl wall DIR-knock.down-CAUS
‘They knocked down the wall.’
Causative is not the last syllable of its clause:
ŋà bùʐà ʂè-qʰàlà-zə́-ɕà.
1sg wall DIR-knock.down-CAUS-PERF:1sg
‘I knocked down the wall.’

Mongsen Ao (Kuki-Chin-Naga; Nagaland, India) has three phonologically and percep-

tually level tones, H(igh), M(id), L(ow). Lexically, /H/ is marked: most non-derived lexical
items occur with /M/ and/or /L/ tones. However, the /L/ toned nominalizing suffix /-pàʔ/
causes the last syllable of the verb stem to be pronounced with /H/: /ləmsa/ ‘distribute’
(past), /ləmsá-pàʔ/ ‘distribution.’ (Coupe 2007)

2.5 The Tone-Laryngeal Interface

Tamangic languages, Tibetan dialects, and Kham show interactions between pitch patterns
and laryngeal characteristics, especially breathiness of the vowel and voicing/aspiration of
Is There a Himalayan Tone Typology? 211

the initial consonant.

In the Tamangic group, as shown by the shading in (5), all languages except Manang
have one to two tone categories that are characterized by breathy voice. In both Tamangic and
some Tibetan dialects, certain tones only occur with a given aspiration state. In Tamangic,

“for all eight dialects, the obstruent initials, on which a contrast of aspiration exists under
tone-1 and tone-2, do not contrast for aspiration under tone-3 and tone-4. In words of tones 3
and 4, only one laryngeal mode is found on word-initial consonants. In the five dialects with a
main High/Low division, the resulting archiphoneme is unaspirated.” (Mazaudon 2005)

In Lhomi Tibetan, high and low register are characterized by the following bundles of
features:

(22) Lhomi Tibetan register features (S. Watters 2002)

High register “tense”

modal voice on the vowel
glottal stop on vowel-initial syllables
strong aspiration
plain stops and affricates are voiceless

Low register “lax”

absence of above
often breathy voice on the vowel
voiceless stops and affricates vary between slight and no aspiration
vowels are longer than their high register counterparts; vowel quality contrasts are maintained
rather than vowel length contrasts

3. Discussion
Of the hundreds of Sino-Tibetan languages, only a small percentage have been described at
all; of these, even fewer have had their tone systems subjected to adequate phonetic, phono-
logical, and morphological scrutiny to be able to say anything definite about their typological
characteristics. In spite of this lacuna, certain patterns emerge from the languages that have
been discussed above.
In order to answer the question raised by the title of this paper, the notion of ‘typology’
must be clarified. In some uses, ‘typology’ refers to features that are found to frequently
occur among the group of languages under consideration, such as the large number of lan-
guages for which primary stress assignment is either (stem-) initial or penultimate (Downing
2004; Hyman 1977). The Greenbergian sense of typology is much more restrictive, and is
intended to have at least probabilistic predictive power, such as his Universal 2: “In languages
with prepositions, the genitive almost always follows the governing noun, while in languages
with postpositions it almost always precedes.” (Greenberg 1963)
212 Jonathan P. Evans

If we allow for tone typology categories that “leak,” (Hyman 2007) then there are some
observations to be drawn from our sample of about 20 languages. Before stating these
observations, we recall that a number of ST languages were excluded from this comparison,
namely those with “omnisyllabic” tone (Matisoff 1999), in which there are minimal word-
level effects on tonal phonology (e.g., Mandarin Chinese, Lahu).
The most commonly found features among these tonal systems are binary oppositions
(e.g., High vs. Falling), culminativity (found everywhere except Kuki-Chin-Naga and Dayang
Pumi (Matisoff 1997)), and restrictions to certain locations.
Binary tone systems increase in number if certain assumptions are made; e.g., if tone
and phonation type are separated, then numerous TGTM languages have binary tone systems,
coupled with a modal/breathy voice distinction.
For languages with restrictions on where tone may be specified, the most common loca-
tions are left and right edges of the phonological word or morphological stem. Contour tones
may be restricted to word-final (Muya, Ikeda 2002, Kuki-Thaadow, Hyman 2007) or clause-
final positions (contour tones on Mianchi Qiang verbs). Final-only contour tones are found
in languages across the world, due to the typical lengthening of final syllables that provides
enough time to articulate a contour (Zhang 2002).
Certain typological relationships are tautologous – tonal polarity is only found in lan-
guages with binary tone systems, and morphological tone deletions are only possible in lan-
guages that permit syllables to remain tonally non-specified.
There are some intriguing sub-typologies within the set of languages examined: mor-
phological manipulation of tones is common in Qiangic and Kuki-Chin-Naga, breathy phona-
tion is found in Bodic and Kham (due to the course of tonogenesis via registrogenesis).
There are no clear geographic tendencies in the structure of tone systems for this group
of languages. For example, within the Qiangic branch, rGyalrongic, which is comprised of
mostly binary tonal dialects, abuts Northern Qiang, which has only a few tonal forms in a few
varieties, which borders Southern Qiang, where most dialects have binary tone systems in
their native lexica. In the same area where rGyalronic languages are spoken, Amdo Tibetan
(no lexical tone) is in widespread use.
In spite of the lack of geographic trends among the languages examined here, we may
observe that “omnisyllabic” languages, such as Lahu, tend to fall more within the Sinosphere,
with influence from Chinese, Thai or other languages with larger tonal inventories and denser
tonal specification than is found among these more westerly ST languages (cf. Matisoff
1999).
In recent publications, two languages of this area, the Mianchi dialect of Qiang (Evans
2008), and Kuki-Thaadow (Hyman 2007) have been noted as having prototypical African
tendencies in their tonal systems:
Is There a Himalayan Tone Typology? 213

(23) African vs. East Asian vs. Mianchi tone properties

African languages (typical) Chinese dialects (typical) Mianchi more like…

Privative /T, Ø/ or binary Rich tonal inventory African (/H, L, Ø/)
contrast (/H, L/ or /H, L, Ø/)
Limits on adjacency of Limited OCP African
marked tone (OCP) or
co-occurrence at word level
(culminativity)
Sparse distribution of tone Every lexical syllable has African
specifications a tone
Toneless syllables receive Pitch of toneless syllables African
default tone (usually L) determined by context
Tones can float Tones are pre-linked to African
syllables. No floating
tones.
One-to-many tone Mostly one-to-one tone Chinese
associations (spreading) associations
Downstep caused by L No downstep Chinese
Level tones primary Level and contour tones are African
both primary
Contour tones are sequences Contour tones form units African
of level tones at a deeper level of
structure
Restricted occurrence of Contour tones occur freely African
contour tones
Morphological tones No morphemes that lack African
segmental content
Tonal polarity of some No tone polarity African
affixes
Tendency toward fixed tone No higher order restrictions African
patterns

Employing Pike’s (1948) characterizations, Hyman (2007) shows the ways that Kuki-
Thaadow is more similar to ‘Register tone systems’ (prototypical Bantu), than it is to ‘Contour
tone systems’ (Mandarin Chinese, Thai, etc.):
214 Jonathan P. Evans

(24) Kuki-Thaadow tonal characteristics (Hyman 2007)

A. “Contour tone systems” B. “Register tone systems” Kuki-T.

Fewer level tones than contours More level tones than contours B
Contour tones = units Contour tones = sequences B
(clusters)
Contour tones have free distribution Contour tones (clusters) are often B
within the utterance limited to the last syllable
Dissimilation of contour + contour Dissimilation of contour tones = rare B
Metathesis of features within Metathesis of contour tones = rare B
a contour
No downstep Downstep B
Floating tones = rare Floating tones = frequent B
Tone spreading = rare Tone spreading = frequent B
Function of tone = lexical Function = lexical and/or B
grammatical
Words are monosyllabic Words come in various sizes A
Tones are restricted by syllable type Tones may occur on any syllable type A

4. Conclusions
In recent years, there have appeared an increasing number of tonological studies of Himalayan
languages. Precise observations made in these works, coupled with unambiguous examples,
have made possible the typological characterizations drawn in this preliminary study. The
preceding discussion presents an attempt to find typological generalizations of tone systems
in “western” Tibeto-Burman, with regard to phonetic, phonological, and morphological
characteristics.
Outside of Qiangic, tone often has its phonetic “fellow travelers” – phonation type in
Tibetan, Kham, and TGTM (as well as in Burmese); a traveling final glottal stop in Bodo-
Garo; syllable length in Boro and Lhomi Tibetan.
In most of these languages, tones are not completely separated from other aspects of
phonology, especially prosodic word-level limitations on numbers and locations of tonemes.
As has been observed in the above table (Hyman 2007), it is not surprising that in lan-
guages in which words have variable lengths (whether measured by morphemes or syllables),
there are interactions between morphology and tone, whether morpho-syntactic (as in the
cases of tonal assignment, deletion, and polarity) or at the level of word-formation (e.g., tonal
realization in Lavrung).
This study has not delved into the histories of the various subgroups; in some cases, the
origins of the tonal systems are clear, as in Tibetan dialects. For groups such as TGTM and
Lolo-Burmese, tonal splits can be detected, but the origin of tones at the proto-subgroup level
are obscure. In other subgroups, such as Qiangic, there does not appear to be any tonal cor-
Is There a Himalayan Tone Typology? 215

respondence between languages.

So, to restate the question: Is there a Himalayan tone typology? Two answers may be
provided:

No: These languages do not provide evidence for strict implicational typologies in the
Greenbergian sense, aside from tautological trivialities.
Yes: There are typological tendencies that are spread across these related languages
(typology in the Downing 2004).

As more studies are done, especially on the little-known and prosodically under-
described languages of Northeast India, the typology(-ies) of tone in this area will become
increasingly clear.

Postscript. A brief digression on the notion of Tone-Bearing Unit

Consider the following three comments about Tone-Bearing Units (TBU):

• “The tone bearing unit is the element in the segmental tier to which tone associ-
ates.” (Gussenhoven 2004: 29)
• “It seems that tone always associates to prosodic entities, but languages can differ
as to whether the syllable or the mora is the TBU.” (Yip 2006; similar observation
in Hyman 2001a)
• “In TGTM languages the tone bearing unit in the lexicon is the morpheme; in the
sentence, it is the word, formed of a tonal lexical item plus its atonal grammatical
affixes. The tonal melody which characterizes the tone of the lexical item fully
determines the pitch of the suffix or string of suffixes. In many cases, the melody
simply extends, or rather deploys itself over the number of syllables available.”
(Mazaudon 2005)

Almost forty years ago, Ilse Lehiste (1970, cited in Hyman 2005) observed: “A certain
degree of vagueness seems to characterize most discussions of prosodic features.” In the
above quotes, the first two authors use TBU as a phonological unit (either segmental or pro-
sodic) that can accept a tone; Yip’s quote appears to be more specific, as Gussenhoven does
not here state what ‘elements’ can appear in the CV tier. For both of these authors, the TBU
may be seen as the phonological domain under the tone, although they disagree as to how
directly tone relates to segments.
By way of contrast, Mazaudon appears to define TBU as the domain over the tone. Her
first observation is that morphological units have tone in their lexical entries (this may be
taken as definitional of a tone language). Second, she notes that when Tamangic morphemes
are combined into a word, suffixes are toneless; their pitch specification is determined by
the tonal category of their lexical morpheme. Similar effects determine which lexical tone
specification gets pronounced when compounds are formed. In Mazaudon’s usage, TBU is
the morphological unit in which only a single tone specification is found in the pronounced
form. For Gussenhoven and Yip, TBU is the phonological unit which receives a tonal pitch
command.
216 Jonathan P. Evans

Acknowledgments
This paper has benefited from discussions with Robbins Burling, Alec Coupe, Scott DeLancey,
Nathan Hill, Larry Hyman, Alexis Michaud, You-Jing Lin, Yasuhiko Nagano and Jackson
T.-S. Sun. Any remaining errors are my own. The research was supported by funding from
the National Science Council (Taiwan), grant #96-2411-H-001-069.

Notes
1) The validity of Naga as a linguistic genealogical subgrouping is not universally accepted (Burling
2003).
2) It should be noted that fieldwork with other speakers of this same dialect have not found evidence
for tonal distinctions (Nagano 2003). Subsequent versions of Lin (forthcoming) arrive at the
surface form via a different derivation. However, the relevant observations – lexical tone can only
be specified on the final syllable, and the remaining tone assignments are not distinctive – are
unchanged.
3) /HL/ and /LH/ have been substituted for Ding’s /F/, /R/.

References
Burling, R.
2003 The Tibeto-Burman Languages of Northeastern India. In G. Thurgood and R. J. LaPolla
(eds.), Sino-Tibetan Languages, 169–192. London/New York: Routledge.
Burling, R. and Joseph, U. V.
2007 Boro tones. Paper presented at the 2nd Annual Meeting of the North East India Linguistic
Society. India, Assam.
Chelliah, S. L.
2003 Meithei. In G. Thurgood and R. J. LaPolla (eds.), Sino-Tibetan Languages, 427–438.
London/New York: Routledge.
Chen, M. Y.
2000 Tone Sandhi: Patterns across Chinese Dialects. Cambridge Studies in Linguistics 92.
Cambridge: Cambridge University Press.
Coupe, A. R.
2007 A Grammar of Mongsen Ao [Mouton Grammar Library 39]. Berlin/New York: Mouton de
Gruyter.
Ding, P.
1998 Fundamentals of Prinmi (Pumi): A Tibeto-Burman language of Northwestern Yunnan.
Ph.D. thesis, Canberra: Australian National Univeristy.
2006 A typological study of tonal systems of Japanese and Prinmi: Towards a definition of pitch-
accent languages. Journal of Universal Language 7(1): 1–35.
Downing, L. J.
2004 What African languages tell us about accent Typology. ZAS (Zentrum für Allgemeine
Sprachwissenschaft) Papers in Linguistics, 37: 101–136.
Is There a Himalayan Tone Typology? 217

Durand, J.
1990 Generative and Non-linear phonology. London: Longman.
Ebert, K. H.
2003 The Kiranti languages. In G. Thurgood and R. J. LaPolla (eds.), Sino-Tibetan Languages,
505–517. London/New York: Routledge.
Evans, J. P.
2006a Vowel quality in Hongyan Qiang. Language and Linguistics 7(4): 937–960.
2006b How do you analyze a three-tone system, when all three tones are high? Paper presented at
39th International Conference on Sino-Tibetan Languages and Linguistics. Sept 14–17,
Seattle, University of Washington.
2008 ‘African’ tone in the Sinosphere. Language and Linguistics 9(3): 463–490.
n.d. Structured privativity and the complexity of simplicity, or The ups and downs of privative
tone. ms.
Genetti, C.
2003 Bolakhā Newār. In G. Thurgood and R. J. LaPolla (eds.), Sino-Tibetan Languages, 355–370.
London and New York: Routledge.
Greenberg, J. H.
1963 Some universals of grammar with particular reference to the order of meaningful elements.
In J. H. Greenberg (ed.), Universals of Language, 73–113. Cambridge MA: MIT Press.
Gong, H.-C.
2003 Tangut. In G. Thurgood and R. J. LaPolla (eds.), Sino-Tibetan Languages, 602–620.
London/New York: Routledge.
Gussenhoven, C.
2004 The Phonology of Tone and Intonation. Cambridge: Cambridge University Press.
Hargreaves, D.
2003 Kathmandu Newar (Nepāl Bhāśā). In G. Thurgood and R. J. LaPolla (eds.), Sino-Tibetan
Languages, 371–384. London/New York: Routledge.
Hildebrandt, K.
2004 A Grammar and dictionary of the Manange language. In C. Genetti (ed.), Tibeto-Burman
Languages of Nepal: Manange and Sherpa, 2–189. Canberra: Pacific Linguistics.
2005 A phonetic analysis of Manange segmental and suprasegmental properties. Linguistics of
the Tibeto-Burman Area 28(1): 1–36.
Huang, B. and Zhou, F.
2006 Qiangyu Yanjiu [Studies on the Qiang Language]. Chengdu: Sichuan People’s Publishing
House.
Hyman, L. M.
1977 On the nature of linguistic stress. In L. M. Hyman (ed.), Studies in Stress and Accent, 37–82.
Southern California Occasional Papers in Linguistics 4. Los Angeles: Department of
Linguistics, University of Southern California.
2001a Tone systems. In M. Haspelmath, E. König, W. Oesterreicher and W. Raible (eds.), Language
Typology and Language Universals: An International Handbook 2: 1367–1380. Berlin/New
York: Walter de Gruyter.
2001b Privative tone in Bantu. In S. Kaji (ed.), Cross-linguistic Studies of Tonal Phenomena,
218 Jonathan P. Evans

237–257. Tokyo: Institute for the Study of Languages and Cultures.

2005 Word-prosodic typology. Paper presented at Stress and Tone conference. The Netherlands,
Leiden.
2007 Kuki-Thaadow: An African tone system in Southeast Asia. UC Berkeley Phonology Lab
Annual Report. Berkeley.
n.d. How (not) to do phonological typology: The case of pitch-accent. Manuscript submitted for
publication.
Ikeda, T.
2002 On pitch accent in the Mu-nya language. Linguistics of the Tibeto-Burman Area 25(2):
27–45.
Joseph, U. V. and Burling, R.
2001 Tone correspondences among the Boro languages. Linguistics of the Tibeto-Burman Area
24(2): 41–55.
Lehiste, I.
1970 Supersegmentals. Cambridge, MA: MIT Press.
Lin, Youjing.
2009 Units in Zhuokeji rGyalrong discourse: Prosody and grammar. Ph.D. dissertation, University
of California, Santa Barbara.
Liu, G.
1998 Mawo Qiangyu Yanjiu [Research on the Mawo dialect of the Qiang language]. Chengdu:
Sichuan Minzu Chubanshe.
Matisoff, J. A.
1997 Dàyáng Pumi phonology and adumbrations of comparative Qiangic. In D. Thomas (ed.),
Mon-Khmer Studies 27: A Journal of Southeast Asian Languages, 171–213. Dallas: Summer
Institute of Linguistics.
1999 Tibeto-Burman tonology in an areal context. In S. Kaji (ed.), Proceedings of the Symposium:
Cross-Linguistic studies of Tonal Phenomena, Tonogenesis, Typology, and Related Topics,
3–32. Tokyo: Tokyo University of Foreign Studies.
2003 Handbook of Proto-Tibeto-Burman: System and Philosophy of Sino-Tibetan Reconstruction.
Berkeley: University of California Press.
Mazaudon, M.
1977 Tibeto-Burman tonogenetics. Linguistics of the Tibeto-Burman Area 3(2): 1–123.
2005 On tone in Tamang and neighbouring languages: synchrony and diachrony. In S. Kaji (ed.),
Proceedings of the Symposium: Cross-Linguistic Studies of Tonal Phenomena: Historical
Development, Tone-Syntax Interface, and Descriptive Studies, 79–96. Tokyo: Tokyo University
of Foreign Studies.
Myers, S.
1997 OCP effects in Optimality Theory. Natural Language and Linguistic Theory 15: 847–92.
Nagano, Y.
1984 A Historical Study of the rGyalrong Verb System. Tokyo: Seishido.
2003 Cogtse Gyarong. In G. Thurgood and R. J. LaPolla (eds.), The Sino-Tibetan Languages,
469–489. London/New York: Routledge.
Is There a Himalayan Tone Typology? 219

Noonan, M.
2003 Nar-Phu. In G. Thurgood and R. J. LaPolla (eds.), The Sino-Tibetan Languages, 336–352.
London/New York: Routledge.
Pike, K. L.
1948 Tone Languages. Ann Arbor: University of Michigan Press.
Qu, A.
1981 Zangyu de biandiao (Tone sandhi in Tibetan). Minzu Yuwen 4: 20–27.
Sun, H.
1981 Qiangyu Jianzhi [A Brief Description of the Qiang Language]. Beijing: Nationalities
Press.
Sun, J. T.-S.
1997 The typology of tone in Tibetan, Chinese Languages and Linguistics IV: Typological Studies
of Languages in China. Symposium series of the Institute of History and Philology,
Academia Sinica; No. 2. Taipei: Academia Sinica.
2003 Tani languages. In G. Thurgood and R. J. LaPolla (eds.), Sino-Tibetan Languages, 456–632.
London/New York: Routledge.
2008 Tonality in Caodeng rGyalrong. In B. Huber, M. Volkart and P. Widmer (eds.), Chomolangma,
Demawend und Kasbek, Festschrift für Roland Bielmeier I: 257–280. Halle: International
Institute for Tibetan and Buddhist Studies.
Weidert, A.
1987 Tibeto-Burman Tonology. Amsterdam/Philadelphia: John Benjamins.
Watters, D. E.
2003 Kham. In G. Thurgood and R. J. LaPolla (eds.), The Sino-Tibetan Languages, 683–704.
London/New York: Routledge.
Watters, S. A.
2002 The sounds and tones of five Tibetan languages of the Himalayan region. Linguistics of the
Tibeto-Burman Area 25(1): 1–65.
Yip, M.
2002 Tone. Cambridge: Cambridge University Press.
2006 Tone. In P. de Lacy (ed.), The Cambridge Handbook of Phonology, 229–252. Cambridge:
Cambridge University Press.
Zhang, J.
2002 The Effects of Duration and Sonority on Contour Tone Distribution: A Typological Survey
and Formal Analysis. New York: Routledge.
Appendix: Properties of Selected Himalayan tone systems

Qiangic Bodic Tani/Mirish Bodo-Garo Kuki-Chin-Naga Kiranti

Qiang Qiang Jiarong Jiarong Jiarong Pumi Pumi Tibetan Kuki-
Muya TGTM Manange Apatani Gallong Tiwa Bodo / Boro Ao Kham
Mianchi Muka Caodeng Lavrung Zhuokeji Dayang Niuwozi Lhasa Thaadow
1. ‘Binary’ tone values + + + + + + + + + -/+ - + - + + +/- - +
2. Culminative + + + + + - + + + + + + + + + - - +
3. Restricted location - + + - + - + + + + + + + - ?; cf. (11) - - +
/-ʔ/ can shift
+
4. Checked tones on right edge - - - - - - + - - - H has /-ʔ/ to right edge - - -
(allophonic)
of word
5a. Suffix extra-metrical - + + - + + - - - - + + - +
5b. Polarity + + + -
5c. Morphological deletions - - + + -
5d. Tone assigned by morph + + + +
some +
6a. Breathy - - - - - - - - - - - - - +
dialects (Manang -)
6b. Voice, aspiration effects - - - - - - + (dial.) + +; < 100% - - - - - - -
Joseph, Burling,
Evans Evans J. Sun J. Sun Y. Lin Matisoff Ding Ikeda J. Sun Mazaudon Hildebrandt Weidert Weidert Hyman Coupe D. Watters
References Burling Joseph
2008 n.d. 2008 2008 2008 1997 1998, 2006 2002 1997 2004 2004, 2005 1987 1987 2007 2007 2003
2001 2007

Notes:
1. Numbers are page numbers in the given reference.
2. Minus (-) means that the source either explicitly states that the property is lacking, or the source is extensive enough to reasonably draw that conclusion.
3. Blanks represent unknown values.