A PHONOLOGY OF STAU
by
A. CHANTEL VANDERVEEN

A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS
in
THE FACULTY OF GRADUATE STUDIES
Master of Arts in Linguistics, Analytic Stream
We accept this thesis as conforming to the required standard

...............................................................................
Dr. Roderic F. Casali, PhD; Thesis Supervisor

................................................................................
Dr. Keith L. Snider, Ph.D.; Second Reader

................................................................................
Dr. Jamin R. Pelkey, Ph.D.; External Examiner

TRINITY WESTERN UNIVERSITY
May 2015
© A. Chantel Vanderveen

A phonology of Stau
A. Chantel Vanderveen
Trinity Western University

ABSTRACT
This thesis is a description of the phonology of Stau, a Rgyalrongic language of the
Tibeto-Burman family, based on original field research. Stau is spoken by approximately
23,000 people in the west of Sichuan province, China. It is an almost unstudied language.
Apart from a sketch of the phonology and grammar by Huang (1991), which provides a phonetic (rather than phonemic) analysis of Stau sounds, lists attested onsets and rhymes, and
discusses tone, there has been virtually no systematic study of the phonology of language.
This thesis provides a more extensive study of Stau phonology, covering segmental phonology, acoustic analysis of stops and of vowels, syllable structure, phonotactics, phonological
processes, and pitch phenomena.
Of particular interest in this phonology are Stau’s large phonemic inventory of fortytwo consonants and eight vowels, its large syllable canon which allows onset clusters of up
to three consonants and single coda consonants, phonotactic constraints among its consonant clusters, and vowel changes in reduplication.

ACKNOWLEDGEMENTS
I am indebted to many people who helped and supported me over the course of
researching and writing this thesis. First, I want to thank all those who advised me academically and linguistically. Jesse Gates’ kindness and his enthusiasm for Stau first inspired me
to ask him to mentor me in a fieldwork practicum (which doubled as thesis research). He
gave me much help and encouragement over the last several years during the long process
of research, analysis, and writing. One example of such help is that he commented on my
draft, saving me from publishing more than one glaring error. I am grateful to my thesis
advisor, Dr. Rod Casali, who led me through the process of writing a phonology, answering all my questions (such as, “What exactly should be in a phonology besides a phonemic
inventory?”). Through his patience and equanimity I always came away from his office
calmer and with renewed motivation. Without Rod’s comments on my drafts, this thesis
would be very much inferior. My second and third readers, Dr. Keith Snider and Dr. Jamin
Pelkey, also gave their time to read and comment on this thesis. Their constructive criticism helped me improve and polish my work. Jamin deserves further gratitude for being my
fieldwork practicum supervisor when I went to China. Katylin Wonnell, a fellow student of
Stau, also commented on my draft, and spent I-don’t-know-how-much time correcting my
verbs to be consistently in third person.
Our IT guys at CanIL, Doug Rintoul and Larry Hayashi, helped me prepare the
equipment and software I needed to do research and assisted me with all my computer
problems. I am also grateful to Andy Black, whom I’ve never met, but who has always been
so prompt and helpful to solve the problems I encountered using his wonderful XLingPaper
(xlingpaper.org). Thank you so much for this program! It has made formatting hassle-free.

iv

v

No more worrying about example numbering or tabs going out of whack (looking at you,
Microsoft Word)!
Many friends encouraged me and prayed for me, more than I can name. My cousin
Heidi encouraged me throughout the thesis-writing process. She helped motivate me by
getting me to talk about my work and by entering into my excitement when I made new
discoveries. My friends and former roommates Rachael and Giselle were there at the very
beginning of this process, when I tried to arrange a fieldwork trip in 2011 but it fell through
(which, looking back, was a blessing in disguise), and they were there for me via Skype and
email as I analyzed and wrote. I am thankful for my friends Ania and Brooke, who gave
this thesis a final proofread. Ania, my memorization buddy, helped keep me sane as I wrote
through our regular Skype calls.
Second most importantly, I thank my family: Dad & Mom, Jeff & Mandie, Hannah,
Reuben & Megan, Jared & Katrina, and Caleb. I am so grateful for my parents’ support and
love. They encouraged me to do the practicum in China to see if I liked field linguistics,
even though it meant going halfway around the world. Their unwavering support, despite
my taking longer than I planned or expected, gave me confidence to press on. At the same
time, they were there to give me a well-timed, gentle kick-in-the-butt to get this thesis
finished.
I thank my Opa for lending me In Search of the Source when I was in first-year at
University of Manitoba, which started me on the track that led me to TWU and to CanIL.
My greatest thanks go to my heavenly Father. I have learned over and over: “Unless
the Lord writes the paper, the student labours in vain” (Psalm 127:1, Revised Vanderveen
Version). SDG.

PREFACE
1

=

first person

2

=

second person

3

=

third person

aprfx

=

adjectival prefix

const

=

constative

cop

=

copula

dem

=

demonstrative

dim

=

diminutive

dirdown

=

directional prefix; downward

dirin

=

directional prefix; inward

dirneut

=

directional prefix; neutral

dirout

=

directional prefix; outward

dirup

=

directional prefix; upward

gen

=

genitive

ins

=

instrumental

neg

=

negation, negative

nmlz

=

nominalizer/nominalization

num1

=

numeral prefix for 1

num2

=

numeral prefix for 2

num3

=

numeral prefix for 3

pers

=

suffix for person

pl

=

plural

prf

=

perfect
vi

vii

red

=

reduplication

Chapter 1
Introduction
The primary goal of this thesis is to describe the phonology of Stau, a Rgyalrongic
language of the Tibeto-Burman language family. Stau phonology has a number of interesting features. Its consonants are one of its most interesting features: Stau has an unusually
large consonant inventory of 42 consonants. This size comes from having three voicing
distinctions in stops and affricates and two in fricatives. Stops, nasals, and fricatives, in
addition to being found at the usual labial, alveolar, and velar places of articulation, are
also found at palatal and uvular places of articulation. Finally, Stau also has two lateral
fricatives.
Stau has a complex syllable canon that allows final consonants and onset clusters
of up to three consonants. Stau retains its syllabic complexity from Proto-Tibeto-Burman,
which also had a complex syllable canon, unlike other current-day Tibeto-Burman languages which have “vastly simpler syllables” (Matisoff 1991:140). A large consonant inventory combined with a complex syllable canon means that Stau has a large number of
licit consonant clusters. The phonotactic constraints that come into play within the clusters
are a further interesting feature of Stau.
Stau vowels show some phonotactic constraints as well. Before codas, the front
tense vowels /i/ and /e/ are prohibited; in this context their lax counterparts [ɪ] and /ɛ/ appear
instead. Vowels play a role in reduplication: In some reduplications, the vowel of the base
changes when it is reduplicated. If the base has /æ/ or /ɛ/ as its nucleus, the vowel will
change to /ə/ in the reduplicant; in bases with /ə/ or /o/, the vowel changes to /ɛ/.

1

2

This phonological description is significant because it adds to the body of research
and analysis on Rgyalrongic languages. As will be made clear in §2.3, the study of Rgyalrongic languages is still young. Many questions are unsettled, such as how many languages
the subgroup comprises and how these languages are related to each other. I hope that this
description will contribute to the growth of Rgyalrongic studies and to an accurate understanding of this subgroup.
The youth of Rgyalrongic studies also means that the relationship of Rgyalrongic
to Tibeto-Burman is as yet undetermined. An understanding of the Rgyalrongic subgroup
will in turn contribute to understanding the Tibeto-Burman family. I do not plan in this
thesis to use my findings to directly address these questions, but I hope that the data in this
description could be used in comparison to that of other Rgyalrongic lects for historical
reconstruction.
Another reason that this phonological description is an important contribution is that
few in-depth descriptions of Rgyalrongic languages exist in English. Much of the literature
on the subgroup is in Chinese or (less frequently) in French. A fair number of articles are
available but, as for in-depth descriptions like theses and dissertations, most of these are
in Chinese or French (e.g. Mansier 1983, Huang 1991, Duo’erji 1998, Jacques 2004, Lai
2013). One exception is the dissertation of Prins (2011), and her work is on a lect from
Rgyalrong proper. This thesis is the first description of Western Rgyalrongic phonology
available in English.
In some ways, the present description reiterates what is described in Huang (1991)’s
description of Gexi Stau. However, that was not known at the outset of data collection. So
one outcome of this work is that it confirms that the varieties of Stau spoken in Gexi and
Mazi are the same or very similar. However, this thesis also adds to what Huang’s work
established. For instance, the segmental phonology here provides a phonemic inventory of
Stau, whereas in Huang’s inventory of sounds she does not differentiate between phoneme

3

and allophone. Where she does describe variation it is not framed as allophony, and the
contexts of variation are not always given. Huang’s description is more concerned with
accurate phonetic transcription. Knowing which Stau sounds are phonemic and which are
allophonic will be helpful if an orthography is developed. This work also adds to that of
Huang an acoustic analysis of the vowels and analysis of phonotactics and phonological
processes.

1.1 Overview
The rest of this thesis is as follows: Chapter 2 provides background information. I
will first introduce the Stau people (§2.1) and linguistic context (§2.2). The latter section
will particularly focus on understanding of Stau’s immediate genetic affiliations (§2.2.2)
and its relationship to Tibeto-Burman (§2.2.3). In §2.3 I will review the previous scholarship on Stau that has been available to me. Section 2.4 provides a description of my fieldwork practices. Finally, I outline what I expect to find in my analysis of Stau phonology
based on previous work on the language and its relatives (§2.5).
Chapter 3 presents Stau’s segmental phonology, beginning with a summary of the
phonemic inventory (§3.1). The next two sections give evidence of contrast for each of the
consonants (§3.2) and vowels (§3.3) introduced in the summary. In addition to evidence
of contrast, allophony and variation in pronunciation are discussed. Section 3.4 discusses
similarities and differences between the segmental phonologies of this thesis and Huang
(1991).
Chapter 4 provides acoustic analysis of voicing in stops and of vowels. Section 4.1
uses acoustic analysis to answer the puzzle of whether stops have two or three contrasts in
voicing. The vowels are also analyzed acoustically in order to determine their qualities as
defined by their first and second formants (§4.2).
In chapter 5, Stau syllable structure is described. First, §5.1 presents the syllable
types that Stau uses, and then §5.2 moves on to discuss syllabification in Stau. Chapter 6

4

deals with phonotactic restrictions in Stau. Section 6.1 looks at phonotactic constraints in
syllable onsets, focusing particularly on restrictions within consonant clusters, and §6.2 at
constraints within the rhyme.
Chapter 7 considers phonological processes. It begins with an excursus on the relationship between the vowels /ɛ/ and /æ/ (§7.1), which plays a part in several phonological
processes. From there, I move into a discussion of voice assimilation and nasalization of
/v/ (§7.2). Then, §7.3 presents several cases of vowel assimilation in Stau. Finally, vowel
changes that occur in reduplication are discussed in §7.4.
Chapter 8 is the final chapter of analysis, and it looks at pitch phenomena. Background information on how pitch phenomena have been analyzed in related languages is
given in §8.1. In the next section, I describe the pitch patterns that I observed in Stau on
words in isolation (§8.2). Then in §8.3, I interpret the data, tentatively proposing that Stau
has a fixed stress system.
Finally, Chapter 9 summarizes all that came before and concludes the thesis.

Chapter 2
Background information
Chapter 2 provides background information to set the stage for the descriptive part
of this thesis. We start by looking at the Stau people and culture in §2.1. In the following
section, there will be an introduction to the linguistic context in which Stau exists: how
Stau figures into the Tibeto-Burman language family and into its own subgroup. Section
2.3 looks at previous work that has been done on Stau, while §2.4 describes the details of
the fieldwork on which this thesis is based. Finally, §2.5 lays out my expectations for Stau
phonology based on previous scholarship.

2.1 The Stau people
The Stau live west of the Sichuan provincial capital, Chengdu, in the People’s Republic of China (PRC), in Daofu/rTau and Luhuo Counties of Ganzi/dKar-mdzes Prefecture.
Most Stau live along the Xianshui River valley in Daofu County. The focal point of the Stau
area is Daofu Xian, the county seat.
Daofu County belongs to an area of Sichuan Province known as the “Ethnic Corridor.” The Ethnic Corridor is a region characterized by mountainous topography and by ethnic diversity. It divides the Tibetan Plateau from the Sichuan Basin, stretching north-south
from “southern Gansu and eastern Qinghai down through western Sichuan and southeastern Tibet to western Yunnan and the northern Burmese and Indian frontiers” (Sun 1990:1).
The extremity of the Corridor’s high peaks and steep river valleys discouraged travel and
resulted in the isolation of communities, which in turn led to increased ethnic and linguistic
diversity. Stau is one of the languages that developed in this linguistic hothouse. The map

5

6

below shows the higher-order ethnolinguistic groups that live in the Ethnic Corridor (Stau
is within “Western Jiarongic”).

Figure 1. Map of the Sichuan Ethnic Corridor (from Gatehouse 2014)
The Ethnic Corridor has for a long time been under Tibetan cultural and religious
influence, and the Stau area is no exception to the rule. Thus, in 1950s, when the new
PRC government undertook to classify all the ethnic minorities (minzu) and confer minority
nationality status on them, the Stau were included under the Tibetan minzu. This decision
is understandable, since the Stau are very much part of the Tibetan world. In regard to
religion, they are Tibetan Buddhist, which involves them in that sphere, where religion and
culture are so closely tied, in myriad ways. Most Stau women wear the traditional Tibetan
chupa, a long robe-like dress that wraps around the body and is cinched by a sash; if they

7

are married, they will also wear a colourful striped apron. Men used to wear the chupa, but
now most wear Western clothing. However, when the weather gets cold, men will put on
their warm, yak fur lined chupas. The Stau eat tsampa, a staple food all over the Tibetan
Buddhist world that is made by mixing barley flour with yak butter tea and kneading it into
dumplings. The religious influence of Tibetan Buddhism is seen in the way the Stau area
is marked by stupas (religious monuments that house sacred relics) and prayer flags, which
emblazon hills and rooftops. Many Stau have prayer wheels in their homes, and sometimes
a large prayer mill can be seen along the road, kept perpetually turning by the power of a
waterwheel. Stau often wear prayer beads with which they count the mantras they recite.
The Stau not only practice culture in a way that makes it easy to call them Tibetan, they
themselves identify as Tibetan, and do not want to be thought of as non-Tibetan (Jesse
Gates, p.c.).
Despite a close identification with the Tibetan world, the Stau also have a strong
identity as a people; Stau people take pride in being from Daofu. They even have a folk
story that their language is an ancient version of Tibetan (Jesse Gates, p.c.). Sometimes
this multiplicity of identifications can frustrate scholars who would prefer neat and discrete
anthropological categories for people groups. Pelkey (2011) notices a similar phenomenon
of multiple identities among the Phula of Yunnan Province and Vietnam. One of the Phula
groups he researched has six “embedded ethnic identities” (2011:7). At first Pelkey tended
to see this as an unjust obscuration of diversity, but over time he learned that “this condensed
status actually represents the continuation of an ancient tradition of ethnic agglomeration
in the region” (2011:7). Rather than expecting people to identify solely with one ethnic
group, Gates (2014:3) suggests that a more helpful model to understand cultural identity in
the Ethnic Corridor is that of the Russian matryoshka doll. As with the matryoshka dolls
that nest one inside the next, multiple ethnic identities exist together as one identity nests
within another identity. For the Stau, the local village would form the first level of a person’s

8

identity, over which is layered Stau, then Western Rgyalrongic, Rgyalrongic (the latter two
may be ethnolinguistic layers of which members are less consciously aware), and finally
Tibetan.
One aspect of Stau culture that is distinctive from other Tibetans is their architecture.
They build large houses out of wood logged from local forests. Massive timbers are used as
columns in the frame of the house. In the style typical of Tibetan architecture, Stau houses
have a pleasing geometrical frieze under the eaves. This frieze is formed by the ends of the
secondary beams (which support the floors and roof) that are painted white to contrast with
the reddish brown of the other timbers, which produces a decorative, chequer-like pattern
at the terminus of each floor.

9

Figure 2. Stau house
Stau houses differ from those of the Rgyalrong, a related ethnolinguistic group who
live north and east of the Stau, in that Stau houses are larger and use more wood in their
construction. Rgyalrong houses are often constructed completely from stone; Stau houses,
like the one pictured above, usually have a stone base while the rest of the structure is made
with wood.

10

The Stau rely primarily for their livelihood on agriculture, both farming crops and
raising livestock. The majority of the Stau are settled agriculturalists and live in the valleys.
Some are pastoralists; these typically live at higher elevations, up the mountainside. Unlike
the Amdo Tibetans, Stau herders are not nomadic. Those who are farmers grow barley,
wheat, buckwheat, and corn (the latter mostly as pig feed). Lately, they have also started
growing canola as a cash crop. Other crops they grow include bok choy, tomatoes, peppers,
hua jiao (Sichuan numbing pepper), apples, and walnuts. The Stau raise cows, mdzo (a
yak-cow hybrid), pigs, sheep, and horses. In addition to agriculture, Daofu County also
hosts lumber and gold mining industries, though jobs in these industries are primarily filled
by Han Chinese workers (Marshall & Cooke 1997). Because farming is often insufficient
as a livelihood and alternative jobs are in short supply, many of the young people leave the
area and migrate to Chengdu to find work. Another strategy that many Stau use to make
ends meet is to supplement their income by hand-collecting and selling cordyceps sinensis
or caterpillar fungus in the spring. Cordyceps sinensis is a fungus found on the Tibetan
Plateau that parasitizes root-boring caterpillars (Winkler 2008:294). The fungus is used in
both traditional Chinese and Tibetan medicine. Because of its scarcity, it fetches a very high
price and contributes substantial cash income (Winkler 2008).
In terms of language vitality, Stau is not in immediate danger. Of the approximately
23,000 Stau (Lewis et al. 2013), the great majority use their mother tongue. Parents still
pass on the language to their children, and the language is used in many domains. The
Ethnologue gives Stau a 6a (Vigorous) status on EGIDS (Lewis et al. 2013),1 meaning that
Stau is sustainable. However, Mandarin Chinese has jurisdiction in the domains of goverment and education, where its use is mandatory, and has de facto control in some areas of
business. No studies are available on Stau language vitality, but the closely related Rgyalrong was included in such a study (Huang X. 2000, cited in Zhou 2003:30). Although the
1

Expanded Graded Intergenerational Disruption Scale

11

findings on Rgyalrong cannot be applied to Stau indiscriminately, they do provide an indication of what the situation might be like in Daofu (though the latter will probably be worse
since Rgyalrong has a larger population). Huang X. (2000)’s study measures the ethnolinguistic vitality of 60 minority languages in ten domains. Rgyalrong is used minimally in
the domains of adminstration, legislature, judiciary, religion, and economics; not at all in
education, publication, media, literature and the arts, or IT. Out of the highest possible score
of 40 (the level assigned to Mandarin Chinese), the following were Rgyalrong’s scores:
Domain

Score

Administration 3.33
Legislature
5.00
Judiciary
3.33
Religion
2.00
Economics
1.25
As Chinese is used in more domains, knowledge of the language has become important for economic success. Although minorities like the Stau may bemoan Chinese’s encroachment, pragmatism stops them from preventing their children from learning it. More
and more people, especially the younger generations who are receiving more schooling, are
bilingual in Chinese. This exerts pressure on Stau, and could cause its EGIDS level to drop
to 6b (Threatened) in the future. It is relevant to note that Stau has no writing system, which
might provide ballast against language attrition.

2.2 Stau’s linguistic context
The study of Stau is young, as will become clear in §2.3 on previous work. For this
reason, very little can be said with certainty or without qualification about it. It is generally
accepted, at least as a working hypothesis (by, e.g., J. Sun 2000a, whom I will follow in
this regard), that there is a Rgyalrongic subgroup of which Stau is a member. However, the
existence of the subgroup has been contested on not unreasonable grounds. Consequently

12

Stau’s immediate sibling relations are not known with certainty. Neither have scholars come
to a consensus on the larger issue of the classification of Stau within Tibeto-Burman (TB),
or the smaller issue of the dialect ramifications of Stau. I will address the question of the
immediate sister and daughter affiliations of Stau in §2.2.2 and the classification of Stau
within Tibeto-Burman in §2.2.3. But first, in §2.2.1, I disgress briefly. The intersection
of languages―English, Chinese, Tibetan, Stau―has caused a profusion of names for Stau
within the literature. To avoid confusion, a discussion of nomenclature is warranted.
(Please note that the discussion of classification issues throughout this chapter is
quite detailed; readers who are not interested in classification should skip ahead.)
2.2.1 Nomenclature
Nomenclature is often an issue for Tibeto-Burman languages. As Matisoff (1996:ix)
notes, “[r]ampant polynymy prevails in the TB family. Rare is the language that is not
known by more than one name.” Stau is no different. In the first place, Stau has many allograms, which can be confusing to the novice. The allogram rTau uses the Wylie transliteration of Tibetan script. Because the pronunciation of the r in Written Tibetan (WT) is misleading to newcomers, other scholars changed this preinitial to the more phonetic s. Other
allograms include Daofu (Chinese pinyin), Dawu (Tibetan pinyin), and Taofu (Wade-Giles
romanization).
Capitalization is also an issue among Rgyalrongic languages because some, like
Stau, have Tibetan names. In Tibetan, it is conventional to capitalize the root letter; if the
first letter it is a prefix, it is written in lowercase. Thus, one gets sTau, rGyalrong, etc. In
order to avoid confusion, I will follow the Western convention of capitalizing the first letter.
Another issue is that Stau belongs to a group of lects, generally believed to include
Geshiza (of Danba County) and Shangzhai (of the southern corner Rangthang County), that
have no general autonym. As J. Sun (2000a) explains, early Western explorers used a variety of names for the language that they briefly recorded. Hodgson (1874) used Hórpa.

13

Other paleonyms that were used include Pawang (Rosthorn 1897), Gešits’a (Laufer 1916),
Bawang Rong-Ke (Edgar 1933-4), and Taofu (Migot 1957).2 Following Ma Changshou
(1944), Sun Hongkai uses Ergong as the higher order name for the language (or language
cluster) spoken in “Danba, Daofu, Luhuo, and Xinlong Counties of the Ganzi Autonomous
Prefecture of the Tibetan Nationality, and at some localities around the Guanyinqiao area of
Jinchuan County of the Aba Autonomous Prefecture of the Tibetan Nationality” (1990:123). However, this name is a Chinese exonym and has since been retired as offensive. Huang
(1991) uses “Daofu” for both the larger group and the particular dialect of the Stau area. To
add another complexity, since the 1990s, the definition of the language cluster has shrunk.
Nyagrong-Minyag, the lect found in Xinlong County, and Khroskyabs (also called Guanyinqiao) are no longer considered part of the cluster. J. Sun (2000b, 2000b) has used “HorpaShangzhai” for the reduced cluster, but now uses the shortened “Horpa” (2005:3) to be consistent with other linguists, such as Jacques & Michaud (2011). Gates (2012:4) suggests
“Western Rgyalrongic” as an appropriate term.
In 2013, Guillaume Jacques initiated a discussion on the Rgyalrongic listserv among
linguists who work on Rgyalrongic languages and any available Rgyalrongic speakers in
order to come to a consensus on an acceptable name for this language group, as well as a
name for the lect until now called Lavrung.3 With the input of several native speakers, a
consensus was reached to call Lavrung “Khroskyabs” (pronounced [tʂʰoscæ] or [tʂʰoscæv]).
However, I do not think one was reached on a name for the language group Stau belongs
to, so I have decided to adopt Gates’ term Western Rgyalrongic.
2

All of these are cited in J. Sun 2000a.
This discussion is not available online.
However, one can subscribe to the listserv at
http://www.lsoft.com/scripts/wl.exe?SL1=RGYALRONG&H=LISTSERV.LINGUISTLIST.ORG.
3

14

2.2.2 Stau’s immediate family
Western Rgyalrongic is usually considered to be sister to a number of lects situated nearby in western Sichuan: Rgyalrong with its several varieties (e.g. Sidaba, Chabao,
and Situ); and Khroskyabs. Together these lects are called Rgyalrongic. The following
map shows where each Rgyalrongic (= Jiarongic) lect is spoken. A note for map-readers:
Lavrung = Khroskyabs, Daofu = Stau, Xinlong-Muya = Nyagrong-Minyak (no longer considered a Rgyalrongic language by many scholars).

Figure 3. Map of Rgyalrongic languages (from Gatehouse 2014)
Within the literature on Rgyalrongic, various subclassifications of these lects have
been proposed. I have chosen four of the major proposals to compare and contrast: those
of J. Sun (2000a), a leading Rgyalrongic linguist, Jacques & Michaud (2011), Qu (1990)

15

and Lin (1993), and Huang (1991).4 The classification I follow mostly closely and use as
point of comparison is that of J. Sun. The proposals can be seen in Figure 4 below.5 The
chart sets the four subgroupings side-by-side in order to show how they correspond to and
differ from each other, including in terms of nomenclature. Languages that are in cells along
the same row are essentially equivalent even if they are called by different names, though
undoubtedly there are differences in the way that each linguist defines them.

Figure 4. Comparison of four subgroupings of Rgyalrongic
Qu (1990) & Lin X. (1993) and Huang (1991) represent two classifications from the
early years of Rgyalrongic scholarship. Qu and Lin considered all the Rgyalrongic lects to
be dialects of one Rgyalrong language. They subdivided it into three dialects: Northern,
Eastern (both included in J. Sun’s “Rgyalrong proper”), and Western (Stau, Shangzhai, and
Khroskyabs). Huang is one of the scholars mentioned in the introduction to this section who
do not (or did not) assume a Rgyalrongic subgroup. Although Huang has since changed her
mind, in her 1991 classification, she classified Rgyalrong and Daofu (in an extended sense)
as direct daughter languages of the Qiangic branch, not more closely related to each other
than to all the other Qiangic languages (see §2.2.3 below for a description of the putative
Qiangic branch of TB).
4

The latter three as cited in J. Sun 2000a.
The reader may be confused to see an ISO 693-3 code referring to a language family―jya referring to the
Rgyalrong family―when the codes are supposed to refer to a single language. The reason for this deviation is
that Rgyalrong used to be considered a single language. Now this and the other ISO codes for Rgyalrongic are
out-of-date but new codes have not yet been approved. Gates (2012) begins to rectify this situation, providing
evidence that Rgyalrong lects are separate languages.
5

16

However, over the ensuing years, further research has changed the proposed classifications. J. Sun proposes a Rgyalrongic that forms its own diachronic subgroup distinct
from Qiangic, and that comprises at least three synchronic languages―Rgyalrong proper,
Horpa-Shangzhai, and Khroskyabs (Lavrung). He gives evidence for this hypothesis in two
articles: J. Sun (2000a, b). J. Sun (2000a) shows, contra Huang (1991), that Khroskyabs is a
Rgyalrongic language, giving evidence that it shares with Sidaba Rgyalrong three instances
of unusual inflexional verb morphology that are not borrowed from a non-Rgyalrongic
source. Further evidence of verb-stem alternations shared by Rgyalrong, Horpa-Shangzhai,
and Khroskyabs demonstrates that these three languages form a stammbaum (i.e. a group
of languages descending from one common ancestor; J. Sun 2000b). Both Huang (2003:60)
and Sun (2004:313) have accepted J. Sun’s hypothesis.6
However, the hypothesis undoubtedly will not last in its current form. Many lects
are insufficiently studied, leaving Rgyalrongic classification still subject to change. One
notable question that needs to be answered is that of where Khroskyabs fits into Rgyalrongic―is it more closely affiliated with Rgyalrong or with Western Rgyalrongic?
Another area under debate is that of the internal ramifications of Western Rgyalrongic. Western Rgyalrongic varieties are spread through Daofu, Danba, Luhuo, and Rangthang counties. However, as J. Sun (2007:212) remarks, “[d]ialect differentiation within
Horpa [i.e. Western Rgyalrongic] is still quite unclear.” J. Sun has shown that Stau (Daofu
Country) and Shangzhai (Rangthang County) are closely related. In addition to sharing core
lexical items, both use the unusual morphological phenomenon of marking tense-aspect
through asperation inversion (2000b:230). He proposes that they comprise two dialects
of one synchronic language, Horpa-Shangzhai (2000a:166). In response to this proposal,
Gates (2012:17) objects that Stau and Shangzhai share low intelligibility, have many phonological and morphosyntactic differences, and they do not share an ethnic identity at a lower
6

As cited in J. Sun 2005:3.

17

level of embedding. He suggests instead that the morphological similarities J. Sun identifies may be “reflexes from a common meso-level proto-language that broke from ProtoRgyalrongic” (2012:18). If this were so, Horpa-Shangzhai would be “a diachronic clade
within Rgyalrongic, consisting of several synchronic languages,” which Gates calls “Western Rgyalrongic” (2012:18).
In regard to the other varieties of Western Rgyalrongic, progress is being made to
classify them into synchronic dialect groups. Jesse Gates has been undertaking sociolinguistic survey in the area, using recorded text tests (RTTs) as well as informal interviews.
The task of defining languages and dialects is not cut and dried, as Gates (2012) acknowledges. It relies on more than simply identifying mutual intelligibility (2012:30). He says,
citing Pelkey (2011:37), that language boundaries are affected by other factors in addition
to linguistic ones, making the boundaries complex and gradient. For this reason, language
definition should involve knowledge from outside the linguist’s purview. Pelkey himself
writes: “Since any given lect constitutes a dynamic phenomenon with underpinnings that
are linked to contact, intelligibility, socio-history, geopolitics, culture, cognition and structure alike, the dialectologist should simply require that a given language or dialect definition
attempt to be equally complex in its application” (2011:37). Gates takes this approach to
identifying the languages and dialects of Western Rgyalrongic. Based on the evidence of
RTTs he has conducted throughout Daofu and and Danba Counties, Gates (2013) argues
that what scholars have been calling one language, Horpa = Daofu = Stau, is actually three
“dialect groups” that form a continuum: Stau, Geshiza, and an as-yet unnamed group provisionally called Bawang-Niega. Gates sees the term “dialect group,” which he adopts from
Tournadre (2008), as more truly representative of the actual linguistic situation than “language,” which gives the impression of definite, unambiguous boundaries between varieties.
Gates hypothesizes that the Stau dialect group extends along the Xianshui River
from Renda Township (reported to have speakers), through Kongse, Jiasekong, Mazi, Daofu

18

Town, Gexi, Wari, Yuexue, Geqia (= Geka), to Muru.7 One village just west of Bian’er,
Dangling, is also part of the Stau dialect group. Bian’er and Dandong, on the other hand,
form the transition zone between Stau and the Geshiza dialect groups; these two could
be classified with Stau or with Geshiza. Within the Geshiza dialect group fall Shazhong,
Geshiza, and two villages of Donggu, Dongma and Maoniu. The third dialect group that
differs distinctly from Geshiza and Stau is found in Niega (= Jiaju) and Bawang.
2.2.3 Stau within Tibeto-Burman
Early in Rgyalrongic studies when Rgyalrong proper was the only language under
study, linguists assumed Rgyalrong was closely related to Tibetan and belonged to the Bodic
branch of TB (e.g. Wolfenden 1936, Chang & Chang 1974). As more languages, such as
Stau, were discovered, linguists began to see past the surface lexical similarities between
Rgyalrongic languages and Tibetan. Today, two competing hypotheses about the classification of Rgyalrongic within Tibeto-Burman are debated. Some scholars place Rgyalrongic in
a Qiangic branch of TB; others make it a daughter of a branch called Rung. The Qiangic hypothesis enjoys more currency, and most literature assumes a version of this classification.
Rung as a branch of TB and as antecedent of Rgyalrongic is a newer and less well-known
classification.
The Qiangic hypothesis proposes that the languages of the peoples of the Ethnic
Corridor form one subgroup. As Chirkova (2012) points out, the Orientalist F.W. Thomas
(1948) was the first to suggest this hypothesis. However, it did not become widespread
until it was propounded by the highly respected Chinese linguist Sun Hongkai. Sun (1962,
as cited in Chirkova 2012:136) proposed Qiangic as a subgroup for the Qiang, Rgyalrong,
and Pumi languages. Then, on field trips taken between 1978 and 1982 in Sichuan, Yunnan, and southern Gansu, Sun identified seven previously undocumented languages―Muya
7

These places are townships (xiàn) unless otherwise noted.

19

(Minyak), Shixing, Ersu, Namuyi, Guiqiong, Ergong (= Stau), and Zhaba (1987:120). These
languages were also subsumed into the subgroup (Sun 1983, trans. 1990).
Though the Qiangic hypothesis has become the default classification for many TibetoBurmanists, it has not gone unchallenged. Chirkova (2012) calls into question the validity
of the subgroup, arguing that the evidence for it is not probative. Qiangic is based on the
evidence of shared vocabulary and shared typological features, such as a large number of
consonant clusters, large consonant and vowel inventories, and directional prefixes (the
latter considered particularly characteristic of Qiangic) (2012:138). However, Chirkova
raises the objection that typological features can be shared through areal diffusion, not only
genetic inheritance. Because the Tibetan borderland that the putative Qiangic languages inhabit is such a melting pot of multilingualism, this explanation is plausible. Corroborating
evidence is found in the fact that the shared features demonstrate a “conspicuous absence of
cognateness” (2012:137). In addition, the evidence offered by shared vocabulary becomes
less significant when one considers that a portion of the already low 20-25% lexical similarity consists of Sino-Tibetan cognates and cultural borrowings from Tibetan (2012:137,
141). Finally, Qiangic as a subgroup has not been substantiated by truly probative evidence, namely that of shared innovations, or rather I would add, following Nichols (1996)
and LaPolla (2013), paradigmatic or idiosyncratic shared innovations.
It is on this type of evidence that the Rung subgroup is based. Thurgood (1985) and
Ebert (1990) were first to propose the Rung hypothesis; it was expanded and improved upon
by LaPolla (2003, 2006), as Thurgood (2003:14) himself acknowledges. In LaPolla’s proposal, Rung comprises “Rgyalrong, T’rung (Dulong), Rawang, Kiranti, Kham, and Western
Himalayan” (2003:30), a set of languages not intuitively grouped together, as they are separated by a great geographical distance and the Himalayas. LaPolla maintains an association
between the “Qiangic” languages and Rgyalrong. He speculates that the “Rung languages
most likely split off from an even higher-level grouping with the Qiangic languages, then

20

Rgyalrong split off from the group as migrations moved south, then Western Himalayan
split off from Kiranti and Rawang, and then these two groups split” (2006:394). The tree
in Figure 5 illustrates LaPolla’s proposed subgrouping:

Figure 5. Qiangic-Rung subgrouping
LaPolla bases Rung, in contrast to Qiangic, on “paradigm-like sets” of shared innovations (2013:464). He argues that the evidence for Rung is methodologically superior to
that for Qiangic, making Rung a better hypothesis, despite its lack of immediate intuitiveness. LaPolla is very concerned about sound methodology within Tibeto-Burman historical
linguistics. In a 2013 article he warns against basing subgroupings on geography or other
inadmissible evidence (such as that used to back up Qiangic). LaPolla recommends following the comparative methodology outlined in Nichols (1996).
To summarize briefly, Nichols (1996:48) calls for claims of genetic relatedness to be
based on what she terms “individual-identifying evidence.” Individual-identifying evidence
contrasts with “type-identifying evidence,” namely shared features, such as the presence of
a grammatical category like ergativity or possessing a large consonant inventory. Such
features can be easily spread among unrelated languages in an area; they can identify a
“type” of language but not an individual proto-language. The kind of evidence needed to
establish a stammbaum is not to “individual items but whole systems or subsystems with a
good deal of internal paradigmaticity, ideally multiple paradigmaticity.” In addition, these

21

paradigmatic systems must show cognancy (1996:48).8 Only this type of evidence can rule
out the possibility that features have been shared by chance or by diffusion, and qualifies
as individual-identifying.
It is this kind of evidence that gives weight and credence to the Rung hypothesis. In
the same 2013 article, LaPolla goes through the evidence for Rung, as a demonstration of
how Nichols’ method works. He shows that the languages of Rgyalrong, Dulong-Rawang,
Kiranti, and Western Himalayan not only share the feature of a person-marking system, but
also―crucially―that these systems are cognate (2013:468, 471):
(1)

1sg

1pl

2pl

dual

Proto-Rgyalrong
*-ŋ
*-i
*-ñ *-tsh
Proto-Dulong-Rawang *-ŋ
*-i
*-n *-si
Proto-Kiranti
*-ŋ
*-i
*-ni *-ci
Proto-W. Himalayan
*-g/ŋ *-ni *-ni *-si
One may have noticed that LaPolla refers to “Rgyalrong,” not “Rgyalrongic.” He
is one of the scholars mentioned in the introduction to this chapter who questions whether
Stau is a sister language of Rgyalrong. LaPolla (2003) splits up the putative Rgyalrongic
subgroup, leaving Ergong and Daofu9 in the Qiangic subgroup and seconding Rgyalrong
to Rung. He has said subsequently that, based on J. Sun & Tian (2013), it appears that Stau
has most of the Rung person-marking system. However, some of the set is not accounted
for in the data given by Sun & Tian. This could be because the data was not relevant to their
thesis, rather than because it does not exist, in which case Stau could yet be incorporated
into Rung (LaPolla, p.c.).
8

LaPolla (2013) points out that TB’s monosyllabicity and dearth of morphology make applying Nichols’
criterion of paradigmaticity to this context difficult. However, the standard can be met within TB if the
criterion is tweaked to include “paradigm-like sets of words and grammatical forms” (2013:465).
9
It is unclear what distinction LaPolla is making with the two language names, which most Rgyalrongic
linguists use interchangeably.

22

To sum up, no definite conclusion has been reached concerning the classification
of Stau within Tibeto-Burman, except that it belongs to the conglomerate of languages that
originate in the Ethnic Corridor.

2.3 Previous work
Little previous work has been done on Stau or on Western Rgyalrongic, to which
it provisionally belongs. What work has been done has for the most part been written in
Chinese, which I cannot read fluently, and is often unobtainable for me in Canada. Some of
the earliest documentation of Stau was collected by the early Western explorers mentioned
in §2.2.1. Some of them took down short wordlists in the Daofu area (e.g. Hodgson 1853),
but they were not professional linguists and made little attempt to systematically study the
language. The first linguistic work done on Stau was an article by Wang (1970) about the
consonant clusters of Tibetan loanwords in Stau. Later, Sun (1983) wrote a paper containing brief and broad sketches of seven recently discovered languages of the Sichuan Ethnic
Corridor, one of these being “Ergong” as spoken in the Dasang District of Danba County
(1990:12-4). An English version of this paper translated by J. Sun was published in 1990.
The first substantial descriptive work on a Western Rgyalrongic lect appeared in
1990―an article on phonology and verb conjugation in Gexi Daofu by Huang. Huang
(1991) also contributed a chapter on Daofu to the collection Zangmianyu Shiwuzhong [Fifteen Tibeto-Burman languages]. Both Huang (1990) and (1991) are written in Chinese.10
Another linguist who has made significant contributions to the study of Western
Rgyalrongic is Duo’erji. He studied his mother tongue, a variety of Western Rgyalrongic
spoken in Geshiza (Danba County). In 1995, Duo’erji published a short phonology on
Geshiza, and in 1998 a more comprehensive description of the language that included grammar as well as phonology. Like Huang’s work, these sources are written in Chinese.11
10
11

I have a copy and English translation of the latter, but not the former.
I have a copy of the later work, though not the earlier one.

23

Sources that deal with varieties of Western Rgyalrongic and are available in English
only discuss the languages in relation to particular classificatory or grammatical issues; none
are thorough descriptions. J. Sun (2000a), (2000b), and (2007) are examples of this type of
work. All three articles compare data from representative Rgyalrongic lects, focusing on
particular grammatical features, in order to prove their status as a related subgroup and to
provide some indication of their internal subgrouping.12
In the last year or two, as more academics have been joining and uploading their
research to the academic social networking site, Academia.edu, resources―especially unpublished working papers―have become available faster. The site provided a platform for
Jacques et al. (2013) to upload an unpublished article on verb inflection in Stau.
Four works on the Wobzi dialect of Lavrung (= Khroskyabs), a Rgyalrongic language sometimes proposed to be more closely related to Western Rgyalrongic than to Rgyalrong proper, have also been made available on Academia.edu by the author Lai. Three are
articles on person agreement, morphology, and consonant clusters, (2013b), (2013c), and
(2013d) respectively, and the fourth is Lai’s M.A. thesis (2013a).
Rgyalrong proper is much more thoroughly described than any other Rgyalrongic
macrolanguage, and in comparison to Western Rgyalrongic has a wealth of resources. A
number of lects within Rgyalrong proper have been studied.13 Mansier (1983) described the
phonology and lexicon of Tsenla (= bTsan-lha, = Xiaojin) Rgyalrong. Nagano (2003) has
written on Cogtse, a variety of Situ Rgyalrong. Prins (2011) for her doctoral dissertation
wrote a grammar on another Situ dialect, Jiaomuzu. Caodeng, a Sidaba variety, is described
in a number of articles by J. Sun, notably (2003). Jacques (2004) wrote his doctoral dissertation on the phonology and morphology of Japhug (a lect of the Chabao variety). Chinese
linguists also have published more on Rgyalrong proper than any other Rgyalrongic language; representative works include Lin X. (1993) and Qu (1984).
12

See §2.2 for a discussion of J. Sun’s conclusions concerning the genetic affiliation and subgrouping of
Rgyalrongic, and of Stau’s genetic classification.
13
See §2.2.2 for more on how these lects are related to Stau.

24

A couple of sources that deal with the Rgyalrongic subgroup generally were published recently, Gates (2012, 2013) and Nagano & Prins (2013). Gates’ (2012) thesis deals
with Rgyalrongic dialectology and the question of how many synchronic Rgyalrongic languages or dialect groups there are.14 He demonstrates that Situ Rgyalrong15 is not just a
dialect of Rgyalrong proper but is a language, according to the criteria defined by the ISO
(International Organization for Standardization), and hypothesizes that Rgyalrong actually
comprises five synchronic dialect groups. Gates (2013), a presentation at the 3rd Workshop
on Sino-Tibetan Languages of Sichuan, continues the work of Rgyalrongic dialectology by
arguing from recorded text tests (RTTs) that the dialects of Western Rgyalrongic form a dialect continuum with three dialect groups (see §2.2.2 for more specifics). Nagano & Prins
(2013) is an online database presenting wordlist and sentence data from 81 lects―primarily
Rgyalrongic ones but also a few non-Rgyalrongic, such as Nyagrong-Minyag and Queyu.
The same 425- or 1200-item wordlist and 200 sentences were used for elicitation at each
location. The database allows several ways of interacting with the data. An interactive
map allows the user to click on a data point, and on the same page search for and listen
to a particular word in the chosen lect. From the map one can also access a PDF of the
wordlist collected at each data point. From another page, one can chose an English gloss
and browse all the forms collected for that item of the wordlist. On a third page one can
browse by language the 200 sentences.

2.4 Fieldwork background
I conducted the fieldwork for this thesis in Sichuan Province, China, between late
June and early August 2012. The bulk of this time I spent in Chengdu, Sichuan’s capital
city. While in Chengdu I worked with a female speaker of Stau in her early 30s named
Tsomo.16 She grew up in Mazi Township of Daofu County, Ganzi Prefecture, and moved
14

Gates uses the terms “languages” and “synchronic languages” throughout his thesis, but now prefers the
term “dialect group(s)” to more accurately represent of the linguistic situation (Gates, p.c.).
15
Refer back to Figures 3 and 4 in §2.2.2 to recall where Situ is situated within Western Rgyalrongic.
16
Tsomo has given me permission to cite her by name.

25

with some of her family to Chengdu some years ago. Tsomo speaks Stau in the home with
her family. She speaks some Sichuanese Mandarin, though she is not fluent. Interestingly,
Tsomo’s family is part of the Stau population that intermingles and intermarries with the
nomadic Amdo Tibetans. For this reason, Tsomo speaks Daofu-style Amdo Tibetan and has
closer ties to the Tibetan-speaking world than other Stau people might. Because she runs a
business in Chengdu that caters to Tibetan customers, she has also learned to communicate
with Tibetans of nearly all dialects.
For one week of the two months spent in Sichuan, I went on a trip to the Stau language area, accompanying a couple of other linguists who were doing sociolinguistic survey. For my part, the purpose of the trip was to meet Stau people, to see how they live in
their own milieu, and to experience the culture of which their language is a part. Secondarily, I hoped to gather wordlist data from other Stau speakers in order to compare it to
Tsomo’s idiolect as represented by the data I had collected with her. With data from other
speakers, more firm conclusions could be drawn about what is part of the language and what
is distinctive of my language consultant’s idiolect. In addition, data from multiple speakers
might provide some clues to the variation that exists within the Mazi Stau dialect.
I gathered parts of the wordlist I had collected with Tsomo with two speakers, a male
in his late 20s or early 30s and a female speaker in her 30s. However, I did not have enough
time to collect sufficient data to make any certain or definite claims about dialect-internal
variation. Only 195 words were gathered with the male language consultant, and 95 with
the female speaker. Because wordlists of such small size are insufficient to base claims on,
I have not incorporated them into the thesis.
In my main work with Tsomo, the work went through several stages. The first
stage was initial data collection, during which Tsomo and I worked in her home. In the
morning, my colleague Jesse Gates accompanied me and would translate between Mandarin
and English for me. In the afternoon I would return by myself and variously use my limited

26

Mandarin, gesture and actions (e.g. pointing to body parts), and images from Google to
elicit words on my list. If Tsomo and I could not negotiate to mutual understanding over a
particular word, I skipped it, to return to it with Jesse the next morning when we could have
more nuanced communication.
The wordlist I used was a pared-down, 425-word version of one created by Yasuhiko
Nagano for the Nagano Rgyalrongic Language Survey Project, given to me by Gates. It
proved very useful, since it includes not only an English gloss column, but also columns
for the Chinese gloss written with characters, the Chinese gloss written in pı̄nyı̄n, and the
Written Tibetan gloss. Because I do not read Tibetan, the latter was not of much use, but
the Chinese glosses were very helpful to me with my limited Chinese vocabulary. Using
the pı̄nyı̄n, I could ask Tsomo for the Stau equivalent of a Chinese word I had never before encountered. If I had used a wordlist with English glosses I could have looked up the
Chinese equivalents, but I would have no assurance that the Chinese word had the semantic connotation I intended, whereas the Chinese glosses in the modified Nagano wordlist
had been chosen by linguists with a high degree of fluency in Chinese. Another perk of the
wordlist was that the version Gates gave me had been filled out by his colleague Tshe-dbang
sGron-ma with a dialect of Stau from Kongse (= Khang gsar). If attempts at communication in Chinese or through gesture failed, I could try pronounce the Kongse word, which
sometimes succeeded in sparking understanding.
While working from Tsomo’s home, I recorded our sessions in full on a Roland
R-09 Edirol portable recorder, while simultaneously marking down what I heard in a data
notebook. However, these recordings were not of high enough quality to use for acoustic
analysis. So after we had elicited as many of the wordlist items as possible, we moved to
the second stage of the work: making high-quality recordings of the wordlist. I borrowed
a colleague’s quiet, muffled apartment and some of their equipment. I used a Shure SM58
cardioid microphone on a stand, which plugged into a CEntrance MicPort Pro USB microphone pre-amplifier, which recorded directly into Adobe Audition 1.5. As we went through

27

the wordlist I would prompt Tsomo with the Stau word as best as I could pronounce it. She
would repeat the word into the recording microphone three times, and then put it in the context of a sentence, which she also said three times. At first, Tsomo would construct these
frame sentences on the spot and they would vary. Over time, a couple sentences became
standard:
(2)

a. tʰɛ
__ ŋə-rə
dem __ cop-const
‘This is __’
b. tʰi
__ kaji ŋə-rə
dem __ pretty cop-const
‘This __ is pretty’

2.5 Expectations
Looking at Stau data from other researchers and the phonologies of related languages, some expectations of what Mazi Stau might look like can be formed. The first
expectation is that Stau will have a large consonant inventory. The second is that the language will have a large number of unusual consonant clusters. These are both typological
features of Qiangic (Chirkova 2012:137), and features found in many Rgyalrongic varieties.
Consonants I expect to find include a series of affricates at three places of articulation: dental or alveolar, alveolo-palatal or palatal, and retroflex; each having three voicing distinctions: tenuis (voiceless unaspirated), voiceless aspirated, and voiced. Many
available phonologies of Rgyalrongic languages describe this series of affricates, including the Zhuokeji variety of Situ (Lin Y.J. 2003), Caodeng Sidaba (J. Sun 1994), Japhug
Chabao (Jacques 2004), Wobzi Khroskyabs (Lai 2013d), Puxi Shangzhai (J. Sun 2000b),
and Gexi Stau (Huang 1991). All these languages, as well as Jiaomuzu Situ (Prins 2011)
and Cogtse Situ (Nagano 2003), also have three voicing distinctions in their stops, of which
they have at least four places of articulation: labial, dental or alveolar, palatal, and velar.

28

Six of these―Caodeng, Japhug, Khroskyabs, Shangzhai, Gexi, Stau (Wang 1970)―also
use voiceless unaspirated and voiceless aspirated uvular stops, and voiced and voiceless
uvular fricatives. Other less common consonants I expect based on their frequent use in
the inventories of the above mentioned nine lects include the lateral fricative (voiceless and
possibly voiced versions) and likely both voiced and voiceless velar fricatives.
Based on the vowel inventories of related lects, I expect Mazi Stau to have over a
vowel inventory of slightly above-average size with seven or eight vowel qualities (Maddieson 2013d). The various Rgyalrong phonologies describe between six (Jiaomuzu) and
nine (Caodeng) vowels in their inventories. The available descriptions of Western Rgyalrongic lects have seven and nine vowels, as shown in the following vowel charts.
(3)

Rtau-Kongse vowel inventory (Jacques et al. 2013)17
i
e

u
o

ə
ɞ
a

(4)

Gexi Stau vowel inventory (Huang 1991)
i
e

ə

ə˞

u
o

ɛ
a
17

ɑ

The vowels in this chart were gleaned from the language data in Jacques et al. 2013. The article did not
have a vowel chart.

29

(5)

sTau vowel inventory (unknown variety) (Wang 1970)
i
e

ɯ
ə
ɛ

u
o
ɔ

a
Although the vowel charts in (3)-(5) all represent varieties of Stau, there are significant differences among their vowel inventories, which make predictions for Mazi less
certain. The lack of consistency may be due to differing transcription judgments on the part
of the linguists. It is worth noting that none of the vowel inventories here is based on an
acoustic plot like that in §4.2.
Another factor that might cause discrepancy is the distance between Gexi and Kongse
(WT = Khang gsar) villages. Unfortunately, where exactly Wang’s language consultants
were from is unknown, which makes his data less helpful. His language consultants, whom
Wang worked with in Kathmandu, Nepal, came from different villages at a significant distance from each other: the first consultant’s “original home was a distance of three hours by
walking east of Stau, while [the second consultant]’s was two days on horseback south of
Stau” (1970:633). Considering the variety and number of languages that are being discovered within short distances in the Ethnic Corridor, treating significantly distant lects as the
same language without prior investigation would not be considered good practice by contemporary linguists, and calls the reliability of the data into question. However, I include
Wang’s analysis here for thoroughness and for potential comparison. The analyses being
what they are, the charts in (3)-(5) show some consistency. Based on the vowel inventories,
I expect also to find /i/, /e/, /u/, /o/, and /a/ in Mazi.
A common feature of languages of the Ethnic Corridor that Mazi is likely to share
is a large number of consonant clusters. Many of these are likely to be unusual. Matisoff
writes that the Qiangic languages (as which he categorizes Stau) are “characterized by initial

30

consonant clusters comparable in complexity to those of Written Tibetan” (1991:482). Various Rgyalrongic dialects have been described with this phenomenon. Lai (2013d) reports
this phenomenon in Wobzi Khroskyabs, Prins (2011) in Jiaomuzu Rgyalrong, and Duanmu
(2008:224-36) notes the same for several other Rgyalrong lects.
In terms of consonant codas, I do not expect Mazi to have many. In the first place,
a lack of them is a characteristic of the putative Qiangic languages (Chirkova 2012:137).
Second, J. Sun (2000b:213) contrasts the way Gexi and Geshiza Stau have lost codas with
the coda-conservative Puxi Shangzhai. However, the lack of final consonants of which J.
Sun speaks may only be in comparison to Puxi. Matisoff (2004:328) notes that in comparison with other Qiangic languages Rgyalrong and Ergong (= Stau) do better at preserving
codas.
A feature Mazi Stau probably does not have is tone. J. Sun remarks that “most
sources on Horpa agree on its lack of lexical tone or contrastive accent” (2000b:222). He
does, however, note in a footnote that Liu (1989) “reports as many as four tones in one
variety of Horpa distributed in Xianshui Township of Daofu County” (2000b:222). Many
related Rgyalrongic varieties are also described as having tonal systems, usually restricted
tone or “pitch accent” systems; Caodeng, Zhuokeji, and Showu (=Zbu) Rgyalrong, Wobzi
and Thugschen Khroskyabs, and Puxi Shangzhai being among them.18 These languages
use tone more frequently to mark grammatical distinctions than lexical ones. For example,
Caodeng, Zhuokeji, Showu, and Puxi all use tone on the verb stem to mark tense-aspect.
Other Rgyalrongic lects do not use tone: Japhug of Chabao Rgyalrong (Jacques 2004:74),
and Njorogs Khroskyabs (=Yelong; Lai 2013a:31).

18

Described or mentioned in, respectively, J. Sun 2008, Lin Y.J. 2003, J. Sun 2004, Lai 2013a (both Wobzi
and Thugschen), and J. Sun 2000b.

Chapter 3
Segmental phonology
In this chapter, I briefly introduce the phonemes of Stau in §3.1. Section 3.2 will
give further detail on the consonants, specifically any allophony and evidence of the contrasts posited in §3.1. Similarly, in §3.3, I will elaborate on the vowel inventory. For both
consonants and vowels, each allophone will be described at the beginning of the section
dealing with its parent phoneme or phonemic category. After discussion of allophony, for
each phoneme I give evidence of its phonemic status in the form of a list of example words
in which the phoneme occurs. A representative variety of contexts will be provided. Presenting lists of examples was chosen over minimal pairs in order to provide more data for
fellow researchers to peruse, so that they could compare Stau to their own language of study
or find evidence for a hypothesis on which they are working. If a reader would prefer to see
minimal pairs as evidence of contrast, I list minimal and near-minimal pairs in Appendix
A. Finally, §3.4 will compare my findings on the phonemic inventory with those of Huang
(1991).1

3.1 Inventory of phonemes
Stau has 42 consonant phonemes, as the table below illustrates. The consonants in
brackets are phones that occur in Stau but are not full phonemes. The reader might expect
based on the tendency toward phonological symmetry that these sounds would be phonemes
1

A couple notes on the data used as evidence: The verbs given as evidence are as much as possible in
third person form. I heartily thank Katylin Wonnell for going through my examples and correcting the verbs
to be consistently third person. It is also relevant to note that, though there are vowel alternations at work
in the verbal morphophonemics of Stau, I still use vowels from verbs as evidence to demonstrate segmental
phonology.

31

32

in Stau. Thus I include the consonants in the inventory to assure the reader that these gaps
in the symmetry of the consonant inventory are not the result of an oversight.

vl unaspirated stop
vl aspirated stop
vd stop
vl unaspirated affricate
vl aspirated affricate
vd affricate
vl fricative
vd fricative
nasal
vl lateral fricative
vd lateral fricative
approximant

Stau consonants
labial alveolar retroflex palatal
p
t
c
pʰ
tʰ
cʰ
b
d
ɟ
ts
tʂ
tɕ
tsʰ
tʂʰ
tɕʰ
dz
dʐ
dʑ
(f)
s
(ʂ)
ɕ
v
z
r
ʑ
m
n
ɲ
ɬ
ɮ
w
l
j

velar
k
kʰ
g

uvular
q
qʰ
(ɢ)

x
ɣ
ŋ

χ
ʁ
(ɴ)

A 42-consonant inventory is typologically remarkable in cross-linguistic comparison. According to the World Atlas of Language Structures, a consonant inventory with 34
or more consonants is categorized as large, in comparison to the average of 19-25 phonemes
(Maddieson 2013a). Although the size of the inventory is unusual cross-linguistically, areally it is not; a large consonant inventory is a feature of languages of the Sichuan Ethnic
Corridor (as noted in §2.2.3, the possession of a large consonant inventory has been cited
as a characteristic of the putative Qiangic subgroup, the extent of which largely coincides
with that of the Ethnic Corridor).
The consonant inventory could even be expanded to 58 consonants, depending on
one’s analysis. There are sixteen stops and affricates preceded by homorganic nasals that
could be analyzed either as sequences of two consonants or as one segment―i.e., as consonant clusters or as an additional “prenasalized” manner of articulation. I have decided to
analyze these as consonant clusters. To see a discussion of my reasons for this decision, see
§6.1.2.

33

Stau’s vowel inventory is a more typical size, consisting of eight vowels:
Stau vowels
front back
high

low

i
e
ɛ
æ

u
o
ə
ɑ

Though the vowel inventory seems tame next to the overflowing consonant inventory, it is considered on the large side in comparison to the average five- or six-vowel inventory that 51% of the world languages have (Maddieson 2013d). Again, this size of vowel
inventory is typical of Corridor languages. Fellow Rgyalrongic languages Japhug Rgyalrong and Puxi Shangzhai have eight and nine vowels, respectively (Jacques 2004, J. Sun
2000b). Qiang has vowels at eight positions, with additional phonemic contrast provided
by vowel length at seven of these positions (LaPolla & Huang 1996:25).

3.2 Consonants
In this section, evidence of contrast for the consonant phonemes proposed in §3.1
is given for each phoneme. Allophony and any particularities of articulation are discussed
at the beginning of each subsection. The subsections are ordered according to manner of
articulation and then by place of articulation.
3.2.1 Stops
Stau has five stop series at bilabial, alveolar, palatal, velar, and uvular places of
articulation. Each of these distinguish between three types of voicing: voiceless aspirated,
voiceless unaspirated, and voiced.
Readers may notice in the lists of examples of voiced stops that there are very few
or even no examples of word-initial voiced stops. This may cause some doubt on the score

34

of their phonemic status. Please refer to §4.1 for acoustic evidence that voiced stops are
indeed separate phonemes from voiceless unaspirated stops.
3.2.1.1 Bilabial stops /pʰ, p, b/
Stau makes a three-way distinction between the voiceless aspirated bilabial stop /pʰ/,
the voiceless unaspirated bilabial stop /p/, and the voiced bilabial stop /b/. The bilabial stops
have no allophonic variation.
The voiceless unaspirated bilabial stop is involved in neutralization. In rapid speech,
[p] sometimes appears as a coda in Stau, e.g. [ɣməlɛp] ‘fire’. However, in more careful
speech the bilabial consonant can also be pronounced as [v], as (6) illustrates:
(6)

[ɣməlɛv] ‘fire’
/p/ is not a contrastive phoneme in coda position; in this context, the contrast be-

tween /p/ and /v/ is neutralized. (Thus the phonological transcription of ‘fire’ is ɣməlɛv.)
See §§3.2.3.1 and 6.2 for further discussion of the neutralization.
The following examples show each of the bilabial stops in a variety of environments,
representative of their distribution.
(7)

/pʰ/
a. pʰi
[pʰi]
‘to run away’
b. pʰe
[pʰe]
‘to vomit’
c. pʰɛ
[pʰɛ]
‘to throw out’
d. pʰæ
[pʰæ]
‘half’
e. pʰəru [pʰəʐu] ‘basket carried on back’
f. pʰupʰa [pʰupʰa] ‘male pig’
g. pʰjɛr
[pʰjɛʂ]
‘to open, unfurl’
h. pʰjəsu [pʰjəsu] ‘outside’
i. pʰrupʰru [pʰʂupʰʂu] ‘white’
j. ræpʰi [ʐæpʰi] ‘mahjong’
k. nə-pʰæ [nəpʰæ] ‘split’ (dirdown -split; adj)
l. tə-pʰɛ [təpʰɛ]
‘to lose’ (dirneut -lose)

35

(8)

(9)

m.
n.
o.
p.

ləpʰu
mpʰrivæ
xpʰə
mpʰi

[ləpʰu]
[m̥pʰʂivæ]
[xpʰə]
[m̥pʰi]

‘tree’
‘prayer beads’
‘butt’
‘to card’

/p/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.

pi
[pi]
‘ball of tsampa’
pɛrspo [pɛrspo] ‘to walk’
pæbə
[pæbə]
‘insect’
pəpə
[pəpə]
‘shallow’
pubæ
[pubæ]
‘Tibetan’
pjɛno
[pjɛno]
‘meat’
plɛ
[plɛ]
‘thigh’
prilærə [pʐilærə] ‘to whinny’
xopi
[xopi]
‘table’
tə-pɛ
[təpɛ]
‘to take out’ (dirneut -take.out)
æpæ
[æpæ]
‘father’
tɕʰəpæ [tɕʰəpæ] ‘clothing’
rtɛpu
[ʂtɛpu]
‘stallion’
scərpə [scəʂpə] ‘soy sauce’
spo
[spo]
‘grassland’
xpoŋ
[xpõ]
‘shoulder’
ɲtɕʰaχpa [ɲ̥tɕʰaχpa] ‘to fold’

/b/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.

bəti
[bəti]
‘cheek’
bjɛrgə [bjɛrgə] ‘pheasant’
bæɴɢe [bæɴɢe] ‘spider, fly’
æ-bɛr [æbɛʂ] ‘one step’ (num1-step)
pubæ [pubæ] ‘Tibetan’
ɲɟaba [ɲɟaba] ‘mud’
rʁəbə [ʐʁəbə] ‘thin’
rŋəbo [ʐŋəbo] ‘roasted barley’
mobrɛ [mobʐɛ] ‘tears’
bjo
[bjo]
‘to fly’
mbjɛmə [mbjɛmə] ‘deaf person’
stoŋbæ [stoŋbæ] ‘empty’
mbo
[mbo]
‘box for grain’
zbəqe [zbəqe] ‘to urge’
rbu
[ʐbu]
‘bee’

36

3.2.1.2 Alveolar stops /tʰ, t, d/
Three alveolar stops are distinguished in Stau, the voiceless aspirated alveolar /tʰ/,
the voiceless unaspirated /t/, and the voiced aspirated /d/.
(10)

(11)

/tʰ/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.

tʰi-ɲi [tʰiɲi]
‘third person plural pronoun’ (3-pl)
tʰævkæ [tʰæfkæ] ‘stove’
tʰadʑi [tʰadʑi] ‘far’
tʰəvæ [tʰəvæ] ‘hammer’
tʰətʰə [tʰətʰə] ‘sweet’
tʰutu [tʰutu] ‘mixed together’
ntʰvæ [n̥tʰfæ] ‘to tread on it’
ɣə-vtʰi [ɣəftʰi] ‘to drink’ (dirin -drink)
metʰɛv [metʰɛv] ‘stove’
vtʰɛ
[ftʰɛ]
‘to take off clothes’
tsəntʰɛv [tsəntʰɛv] ‘scissors’
kɛ-mtʰu [kɛmtʰu] ‘high’ (aprfx-high)
ntʰævæ [n̥tʰævæ] ‘decorative apron’
tʰi
[tʰi]
‘to hang’
stʰjæ [stʰjæ] ‘to support, prop up’
rʰærtʰə [ʂtʰæʂtʰə] ‘right’
xtʰo
[xtʰo]
‘ground, plain’

/t/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.

tæmbə
tadʐu
tətkɛr
tə-vævtʰi
təvdəze
tutu
æti
bəti
vɛtɛm
kʰətæ
rata
və
mito
rtɛvrtɛv
ŋosti

[tæmbə]
[tadʐu]
[tətkɛʂ]
[təvæftʰi]
[təvdəze]
[tutu]
[æti]
[bəti]
[vɛtɛm]
[kʰətæ]
[ʐata]
[və]
[mito]
[ʂtɛfʂtɛv]
[ŋosti]

‘bottle’
‘silk’
‘sesame’
‘to smoke’ (dirneut -smoke)
‘boy, son’
‘basket carried on back’
‘older brother’
‘cheek’
‘peanuts’
‘dog’
‘mill’
‘to make’
‘flower’
‘fine’
‘front’

37

p. xɕɛt
[xɕɛt̚]
‘to whip’
q. patmamɛto [patmamɛto] ‘type of flower’
r. wut
[wut̚]
‘light’ (n.)
(12)

/d/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.

doləmæ [doləmæ] ‘eggplant’
doŋbə [doŋbə] ‘stem’
dordʑe [dordʑe] ‘religious object monks hold while chanting’
kɛdi
[kɛdi]
‘child’
kɛ-de
[kede]
‘small’ (aprfx-small)
kʰɛdɛʂ [kʰɛdɛʂ] ‘scarf’
adɛ
[adɛ]
‘this’
ædæ
[ædæ]
‘older sister’
vədæ
[vədæ]
‘wife’
tɕedə
[tɕedə]
‘time’
ɣdæmbæ [ɣdæmbæ] ‘because’
mdæ
[mdæ]
‘arrow’
ndjɛv
[ndjɛv]
‘to sleep’
rdə
[ʐdə]
‘lower leg’
ʁdu
[ʁdu]
‘wood pail’
xavdu [havdu] ‘now’
zdi
[zdi]
‘stone wall’

3.2.1.3 Palatal stops /cʰ, c, ɟ/
Stau distinguishes among three palatal stops: the voiceless aspirated /cʰ/, voiceless
unaspirated /c/, and voiced /ɟ/. It is good to be aware when comparing Stau with related
languages that many Chinese linguists transcribe these sounds with the symbols <cçh>,
<cç>, and <ɟj>, respectively.
The voiceless aspirated palatal stop /cʰ/ can be difficult to distinguish from the voiceless aspirated alveolo-palatal affricate /tɕʰ/. The palatal stop is often heavily aspirated, and
its aspiration can sound like frication of the fricative portion of /tɕʰ/. Likewise the difference between /ɟ/ and /dʑ/ is sometimes difficult to recognize. The production of a voiced
stop at the palatal place of articulation creates a somewhat sticky sound, not unlike a voiced
fricative, leading to confusion between /ɟ/ and /dʑ/.

38

Below are examples of each of the three palatal stops.
(13)

/cʰ/
a. cʰəpʰɛv [cʰəpʰɛv] ‘stick on millstone’
b. cʰu-rə
[cʰuʐə]
‘hot (of weather)’ (hot-const)
c. wocʰi
[wocʰi] ‘lower abdomen’
d. kɛ-cʰɛ
[kɛcʰɛ]
‘big’ (aprfx-big)
e. mə-cʰɛcʰɛ [məcʰɛcʰɛ] ‘busy’ (neg-idle)
f. necʰa
[necʰa] ‘good morning’
g. loŋcʰə
[lõcʰə]
‘to plow (in the eighth month)’
h. xæcʰo
[hæcʰo] ‘sneeze’
i. vcʰə
[fcʰə]
‘to weigh’
j. ɲcʰə
[ɲ̥cʰə]
‘to hit, beat; thresh’
k. ɲcʰæræ [ɲ̥cʰæræ] ‘to play, have fun’
l. kə-ɲcʰɛr [kəɲcʰɛr] ‘to hide’
m. rcʰɛ
[ʂcʰɛ]
‘to bite’
n. xcʰi
[xcʰi]
‘to puncture’

(14)

/c/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.

coŋ
[coŋ]
‘clay wall’
scici
[scici]
‘to look at, see’
pəcæ
[pəcæ]
‘stick of wood’
wocæ [wocæ] ‘navel’
ɲcʰəcə [ɲ̥cʰəcə] ‘to fight’
gaca
[gaca]
‘goodbye (evening)’
scʰɛco [scʰɛco] ‘to chase’
vcə
[fcə]
‘rat, mouse’
rcuqu [ʂcuqu] ‘between’
sciskɛr [sciskɛʂ] ‘birthday’
ɕərʁɛscɛ [ɕərʁɛscɛ] ‘toothbrush’
scævæ [scævæ] ‘paddle’ (n.)
tɕəsco [tɕəsco] ‘to paddle’

39

(15)

/ɟ/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.

ɟirə
[ɟiʐə]
‘in’
ɟɛuræ [ɟɛuræ] ‘facial hair’
ɟɛzo [ɟɛzo] ‘potato’
ʑæʁɟæ [ʑæʁɟæ] ‘lame person’
sɲəɣɟi [sɲəɣɟi] ‘afternoon’
mɟo-rə [mɟoʐə] ‘fast’ (fast-aprfx)
ɲɟɛm [ɲɟɛm] ‘wall’
ɲɟəræ [ɲɟəræ] ‘to run’
ɲuɲɟu [ɲuɲɟu] ‘left’
ŋarɟɛr [ŋarɟɛʂ] ‘roar’ (n.)
pərɟə [pərɟə] ‘grandchild’
rɟæ
[ʐɟæ]
‘Chinese’
rɟæmæ [rɟæmæ] ‘scale’
ʁɟi
[ʁɟi]
‘hole’
vɟə
[vɟə]
‘saliva’

3.2.1.4 Velar stops /kʰ, k, g/
There are three stop phonemes at the velar place of articulation: the voiceless aspirated velear stop /kʰ/, the voiceless unaspirated velar stop /k/, and the voiced velar stop /g/.
In my dataset, the voiced /g/ never appears word-initially. However, the contrast between
/g/ and /k/ is present word-medially. See §4.1 for more evidence of the contrast between
voiced and voiceless unaspirated stops.
(16)

/kʰ/
a. kʰɛgɛ [kʰɛgɛ] ‘after’
b. kʰæɕjæ [kʰæɕjæ] ‘lips’
c. kʰambo [kʰambo] ‘bag’
d. kʰətæ [kʰətæ] ‘dog’
e. kʰətsʰi [kʰətsʰi] ‘water channel’
f. kʰri
[kʰʂi]
‘chair’
g. tsʰukʰæ [tsʰukʰæ] ‘colour’
h. ækʰə [ækʰə] ‘paternal uncle’
i. zɛkʰoŋ [zɛkʰõ] ‘restaurant’
j. pukʰu [pukʰu] ‘mosquito’
k. mkʰə [m̥kʰə] ‘smoke’
l. ɣrərkʰu [ɣʐəʂkʰu] ‘cold’

40

m. ŋkʰvo
n. skʰro
(17)

(18)

[ŋ̥kʰfo]
[skʰʂo]

‘key’
‘ant’

/k/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.
r.
s.
t.
u.

kindækindæ [kindækindæ] ‘dripping’
kɛ-de
[kede]
‘small’ (aprfx-small)
kɛ-dʑi
[kɛdʑi]
‘long’ (aprfx-long)
kɛ-skvɛ
[kɛskvɛ]
‘sharp’ (aprfx-sharp)
kæɕæ
[kæɕæ]
‘morning’
kapəla
[kapəla]
‘forehead’
kə-xæ
[kəxæ]
‘to come out’ (dirout -come.out)
kəχo
[kəχo]
‘bark’
koŋ-kɛcʰɛ [koŋkɛcʰɛ]
‘expensive’ (price-big)
ku-rə
[kuʐə]
‘to understand’ (understand-const)
krə
[kʐə]
‘boat’
ɮɛki
[ɮɛki]
‘bracelet’
tsʰeke
[tsʰeke]
‘hot’
rekɛ
[ʐekɛ]
‘and’
tɕʰəkəv
[tɕʰəkəv]
‘watermelon’
vɛko
[vɛko]
‘pigsty’
ʁnɛkuku
[ʁnɛkuku]
‘dark’
rekwe
[ʐekwe]
‘foal’
tɕʰəvka
[tɕʰəfka]
‘tap’ (n.)
skɛ
[skɛ]
‘language’
tə-skrə-sə [təskrəsə]
‘late’ (dirneut -late-prf)

/g/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.

kʰɛgɛ
[kʰɛgɛ] ‘after’
ægəjo
[ægəjo] ‘to grow up’
tsəgə
[tsəgə] ‘clothing’
ʑigə
[ʑigə]
‘around’
mə-gu-rə [məguʐə] ‘not understand’ (neg-understand-const)
vzu
[vzu]
‘to take’
ŋgə
[ŋgə]
‘9’
kɛ-rgi
[kɛrgi] ‘hard’ (aprfx-hard)
rgəmɛ
[ʐgəmɛ] ‘stone (for building house)’
dzo-vge [dzovge] ‘to cross bridge’ (bridge-cross)
vgə
[vgə]
‘to clothe’
zgri
[zgʐi]
‘star’
mgrə
[mgʐə] ‘wall (that one dries barley against)’

41

3.2.1.5 Uvular stops /qʰ, q, (ɢ)/
Unlike the preceding stop series, the uvular series has two definite voicing distinctions, between the voiceless aspirated /qʰ/ and the voiceless unaspirated /q/. The voiced
uvular stop /ɢ/ is a doubtful phoneme. It only occurs following the uvular nasal ɴ in my
data. All of its occurrences are listed below in (21). Examples of the two demonstrable
phonemes are given in (19) and (20).
(19)

/qʰ/
a. qʰegəɟirə [qʰegəɟiʐə] ‘raining’
b. qʰəsji
[qʰəsji]
‘tomorrow’
c. qʰəzi
[qʰəzi]
‘bowl’
d. qʰæ
[qʰæ]
‘to laugh’
e. qʰæsɬon-rə [qʰæsɬõdʑə] ‘happy’ (happy-const)
f. qʰosto
[qʰosto]
‘back’
g. qʰræqʰræ [qʰʂæqʰʂæ] ‘coarse’
h. qʰre
[qʰʂe]
‘to pull down’
i. rqʰwa
[ʂqʰwa]
‘Adam’s apple’
j. məqʰi
[məqʰi]
‘rain’
k. zuqʰi
[zuqʰi]
‘ugly’
l. ŋaŋeqʰo [ŋaŋeqʰo] ‘myself’
m. sqʰi
[sqʰi]
‘younger sister’
n. nəsqʰa
[nəsqʰa]
‘20’
o. sqʰə
[sqʰə]
‘to extinguish’

(20)

/q/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.

qoqo [qoqo] ‘indent’
qavla [qavla] ‘branch’
qur
[quʂ]
‘to snore’
qrə
[qʐə]
‘female yak’
vqe
[fqe]
‘to throw’
səqə [səqə] ‘small piece of machinery’
rdəqu [ʐdəqu] ‘mortar bowl’
tærqæ [tæʂqæ] ‘ladle’
rqo
[ʂqo]
‘(tree) trunk’
rqwa [ʂqwa] ‘Adam’s apple’
jæɴqjo [jæɴqjo] ‘palate’

42

(21)

/ɢ/
a.
b.
c.
d.

ɴɢwɛ [ɴɢwɛ] ‘5’
ʁaɴɢwɛ [ʁaɴɢwɛ] ‘15’
ɴɢwi [ɴɢwi] ‘hoe’
bæɴɢe [bæɴɢe] ‘spider, fly’

3.2.1.6 Glottal stop [ʔ]
The glottal stop is not a phoneme in Stau, only occurring before word-initial onsetless V syllables.
3.2.2 Affricates
Stau has three series of affricates: alveolar, retroflex, and alveolo-palatal. The threeway voicing contrast found in the stops is also present in the three affricate series. Another
similarity between affricates and stops is that, like voiced stops, voiced affricates also occur
infrequently word-initially.
3.2.2.1 Alveolar affricates /tsʰ, ts, dz/
Stau distinguishes three alveolar affricates, the voiceless aspirated avleolar /tsʰ/, the
voiceless unaspirated /ts/, and the voiced aspirated /dz/. Examples of each phoneme follow
below.
(22)

/tsʰ/
a. tsʰeke [tsʰeke] ‘hot’
b. tsʰɛko [tsʰɛko] ‘sheepfold’
c. tsʰædəm [tsʰædəm] ‘pitcher, thermos’
d. tsʰə
[tsʰə]
‘salt’
e. tsʰosusu [tsʰosusu] ‘idle’
f. vtsʰu
[ftsʰu]
‘to milk (cow)’
g. ætsʰe [ætsʰe] ‘a little’
h. kʰətsʰi [kʰətsʰi] ‘water channel’
i. bətsʰɛl [bətsʰɛl] ‘water spinach’
j. xotsʰɛv [xotsʰɛv] ‘pepper (vegetable)’

43

k.
l.
m.
n.
o.
p.
q.
r.

rgɛrtsʰo
mtsʰu
ntsʰɛm
rtsʰe
rtsʰə
rtsʰu
xtsʰə
vtsʰu

[ʐgɛ˞tsʰo]
[m̥tsʰu]
[n̥tsʰɛm]
[ʂtsʰe]
[ʂtsʰə]
[ʂtsʰu]
[xtsʰə]
[ftsʰu]

‘spine’
‘lake’
‘between, around’
‘lung’
‘cough’
‘to kick’
‘earth, soil’
‘to take out (of water)’

(23)

/ts/
a. tsi
[tsi]
‘grass’
b. tsəgə
[tsəgə]
‘clothing’
c. tsələ
[tsələ]
‘cat’
d. tsəntʰɛv [tsəntʰɛv] ‘scissors’
e. tsoŋ
[tsoŋ]
‘scallion’
f. atsawatsa [atsawatsa] ‘locust’
g. tə-tsə-sə [tətsəsə]
‘rotten’ (dirneut -rot-prf)
h. nətso
[nətso]
‘sun’
i. tɕʰatsoŋ [tɕʰatsoŋ] ‘all’
j. rə-rŋutsu [ʐəʐŋutsu] ‘to kneel’ (dirup -kneel)
k. rtsɛ
[ʂtsɛ]
‘deer’
l. rtsæmbræ [ʂtsæmbræ] ‘bowl that catches milled tsampa’
m. rtso
[ʂtso]
‘floor’
n. rtsudʑu [ʂtsudʑu] ‘number’
o. xtsoŋma [xtsoŋma] ‘clean’

(24)

/dz/
a. dzəvə
[dzəvə]
‘husband’
b. dzo
[dzo]
‘bridge’
c. rdzulu
[ʐdzulu]
‘to crawl’
d. mdzemdze [mdzɛmdze] ‘polite’
e. ndzə
[ndzə]
‘to hide it’
f. ndzæ
[ndzæ]
‘third floor room’
g. ndzu
[ndzu]
‘to sit down’
h. mtɕʰurdzæ [mtɕʰurdzæ] ‘teapot’
i. rdzærə
[ʐdzæʐə]
‘peak of mountain’
j. wərdzə
[wərdzə]
‘yak tail’
k. vdzi
[vdzi]
‘human’

44

3.2.2.2 Retroflex affricates /tʂʰ, tʂ, dʐ/
Although retroflex affricates are attested in Stau, they are fairly infrequent and
mainly occur in borrowings from either Tibetan or Chinese. All retroflex affricates that
appear in my dataset are included in the examples below. If an example is a known loanword, a word proving its loanword status will be included underneath it.
Please note, with reference to the borrowings from Tibetan, that I am not proposing
that the varieties mentioned below are the precise donor languages from which Stau has
borrowed. The Tibetan form that Stau most closely resembles was chosen from the STEDT
database in order to clearly demonstrate loanword status. However, one could note that the
Tibetan varieties which the Stau loanwords resemble most closely chiefly belong to Amdo,
a dialect spoken in Daofu County.
Here follow all occurrences of the retroflex affricates /tʂʰ/, /tʂ/, and /dʐ/ in my
dataset:
(25)

/tʂʰ/
a. tʂʰeɕaŋ [tʂʰeɕaŋ]
‘wagon’
b. tʂʰatʂoŋ [tʂʰatʂõ]
‘tea cup’
Chinese: <chā zhong> ‘tea cup’

(26)

/tʂ/
a. tʂəma
[tʂəma]
‘dirty’
13
53
Batang Tibetan: tʂɪ mɑʔ
‘dirt, filth’
b. tʂæ
[tʂæ]
‘to cut with scissors’
Alike Tibetan: ndʐak
‘to cut, sever’
c. tʂərvə
[tʂə˞və]
‘bell’
Alike Tibetan: tʂər wə
‘bell’
d. tʂʰatʂoŋ
[tʂʰatʂõ]
‘tea cup’
Chinese:
<chá zhǎn> ‘tea cup’

45

(27)

/dʐ/
a. ndʐændʐæ
[ndʐændʐæ] ‘same’
Alike Tibetan:
ndʐa
‘look like, resemble’
b. ndʐə
[ndʐə]
‘time’
c. æ-ndʐə
[ændʐə]
‘first time’ (num1-time)
d. su-ndʐə
[sundʐə]
‘third time’ (num3-time)
e. vdʐər
[vdʐə˞]
‘to roll up’
55
Batang Tibetan:
dʐi
‘roll up (cloth)’
f. mdʐu
[mdʐu]
‘dragon’
Zeku Amdo Tibetan: mdʐək
‘dragon’
g. mdʐu
[mdʐu]
‘thunder’
Zeku Amdo Tibetan: mdʐək
‘thunder’
h. mdʐu
[mdʐu]
‘wild yak’
i. ʁdʐu
[ʁdʐu]
‘to mill’
j. ʁdʐu
[ʁdʐu]
‘tsampa’
k. tadʐu
[tadʐu]
‘silk’
Zeku Amdo Tibetan: tʂhu tsə
‘silk’

3.2.2.3 Alveolo-palatal affricates /tɕʰ, tɕ, dʑ/
Three alveolo-palatal affricates are phonemic in Stau, the voiceless aspirated /tɕʰ/,
the voiceless unaspirated /tɕ/, and the voiced /dʑ/. Below are examples to demonstrate the
contrast between the three phonemes.
(28)

/tɕʰ/
a. tɕʰɛʁloŋ
[tɕʰɛʁlõ]
‘bull’
b. tɕʰæ
[tɕʰæ]
‘on’
c. tɕʰæɣwə
[tɕʰæɣwə]
‘millstone’
d. ɲtɕʰaχpa
[ɲ̥tɕʰaχpa]
‘to fold’
e. tɕʰətɕʰə
[tɕʰətɕʰə]
‘salty’
f. tɕʰutsʰu
[tɕʰutsʰu]
‘watch’
g. ætɕʰə
[ætɕʰə]
‘with’
h. ɮətɕʰu rgəmɛ [ɮətɕʰu ʐgəmɛ] ‘stones’
i. lutɕʰoŋ
[lutɕʰõ]
‘young’
j. noptɕʰo
[noptɕʰo]
‘side’
k. mtɕʰutɕu
[m̥tɕʰutɕu]
‘tassel’
l. mtɕʰurtin
[m̥tɕʰuʂtɪn]
‘tower’
m. kɛntɕʰi-rə
[kɛntɕʰiʐə]
‘good to look at’ (good.looking-const)
n. noŋtɕʰə
[nõtɕʰə]
‘guts’
o. rtɕʰæmbəqolu [ʂtɕʰæmbəqolu] ‘bubble’

46

p. rtɕʰu
q. ʁavtɕʰo
r. xtɕʰo

[ʂtɕʰu]
[ʁaftɕʰo]
[xtɕʰo]

‘bottle for making wine’
‘16’
‘6’

(29)

/tɕ/
a. tɕi
[tɕi]
‘hat’
b. tɕitɕæ
[tɕitɕæ]
‘rind’
c. tɕeʁoŋ
[tɕeʁõ]
‘kitchen’
d. tɕɛ
[tɕɛ]
‘road’
e. tɕæ
[tɕæ]
‘tea’
f. tɕæmtsʰæ
[tɕæmtsʰæ]
‘tea strainer’
g. tɕaχpa
[tɕaχpa]
‘to steal’
h. tɕədə
[tɕədə]
‘book’
i. tɕo
[tɕo]
‘waist’
j. tɕonzɛr
[tɕonzɛʐ]
‘nail’
k. tɕutɕæ
[tɕutɕæ]
‘metal’
l. ætɕɛ
[ætɕɛ]
‘together’
m. rə-tɕe
[ʐə-tɕe]
‘to come up’ (dirup -come.up)
n. zamasotɕi
[zamasotɕi] ‘strainer’
o. rŋərŋæsətɕe [rŋərŋæsətɕe] ‘flipper’
p. ŋkʰratɕa
[ŋ̥kʰʂatɕa]
‘to shiver’
q. tsʰɛtɕə
[tsʰɛtɕə]
‘cleaver’
r. pətɕo
[pətɕo]
‘to wreck, tear’
s. mtɕʰutɕu
[m̥tɕʰutɕu]
‘tassel’
t. ʁəptɕæ
[ʁəptɕæ]
‘hair on head’
u. zɣartɕa
[zɣaʂtɕa]
‘celery’
v. vtɕækʰæ-zŋo-re [ftɕækʰæzŋore] ‘rack for hanging things on’ (?-hang-nmlz)
w. xtɕərscɛ
[xtɕərscɛ]
‘clip, pin’

(30)

/dʑ/
a. dʑuæ
[dʑuæ]
‘to swim’
b. dʑə
[dʑə]
‘to meet, run into’
c. ædʑæpædʑæ [ædʑæpædʑæ] ‘sandals’
d. kɛ-dʑi
[kɛdʑi]
‘long’ (aprfx-long)
e. ndʑə
[ndʑə]
‘to pull’
f. ɣədʑelə
[ɣədʑelə]
‘after’
g. kɛ-dʑɛdʑi
[kɛdʑɛdʑi]
‘far’ (aprfx-far)
h. xakon-rə
[xakõdʑə]
‘to know’ (know-const)
i. kɛ-dʑoŋ
[kɛdʑoŋ]
‘straight’ (aprfx-straight)
j. ʑɛdʑuntɕʰɛm [ʑɛdʑuntɕʰɛm] ‘to dance’
k. ɕɑndʑu
[ɕãdʑu]
‘worm’

47

l.
m.
n.
o.
p.

mdʑəsnæ
kɛ-ndʑɛm
landʑə
snopdʑə
vdʑəvdʑə

[mdʑəsnæ]
[kɛndʑɛm]
[landʑə]
[snopdʑə]
[vdʑəvdʑə]

‘seed’
‘soft’ (aprfx-soft)
‘railing’
‘hanging (n.)’
‘friend’

3.2.3 Fricatives
Stau has voiced and voiceless pairs of fricatives for the same five places of articulation at which the stops are found. There is one exception: at the labial place of articulation,
the voiceless fricative is allophonic.
Some linguists find a third contrasting, voiceless aspirated fricative in the alveolar
and palatal series. Huang (1991) includes /sʰ/, /ɕʰ/, and /ɬʰ/ in her consonant inventory of
Daofu (=Stau), although she puts /s/ in brackets, possibly indicating that it is a marginal
phoneme. J. Sun (2000b) finds /sʰ/, /ʃʰ/, and /çʰ/ (contrasting respectively with /s/, /ʃ/, and
/ç/) in the related Puxi Shangzhai. In this article, he notes the way aspirated and unaspirated
consonants are inverted in verb stems to mark tense-aspect. In the same article, he shows
that Huang’s Daofu also uses the same aspiration inversion as tense-aspect marking. In my
own data I did not find the presence of aspirated fricatives in the lexicon, except perhaps
in the lateral fricatives (see §3.2.6.2 for more on this). However, I did not make a study of
tense-aspect; aspirated fricatives may yet be found at work in that grammatical function.
3.2.3.1 Labial fricative /v/
The labial fricative, as just mentioned, has no phonemic voiceless counterpart. The
voiceless labial fricative [f] is an allophone of /v/. The voiced labial fricative becomes [f]
next to voiceless consonants, e.g. (31f) and (31u). One exception to this rule seems to be
(31q), in which the /v/ remains fairly voiced.
In coda position, contrast between /v/ and the voiceless unaspirated labial stop /p/
is neutralized in favour of /v/. Phonetically, [v], [f], and [p] are variously pronounced in

48

this position. The fricatives appear in more careful speech, and the stop in faster speech
(Katylin Wonnell, p.c.). See (31x-31z) for examples of the labial fricative in syllable-final
position.
(31)

/v/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.
r.
s.
t.
u.
v.
w.
x.
y.
z.

vivəx
[vivəx]
‘pressure cooker’
vɛko
[vɛko]
‘pigsky’
væ
[væ]
‘pig’
vəvæ
[vəvæ]
‘to repair, build’
vo
[vo]
‘stomach’
vcʰə
[fcʰə]
‘to weigh’
vɕi
[fɕi]
‘to need’
vdzi
[vdzi]
‘person’
vɟə
[vɟə]
‘saliva’
vlɛ
[vlɛ]
‘to put something’
vqo
[fqo]
‘sky’
vsɛ
[fsɛ]
‘to kill’
rɛvræ
[ʐɛvʐæ]
‘to thresh’
vzɛvzə
[vzɛvzə] ‘to scratch’
vʑo
[vʑo]
‘to plane’
spəvji
[spəvji]
‘sores’
stʰvæ
[stʰvæ]
‘to press down’
ʑɛrvæ
[ʑɛrvæ]
‘blind’
me-skvɛ-rə [meskvɛʐə] ‘blunt’ (neg-sharp-const)
ɣvə
[ɣvə]
‘oats’
svo
[sfo]
‘bright’
ntʰvæ
[ntʰfæ]
‘to tread on it’
ŋkʰvo
[ŋ̥kʰfo]
‘key’
ʑəzɛv
[ʑəzɛv]
‘turtle’
ɣməlɛv
[ɣməlɛp] ‘fire’
ɬɛvsə
[ɬɛpsə]
‘lightning’

3.2.3.2 Alveolar fricatives /s, z/
Stau has two alveolar fricatives, voiceless /s/ and voiced /z/. The following examples show each phoneme in a variety of environments, representative of their distribution.

49

(32)

(33)

/s/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.
r.

si
[si]
‘liver’
sɛpʰjo
[sɛpʰjo]
‘direction’
səli
[səli]
‘to roll’
səm
[səm]
‘mood, seat of emotions’
su-ndʐə [sundʐə]
‘third time’ (num3-time)
skɛ
[skɛ]
‘voice, language’
smi-ze
[smeze]
‘daughter, girl’ (woman-dim)
sni
[sni]
‘nose’
sɲu
[sɲu]
‘beans (for pig feed)’
sŋuscæ [sŋuscæ] ‘blue’
spə
[spə]
‘pus’
sxɛsxo
[sxɛsxo]
‘to shake’
æsɛ
[æsɛ]
‘full’
zamasotɕi [zamasotɕi] ‘pot’
tsʰosusu [tsʰosusu] ‘idle’
qʰosto
[qʰosto]
‘back’
qʰəsji
[qʰəsji]
‘tomorrow’
xsər
[xsəʂ]
‘to stir-fry’

/z/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.
r.
s.
t.

zivæ
[zivæ]
‘mane’
zɛkʰoŋ [zɛkʰõ] ‘restaurant’
zæntɕʰæ [zæntɕʰæ] ‘to feel itchy’
zama [zama] ‘everything’
zəli
[zəli]
‘to fall’
zondo [zondo] ‘horn’
v-zu
[vzu]
‘to sew’
zdo
[zdo]
‘cloud’
zgəŋæ [zgəŋæ] ‘egg’
zgri
[zgʐi]
‘star’
zjɛ
[zjɛ]
‘to rake’
zɮæ
[zɮæ]
‘to chant’
zɲe
[zɲe]
‘7’
zʁa
[zʁa]
‘10’
ʁja-ze [jaze]
‘calf’
ʑəzɛv [ʑəzɛv] ‘turtle’
ɣzi
[ɣzi]
‘shoe’
rzelo [ʐzelo] ‘to lift skirts’
ʁzaʁzav [ʁzaʁzav] ‘thank you’
vzɛvzə [vzɛvzə] ‘to scratch’

50

3.2.3.3 Alveolo-palatal fricatives /ɕ, ʑ/
Two fricatives are found at alveolo-palatal place of articulation, /ɕ/ and /ʑ/. As
Jacques (2004:19) notes, many Rgyalrongic linguists use the postalveolar symbols <ʃ> and
<ʒ> for these sounds. However, I will follow Jacques (2004) and Huang (1991) and use the
more accurate alveolo-palatal <ɕ> and <ʑ>.
(34)

(35)

/ɕ/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.

ɕi
[ɕi]
‘highland barley’
ɕɛr
[ɕɛʂ]
‘glass’
ɕækɛcʰɛ [ɕækɛcʰɛ] ‘fat’
ɕaɣrə
[ɕaɣʐə]
‘small sickle’
ɕə
[ɕə]
‘teeth’
ɕovə-rgɛm [ɕovəʐgɛm] ‘cardboard box’ (paper-box)
ɕu
[ɕu]
‘strength’
ɕwæ
[ɕwæ]
‘night’
vɕjæ
[fɕjæ]
‘to seek something’
æɕəm
[æɕəm]
‘corn’
ɕəɕæ
[ɕəɕæ]
‘to wipe’
tʂʰeɕaŋ
[tʂʰeɕaŋ] ‘wagon’
laɕəv
[laɕəv]
‘rubber gloves’
ɕuɕu
[ɕuɕu]
‘behind’
vɕoxpæ [fɕoxpæ] ‘wing’
xɕi
[xɕi]
‘sweat’

/ʑ/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.

ʑigə
[ʑigə]
‘around’
ʑele
[ʑele]
‘turnip’
ʑɛrvæ [ʑɛrvæ] ‘blind’
ʑæ
[ʑæ]
‘to limp’
ʑæʁɟæ [ʑæʁɟæ] ‘lame person’
ʑəvə
[ʑəvə]
‘village’
ʑoŋər [ʑoŋə˞] ‘sweet potato’
ʑu
[ʑu]
‘yogurt’
æʑu
[æʑu]
‘maternal uncle’
xeʑi
[heʑi]
‘how many’
tə-ʑə [təʑə]
‘to melt’ (dirneut -melt)
jɛʑoŋbu [jɛʑoŋbu] ‘often, always’
zʁəʑjæ [zʁəʑjæ] ‘comb’

51

n. vʑo
o. ɣʑi
p. ʁʑə

[vʑo]
[ɣʑi]
[ʁʑə]

‘to plane’
‘to teach’
‘bow’

3.2.3.4 Velar fricatives /x, ɣ/
Stau distinguishes between two velar fricatives /x/ and /ɣ/. The voiceless /x/ wordinitially before vowels has two allophones, [x] and [h], and they are in free variation, as
evidenced in the following examples.
(36)

a. xævʑi ‘yet, still’
hævʑi ‘yet, still’
b. xaji ‘also’
haji ‘also’
In some cases preinitial /ɣ/ is disappearing word-initially. In words such as ɣdæmbæ

‘because’ and ɣmɛɕu ‘scar’, the voiced velar fricative is not clearly heard in isolation, but
appears more distinctly in sentence context.
Examples (37) and (38) provide a representative set of the occurrences of /x/ and /ɣ/
that demonstrates their phonemic status.
(37)

/x/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.

xændʑe [xændʑe]
‘scarf’
xaji
[xaji]~[haji] ‘also’
xə
[xə]
‘bull-yak crossbreed (Chinese: pian niu)’
xovɛxovɛ [xovɛxovɛ] ‘a minute ago’
xpʰə
[xpʰə]
‘butt’
xsærpæ [xsærpæ]
‘new’
xtʰəxtʰə [xtʰəxtʰə]
‘behind’
xtsʰə
[xtsʰə]
‘earth, soil’
maxe
[maxe]
‘water buffalo’
tə-xæ
[təxæ]
‘to come out’ (dirneut -come.out)
moxkɛr [moxkɛʂ]
‘white fungus’
rcaxpa [ʂcaxpa]
‘excrement’
sxɛsxo [sxɛsxo]
‘to shake’
vivəx
[vivəx]
‘pressure cooker’

52

(38)

/ɣ/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.

ɣə-pʰri [ɣəpʰʂi] ‘to untie (unvolitional)’ (dirin -untie)
ɣəræ
[ɣəʐæ]
‘chicken’
ɣəzə
[ɣəzə]
‘bird’
ɣmɛ
[ɣmɛ]
‘wound’
ɣrəkɛnəv [ɣʐəkɛnəv] ‘flood’
ɣzi
[ɣzi]
‘shoe’
ɣwæ
[ɣwæ]
‘to hug’
ɣzəɣzə [ɣzəɣzə] ‘some’
ɣe
[ɣe]
‘to touch’
maɣər [maɣə˞] ‘type of tree’
kɛ-ɣji
[kɛɣji]
‘light’ (aprfx-light)
sɲəɣɟi [sɲəɣɟi] ‘afternoon’
sɬəɣnə [sɬəɣnə] ‘moon’
ɣro
[ɣro]
‘to wither’
rɣi-ru-re [ʐɣiʐuʐe] ‘stable’ (horse-pen.in-nmlz)
zɣartɕa [zɣaʂtɕa] ‘whip’
moroɣ [moʐoɣ] ‘black fungus’

3.2.3.5 Uvular fricatives /χ, ʁ/
Two uvular fricatives occur in Stau, the voiceless /χ/ and voiced /ʁ/.
Like the voiced velar fricative, /ʁ/ is also deteriorating word-initially before consonants. For example, the word ‘pretty’ is pronounced [ajerə] in isolation. In sentence
context, ajerə no longer has a word-initial vowel, as one would expect. Instead it has a
word-initial uvular fricative.
(39)

*ŋa ʔaje-rə
ŋa ʁje-rə
I
pretty-const
‘I am pretty.’
Other examples include ʁcʰi ‘hole’ and ʁnɛmju ‘roof’.
/ʁ/ is found much more frequently in Stau than /χ/. /χ/ occurs 13 times in my dataset,

all of which are included in (40), whereas /ʁ/ occurs 79 times.

53

(40)

(41)

/χ/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.

χodzu-pare [χodzupare] ‘cloth worn on head’ (?-cloth)
χjə
[χjə]
‘to destroy’
æχe
[æχe]
‘a little’
læχape
[læχape]
‘cabbage’
kəχo
[kəχo]
‘bark’
ɲtɕʰaχpa [ɲ̥tɕʰaχpa] ‘to fold’
tɕaχpa
[tɕaχpa]
‘to steal’
χtɛ
[χtɛ]
‘to return (something to someone)’
χtsa
[χtsa]
‘to cut oneself’
tærtʰaχ
[tæʂtʰaχ]
‘prayer flags’
toχtoχ
[toχtoχ]
‘narrow’
tsʰæsnəχ [tsʰæsnəχ] ‘spinach’

/ʁ/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.

ʁa
[ʁa]
‘door’
ʁalo
[ʁalo]
‘chest’
ʁərja
[ʁərja]
‘bone’
ʁəvdæ
[ʁəvdæ]
‘to nod’
ʁo
[ʁo]
‘to help’
ʁdu
[ʁdu]
‘pail’
ʁja
[ʁja]
‘male yak’
ʁjə
[ʁjə]
‘fish’
ʁlə-və
[ʁləvə]
‘to sing’ (song-make)
ʁmæʁmæ [ʁmæʁmæ] ‘low’
jəʁo
[jəʁo]
‘upstairs’
sɲiʁə
[sɲiʁə]
‘pen’
ɲəʁzɛn
[ɲəʁzɛn]
‘wool’
mbre-zʁe-jo [mbrezʁejo] ‘rice cooker’ (rice-boil-container)
rʁa-mə
[ʐʁamə]
‘crazy person’ (crazy-pers)

54

3.2.4 Rhotic
The rhotic /r/ has four allophones in Stau: the trill [r], the voiced and voiceless
retroflex fricatives [ʂ] and [ʐ], and rhotacization on the vowel. They largely occur in complementary distribution.
The rhotic is realized as the voiceless retroflex fricative [ʂ] when it occurs as a
preinitial before voiceless consonants, as a medial after voiceless aspirated consonants, and
often it occurs as a coda. For example:
(42)

rcaxpa [ʂcaxpa] ‘excrement’
kʰri
[kʰʂi]
‘chair’
spəɲcʰɛr [spəɲcʰɛʂ] ‘frog’
The voiced [ʐ] has the most widespread distribution. It occurs as a simple initial,

in preinitial position with voiced consonants, and as a medial with voiced and voiceless
unaspirated consonants. For example:
(43)

rekwe [ʐekwe] ‘foal’
rdzærə [ʐdzæʐə] ‘peak of mountain’
ɣrəɲcʰɛr [ɣʐəɲcʰɛʐ] ‘puddle’
prilæ [pʐilæ] ‘to whinny’
The occurrence of the trill [r] is not completely predictable. It overlaps in distribu-

tion with [ʂ] and especially [ʐ], occurring in place of these in longer utterances. For instance,
the /r/ in the word ‘wine’ is [ʐ] in isolation, but becomes [r] in the compound ‘cigarette and
wine shop’:
(44)

ara
[aʐa]
‘wine’
təvæ-ara-zjəre [təvæarazjəre] ‘cigarette and wine shop’ (tobacco-wine-shop)

55

Often this change happens between words in isolation and in context: ‘to call, shout’
is [skəʐi] in isolation and [skəri] in context.
Trill allophone [r] also alternates with [ʂ] but not in all contexts as with [ʐ]. In
voiceless consonant clusters [ʂ] always occurs whether in isolation or context, but in context
word-final [ʂ] does become [r]. For instance, [pɛʂ] ‘photo’ is realized as [pɛr] when a vowel
follows it in a sentence.
The trill allophone also occurs word-medially after a vowel and before a consonant,
e.g.,
(45)

sərʁə [sərʁə] ‘rotary paddle’
The final allophone of /r/ to be mentioned is rhotacization on a preceding vowel.

Most frequently rhotacization happens word-finally, as in (46), but also occurs word-medially
before the lateral approximant, as in (47).
(46)

musɛr [musɛ˞] ‘young woman’

(47)

barlu [ba˞lu] ‘leaf’
Rhotacization is attested on the vowels /ɛ/, /ə/, /u/, and /ɑ/. Word-finally, [ʂ] is also

an attested realization of /r/ after all these vowels (except /ɑ/ as it never occurs in a wordfinal syllable before /r/), and there seems to be no rule dictating which allophone is used. In
fact, rhotacization can alternate with a voiceless fricative realization, as (48) demonstrates:
(48)

[ɣmu˞] ~ [ɣmuʂ] ‘mouth, beak’
The rhotic, like /ɣ/ and /ʁ/, in word-initial preinitial position often appears only as

voicing or is not articulated at all. The only way to tell an /r/ exists is to hear the word
in a sentence. There the /r/ betrays itself as [r], [ʐ], or rhotacization. For instance, rɟætɕæ

56

‘looseleaf tea’ in isolation shows a voicing bar but no other indication of a word-initial /r/.
In a sentence, the /r/ becomes apparent, being heard as a trill:
(49)

tʰɛ rɟætɕæ
ŋə-rə
this looseleaf.tea cop-const
‘This is looseleaf tea.’
Below follows a representative set of examples of the occurrence of /r/, which in-

cludes all its four allophones:
(50)

/r/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.
r.
s.
t.
u.
v.
w.
x.
y.
z.
aa.
ab.
ac.

rekwe
[ʐekwe] ‘foal’
rɛmɛ
[ʐɛmɛ]
‘mare’
ræpʰi
[ʐæpʰi] ‘mahjong’
rata
[ʐata]
‘mill’
rərɮɛv [ʐəʐɮɛv] ‘wave’
ro
[ʐo]
‘1’
rcaxpa [ʂcaxpa] ‘excrement’
rdzærə [ʐdzæʐə] ‘peak of mountain’
rɣi-ru-re [ʐɣiʐuʐe] ‘stable’ (horse-pen.in-nmlz)
rjɛ
[ʐjɛ]
‘8’
rko
[ʂko]
‘ankle’
rŋæ
[ʐŋæ]
‘face’
rtsʰe
[ʂtsʰe]
‘lung’
rʁa-mə [ʐʁamə] ‘crazy person’ (crazy-pers)
rwo
[ʐwo]
‘ice’
skəri
[skəʐi]
‘to call, shout’
ɣəræ
[ɣəʐæ]
‘chicken’
ara
[aʐa]
‘wine’
ɮo-rə
[ɮoʐə]
‘tasty’ (tasty-const)
ɣrə-ɲcʰɛr [ɣʐəɲcʰɛʐ] ‘puddle’ (water-?)
kʰri
[kʰʂi]
‘chair’
mbre
[mbʐe]
‘rice’
prilæ
[pʐilæ] ‘to whinny’
ŋkʰurlu [ŋ̥kʰu˞lu] ‘wheel’
mormi [mormi] ‘eyebrow’
tærqæ [tæʂqæ] ‘ladle’
pʰusɛr [pʰusɛ˞]
‘young man’
pətɛr
[pətɛʂ]
‘rash’
spəɲcʰɛr [spəɲcʰɛʂ] ‘frog’

57

ad.
ae.

ɲɛr
ʁərʁər

[ɲɛʐ]
‘to taste’
[ʁərʁər] ‘round’

3.2.5 Nasals
Stau has nasals at all places of articulation that stops and fricatives are found at. The
one exception is that the uvular nasal [ɴ], like the voiced uvular stop, is not a full phoneme.
Word-initially before voiceless aspirated consonants, nasals are voiceless, only articulated as a closing of the lips together before the stop, sometimes with airflow through
the nasal cavity. Several examples of voiceless nasals are provided in (51):
(51)

mkʰərjɛ [m̥kʰərjɛ] ‘pipe’
mpʰi [m̥pʰi] ‘to card’
ntsʰɛm [n̥tsʰɛm] ‘between, around’
ɲcʰə
[ɲ̥cʰə]
‘to hit, beat; thresh’

3.2.5.1 Bilabial nasal /m/
The bilabial nasal /m/ occurs most frequently in my data―188 times (followed
closely by /n/ which occurs 172 times). One of the reasons for its frequency of occurrence may be that it has the widest distribution of all the nasals. In addition to appearing as
a simple initial and as a preinitial in homorganic consonant clusters, it occurs as a preinitial
before alveolar, palatal, and velar consonants.2
Below is a sample of the occurrences of /m/ in my dataset.

2

See Table 4 for a chart that shows the co-occurrence of preinitial nasals with consonants in consonant
clusters.

58

(52)

/m/
a. mi
[mi]
‘mole’
b. metʰɛv [metʰɛv] ‘stove’
c. mɛji
[mɛji]
‘butter’
d. mæmæ [mæmæ] ‘grandmother’
e. marna [marna] ‘oil’
f. mə
[mə]
‘younger brother’
g. mo
[mo]
‘eye’
h. mupʰa [mupʰa] ‘sow’
i. mbjɛ-mə [mbjɛmə] ‘deaf person’ (deaf-pers)
j. mkʰə-rjɛ [m̥kʰərjɛ] ‘pipe’ (smoke-ceramic)
k. mɲə
[mɲə]
‘can, able to’
l. ŋəmɛ
[ŋəmɛ]
‘cow’
m. æmə
[æmə]
‘mother’
n. rŋæma [ʐŋæma] ‘horse tail’
o. ʁnɛmju [ʁnɛmju] ‘roof’
p. ʁəmɬe
[ʁəmɬe]
‘braid’
q. tɕæmtsʰæ [tɕæmtsʰæ] ‘tea strainer’
r. ɣmɛɕu [ɣmɛɕu] ‘scar’
s. smi
[smi]
‘woman’
t. ʁmɛʑi
[ʁmɛʑi]
‘pot without handles’
u. ɲəlɛm
[ɲəlɛm]
‘dream (n.)’
v. sənəm [sənəm] ‘farming’
w. səmnɛ
[səmnɛ]
‘worry’

3.2.5.2 Alveolar nasal /n/
Like most languages, Stau has an alveolar nasal /n/. In (53), examples demonstrating
the phonemic status of this sound are provided.
(53)

/n/
a.
b.
c.
d.
e.
f.
g.
h.
i.

nə-ndʐə
nene
nɛmɛ
na
nə
noptɕʰo
ndjindji
æne
kɛ-nəv

[nəndʐə]
[nene]
[nɛmɛ]
[na]
[nə]
[noptɕʰo]
[ndjindji]
[æne]
[kɛnəv]

‘second time’ (num2-time)
‘breast’
‘toe’
‘post (for building house)’
‘to rest’
‘side’
‘red’
‘paternal aunt’
‘deep’ (aprfx-deep)

59

j.
k.
l.
m.
n.
o.
p.
q.
r.
s.
t.
u.

pjɛno [pjɛno] ‘meat’
ɕɛmnu [ɕɛmnu] ‘glasses’
kɛ-nɮə [kɛnɮə] ‘heavy’ (aprfx-heavy)
ɕontʰo [ɕõntʰo] ‘fruit’
kæntɕʰæ [kæntɕʰæ] ‘rolling pin’
tʰi-ɣne [tʰiɣne] ‘they (dual)’ (3-two)
rni
[ʐni]
‘to mix’
ʁnɛʁnɛ [ʁnɛʁnɛ] ‘dark’
tə-ji-n [təjɪn]
‘to say’ (dirneut -say-2)
ʁɮəstɛn [ʁɮəstɛn] ‘sleeping mat’
rgɛrgən [rgɛrgən] ‘teacher’
səŋun [səŋun] ‘who’

3.2.5.3 Palatal nasal /ɲ/
In addition to the typical bilabial and alveolar nasals included in the consonant inventories of most languages, Stau also has the palatal nasal /ɲ/, examples of which follow
below:
(54)

/ɲ/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.

ɲi
[ɲi]
‘you (sg)’
ɲɛr
[ɲɛʐ]
‘to taste’
ɲæ
[ɲæ]
‘fish’
ɲæɲæ
[ɲæɲæ]
‘black’
ɲəmæ-mɛto [ɲəmæmɛto] ‘sunflower’ (?-flower)
ɲoŋpæ
[ɲoŋpæ]
‘old’
ɲcʰæræ
[ɲ̥cʰæræ]
‘to play’
ɲcʰə
[ɲ̥cʰi]
‘to hit, beat; to thresh’
ɲɟɛrdo
[ɲɟɛrdo]
‘change’
tʰiɲi
[tʰiɲi]
‘they’
lɛɲu
[lɛɲu]
‘evening’
kovɲu
[kofɲu]
‘garlic shoot’
mɲærə
[mɲæʐə]
‘not’
mtsʰərɲi
[mtsʰərɲi] ‘to pull’
sɲæsɲæ
[sɲæsɲæ] ‘bitter’
ʁazɲe
[ʁazɲe]
‘17’
ʁɲæ
[ʁɲæ]
‘cow dung’

60

3.2.5.4 Velar nasal /ŋ/
Stau has a phonemic velar nasal /ŋ/. Most languages with a phonemic /ŋ/ restrict
its distribution to initial, medial, or final position, or a combination these (Anderson 2013).
In contrast, Stau /ŋ/ has quite an unrestricted distribution. It appears as a simple initial,
word-medially before vowels, word-initially and -medially as a preinitial, and as a coda, as
the examples in (55) demonstrate.
The velar nasal has one allophone, nasalization. If the sound preceding a /ŋ/ is a
vowel, particularly if the nasal is word-final, it will be realized as nasalization on the vowel,
e.g. (55p).
(55)

/ŋ/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.

ŋe-rə
[ŋeʐə]
‘to be okay’ (be.okay-const)
ŋæ
[ŋæ]
‘I’
ŋarɟɛr
[ŋarɟɛʂ]
‘roar’ (n.)
ŋəzu
[ŋəzu]
‘to howl’
ŋorə
[ŋoʐə]
‘illness’
ŋgəja
[ŋgəja]
‘ring’
ŋkʰrɛ
[ŋ̥kʰʂɛ]
‘to shake something’
æŋæze
[æŋæze]
‘baby’
tʰoŋbe
[tʰoŋbe]
‘pot’
tʰoŋkæ
[tʰoŋkæ]
‘thangka’
toŋskə
[toŋskə]
‘thread’
rŋarʁe-pare [rŋarʁepare] ‘cloth for washing’ (wash-cloth)
rŋe
[ʐŋe]
‘to hear’
sŋuscæ
[sŋuscæ]
‘blue’
zŋo
[zŋo]
‘to hang’
coŋ
[cõ]
‘clay wall’
ɣrətoŋ
[ɣʐətõ]
‘well (n.)’

3.2.5.5 Uvular nasal [ɴ]
Unlike the aforementioned four nasals, the uvular nasal [ɴ] does not have a wide distribution. It only appears together with the uvular consonants /q/ and /ɢ/. All the examples
of [ɴ] available in my data are included below:

61

(56)

/ɴ/
a.
b.
c.
d.
e.

bæɴɢe [bæɴɢe] ‘spider, fly’
ɴɢwi [ɴɢwi] ‘hoe’
ɴɢwɛ [ɴɢwɛ] ‘5’
ʁaɴɢwɛ [ʁaɴɢwɛ] ‘15’
jæɴqjo [jæɴqjo] ‘palate’

3.2.6 Laterals
Stau has three lateral consonants: the approximant /l/, and the voiceless and voiced
fricatives /ɬ/ and /ɮ/. Examples of each are given in the subsections §3.2.6.1 and §3.2.6.2.
3.2.6.1 Lateral approximant /l/
The lateral approximant /l/ appears as an onset in word-initial and -medial position,
as a medial in consonant clusters, and as a coda, as the following examples show:
(57)

/l/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.

leskæ
[leskæ]
‘work’
lɛpu-rɲərɲə [lɛpurɲərɲə] ‘carrot’ (radish-yellow)
lærɲə
[lærɲə]
‘asparagus lettuce’
landʑə
[landʑə]
‘railing’
ləpʰu
[ləpʰu]
‘tree’
loŋbutɕe [lõbutɕe]
‘elephant’
lu
[lu]
‘pole of tool’
kʰælɛv
[kʰælɛv]
‘lid’
prilærə
[pʐilærə]
‘to whinny’
jələ
[jələ]
‘saying, expression’
qavla
[qavla]
‘branch’
plɛ
[plɛ]
‘thigh’
v-lɛ
[vlɛ]
‘to put, leave something’
ʁlə-və
[ʁləvə]
‘to sing’ (song-make)
snəmtsʰɛl [snəmtsʰɛl] ‘cole’
tsoŋrəl
[tsõʐəl]
‘onion’

62

3.2.6.2 Lateral fricatives /ɬ, ɮ/
Stau has two lateral fricatives, /ɬ/ and /ɮ/. These lateral fricatives only occur as
initials, both as simple initials and in clusters with preinitials. The voiceless /ɬ/ occurs with
voiceless preinitials and the voiced /ɮ/ with voiced preinitials.
Although lateral fricatives are cross-linguistically infrequent, occurring in nine percent of the world’s languages (Maddieson 2013b), they are not infrequent areally. Most
Rgyalrongic languages have the voiceless lateral fricative. A few inventories include it as a
marginal phoneme. The voiced counterpart is not as prevalent. Of Rgyalrongic languages
other than Stau, the voiced lateral fricative only occurs in Puxi Shangzhai. Apart from
Rgyalrongic, Ethnic Corridor languages (such as Qiang and Ersu) and Southeastern Ngwi
languages (e.g. Phola and Azha) also have lateral fricatives.
The reader may recall from §3.2.3 that aspirated fricatives have been found in some
languages related to Stau, but that in my data I did not find any except possibly among
the lateral fricatives. Lateral fricatives in Stau have an unusual phonetic articulation. They
have a stop release, almost as if they were reverse affricates. The voiced /ɮ/ release sounds
like a [d], and the voiceless /ɬ/ sounds like an aspirated [tʰ]. In addition, one initial voiceless
lateral fricative and all the voiceless lateral fricatives in consonant clusters have a release
that sounds like an unaspirated [t].
If we compare spectrograms of the two word-initial voiceless lateral fricatives, one
from ɬɛɬə ‘to get wet’ with a [tʰ]-like release (Figure 6) and one from ɬɛvsə ‘lightning’ with a
[t]-like release (Figure 7), we can see the difference in the length of VOT and in the amount
of aspiration.

63

Figure 6. Spectrogram of /ɬ/ with [tʰ]-like release; aspiration from release burst to onset of
voicing shown between vertical lines
In Figure 6 above, we can see on the left the expected frication of a lateral fricative,
and in the center between vertical lines a long section of aspiration. This aspiration, from
release burst to onset of voicing, lasts 60 ms, and is what gives this /ɬ/ a [tʰ]-like sound.
In Figure 7 below, we see a similar section of frication on the left. Then at the vertical
line there is a release burst (the [t]-like release), after which the vowel follows immediately
without any aspiration.

64

Figure 7. Spectrogram of /ɬ/ with [t]-like release; dotted line on release burst, displays no
aspiration
The two spectrograms show a clear phonetic difference between the two fricatives.
If the lateral fricatives with [tʰ] release are phonemically different from those with a [t]
release, then we would have /ɬʰ/ and /ɬ/. However, since I only have one example of [t]release /ɬ/ as a simple initial and no longer have access to a Stau speaker to consult native
speaker intuition on these sounds, there is not enough evidence to claim the existence of a
phonemic aspirated lateral fricative.
Examples of /ɬ/ and /ɮ/ in a variety of contexts are found below:
(58)

/ɬ/
a.
b.
c.
d.
e.
f.
g.
h.

ɬɛɬə
[ɬɛɬə]
‘wet’
ɬɛvsə
[ɬɛpsə]
‘lightning’
ɬæ
[ɬæ]
‘god’
ɬækoŋ
[ɬækõ]
‘lama’s house’
ɬə
[ɬə]
‘cow’s milk’
ɬokəv
[ɬokəv]
‘pumpkin’
æ-mpʰæɬinə-re [æmpʰæɬinəʐe] ‘inside out’ (num1-inside.out-nmlz)
nə-mpʰæɬinə-re [nəmpʰæɬinəʐe] ‘inside out’ (num2-inside.out-nmlz)

65

(59)

i.
j.
k.
l.
m.
n.
o.
p.
q.

mɬe
ʁəmɬe
rɬə
sɬə
sɬəkʰro
æ-sɬə
qʰætəsɬoŋ
sɬopræ
vɬɛ

[mɬe]
[ʁəmɬe]
[ʂɬə]
[sɬə]
[sɬəkʰʂo]
[æ-sɬə]
[qʰætəsɬõ]
[sɬopræ]
[fɬɛ]

‘to braid’
‘braid’
‘wheat flour’
‘stairs’
‘step (n.)’
‘one month’ (num1-month)
‘glad’
‘university’
‘ashes’

/ɮ/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.

ɮi
[ɮi]
‘wheat’
ɮɛ
[ɮɛ]
‘to come, return’
ɮɛvɮi
[ɮɛvɮi]
‘wristbone’
ɮæ
[ɮæ]
‘hand’
ɮæbjænoŋ [ɮæbjænõ] ‘palm (of hand)’
ɮə
[ɮə]
‘field’
ɮəɮæ
[ɮəɮæ]
‘to plow’
ɮərji
[ɮərji]
‘to mix’
ɮo-rə
[ɮoʐə]
‘tasty’ (tasty-const)
ɣɮə
[ɣɮə]
‘4’
ɣɮəli
[ɣɮəli]
‘pestle’
ʁɮi
[ʁɮi]
‘to roll’
ʁɮərko
[ʁɮəʂko] ‘bamboo shoot’
vɮi
[vɮi]
‘neck’
vɮɛ
[vɮɛ]
‘tongue’
vɮæzæ
[vɮæzæ] ‘sleeve’
zɮæræ
[zɮæræ] ‘to winnow’

3.2.7 Glides
Stau has two glides, the palatal /j/ and labio-velar /w/. Both glides have similar
distribution, appearing as simple onsets, word-initially and word-medially, and as medials
in consonant clusters.

66

(60)

(61)

/j/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.
r.
s.
t.
u.
v.

jɛləgə
[jɛləgə]
‘so-called’
jæ
[jæ]
‘mouth’
jæ-kʰæ ɲcʰɛɲcə [jækʰæ ɲcʰɛɲcə] ‘to argue’ (mouth-ins to.fight)
jəʁo
[jəʁo]
‘upstairs’
jo
[jo]
‘house’
jovə
[jovə]
‘wife’
kɛ-ji
[kaji]
‘pretty’ (aprfx-pretty)
ŋæji
[ŋæji]
‘we’
jə
[jə]
‘to say’
kɛ-jɛ
[kɛjɛ]
‘easy’ (aprfx-easy)
æjæ
[æjæ]
‘maternal aunt’
ŋgəja
[ŋgəja]
‘ring’
mojo
[mojo]
‘eyelid’
bjolæ
[bjolæ]
‘to float’
ɕjæ
[ɕjæ]
‘to seek something’
ndjɛvji
[ndjɛvji]
‘to doze’
ɣræmjæ
[ɣʐæmjæ]
‘shadow’
s-bjæ
[spjæ]
‘to split’
kɛrjɛ
[kɛrjɛ]
‘ceramic’
qʰəsji
[qʰəsji]
‘tomorrow’
zʁəʑjæ
[zʁəʑjæ]
‘comb’
zjəre
[zjəre]
‘shop (n.)’

/w/
a. weqe
[weqe]
‘rabbit’
b. wɛrtsʰi
[wɛʂtsʰi]
‘lard’
c. wərdzə [wərdzə] ‘yak tail’
d. wo
[wo]
‘again’
e. wocʰi
[wocʰi]
‘lower abdomen’
f. wocæ
[wocæ]
‘navel’
g. wur
[wuʂ]
‘pillow’
̚
h. wut
[wut]
‘light’
i. kʰawa
[kʰawa]
‘snow’
j. atsawatsa [atsawatsa] ‘locust’
k. ɕwæ
[ɕwæ]
‘night’
l. ɴɢwi
[ɴɢwi]
‘hoe (n.)’
m. rekwe
[ʐekwe]
‘foal’
n. ɣwæ
[ɣwæ]
‘to hug’
o. rqwa
[ʂqwa]
‘Adam’s apple’
p. rwo
[ʐwo]
‘ice’

67

3.3 Vowels
As introduced in §3.1, Stau has eight vowels, evidence for which is given in the following subsections, beginning with the front vowels from high to low, and ending with the
back vowels from low to high.3 Stau vowels at times exhibit secondary qualities, specifically nasalization and rhotacization. However, these are not phonemic, but are allophonic
realizations of the consonants /ŋ/ and /r/. Refer to §§3.2.5.4 and 3.2.4, respectively, for
more information.
In the next chapter, an acoustic analysis of Stau vowels is provided to give a better
sense of their qualities. Please see §4.2 for this.
3.3.1 High front vowel /i/
Stau has one high front vowel /i/. /i/ has two allophones, [i] and [ɪ]. The high front
lax [ɪ] occurs when /i/ is followed word-finally by a coda consonant. The examples in (62)
are all that appear in my data. The latter two verbs are in second-person conjugation.
(62)

mtɕʰurtin [mtɕʰuʂtɪn] ‘tower’
rtin
[ʂtɪn]
‘to stop’
nə-pʰiv [nəpʰɪv] ‘to close’ (dirdown -close)
tə-ji-n [təjɪn]
‘to say’ (dirneut -say-2)
The following examples demonstrate /i/ in a cross-section of contexts:

3

Verbs as evidence: The reader will remember from the introduction to §3 that, though there are vowel
alternations at work in the verbal morphophonemics of Stau, I still use vowels from verbs as evidence to
demonstrate segmental phonology.

68

(63)

/i/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.
r.
s.
t.
u.
v.

ki-lu
[kilu]
‘each, some’
zivæ
[zivæ]
‘mane’
ndjindji [ndjindji]
‘red’
rɣirɣi
[ʐɣiʐɣi]
‘short’
scici
[scici]
‘to look at, see’
rə-rmi-sə [ʐəʐmisə]
‘to sprout’ (dirup -sprout-prf)
səm-sci-rə [səmsciʐə]
‘happy’ (mood.spirit-have-const)
tsəgə-tʰi-re [tsəgətʰire]
‘clothesline’ (clothes-hang-nmlz)
mi
[mi]
‘mole’
pʰi
[pʰi]
‘to run away’
tɕi
[tɕi]
‘hat’
kɛdi
[kɛdi]
‘child’
kʰri
[kʰʂi]
‘chair for lamas’ (honorific form)
tʰani
[tʰani]
‘near’
tʰadʑi
[tʰadʑi]
‘far’
məqʰi
[məqʰi]
‘rain’
pərzi
[pərzi]
‘knife’
vɕi
[fɕi]
‘to want’
vdzi
[vdzi]
‘human’
vɛrvi
[vɛrvi]
‘slowly’
zəli
[zəli]
‘to fall’
xævʑi
[hævʑi]~[xævʑi] ‘yet, still’

3.3.2 Mid front tense vowel /e/
After /i/, /e/ is the next highest front in the Stau vowel inventory. Examples follow
below:
(64)

/e/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.

kɛ-de
nene
ŋe-rə
tsʰeke
væ-ze
xseskə
ʑele
ɣədʑelə
mɬe
tə-ŋe-sə

[kede]
[nene]
[ŋeʐə]
[tsʰeke]
[veze]
[xseskə]
[ʑele]
[ɣədʑelə]
[mɬe]
[təŋesə]

‘small’ (aprfx-small)
‘mother’s milk’
‘to be okay’ (be.okay-const)
‘hot’
‘piglet’ (pig-dim)
‘buddha’
‘turnip’
‘after’
‘to braid’
‘correct’ (dirneut -correct-prf)

69

k.
l.
m.
n.
o.
p.
q.
r.
s.

zɲe
[zɲe]
‘7’
pʰe
[pʰe]
‘to vomit’
ækʰe
[ækʰe]
‘paternal uncle’
pəqe
[pəqe]
‘cow’
qʰre
[qʰʂe]
‘to pull down’
ʁə-vəvæ-re [ʁəvəvæʐe] ‘hair salon’ (head-do-nmlz)
rə-tɕe
[ʐətɕe]
‘to come up’ (dirup -come.up)
nə-vəve
[nəvəve] ‘to do’ (dirdown -do)
maxe
[maxe]
‘water buffalo’

3.3.3 Mid front lax vowel /ɛ/
In addition to the mid front tense /e/, Stau also has the lax counterpart of that vowel,
/ɛ/:
(65)

/ɛ/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.
r.
s.
t.
u.
v.

nɛrdzə
[nɛʐdzə] ‘fingernail’
ɟɛuræ
[ɟɛuræ]
‘facial hair’
tə-vsɛ-sə [təfsɛsə] ‘to be killed’ (dirneut -be.killed-prf)
pɛrspo
[pɛrspo] ‘to walk’
sxɛsxo
[sxɛsxo] ‘to shake’
kɛ-ndʑɛm [kɛndʑɛm] ‘soft’ (aprfx-soft)
ɣmɛɕu
[ɣmɛɕu] ‘scar’
cʰɛ-rə
[cʰɛʐə]
‘idle’ (idle-const)
mɛji
[mɛji]
‘butter’
pɛrvə
[pɛrvə]
‘container’
ɣməlɛv
[ɣməlɛp] ‘fire’
ɲəlɛm
[ɲəlɛm]
‘dream (n.)’
pʰɛ
[pʰɛ]
‘to throw out’
rjɛ
[ʐjɛ]
‘8’
adɛ
[adɛ]
‘this’
mobrɛ
[mobʐɛ] ‘tears’
tɕitɕæ-ɣʑɛ [tɕitɕæɣʑɛ] ‘to peel’ (rind-peel)
tɕazjɛ
[tɕazjɛ]
‘rake’
ɕu-kɛcʰɛ [ɕukɛcʰɛ] ‘strong’ (strength-big)
kɛ-skvɛ [kɛskvɛ] ‘sharp’ (aprfx-sharp)
mə-cʰɛcʰɛ [məcʰɛcʰɛ] ‘busy’ (neg-busy)
ʁə-rʁɛ-scɛ [ʁərʁɛscɛ] ‘shampoo’ (head-wash-tool.instrument)

70

3.3.4 Low front vowel /æ/
Stau has a low front vowel /æ/, of which here are examples:
(66)

/æ/
a. æ-bɛr
[æbɛʂ]
‘first step’ (num1-step)
b. æ-ndʐə
[ændʐə]
‘first time’ (num1-time)
c. æ-sɲi
[æsɲi]
‘day’ (num1-day)
d. ækʰə
[ækʰə]
‘paternal uncle’
e. æŋæze
[æŋæze]
‘baby’
f. mæŋgɛ
[mæŋgɛ]
‘chin’
g. xæcʰo
[hæcʰo]
‘sneeze’
h. spæ-rə
[spæʐə]
‘thirsty’ (thirsty-const)
i. læʁju
[læˠju]
‘wave’
j. xsærpæ
[xsærpæ] ‘new’
k. rkærə
[ʂkæʐə]
‘beautiful’
l. scævæ
[scævæ]
‘paddle (n.)’
m. zætər
[zætəʂ]
‘chopsticks’
n. rŋæ-rʁɛ-scɛ [ʐŋærʁɛscɛ] ‘face wash’ (face-wash-instrument.tool)
o. qʰʂæqʰʂæ [qʰʂæqʰʂæ] ‘coarse’
p. nə-ndjælæ [nəndjælæ] ‘to lick’ (dirdown -lick)
q. leskæ-və
[leskævə] ‘to work’ (work-do)
r. kæɕæ
[kæɕæ]
‘morning’
s. ɮæ
[ɮæ]
‘hand’
t. zʁoræ
[zʁoʐæ]
‘yawn’
u. ɲɟəræ
[ɲɟəræ]
‘to run’
v. pəŋæ
[pəŋæ]
‘man’
w. pubæ
[pubæ]
‘Tibetan’

3.3.5 Low central vowel /ɑ/
Stau has a second low vowel, the central /ɑ/. The phonetic difference between /ɑ/
and the low front /æ/ is not large. However, as you can see in the vowel formant chart
in Figure 20 of §4.2, the two vowels have distinct formant distributions, though they are
nearby each other.

71

(67)

/ɑ/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.
r.
s.
t.
u.

atsawatsa [atsawatsa] ‘locust’
adɛ
[adɛ]
‘this’
ara
[aʐa]
‘wine’
ʁzaʁzav [ʁzaʁzav] ‘thank you’
ʁav
[ʁav]
‘needle’
maɣər
[maɣə˞]
‘type of tree’
barlu
[ba˞lu]
‘leaf’
kaɣɮæ
[kaɣɮæ]
‘thin, flat’
pʰjamdə [pʰjamdə] ‘necklace’
kʰambo [kʰambo] ‘bag’
pʰajo
[pʰajo]
‘together’
rata
[ʐata]
‘mill’
ɲtɕʰaχpa [ɲ̥tɕʰaχpa] ‘to fold’
tʰavtɕa
[tʰaftɕa]
‘bad’
kʰawa
[kʰawa]
‘snow’
qavla
[qavla]
‘branch’
læχape [læχape] ‘cabbage’
ʁa
[ʁa]
‘window’
zʁa
[zʁa]
‘10’
rʁa
[rʁa]
‘to become crazy’
ʁərja
[ʁərja]
‘bone’

3.3.6 Mid central unrounded lax vowel /ə/
The mid central unrounded lax vowel /ə/ occurs by far the most often in Stau. It appears in 541 words in my database; the next closest are /æ/ at 289 and /ɛ/ at 261. One reason
for this inordinate frequency of use is that sometimes other vowels lose their character in
unstressed syllables and become centralized. For instance, on two occasions I recorded the
word ‘cover, lid’; once it was pronounced kʰælɛv and another time kʰæləv.
Below are provided representative examples of /ə/.

72

(68)

/ə/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.
r.
s.
t.
u.

mə-gu-rə [məguʐə] ‘to not understand’ (neg-understand-const)
tsə
[tsə]
‘to rot’
rŋə
[ʐŋə]
‘to become green’
tə-skrə-sə [təskrəsə] ‘late’ (dirneut -late-prf)
bjæ
[bjæ]
‘to split (by itself)’
tʰəɣə
[tʰəɣə]
‘rope’
dzəvə
[dzəvə] ‘husband’
pəsɲi
[pəsɲi] ‘today’
ndərjæ [ndərjæ] ‘to sweep’
ndʑəv
[ndʑəp] ‘to suck’
ʁəvdæ
[ʁəvdæ] ‘to nod’
kɛdərə
[kɛdəʐə] ‘early’
ævəsɲi [ævəsɲi] ‘yesterday’
mə
[mə]
‘younger brother’
ɲə
[ɲə]
‘ear’
ɕə
[ɕə]
‘teeth’
mospə
[mospə] ‘eyelashes’
sqevcə [sqefcə] ‘squirrel’
mdɛrə
[mdɛʐə] ‘drum’
pʰrə
[pʰʂə]
‘tangled’
tæmbə
[tæmbə] ‘bottle’

3.3.7 Mid back rounded vowel /o/
Stau has two round vowels, one being the mid back rounded vowel /o/, as demonstrated by the following examples:
(69)

/o/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.

ɣroɣro [ɣʐoɣʐo] ‘dry’
zondo [zondo] ‘horn’
tɕoji [tɕoji] ‘spoon’
xoscæ [xoscæ] ‘hot pepper sauce’
joskə [joskə] ‘garlic’
xopi
[xopi] ‘table’
ʁoɲu [ʁoɲu] ‘back’
vɕoxpæ [fɕoxpæ] ‘wing’
zdomə [zdomə] ‘fog’
mojo [mojo] ‘eyelid’
mɟo-rə [mɟoʐə] ‘fast’ (fast-const)

73

l.
m.
n.
o.
p.
q.
r.
s.
t.
u.
v.
w.
x.

ro
toχtoχ
ʁavro
rko
no
ɕontʰo
mbjo
kəχo
jəʁo
vɛko
ɟɛzo
rwo
spo

[ʐo]
[toχtoχ]
[ʁavʐo]
[ʂko]
[no]
[ɕõntʰo]
[mbjo]
[kəχo]
[jəʁo]
[vɛko]
[ɟɛzo]
[ʐwo]
[spo]

‘to swell’
‘narrow’
‘11’
‘ankle’
‘to smell’
‘fruit’
‘cutting edge’
‘bark’
‘upstairs’
‘pigsty’
‘potato’
‘ice’
‘grassland’

3.3.8 High back rounded vowel /u/
The second round vowel is the high back rounded vowel /u/, shown in the examples
below:
(70)

/u/
a.
b.
c.
d.
e.
f.
g.
h.
i.
j.
k.
l.
m.
n.
o.
p.
q.
r.
s.
t.

mdʑu-rə [mdʑuʐə] ‘hungry’ (hungry-const)
ŋkʰurlu [ŋ̥kʰu˞lu] ‘wheel’
tur-u
[tuʐu]
‘can, able to’ (can-1)
pukʰu
[pukʰu] ‘mosquito’
kudʑən
[kudʑən] ‘satin’
pʰusɛr
[pʰusɛ˞]
‘young man’
mupʰa
[mupʰa] ‘hoof’
sprurə
[spʐuʐə] ‘butter churn’
rkurɟɛv [ʂkurɟɛv] ‘foot stamp’
pʰu
[pʰu]
‘to cover’
səŋun
[səŋun] ‘who’
vku
[fku]
‘to bend something’
kedi-zu-rə [kedizuʐə] ‘to be pregnant’ (child-hold.carry-const)
ɣrərkʰu [ɣʐəʂkʰu] ‘cold’
xavdu
[havdu] ‘now’
mdʐu
[mdʐu]
‘dragon’
mkʰu
[m̥kʰu]
‘cowshed’
ndzu
[ndzu]
‘to sit down’
ŋəzu
[ŋəzu]
‘to howl’
ʁnɛmju [ʁnɛmju] ‘roof’

74

u.
v.
w.
x.

rtu
rbu
rwu
spu

[ʂtu]
[ʐbu]
[ʐwu]
[spu]

‘to cut hair’
‘bee’
‘breath’
‘incense stick’

3.4 Comparison with Huang (1991)
My findings coincide largely with those of Huang (1991). As regards the vowel
inventory, we are in agreement, both analyzing Stau as an eight-vowel language. When it
comes to the consonant inventory, we differ somewhat. Huang includes more phones than
I do. In addition to the phonemes listed in §3.2, Huang has <sh>, <ɬh>, <ç>, <ɕh>, and
<ɦ>. As well, <ʂ>, <f>, and <h> are also included in her inventory of consonants, but no
mention is made of their status as allophones of /r/, /v/, and /x/. Perhaps the reason for the
difference is that Huang’s concern was to be phonetically accurate, rather than phonologically precise. At any rate, despite these differences, we agree in essentials concerning the
phonemic inventory of Stau.

Chapter 4
Acoustic analysis
This chapter provides acoustic analysis of voicing contrasts in stops in §4.1 and
acoustic analysis of vowel qualities in §4.2.

4.1 Acoustic analysis of voicing in stops
As I mentioned above in §3.2.1 and §3.2.2, voiced stops and affricates occur much
less frequently than voiceless ones in Stau. Word-initially they are rare, and word-medially
they often follow a voiced continuant such as a nasal or fricative. This throws suspicion on
the phonemic status of voiced stops and on the validity of a contrast between voiced and
voiceless unaspirated (or “tenuis”) stops and affricates. In order to investigate the voicing
contrasts, I measured the voice onset time (VOT) of voiceless aspirated, voiced, and tenuis
stops1 in word-initial and intervocalic positions, both in isolation and in sentence context.
The acoustic measurements showed a three-way phonemic distinction in onset of voicing
in Stau stops. In the following sections this evidence will be presented, first for word-initial
stops (§4.1.1) and then word-medial intervocalic stops (§4.1.2).
First, a word on methodology. For acoustic analysis of stops I used the program
Speech Analyzer 3.1. To make VOT duration measurements, I used the waveform-spectrogram
display. Word-initial stops were measured from the beginning of the waveform (from the
stop burst from voiceless stops, and from the beginning of prevoicing for voiced stops) until after the first full sine wave signalling the onset of the vowel. Word-medial stops were
1

Only stops, and no affricates, were measured because my dataset doesn’t include enough affricates to
make a reliable sampling. I hypothesized that if a three-way contrast can be found in the stops, the contrast
will also appear with the same acoustic correlates in the affricates. Although I did not systematically measure
all the affricates, I informally checked to see if they follow the same pattern as the stops, and they do.

75

76

measured from (and including) the last full sine wave of the preceding vowel, up to and
including the first full wave of the following vowel.
In my analysis I measured stops both in words in isolation and words in context.
The sentence context used, as mentioned above in §2.4, varied somewhat, but was usually
one of the following two sentences:
(71)

a. tʰɛ
__ ŋə-rə
dem __ cop-const
‘This is __’
b. tʰi
__ kaji ŋə-rə
dem __ pretty cop-const
‘This __ is pretty’

4.1.1 Voicing in word-initial stops
Word-initial voiceless aspirated stops in Stau evince the acoustic characteristics typical to their voicing class, the principal of which is the delay of the onset of voicing. There
are 72 tokens of word-initial voiceless aspirated stops in the dataset. They all show long
lag before the onset of voicing, with an average VOT of 91 ms in citation form. Figure 8
below shows a typical example of an aspirated alveolar stop.

77

Figure 8. Spectrogram of tʰɛ ‘to take off (clothes)’ demonstrating the articulation of a
voiceless aspirated /tʰ/ with 90 ms VOT
In sentence context, word-initial voiceless aspirated stops look very much similar.
They have a long stop closure, followed by a release burst, and then long lag before the
onset of voicing. They have an average VOT of 63 ms, which is 28 ms shorter than that of
word-initial aspirated stops in words in isolation, but it is still substantial. Figure 9 shows
an example of word-initial /tʰ/ in sentence context.

78

Figure 9. Spectrogram of tʰɛ ‘to take off (clothes)’ in sentence context
Since the acoustic realization of voiceless aspirated stops is very straightforward, I
will not go into further detail.
Tenuis stops also show acoustic correlates typical of their category. They have
shorter VOTs than voiceless aspirated stops. On average, voicing starts 30 ms after the
stop release burst when the word is said in isolation, and 35 ms after the release burst when
in context. Figure 10 demonstrates an unaspirated bilabial stop with a 25 ms VOT.

79

Figure 10. Spectrogram of pæbə ‘insect’ demonstrating the articulation of a voiceless
unaspirated /p/ with 25 ms VOT
Identifying acoustic correlates of voiced stops is more puzzling. In words in citation
form, some word-initial voiced stops show negative VOT while others demonstrate lag
comparable to that of a tenuis stop. One might suggest that the latter are tenuis stops, but
it is clear that the two acoustic manifestations are both possible for phonologically voiced
stops, because some voiced stops in my dataset are produced in one token with voicing
during the stop closure and in another with short lag.
A good example of this in my dataset is the word bæɴɢe. I recorded it on two
separate occasions, once glossed as ‘fly’ and another time as ‘spider’. The stop in ‘fly’ was
produced with 10 ms VOT:

80

Figure 11. Spectrogram of bæɴɢe demonstrating the articulation of a voiced /b/ with 10
ms VOT
The same word on another occasion (when I recorded it glossed as ‘spider’) had
voicing during the stop closure, as seen in the waveform and spectrogram of Figure 12
before the release burst. The reader will notice that voicing shuts off during the closure
before the stop is released; this does not call into question the stop’s voicedness. It is simply
that voicing is difficult to maintain.

81

Figure 12. Spectrogram of bæɴɢe demonstrating the articulation of a voiced /b/ with -120
ms VOT
However, unlike /b/ in bæɴɢe, some word-initial stops with short VOT lag that I
take to be underlyingly voiced never occur in citation-form words in my data with closureinternal voicing. Because it is not always possible to tell the voicing of a word-initial stop
when no words precede it, another method of differentiating phonologically voiced stops
with short lag from true voiceless unaspirated stops is needed. The solution to this problem is to look at the voicing of these stops sentence-medially, particularly in intervocalic
position, where they show closure-internal voicing, whether complete or partial.
For instance, when ɟirə ‘in’ is spoken in isolation its word-initial stop has a positive
VOT of 35 ms, as the spectrogram below demonstrates:

82

Figure 13. Spectrogram of ɟirə‘in’ in isolation; /ɟ/ shows a VOT of 35 ms
In a sentence, though, the phonologically voiced /ɟ/ is also phonetically voiced.
However, the voicing is not sustained throughout the full closure. I call this “partial negative VOT”: when the voicing of a voiced stop cuts out partway through a stop’s closure.
Figure 14 shows ɟirə in sentence context, following a word that ends in a vowel. You can
see in the spectrogram that the voicing bar of /ɟ/ cuts out after 60 ms.

83

Figure 14. Spectrogram of ɟirə‘in’ in sentence context; /ɟ/ shows closure-internal voicing.
A large portion of the voiced stops of the Stau speaker I worked with has only partial
voicing word- and sentence-medially. This type of voicing is still considered negative VOT,
and is still measured from the stop burst leftward to the end of the preceding segment.
Incomplete voicing is noted in (72) below with an asterisk (*), and the length of the partial
voicing bar is also given. On average, voiced stops with short lag which are produced at
the beginning of an isolated word have a positive VOT of 21 ms; if a negative VOT, one of
-149 ms. In sentence context they have an average -122 ms VOT.

84

Voice onset times of word-initial voiced stops

(72)

Word
b

d

ɟ

Gloss

bæɴɢe ‘fly (n.)’
bæɴɢe ‘spider’
ba˞la
‘leaf’
ba˞lu
‘leaf’
bəlætəsɲi‘every day’
bətsʰɛl ‘water
spinach’
bəti
‘cheek’
doləmæ ‘?’
doŋbə ‘stem’
dordʑe ‘?’
ɟɛuræ
‘facial hair’
ɟɛzo
‘potato’
ɟɛzoftəpɛ ‘to harvest
potatoes’
ɟirə
‘in’

Average

VOT isolation (ms)

VOT con- Partial voictext (ms)
ing bar (ms)

10
-120
-240
-60
-155
15

-120
-105*
-110
-115
-120
-165*

-200
-170
20
-130
-140
-130
25

-150
-90
-110
-125
-145
-115

35
21; -149

-120*
-122

80

135

60

4.1.2 Voicing in intervocalic stops
As for word-medial intervocalic voiced stops, the problem was less determining
that two types of voicing were allowed for one phoneme, and more determining what is a
salient length of voicing bar to indicate the presence of a voiced stop. Voiceless unaspirated
stops usually do have a short voicing bar that persists into the stop closure after the vowel
formants disappear. This is exemplified in Figure 15, where word-medial /t/ has a voicing
bar that continues 45 ms into its stop closure.

85

Figure 15. Spectrogram of word-medial voiceless unaspirated /t/ in bəti ‘cheek’
When one looks at the word-medial unaspirated alveolar stop in bəti in sentence
context, one sees that the length of the voicing bar does not change or go much beyond 40
ms. The voicing bar of /t/ in context is still 40 ms:

86

Figure 16. Spectrogram of word-medial voiceless unaspirated /t/ in bəti ‘cheek’ in context
On the other hand, although a phonologically voiced word-medial stop may have a
short bar of closure-internal voicing in a citation-form word that makes the interpretation
of whether or not it is voiced ambiguous, in context it will have clear negative VOT. For
instance, the bilabial stop in nɟaba ‘mud’ has a partial voicing bar of only 40 ms out of a
negative VOT of -175, as seen in Figure 17, but in sentence context it is fully voiced, as
Figure 18 displays.

87

Figure 17. Spectrogram of word-medial voiced /b/ in ɲɟaba ‘mud’ in isolation

Figure 18. Spectrogram of word-medial voiced /b/ in ɲɟaba ‘mud’ in context

88

In general a word-medial voiced stop can be identified by a voicing bar of at least 45
ms, and more frequently of 60-100 ms. As with word-initial voiced stops, voicing becomes
more obvious (i.e. less frequently partially voiced) when in the context of a sentence. This
might be surprising and unexpected since the immediate surrounding context remains the
same. It is not clear why putting words with word-medial voiced stops in a sentence would
trigger this type of change. One idea why this difference in phonetic voicing exists between
stops in isolated words and those in context is that in a sentence the faster rate of speech
allows less time for voicing to cut out. It is worth noting that this effect is very robust; it
happens in virtually all cases. I have ample examples, surprising though it may be.
Table (73) below lists all the words with intervocalic voiced stops, and gives the
VOTs of each voiced stop in isolation and in sentence context. If the stop is partially voiced
either in isolation or context, it is marked with an asterix and the length of the voicing bar
is given in the next column to the right. On average, word-medial voiced stops have a VOT
of -142 ms in isolation and -103 ms in context.
(73)

Voice onset times of word-medial voiced stops
Word

b

Gloss

æbɛr
‘one step’
kʰæbɛrəɲcʰəre‘phone
shop’
nəbɛr
‘second
step’
nɟaba
‘mud’
pæbə
‘insect’
pæbəle
‘butterfly’
pæbəsnærŋæ ‘insects’
pubæ
‘Tibetan’
rŋəbo
‘roasted
barley’
rʁəbə
‘thin’

VOT
isolation
(ms)

Partial
voicing
bar
(ms)

VOT
context
(ms)

-140*
-175*

100
70

-140
-110

-155*

100

-120

-175*
-145*
-115
-160*
-125*
-180*

40
105
50
80
100

-95
-105
-90
-115
-100
-115

-250*

50

-105

Partial
voicing
bar
(ms)

89

d

adɛ
ædæ
ækʰivədæ
æʑuvədæ
ɕukede
kede
kedizurə
kɛdərə
kɛdi
kʰɛdɛr
koŋkede
ludɛ
qʰəzikede
sqədi
tɕedə
tɕədə
tsʰædəm

ɟ

vədæ
naɟɛcʰɛrə
smənvəgəɟi
ʑæʁɟæ

g

ægəjo
kəgəmæ
kʰɛgɛ
məgondʑə
mugurə
tɕəgə
tsəgə
rəgi
ʑigə
zugu

‘this’
‘older sister’
‘paternal
uncle’
‘maternal
uncle’
‘weak’
‘small’
‘to be pregnant’
‘early’
‘child’
‘scarf’
‘cheap’
‘which’
‘small
bowl’
‘to ring’
‘time’
‘book’
‘pitcher,
thermos’
‘wife’
‘many
trees’
‘to make
medicine’
‘lame person’
‘to grow
up’
‘naked’
‘after’
‘not understand’
‘not understand’
‘something’
‘clothing’
‘to clothe’
‘around’
‘to take’

-160*
-135

140

-140*

75

-100

-115*

60

-105*

80

-125
-160*
-130

50

-90
-100*
-95

80

-100
-135*
-205*
-135*
-120
-145*
-165*
-110
-165*
-155*

55
120
105
120
130

-120
-90*

-100
-75
-100
-100
-115*
-100

65

95

70
50

-125
-75
-90*
-90*

70
45

-130*
-170*

115
50

-70
-160*

70

-170*

70

-140*

55

-170*

60

-125

-100

-105

-75*
-125
-95*

45

-110*

50

-100*
-135*
-145*
-125*
-135*

50
85
65
65
85

40

-60*
-75
-80*

-70
-85*
-160*
-90
-135*

40
35

75
70
60

90

Average

-142

77

-103

65

4.2 Acoustic analysis of the vowels
Because vowels, even those transcribed with the same symbol, can vary so much in
their pronunciation from language to language, in this section I analyze the vowels acoustically to provide a more scientific description of them. Ladefoged (2003) recommends as
best practice for an accurate and representative vowel plot that 4-10 speakers of both sexes
and a variety of ages should be recorded, each articulating 5-10 tokens of each vowel. Unfortunately, since my fieldwork was limited by time and location, the acoustic descriptions
here are based on only one speaker. Thus, my analysis of the vowel identities should be
treated as an approximation, and not a representative picture of the vowel production of the
whole speech community.
F1 and F2 measurements were made on recordings sampled at a rate of 48 kHz. Vowels were measured in word-final position to minimize consonant interference. In accordance
with Ladefoged (2003)’s recommendations, vowels with nasals preceding them were excluded from the sample. I made measurements manually in Speech Analyzer 3.1, using the
waveform-spectrogram-spectrum screen configuration, avoiding consonant transitions and
with an eye to steady formant states. Then the formant values were run through the software
program FPlot,2 in order to produce vowel formant plots.
The measurements produced the following F1 and F2 averages:
Average formant values
i
e
ɛ
æ
ɑ
2

Available at casali.canil.ca.

F1

F2

Tokens

336
475
537
804
859

2267
2151
1957
1737
1444

38
30
40
40
35

91

Average formant values
F1
ə
o
u

F2

556 1501
471 1048
336 1181

Tokens
40
39
41

Plotting the average formant values produces the following representation of the
vowel characters and relationships. F1 is plotted on the y-axis against F2 on the x-axis.

Figure 19. Average vowel formant plot
The vowel chart below plots all the individual tokens that were measured, and shows
the variation within the F1 and F2 of each vowel.

92

Figure 20. Formant plot of all tokens
Most of the vowels show consistency in their formant values. /i/, /æ/, and /ɑ/ in
particular are quite distinct in their distribution. /e/ and /ɛ/, whose contrast one might be
suspicious of, do overlap, but /ɛ/ does have a wider distribution into lower F2 and higher F1
values.
/u/ varies widely in terms of its F2 , moving into F2 territory that is considered more
typical of central vowels. However, most of /u/ tokens do not vary significantly in their
aural quality. There is only one vowel (not included in the analysis as it is not word-final)
whose quality I am uncertain about, and that is the back vowel in the word ludɛ ‘which’. It

93

has an F1 of 523Hz and an F2 of 1449Hz. Aurally it sounds like [ʊ], something between /u/
and /o/. However, since it is the only such vowel in my corpus, I am making no conclusions
about it and have transcribed it with the high back vowel /u/.

Chapter 5
Syllable structure
This chapter presents the syllable types that Stau uses (§5.1), and then moves on to
syllabification in Stau (§5.2).

5.1 Syllable canon
Maddieson (2013c) points out that a “significant though not strong” correlation exists “between small consonant inventories and simple syllable structure and large consonant
inventories and complex syllable structures.” Maddieson can count Stau as another language
to bolster the validity of this typological tendency: in addition to having a 42-consonant inventory, the language has a complex syllable canon. The Stau syllable consists of: an initial
consonant (Ci ), which can be preceded by a preinitial consonant (Cp ) and followed by a medial consonant (Cm ); an obligatory vowel nucleus; and an optional final consonant (Cf ). The
syllable canon can be represented as follows:
Table 1. Stau syllable canon
(Cp )

Ci

(Cm ) V

(Cf )

Permit me a brief aside to discuss terminology: the terms initial, preinitial, medial,
and final may be familiar to some readers but new to others. In Sino-Tibetan linguistics,
particularly in the field of historical phonology, these terms are used to refer to the various
slots in a syllable. The reason specific terms are used rather than C1 , C2 , etc., is that different syllable positions license specific sets of consonants. Initial position licenses the largest
number of consonants; in fact, it allows most of the phonemes in the consonant inventory.
94

95

Preinitial, medial, and final position each license a restricted subset of the consonants allowed in initial position. In English, for instance, all consonant phonemes except ŋ are
allowed in initial position, while medial position licenses the liquids r and l and preinitial
position only s. Although Stau allows much freer combination of consonant phonemes, the
types of phonemes the respective positions license are similar: in Stau, medial position licenses only liquids and glides, while one main phonemic class allowed in preinitial position
is that of fricative. Much more can be said on Stau phonotactic restrictions, and we will return to these in §6. Still, the descriptive utility of these terms for cross-linguistic typology
is evident in the way that Stau’s restrictions mirror and expand on those of English. Now
let us return to the syllable canon.
5.1.1 Syllable canon in genetic and areal context
That Stau has a complex syllable canon is not surprising when we consider its
areal context and genetic inheritance. Proto-Tibeto-Burman (PTB) also had a large syllable
canon. Below is the PTB syllable canon as reconstructed by Matisoff (2003:12):
Table 2. Proto-Tibeto-Burman syllable canon
(P2 ) (P1 ) Ci

[T]
(G) V
(Cf ) (s)

We can see strong similarities between this syllable canon and the Stau syllable
canon in Table 1. The PTB syllable obligatorily has an initial consonant (Ci ) and vowel
(V), just as Stau does. As in Stau, the initial consonant can be preceded by consonantal
“prefixes” (P1, 2 ; nearly equivalent to “preinitials”) and can be followed by a liquid or glide
(G) (namely, a medial), and the nucleus can be followed by a final consonant (Cf ) (Matisoff
2003:11).

96

There are differences between the two syllable canons. The PTB syllable is more
complex: It allows two prefixes before the initial consonant, whereas Stau permits only
preinitial. PTB allows a second suffix, *-s; Stau allows only one final consonant. PTB
may also have had tone (whether or not it did is still an open question) (Matisoff 2003:11).
Despite these differences, the similarity between the parent and daughter language can be
clearly seen.
The complexity of Stau’s syllable canon can also be seen as part of an areal phenomenon. It is an areal tendency of East and Southeast Asian languages to have “bulging
monosyllables,” as Matisoff (1990) writes. He explains that SE Asian languages are “overwhelmingly” monosyllabic, and tend to have complex syllable canons (1990:543). Matisoff
(1990:547) argues that bulging monosyllabicity is a persistent diachronic tendency because,
over the course of time, SE Asian languages “oscillate cyclically” through a set of syllable structures―complex monosyllables, (consonantally) simple monosyllables, disyllabic
compounds, and sesquisyllables―a cycle that always returns to monosyllabicity.
In our present ignorance of Rgyalrongic historical phonology, we cannot be sure
how Stau came to be in its current syllabic state―whether it essentially conserves PTB’s
complex monosyllabicity or is cycling back to this type of syllable structure. I would speculate the former is true. One reason I favour this hypothesis is that Shangzhai, a language
closely related to Stau, retains an old PTB causative prefix *s- as a semi-productive morpheme (J. Sun 2007, Matisoff 2003:89). There is some evidence that Stau retains the same
prefix, as (74) and (75) show. In (74), we see that when the prefix s- is added to the verb
ŋkʰərvæ ‘to turn, circle’, the verb becomes causativized. Instead of the verb’s subject doing
the turning as in ŋkʰərvæ, the subject is causing an object to turn.
(74)

ŋkʰərvæ ‘to turn, circle’
s-kərvæ ‘to turn (prayer wheel)’

97

Example (75) demonstrates the same causativization. The verb pjæ ‘to split’ has the
sense that what is splitting is doing so by itself, but when s- is prefixed to it, its meaning
changes to include an agent who is causing the splitting (as when a person is chopping
wood).
(75)

pjæ ‘to split (by itself)’
s-pjæ ‘to split (wood)’
There are other verbs in the database that begin with s and might be examples of

causativized verbs, but no corresponding prefix-less verbs were found to provide proof.1
J. Sun (2007) presents an analysis that opposes my theory. Sun notes that it is characteristic of Rgyalrongic languages to make causatives with the prefix s(ə)-. In Shangzhai,
the causative prefix appears “consistently” as the non-syllabic s-, while Caodeng Rgyalrong and Guanyinqiao Lavrung (Khroskyabs) use “both syllabic (sə-) and non-syllabic allomorphs (s-)” (2007:226). Based on these languages (Sun does not mention including any
other Rgyalrongic varieties in his sampling), he hypothesizes that Proto-Rgyalrongic had a
syllabic causative prefix*sə-. If his analysis is correct, it is likely that Stau has not inherited the complexity of its syllable canon from PTB, but is returning to monosyllabicity from
sesquisyllabicity. The causative prefix would have expanded in the timescale between PTB
and Proto-Rgyalrongic from *s- to *sə-, and then contracted again to a non-syllabic prefix
in contemporary Stau and Shangzhai.
My interpretation of the synchronic data differs. In his hypothesis, Sun does not
make reference to PTB and its causative prefix *s-, which is clearly the source from which
Proto-Rgyalrongic derives its morphological causative. When one takes PTB into account,
my analysis becomes more probable: Proto-Rgyalrongic retained PTB’s non-syllabic prefix and passed it on to its daughter languages. Some Rgyalrongic daughter languages, such
1

E.g. stʰvæ ‘to press down’, stʰi ‘to hang’, scor ‘to support, prop up’.

98

as Caodeng and Guanyinqiao, began inserting a schwa after the prefix to avoid undesirable consonant clusters, creating a second allomorph of the causative morpheme in these
varieties. This interpretation of the data suggests that Stau is not cycling back to the monosyllable but has retained PTB’s monosyllabicity.
Having considered the Stau syllable canon in areal and genetic context, let us now
look at the specific syllable types comprised within the syllable canon.
5.1.2 Syllable types
The syllable canon in Table 1 licenses six syllable types: CV, CCV, CCCV, CVC,
CCVC, and CCCVC. These six types can be generalized into two primary syllable shapes.
The first three can be characterized as syllables formed by an onset and nucleus, the latter
three as onset and nucleus followed by a coda. Examples of each are provided below:
(76)

CV
a. tsʰə
‘salt’
b. he.ʑi ‘how many’
c. ʑə.lə.ze ‘son, boy’

(77)

CCV
a. sqʰi
‘sister’
b. mdʑu.dʑu ‘late afternoon’
c. spə.spæ ‘body hair’

(78)

CCCV
a. ɴɢwɛ
‘five’
b. ndji.ndji ‘red’
c. spru.rə ‘butter churn’

99

(79)

CVC
a. ʁav
‘needle’
b. ɬɛv.sə ‘lightning’
c. tɕʰə.kəv ‘watermelon’

(80)

CCVC
a. ɲɟɛm
‘wall’
b. zjær.kʰu
‘to hurt emotionally’
c. snəm.rʁɛ.scɛ ‘dish detergent’

(81)

CCCVC
a. ndjɛv
‘to sleep’
b. spjoŋ.kʰə ‘wolf’
The six syllable types above appear in Stau roots. In affixes, an additional syllable

type is allowed―a bare nucleus, as (82) demonstrates.
(82)

V
a. æ.pæ
‘father’
b. a.dɛ
‘this’
c. a.tsa.wa.tsa ‘locust’
Unlike the previous six syllable types, the V syllable does not occur in roots, it only

appears in prefixes. The syllable type is further restricted in that it only licenses low vowels,
either /æ/ or /ɑ/. In addition, V syllables are relatively rare. The vowel /ɑ/ only occurs as a
V syllable three times in my dataset, and one instance, (83a), is a Tibetan loanword:
(83)

a. a.ra
‘wine’
b. a.dɛ
‘this’
c. a.tsa.wa.tsa ‘locust’
Most of the words in which /æ/ occurs as a V syllable fall into three identifiable

categories: kinship terms, quantifier nouns and classifiers, and quantifiers and grammatical

100

words. Based on this and other evidence discussed below, it is possible to hypothesize that
all V syllables are historically prefixes.
Several examples of kinship terms beginning with /æ/ are shown in (84):
(84)

Kinship terms
a. æ.mə ‘mother’
b. æ.ti
‘older brother’
c. æ.dæ ‘older sister’
d. æ.ŋæ.ze ‘baby’
When we look at the literature, we find a historical reason for the plethora of kinship

terms that start with /æ/. Matisoff (2003:104-5) reveals that PTB had a prefix *a-, one
semantic function of which is to mark kinship terms. Chirkova (2012:143) mentions that
the use of a- as a kinship prefix is one of the putative characteristics of the Qiangic subgroup,
and that, for the languages of the Ethnic Corridor, the a- prefix marks older kin (example
(84d) being an exception here; perhaps because it has the diminutive suffix -ze).
The æ- that occurs with quantifier nouns and classifiers is a numeral prefix for ‘one’.
In Stau, numbers used for counting (ro ‘one’, ɣne ‘two’) cannot be used with nouns; instead
numeral prefixes must be affixed to the nouns.2
(85)

Quantifier nouns and classifiers
a. æ-bɛr ‘a step; first step’
b. æ-ndʐə ‘a time/instance; first time/instance’
c. æ-sɬə ‘one month’
d. æ-sɲi ‘one day’
æ- can also be prefixed to some action words as a classifier, as in (86). For instance,

when æ- is prefixed to the verb ndə ‘to stab’, the resulting word means ‘a stab’.
2

I am beholden to Jacques (2014) for making clear to me what was going on here, as he notes the same
requirement in Japhug Rgyalrong.
Two other numeral prefixes that occur in my data are nə- ‘two’ and su- ‘three’.

101

(86)

a. æ-ndə ‘a stab’
b. æ-xtɕæ ‘a cut’
The words in (87) I have tentatively grouped together as being of similar character,

quantifiers and grammatical words. If these words do belong to one class, the function or
origin of the prefix they begin with is not clear.
(87)

Quantifiers and grammatical words
a. æ.sɛ ‘full’
b. æ.tɕɛ ‘together’
c. æ.tɕʰə ‘what’
d. æ.tɕʰə ‘with’
e. æ.tsʰe ‘a little’
f. æ.χe ‘a little’
The three words beginning with a V syllable that don’t fit into any pattern of the

above patterns comprise one verb (88a) and two nouns (88b-c).
(88)

a. æ.gə.jo
‘to grow up’
b. æ.ɕəm
‘corn’
c. æ.dʑæ.pæ.dʑæ ‘sandals’
Moving on from licit syllable types, VV sequences seem to be illicit in Stau. Yet,

two such words do occur in the data I gathered:
(89)

a. ɟɛ.u.ræ ‘beard’
b. dʑu.æ ‘to swim’
Perhaps these exceptions are borrowings, or perhaps they are morphologically com-

plex, since VV sequences are permitted in phrases, as the possessive noun phrases below
demonstrate.

102

(90)

a. æjæ-i-dzəvə
maternal.aunt-gen-husband
‘aunt’s husband’
b. æʑu-i-vədæ
maternal.uncle-gen-wife
‘uncle’s wife’

5.2 Syllabification
As for syllabification, Stau follows the principle that consonants and consonant clusters attested at word edges are allowed word-medially. In many cases this means that the
Sonority Sequencing Principle (SSP) is violated (Clements 1990). For instance, according
to the SSP, the word təskrəsə ‘late’ should be syllabified as it is in (91a).
(91)

a. *təs.krə.sə
b. tə.skrə.sə
However, (91b) demonstrates the best syllabification, despite the fact that it violates

the SSP, because it follows the principle of attested word-edge clusters. The cluster skr
is attested at the beginning of the word skʰro ‘ant’ (discounting the minor difference of
aspiration on the velar stop in ‘ant’), while /s/ is never found as a word-edge coda in Stau.
Based on this evidence, (91b) is its correct syllabification.
On the other hand, where a word-edge onset consonant cluster and a word-edge coda
are both attested, the SSP comes in to decide between the two possible syllabifications. For
example, ɕovərgɛm ‘cardboard box’ is a compound of ɕovə ‘paper’ and rgɛm ‘box’. Despite
the morpheme break, the word is syllabified as ɕo.vər.gɛm, because ər is an attested wordedge rhyme, e.g. ʁərʁər ‘round, circular’.

Chapter 6
Phonotactics
In this chapter, I discuss phonotactic restrictions in Stau. The chapter is divided into
two sections, beginning with §6.1 which looks at phonotactic constraints in syllable onsets.
Section 6.1 focuses particularly on restrictions within consonant clusters. Section 6.2 deals
with phonotactic constraints within the rhyme, i.e. both the nucleus and coda.

6.1 Onsets
Because of its large consonant inventory and large number of consonant clusters,
Stau consequently has a large number of permissible onsets. This section deals with the
phonotactic restrictions that apply to onsets. In §6.1.1, I will describe and list the type of
consonants each position of the Stau onset allows. In §§6.1.2-6.1.5, we will look in detail at
the consonant cluster combinations that are licit in Stau, beginning with NCi , then moving
to other Cp Ci , Ci Cm , and finally Cp Ci Cm clusters.
6.1.1 Positional restrictions
All Stau consonantal phonemes are permissible as initials, apart from /ɴ/ and /ɢ/
which only occur together in a cluster. When consonants combine into clusters, the sounds
permissible in onsets are curtailed. Initial position in clusters remains the most unrestricted
position in the syllable, followed by preinitial position. Medial position is the most restricted position in the onset. Let us look at the phonotactics of each type of consonant
cluster individually, starting with Cp Ci , followed by Ci Cm , and then Cp Ci Cm .

103

104

Cp Ci clusters are the most common consonant clusters in Stau. The Ci position licenses voiced and voiceless stops, fricatives, affricates, and nasals. In preinitial position,
fricatives, the rhotic, and nasals are allowed.
In Ci Cm clusters, initial consonants can be stops, fricatives, the rhotic, and /m/. In
medial position, the glides and rhotic are most common, but /v/ and /l/ are also attested.
Within Cp Ci Cm clusters, each position is more restricted than it is in two-consonant
clusters. Initial position only licenses stops, and this does not include palatal stops. Preinitial position allows the rhotic and nasals, as the Cp Ci preinitial position does. However,
from among the fricatives only alveolar fricatives are allowed, whereas in Cp Ci clusters
fricatives from all five places of articulation are licensed. In medial position, most of the
same consonants are attested as are in Ci Cm clusters―glides, rhotic, labial fricative. The lateral approximant is not attested, but this could be an accidental gap. Lateral approximants
are relatively uncommon in medial position of Ci Cm clusters, and three-consonant clusters
are also fairly uncommon. The absence of the lateral approximant in this context could be
a statistical effect.
Table 3 details the specific phonemes that occur in each position of these clusters.
Table 3. Licit consonants by position and cluster type
Cluster type Position

Attested consonants

Cp Ci

Cp
Ci

Ci Cm

Ci
Cm
Cp
Ci
Cm

vszxɣʁrmnŋɴ
pʰ p d tʰ t d cʰ c ɟ kʰ k g qʰ q tsʰ ts dz tɕʰ tɕ dʑ dʐ
vszɬɮɕʑxɣʁrmnɲŋ
pʰ p b tʰ kʰ k qʰ q v s z ɕ ʑ ɣ χ ʁ r m
vlrjw
szrmnŋɴ
pʰ p b tʰ d kʰ k g q ɢ
vrjw

Cp Ci Cm

These are the sounds attested in each position of the three cluster types, but not all
the combinatorial possibilities of these sounds are permitted in Stau.

105

In the subsections that follow, I will describe the particular restrictions that govern
the combination of the phonemes listed for each cluster type in Table 3. I will list with
examples the attested consonant co-occurrences, and in addition discuss the restrictions
that they display.
6.1.2 [Nasal + initial] clusters
The chart below gives all the [nasal + initial] combinations found in Stau. The
five nasals are listed on the vertical axis, and initial consonants on the horizontal. At the
intersection of the row and column that represents an attested [nasal + initial] combination
is entered the number of words in the database in which said cluster occurs.
Table 4. Nasal + consonant co-occurrence chart
pʰ b
m 8
n
ɲ
ŋ
ɴ

21

tʰ d
1
5

tsʰ

6
4
16 3

dz

dʐ

tɕʰ

dʑ

1
7

3
7

5
5

7
16

cʰ

ɟ

kʰ

1 5
8

g

qʰ

q

ɢ

1

5

1

6
17

13

1

Table 4 reveals a number of things. First, preinitial nasals are only found in clusters
with stops and affricates. Second, the stops and affricates are either voiced or voiceless aspirated,1 which is consistent with cross-linguistic tendencies. Obstruents tend to be voiced
following nasals in languages across the world (Hayes & Stivers 2000:1). Languages that
against the grain have NC̥ clusters must “preserve the [voicing] contrast [between the nasal
and obstruent] in spite of the pressure to obliterate it” (Hayes & Stivers 2000:31). Stau preserves the contrast by aspirating the voiceless consonant, thus making the contrast between
[nasal + voiced obstruent] and [nasal + voiceless obstruent] perceptually clearer. In addition
1
This pattern does not hold in Huang (1991)’s description of Gexi Stau. She finds several voiceless
unaspirated stops and affricates in NCi clusters: <nt>, <nts>, <ntɕ>, <ncç> (i.e. /ɲc/), and <nq> (i.e. /ɴq/).

106

to the perceptual benefits of aspirating the voiceless consonant, there is a possible articulatory reason that voiceless stops tend to be aspirated after nasal consonants: Stops naturally
tend to voice in the post-nasal environment, and abduction of the vocal folds (i.e. moving
them apart) is a primary mechanism of “cessation of voicing” (Hayes & Stivers 2000:3).
At the same time, vocal fold abduction is also a primary mechanism of aspiration. Thus,
the abduction deployed to prevent a post-nasal consonant from becoming voiced might also
naturally lead to aspiration.
A third thing to notice is that, whereas many languages that allow [nasal + initial]
clusters stipulate that the nasal must be homorganic with the initial (i.e. having the same
place of articulation), Stau allows other combinations than homorganic ones, at least for
clusters in which /m/ is the preinitial. Alveolar, palatal, velar, and uvular nasal preinitials
all only occur in clusters with consonants from the same place of articulation.
Below, (92) demonstrates each of the [m + initial] clusters with an example from
the database.
(92)

Attested [m + initial] clusters
Cluster Example

Gloss

mpʰ
mb
mtʰ
md
mtsʰ
mdz
mdʐ
mtɕʰ
mdʑ
mɟ
mkʰ
mg

to card (wool)
cushion
high
drum
lake
polite
dragon
teapot
seed
fast
pipe
wall (that one dries barley against)

mpʰi
mbɛr
kɛmtʰu
mdɛrə
mtsʰu
mdzɛmdze
mdʐu
mtɕʰurdzæ
mdʑəsnæ
mɟo-rə
mkʰərjɛ
mgrə

107

As I just mentioned, apart from /m/, the nasals only occur preinitially with homorganic stops and affricates. Some phonologists would for this reason analyze the homorganic clusters as prenasalized consonants. For instance, Jacques (2004) and J. Sun (2004)
interpret homorganic nasal clusters in Japhug and Showu Rgyalrong (respectively) as single prenasalized segments rather than as sequences. In Jacques and Sun’s analyses, the
prenasalized series forms its own natural class in the segmental inventory. The principle
of Ockham’s razor has been cited to support this type of interpretation. Analyzing these
sounds as a single manner of articulation rather than many clusters could be considered
simpler and more elegant.
However, Ockham’s razor can also be used to argue against the prenasalization analysis. While the analysis eliminates some consonant clusters, it does not simplify the syllable
canon. The presence of non-homorganic [m + initial] clusters requires that we still recognize the [nasal + initial] onset type. In addition, the analysis adds sixteen extra consonants
to an already full consonant inventory. For these reasons, I have decided to treat the sounds
as phonological sequences.
The table in (93) provides an example of each of the attested [n + initial] clusters.
(93)

Attested [n + initial] clusters
Cluster Example

Gloss

ntʰ
nd
ntsʰ
ndz
ndʐ
ntɕʰ
ndʑ

to grind
girdle for coat
between, around
to hide something
same
to feel itchy
woollen cloth

ntʰətæ
ndərci
ntsʰɛm
ndzə
ndʐændʐæ
zæntɕʰæ
ndʑætsə

As a preinitial /ɲ/ occurs before voiceless aspirated and voiced palatal stops:

108

(94)

Attested [ɲ + initial] clusters
Cluster Example

Gloss

ɲcʰ
ɲɟ

to hit
mud

ɲcʰə
ɲɟaba

The velar nasal appears as a preinital before voiceless aspirated and voiced velar
stops, as the examples in (95) demonstrate.
(95)

Attested [ŋ + initial] clusters
Cluster Example

Gloss

ŋkʰ
ŋg

wheel
food stuff

ŋkʰurlu
ŋgələ

The uvular nasal is attested in one Cp Ci cluster, before the voiced uvular stop, as
(96) shows.
(96)

Attested [ɴ + initial] clusters
Cluster Example

Gloss

ɴɢ

fly, spider

bæɴɢe

6.1.3 Other [preinitial + initial] clusters
Unlike NCi clusters, other Cp Ci co-occurrences show less systematicity concerning
which preinitials will co-occur with which initials. For this reason I have decided not to
include a Cp Ci co-occurrence chart like the NCi one provided by Table 4―such a chart would
lack the pattern that Table 4 shows, making it harder to read and interpret. In addition, so
many initials are involved in Cp Ci clusters that a co-occurrence chart would be unreadably
wide.

109

In total (excluding NCi clusters), 101 Cp Ci cluster combinations are attested in the
data: twenty-three with /v/ as preinitial, thirteen with /s/, eight with /z/, eleven with /x/,
seven with /ɣ/, three with /χ/, nine with /ʁ/, and twenty-seven with /r/.
One phonotactic restriction that applies to almost all Cp Ci clusters is that the two
consonants must sharing the same voicing. For most preinitials this means they do not cooccur with initials of different voicing: e.g. /x/ only occurs with voiceless consonants, /ɣ/
only with voiced. For /v/ and /r/, which do not have voiceless phonemic counterparts, this
means that when they precede a voiceless consonant, they assimilate in voicing and are
realized respectively as [f] and [ʂ]. There is an exception to this rule: /s/ can be a preinitial
before nasals without assimilating in voicing.
The labiodental fricative /v/ can occur preinitially before aspirated, tenuis, and voiced
stops and affricates, and alveolar, lateral, and palatal fricatives:
(97)

Attested [v + initial] clusters
Cluster Example

Gloss

vtʰ
vt
vd
vcʰ
vc
vɟ
vk
vg
vq
vtsʰ
vts
vdz
vdʐ
vtɕʰ
vtɕ
vdʑ
vs
vz
vɬ

to smoke (a cigarette)
to harvest
to nod
to weigh
rat, mouse
saliva
to eat one’s fill
to cross (a bridge)
sky
to take out of water
to cut
human
to roll up
sixteen
rack for hanging things on
friend
to spin (wool)
to scratch
ashes

vtʰi [ftʰi]
vtəpɛ [ftəpɛ]
ʁəvdæ
vcʰə [fcʰə]
vcə [fcə]
vɟə
vkə [fkə]
vge
vqo [fqo]
vtsʰu [ftsʰu]
nə-vtso [nəftso]
vdzi
vdʐər
ʁavtɕʰo [ʁaftɕʰo]
vtɕækʰæzŋore [ftɕækʰæzŋore]
vdʑəvdʑə
vsu [fsu]
vzɛvzə
vɬɛ [fɬɛ]

110

vɮ
vɕ
vʑ

vɮæzæ
vɕoxpæ [fɕoxpæ]
vʑo

sleeve
wing
to plane

Example (97) shows twenty-two clusters. There is one additional cluster, [fɲ], that
occurs in two words, both compounds created with the stem ɲu ‘shoot, sprout’:
(98)

a. kov-ɲu [kofɲu] ‘garlic shoot’ (garlic-shoot)
b. tɕɛlɛv-ɲu [tɕɛləfɲu] ‘bean sprout’ (bean-shoot)
The cluster does not follow the rule that [f] only occurs before voiceless consonants.

Its departure from the rule may lead one to suspect that the voicing difference between the
two sounds is only perceptual, but the labiodental fricative is indeed voiceless in spectrograms. I am not sure how to account for the appearance of the voiceless [f] here. In Stau
phonology /v/ is phonemic and [f] is its allophone (restricted to preceding a voiceless consonant), and in (98a) and (98b) the labial fricative precedes a voiced nasal. Based on these
premises, we would expect the labial fricative to be realized as [v].
In another compound (99), which uses tɕɛlɛv ‘bean’ as a stem like (98b) does, the
labial fricative is realized as [v] when followed by the voiced palato-alveolar affricate dʑ:
(99)

tɕɛlɛv-dʑoŋ [tɕɛlɛvdʑõ] ‘mung bean sprout’ (bean-?)
One possible explanation for the difference between the labial fricatives in (98) and

(99) is that nasals do not cause voice assimilation. A supporting piece of evidence is that
nasals occur after the voiceless alveolar fricative /s/ without causing it to assimilate for
voicing (see (100) below for examples). However, this would presuppose that the words
for ‘garlic’ and ‘bean’ end with an underlyingly voiceless labial fricative, which contradicts
Stau phonology. Perhaps these words are Tibetan borrowings; if not, I currently have no
satisfactory explanation for their departure from Stau phonological rules.

111

In her more extensive database on Gexi Stau, Huang (1991:4) finds three additional
clusters that were not in my data: <fkh>, <fx>, and <fqh> (which I would write phonemically as /vkʰ/, /vx/, and /vqʰ/). She does not find any clusters with the palatal nasal, neither
<vɲ> or <fɲ>.
The alveolar fricative /s/ occurs before voiceless stops, aspirated and tenuis at all
places of articulation, and before nasals and the voiceless lateral fricative.
(100) Attested [s + initial] clusters
Cluster Example

Gloss

sp
stʰ
st
scʰ
sc
sk
sqʰ
sq
sm
sn
sɲ
sŋ
sɬ

sores
to hang
tiger
to chase
scoop for water
to call, shout
younger sister
to ring
pharmacy
bridge of nose
green pea
blue
university

spəvji
stʰi
sta
scʰɛco
sco
skəri
sqʰi
sqədi
smənzjəre
sneʁdoŋ
sɲurbu
sŋuscæ
sɬopræ

Two clusters which are licit according to the phonotactic rules governing [s + stop]
sequences, /spʰ/ and /skʰ/, are missing in my data. Likely this can be attributed to a lack of
data, especially since Huang (1991:4) finds them both in her data, as well as <sɬh>, <stɕ>,
and <stɕh>.
The voiced alveolar fricative /z/ occurs preinitially before voiced consonants, palatal
and velar nasals, and voiced lateral, velar, and uvular fricatives.

112

(101) Attested [z + initial] clusters
Cluster Example

Gloss

zb
zd
zg
zɲ
zŋ
zɮ
zɣ
zʁ

to urge
claw
sour
seven
to hang
to winnow
whip
to yawn

zbəqe
zdɛrmoŋ
zgozgo
zɲe
zŋo
zɮæræ
zɣartɕa
zʁoræ

The same /z/ clusters are attested in Gexi Stau, and additionally <zm> and <zɟ>
(Huang 1991:4).
Co-occurring with preinitial /x/ we find voiceless stops at all but uvular place of
articulation, and voiceless alveolar and palatal affricates and fricatives.
(102) Attested [x + initial] clusters
Cluster Example

Gloss

xpʰ
xp
xtʰ
xcʰ
xk
xtsʰ
xts
xtɕʰ
xtɕ
xs
xɕ

butt
wind
behind
to puncture
white gourd
thin (like thread)
clean
six
clip, pin
three
to float

xpʰə
xpurju
xtʰəxtʰə
xcʰi
kəxkɛr
xtsʰoxtsʰo
xtsoŋma
xtɕʰo
xtɕərscɛ
xsu
xɕɛr

Two clusters, /xt/ and /xc/, are phonotactically possible but do not appear in the data.
Not unexpectedly, they are attested in Huang’s data (1991:4). She also finds <xɬ>.
The voiced velar fricative /ɣ/ occurs before voiced alveolar and palatal stops, labial
and alveolar nasals, and alveolar, lateral, and palatal fricatives.

113

(103) Attested [ɣ + initial] clusters
Cluster Example

Gloss

ɣd
ɣɟ
ɣm
ɣn
ɣz
ɣɮ
ɣʑ

because
afternoon
to blow
two
shoe
pestle
to teach

ɣdæmæ
sɲəɣɟi
ɣmɛ
ɣne
ɣzi
ɣɮəli
ɣʑi

Huang (1991:4-5) additionally finds /ɣ/ with the voiced labial stop and with voiced
alveolar and alveolo-palatal affricates.
The voiceless uvular fricative is found preceding the voiceless stops /p/ and /t/, and
the affricate /ts/. The /χp/ cluster only occurs word-medially in my data.
(104) Attested [χ + initial] clusters
Cluster Example

Gloss

χp
χt
χts

to fold
to return (something to someone)
to cut oneself

ɲtɕʰaχpa
χtɛ
χtsa

None of the [χ + initial] clusters Huang finds in Gexi overlap with those in my Mazi
Stau data. She finds <χs>, <χth>, and <χɕ> (1991:5).
The voiced uvular fricative appears as a preinitial in nine clusters following voiced
alveolar and palatal stops, nasals, the voiced retroflex affricate, and several voiced fricatives.

114

(105) Attested [ʁ + initial] clusters
Cluster Example

Gloss

ʁd
ʁɟ
ʁm
ʁn
ʁɲ
ʁdʐ
ʁz
ʁɮ
ʁʑ

pail
hole
low
dark
cow dung
tsampa
lama’s clothes
sleeping mat
bow

ʁdu
ʁɟi
ʁmæʁmæ
ʁnɛʁnɛ
ʁɲæ
ʁdʐu
ʁzɛn
ʁɮəstɛn
ʁʑə

The cluster /ʁŋ/ occurs once in the data as well, in the word ʁŋənərkuwu, but unfortunately I did not get its definition. I find a greater number of [ʁ + initial] combinations
than Huang, who does not have /ʁɟ/, /ʁʑ/, or /ʁdʐ/.
The rhotic is the most common preinitial in Stau, occurring before stops of all three
voicing types and all five places of articulation, before nasals, affricates, and fricatives.
(The reader is reminded that when /r/ precedes a voiceless consonant it assimilates to the
latter’s voicelessness and is realized as [ʂ].)

115

(106) Attested [r + initial] clusters
Cluster Example

Gloss

rp
rb
rtʰ
rt
rd
rcʰ
rc
rɟ
rkʰ
rk
rg
rq
rm
rn
rɲ
rŋ
rtsʰ
rts
rdz
rtɕʰ
rtɕ
rz
rɬ
rɮ
rɣ
rʁ

new
bee
right
stallion
mortar bowl
to bite
excrement
scale
cold
sock
old man
trunk
name
to mix
to wait
face
bowl that catches tsampa (at mill)
deer
to crawl
bubble
bike
with lifted skirts
wheat flour
wave
short
to be crazy

xsærpæ [xsæʂpæ]
rbu
rtʰærtʰə [ʂtʰæʂtʰə]
rtɛpu [ʂtɛpu]
rdəqu
rcʰɛ [ʂcʰɛ]
rcaxpa [ʂcaxpa]
rɟæmæ
ɣrərkʰu [ɣrəʂkʰu]
rkombjo [ʂkombjo]
rgɛvʑə
rqo [ʂqo]
rmə
rni
rɲi
rŋæ
rtsʰæmbræ [ʂtsʰæmbræ]
rtsɛ [ʂtsɛ]
rdzulu
rtɕʰæmbəqolu [ʂtɕʰæmbəqolu]
rtɕarta [ʂtɕaʂta]
rzelo
rɬə [ʂɬə]
rərɮɛv
rɣirɣi
rʁɑ

In Gexi, Huang (1991:4) finds several more [r + initial] clusters: <rl>, <rʑ>, <rdʑ>,
<rɢ>, <ʂph>, <ʂqh>, <ʂs>, and <ʂl>. She does not find several that are attested in Mazi
Stau, namely /rtʰ/, /rq/, /rtɕʰ/, /rtɕ/, and /rɬ/.
6.1.4 [Initial + medial] clusters
In total, twenty-nine Ci Cm cluster combinations are attested in my database: two
with /v/ as medial, nine with /r/, four with /l/, thirteen with /j/, and five with /w/. In this
subsection, we will look at each of these medials one at a time, in the order just given.

116

The labiodental fricative /v/ is found as a medial in cluster with two initials: /s/, and
/ɣ/. (107) provides an example of each cluster. Like Cp Ci clusters, consonants in [initial +
v] clusters always share the same type of voicing. If the initial of a cluster is voiceless, /v/
will assimilate in voicing and be realized as [f]. For instance, svo is pronounced as [sfo].
(107) Attested [initial + v] clusters
Cluster Example

Gloss

sv
ɣv

bright
oats

svo [sfo]
ɣvə

In her phonology of Gexi Stau, Huang (1991:5) finds several more clusters than
I do, namely: <dv>, <kv>, <khv>, <qv>, and <qhv>. She also finds <ʑv>, <zv>, <lv>,
and<rv>, but analyzes them as Cp Ci clusters.
The rhotic is medial in nine clusters following stops of all three voicing types, and
labiodental and velar voiced fricatives.
(108) Attested [initial + r] clusters
Cluster Example

Gloss

pʰr
pr
br
kʰr
kr
qʰr
qr
vr
ɣr

tangled
to whinny
tears
bed
horm
coarse
female yak
to pour
dry

pʰrə [pʰʂə]
prilæ [pʐilæ]
mobrɛ [mobʐɛ]
ɲɛkʰri [ɲɛkʰʂi]
krəmbə [kʐəmbə]
qʰræqʰræ [qʰʂæqʰʂæ]
qrə [qʐə]
vrə [vʐə]
ɣroɣro [ɣʐoɣʐo]

Similar to medial /v/, medial /r/ assimilates in voicing to the consonant it follows―not,
however, whenever the initial is voiceless, but only when it is voiceless and aspirated. So,

117

while pʰrə ‘tangled’ is pronounced [pʰʂə], prilæ ‘to whinny’ is pronounced [pʐilæ] with a
voiced /r/.
Huang in her work found the same clusters with a rhotic medial, and /gr/ in addition.
While the latter sequence does appear in my data, it is in a three-consonant cluster: zgri
‘star’. If my dataset were larger, it is likely an example of the cluster /gr/ would be found.
The alveolar lateral approximant is found following the stop /p/ and fricatives /v/
and /ʁ/:
(109) Attested [initial + l] clusters
Cluster Example

Gloss

pl
vl
ʁl

thigh
to put, leave something
to sing

plɛ
vlɛ
ʁləvə

/ɣl/ is attested as well, but the only word in which it occurs, ɣlo, is one for which I
don’t have a definition. Huang finds the basically same four [initial + l] clusters plus <sl>,
though she analyzes the [fricative + l] clusters as Cp Ci clusters. She also finds <bl> rather
than /pl/.
Clusters with medial /l/ are not restricted by any constraint concerning uniformity
of voicing. The same applies to clusters with medial /j/ or /w/.
The palatal glide is the most common medial in Stau, occurring after stops, fricatives, /r/, and /m/:

118

(110) Attested [initial + j] clusters
Cluster Example

Gloss

pʰj
pj
bj
mj
vj
sj
zj
ɕj
ʑj
ɣj
χj
ʁj
rj

outside
meat
pheasant
smooth
sores
to think
heart
to seek
comb
light
to destroy
pretty
horse

pʰjɛsu
pjɛno
bjɛrgə
mjɛmmjɛm
spəvji
ntsʰəsji
zjær
ɕjæ
zʁæʑjæ
kɛɣji
χjə
ʁjerə
rji

In my data I only find labial stops occurring with a medial palatal glide. Huang
(1991:5) finds <thj> and <kj> as well.
Finally, the labio-velar glide is attested with an eclectic set of initials: /k/, /ɕ/, /ɣ/,
and /r/.
(111) Attested [initial + w] clusters
Cluster Example

Gloss

kw
ɕw
ɣw
rw

foal
night
to hug
breath

rekwe
ɕwæ
ɣwæ
rwu

The cluster /qʰw/ occurs once in the data as well, in the word qʰwa, but I do not
know its definition.
Huang (1991) does not find /w/ as a medial, but she does find clusters with /v/ that
look suspiciously like those in (111) which I transcribe with /w/, e.g. re kve ‘pony’, ɕva
‘night, at’ and ɣva ‘hug, embrace’ (Huang 1992).

119

6.1.5 Three-consonant clusters
As was described in §6.1.1, when a cluster has three consonants the phonemes that
can fill each position are especially limited. To recap: preinitial position allows alveolar fricatives, nasals, and the rhotic; initial position allows stops, though no palatal stops;
and medial position allows glides, /r/, and /v/. Table 5 shows the combinations of these
phonemes that are attested in my dataset. Preinitials are on the vertical, medials on the horizontal. To represent attested three-consonant clusters in the chart, the initials are written at
the intersection of preinitial and medials that they are found in a cluster with. For instance,
the preinitial /s/ and the medial /v/ are found with the initials /tʰ/ and /k/, i.e. the clusters
/stʰv/ and /skv/ are attested in Stau.
Table 5. Cp Ci Cm co-occurrence chart
v

r

s tʰ k
z
m
n
ŋ kʰ
ɴ
r

j

p kʰ k q p tʰ
bg
b
pʰ b g
b
d
kʰ g
q

w

ɢ
q

Below, (112) demonstrates each of attested [s + initial + medial] cluster with an
example from the database.

120

(112) Attested [s + initial + medial] clusters
Cluster Example

Gloss

spj
stʰj
spr
skʰr
skr
sqr
stʰv
skv

wolf
to support, prop up
to churn
ant
late
broom
to press down
sharp

spjoŋkʰə
stʰjæ
spru
skʰro
təskrəsə
zəsqræ
stʰvæ
kɛskvɛ

Preinitial /s/ can be found with voiceless stops of aspirated and unaspirated persuasion from all places of articulation that occur in three-consonant clusters. Though far
from all the [s + initial + medial] combinatorial possibilities are attested here, there are no
obvious phonotactic restrictions concerning which medials can occur with which initials.
Huang (1991:5) attests to the existence of another of the combinatorial possibilities /sphr/
in Gexi Stau. She also finds /l/ as a medial in three-consonant clusters, listing <skl> and
<skhl> as confirmed clusters.
Attested three-consonant clusters with preinitial /z/ are much more restricted than
those with /s/. The two in (113) are confirmed, and a third is attested in Stau, /zbj/. Unfortunately I do not know the meaning of the only word in which it occurs, zbji.
(113) Attested [z + initial + medial] clusters
Cluster Example

Gloss

zbr
zgr

tent
star

zbræ
zgri

Although examples are limited, it is clear that preinitial /z/ only occurs with voiced
initials. Possibly it is restricted in three-consonant clusters to occurring with labial and velar
stop initials. Huang (1991:5) records the same three clusters, as well as a fourth <zɢv>.

121

Like in Cp Ci clusters, in three-consonant clusters preinitial /m/ can precede initials
other than only homorganic ones. In this case, the only attested non-homorganic initial is
/g/ in /mgr/; nevertheless, this is a consistency for all /m/ preinitials.
(114) Attested [m + initial + medial] clusters
Cluster Example

Gloss

mbj
mpʰr
mbr
mgr

cutting edge
snake
rice
wall (that one dries barley against)

mbjo
mpʰri
mbre
mgrə

Huang’s findings confirm my own: she additionally finds <mphj> and <mbl>, but
only finds clusters in which /m/ precedes labial stops and the voiced velar stop.
More consistencies can be noticed in nasal initials between two- and three-consonant
clusters: First, nearly all nasal initials in three-consonant clusters (see (114) and (115)) are
found preceding voiced and voiceless aspirated stops. This corresponds to the way tenuis
stops were excluded as initials from [nasal + initial] clusters. There is a three-consonant
exception to this rule: /ɴqj/.
Second, just as in two-consonant clusters non-labial nasals are only found with homorganic initials, the same rule is at work among three-consonant clusters, as (115) demonstrates.

122

(115) Attested [n/ŋ/ɴ + initial + medial] clusters
Cluster Example

Gloss

ndj
ŋkʰr
ŋgr
ŋkʰv
ɴqj
ɴɢw

red
to shake something
area, region
key
palate
hoe

ndjindji
ŋkʰrɛ
ŋgraji
ŋkʰvo
jæɴqjo
ɴɢwi

Finally, we look at [r + initial + medial] clusters, of which there is only one attested
in my data:
(116) Attested [r + initial + medial] clusters
Cluster Example

Gloss

rqw

necklace

rqwarzo [ʂqwarzo]

The scarcity of /r/-preinitial clusters is surprising considering the large number of
two-consonant clusters in which /r/ was a preinitial (see (106) in §6.1.3). However, my
findings are confirmed by those of Huang, who only finds <ʂqv> and <ʂqhv> (1991:5, 12):
(117) ʂqva
‘(house) dilapidated’
nə-ʂqhva ‘have torn’
The former cluster is essentially the same as the one I cite in (116), if one recalls
that /r/ is realized as [ʂ] before voiceless consonants, and that Huang often transcribes as
medial <v> what I transcribe as <w>.

6.2 Rhymes
Stau rhymes are much less complex than onsets. For one, the number of licit codas is much more restricted. As the syllable canon in §5.1 shows, only singletons and no
consonant clusters are allowed in coda position. In addition, final position licenses a small

123

subset of the consonants licit in initial position. It is a fairly eclectic array of consonants:
/t/, /v/, /x/, /ɣ/, /χ/, /m/, /n/, /ŋ/, /r/, and /l/. An example of each is provided below:
(118) Coda Example
t
v
x
ɣ
χ
m
n
ŋ
r
l

wut
ʁɑv
vivəx
moroɣ [moʐoɣ]
tsʰæsnəχ
ɣrərgɛm [ɣʐəʐgɛm]
kudʑən
tsoŋ
kʰɛsɛr [kʰɛsɛʂ]
snəmtsʰɛl

Gloss
light
needle
pressure cooker
black fungus
spinach
sink
satin
scallion
cloth
cole

Codas occur infrequently in the data. Word-finally, they only occur in 71 words of
a 1155-word database. In fact, only /v/, /n/, /ŋ/, and /r/ are native Stau codas. All words
ending with /t/, /x/, /ɣ/, /χ/, /m/, and /l/ are Tibetan loanwords (Katylin Wonnell, p.c.). This
is partially confirmed by Huang (1991), who does not find /t/ or any of the fricatives in her
Gexi Stau data (possibly Huang was more familiar with Tibetan and was able to exclude
loanwords from her data).
There is a case of neutralization in coda position, as is typologically common because of the position’s lack of prominence. In Stau, the contrast between /v/ and /p/ is
neutralized. Phonetically, both do occur as finals but as variant pronunciations in the same
word; the phonemic contrast between the two phones is not upheld, as (119) demonstrates.
The fricative appears in more careful speech, and the stop in faster speech (Katylin Wonnell,
p.c.).
(119) ɣməlɛv [ɣməlɛv ~ ɣməlɛp] ‘fire’
kʰælɛv [kʰæləv ~ kʰæləp] ‘cover, lid’

124

As for the vowels, few restrictions constrain their distribution.
One area where the distribution of vowels is constrained is in vowel and coda cooccurrences. The table below counts all the vowel and coda co-occurrences in the database.2
Table 6. [Vowel + coda] co-occurrences
t
i
e
ɛ
æ
ɑ
ə
o
u

v

x

ɣ

χ

m n

1

r

l

3

4

3

1 17

8

2
10

1
1 4
1

1

ŋ

1
1

1

7
7
4

10 1
33 1
6

The table brings to light several patterns and rules. There are a few co-occurrence
restrictions: /æ/ never occurs before codas; the dorsal fricatives only occur after non-high
vowels; /ŋ/ only appears after /o/ and /ɑ/; and /l/ only occurs following /ə/ and /ɛ/.
An obvious pattern that Table 7 shows is the absence of /e/ pre-coda. The reason /e/
never occurs before a coda is that, in this context, the contrast between /e/ and /ɛ/ neutralizes
in favour of the latter vowel. We can see the process at work in (120):
(120) [mdzɛmdze] mdzemdze ‘polite’
mdzemdze is a reduplication. As we will see later in §7.4, the second part of a
reduplicated word is the base, the first part is the reduplicant. With this knowledge, we
can eliminate the possibility that /ɛ/ is the base vowel and has changed to /e/ word-finally.
Instead, what is happening in this word is when mdze reduplicates, the /m/ of the base
2

Notes on Table 6: In closed rhymes, /i/ appears as its allophone [ɪ]. Appearances of /ŋ/ as coda include
allophonic nasalization on the vowel. Appearances of /r/ as coda include all allophones of /r/: [ʐ], [ʂ], and
rhotacization on the vowel.

125

resyllabifies to become a coda on the first syllable. With a coda consonant next to it, the /e/
of the reduplicant is realized as [ɛ], the contrast between the two vowels neutralizing.
Something similar is going on with the vowel /i/. Despite the fact that the table says
that it does occur before codas, the high front tense vowel never occurs phonetically before
a coda. Instead, in this context, /i/ is realized as its lax allophone [ɪ]. No examples of precoda [i] appear in the data, while all the examples of [ɪ] that appear are all before codas, as
seen below:3
(121) mtɕʰurtin [mtɕʰuʂtɪn] ‘tower’
nə-pʰiv [nəpʰɪv] ‘to close’
rti-n
[ʂtɪn]
‘to stop’
tə-ji-n [təjɪn]
‘to say’
A pattern is apparent here: both front non-low tense vowels /i/ and /e/ are prohibited
from occurring before codas, and their lax counterparts occur in their place.
Finally, let us look at the co-occurrence of vowels ignoring intervening consonants.
Table 7 gives the number of words in which vowels co-occur.
Table 7. Vowel co-occurrence chart

i
e
ɛ
ə
æ
ɑ
o
u
3

i

e

ɛ

ə

æ

ɑ

o

u

9
6
22
58
16
7
7
3

5
17
7
39
18
5
5
7

2
3
42
53
17
8
8
9

23
15
36
98
49
32
32
31

12
2
7
68
56
4
14
13

3
4
2
19
5
3
3
6

1
5
27
33
12
14
14
2

5
2
15
24
11
13
13
30

Note that the /n/ in the latter two examples is the second-person suffix.

126

Although some vowels do not co-occur very often, each Stau vowel does occur with
all other vowels. Vowels are not constrained in their phonotactic distribution with reference
to other vowels.

Chapter 7
Phonological processes
This chapter deals with phonological processes. Since the length of my fieldwork
was limited and since my focus was primarily on static generalizations and phonological
structure, the amount of attention to phonological processes was somewhat limited. However, several phonological processes could be observed within the data I did collect. Section
7.2 deals with voice assimilation and nasalization of /v/. Next, §7.3 presents several cases
of vowel assimilation in Stau. Finally, vowel changes that occur in reduplication will be
discussed in §7.4. First, it is necessary to digress in §7.1 to the relationship between the
vowels /ɛ/ and /æ/, which affects several phonological processes. Explaining it in advance
will simplify the description of these processes.

7.1 Excursus: Vowel shift /ɛ/ > /æ/
I hypothesize that a diachronic vowel change is in progress in Stau, in which the
mid front lax vowel /ɛ/ is lowering and becoming /æ/ in some cases. The hypothesis is
speculative but not without evidence. The vowel shift can be seen when one compares the
data Huang gathered in the early ’90s with my own. A few times, where she transcribes /ɛ/,
I transcribed /æ/.

127

128

(122) Mazi Stau Daofu (Huang 1992)
rŋæmoŋ
qʰæ
ɕəɕæ

rŋɛmu
qhɛ
ɕəɕɛ

Gloss
‘camel’
‘to laugh’
‘to wipe’

Now these few examples could just be accidents of the varying pronunciation of
different speakers or the varying transcription of different linguists. However, further evidence indicates that the difference between my data and Huang’s is one affected by the
passage of time. This evidence is found in compounds. A number of compounds in my
data are formed with the words væ ‘pig’ and ɮæ ‘hand’ as stems. Synchronically these
words have /æ/ as a nucleus (obviously); yet as stems in compounds we find them as vɛ and
ɮɛ. Examples (123) and (124) demonstrate this.
(123) vɛko
vɛ + ko
pig + pen
‘pigsty’
(124) ɮɛvɮi
ɮɛ
+ vɮi
hand + neck
‘wrist bone’
One might try to explain this discrepancy as /æ/ changing to /ɛ/ in response to some
context created by the second stem, but any attempt fails. /æ/ occurs in word-initial CV
syllables, like /ɛ/ does in (123) and (123). Likewise /ɛ/ occurs in word-final syllables, as
/æ/ does in væ and ɮæ. I found no distributional constraint that could account for the vowel
change. Vowel assimilation, the process that would make most sense, has no motivation.
The /o/ in (123) and /i/ in (124) are too dissimilar to both cause /æ/ to change to /ɛ/.
Four more examples―one with vɛ (125) as a stem,1 three with ɮɛ (126-128)―corroborate
this conclusion.
1

The /ɛ/ of vɛ in (125) assimilates to the /e/ of -ze, as will be explained in §7.3.

129

(125) veze
vɛ + ze
pig + DIM
‘piglet’
(126) ɮɛɲu
ɮɛ
+ ɲu
hand + back
‘back of hand’
(127) ɮɛrko
ɮɛ
+ rko
hand + foot
‘arm’
(128) ɮɛki
ɮɛ
+ ki
hand + ?
‘bracelet’
The explanation I propose is that these compounds were formed some time ago (as is
likely with fully lexicalized compounds), a time when væ and ɮæ were still vɛ and ɮɛ.2 Since
then, the vowels of these words changed and became lower, perhaps because they were more
open to vowel shift in their word-final position than their word-medial counterparts.
One compound3 with ‘hand’ as a stem looks like a counterexample, as it appears
with an /æ/ nucleus:
(129) ɮæ-bjænoŋ ‘palm (of hand)’
Still, it is only one example, against four in which ‘hand’ is produced with an /ɛ/.
Perhaps it is a more recent compound, coined after the vowel change.
2

This time was prior to Huang 1992, because ‘pig’ and ‘hand’ are væ and ɮæ in her data as well (though
she transcribes the vowel as <a>.
3
I am not sure what the meaning of the second stem bjænoŋ in (129) is. One idea is that it is pjɛno
‘meat’, which would make sense semantically, the palm being the meaty part of the hand. However, I have
no explanation for the addition of the velar nasal to the word’s end.

130

Based on this evidence, I think it is reasonable to claim that /ɛ/ shifting to /æ/ in
some cases. There is no clear context in which the change is happening, but it occurs most
often word-finally.

7.2 Voice assimilation and nasalization of /v/
The phonological processes we will look at first are those in which the phoneme /v/
figures. /v/ is involved in two types of assimilatory processes .
First, /v/ undergoes regressive voice assimilation when it occurs next to a voiceless
consonant in instances of derivation.
One type of derivation in which voice assimilation occurs is reduplication. In the
reduplicated adjective rtɛvrtɛv ‘fine’ in (130), the labial fricative assimilates in voicing to
the following voiceless consonant, which is itself agreeing with the voicing of the consonant
that follows it.
(130) [ʂtɛfʂtɛv]
rtɛv-rtɛv
red-fine
‘fine’
We can also see voice assimilation in the formation of some of the cardinal numbers
between eleven and nineteen. The morpheme for ‘ten’ is ʁav- (a bound morpheme never
seen by itself like -teen in English; the word for ‘ten’ is zʁa). When it combines with
numerals to form the numbers eleven through nineteen the morphemes that begin with a
voiceless consonant cause /v/ to devoice. For instance, in the derivation of ‘thirteen’, the
final /v/ of ʁav- devoices in assimilation to the voiceless /s/ of -su ‘three’.4
4

A person familiar with Stau will know that the word for three is not su but xsu, and that ‘two’ is ɣne, not
ne (133), and ‘six’ is xtɕʰo, not tɕʰo (132). My analysis is that there is a velar fricative prefix on four of the
cardinal numbers from one-ten. The other such number is ɣɮə ‘four’. When it participates in derivation with
ʁav-, it also drops the prefix: ʁavɮə ‘fourteen’.

131

(131) [ʁafsu]
ʁav-su
ten-three
‘thirteen’
The same process occurs in the derivation of ‘sixteen’. The /v/ in ʁav- assimilates
to the voicelessness of the initial consonant of tɕʰo.
(132) [ʁaftɕʰo]
ʁav-tɕʰo
ten-six
‘sixteen’
In addition to voice assimilation, I would tentatively claim that /v/ also undergoes
nasalization in front of a nasal. As established previously, the morpheme for the cardinal ‘two’ is -ne. When ʁav-, the morpheme for ‘ten’, is compounded with -ne to become
‘twelve’, the result is ʁamne.
(133) [ʁamne]
ʁav-ne
ten-two
‘twelve’
As this is the only example in my data of /v/ preceding a nasal, this rule must be
tentative.

7.3 Vowel assimilation
Several types of vowel assimilation were observed in my data. Regretfully, there
are very few examples of each type of vowel assimilation.
First, we find assimilation of roundness and height in the vowel of the negative
prefix mə- when the first vowel of the root to which it affixes is the high back round vowel
/u/. An example of the process is shown in (134), where mə- assimilates to the vowel of gu
‘to understand’.

132

(134) [mu-gu-ʐə]
mə-gu-rə
neg-understand-const
‘not understand’
The next vowel assimilation we will look at happens in the formation of diminutive
nouns. Stau has a suffix -ze, originally meaning ‘son’ (Huang 1991:20), that affixes to nouns
and derives diminutives. As an aside, it is interesting to note that possessing a diminutive
suffix that is derived from the word for ‘child’ or ‘son’ is common among the putative
Qiangic languages (Chirkova 2012:143). Assimilation happens in diminutive formation
when the non-low front vowels /i/ and /ɛ/ preceding the suffix assimilate to its vowel, /e/.
The high vowel /i/ lowers to assimilate with /e/. In (135), smi ‘woman’ becomes
sme-.
(135) [sme-ze]
smi-ze
woman-dim
‘daughter, girl’
The mid lax vowel /ɛ/ assimilates to the tenseness of /e/. In (136), vɛ- ‘pig’5 becomes
ve-.
(136) [ve-ze]
vɛ-ze
pig-dim
‘piglet’
The word for ‘uvula’, shown in (137), looks like a counterexample to this rule. Its
root, ‘tongue’, has /ɛ/ as its nucleus, but this vowel doesn’t assimilate to the /e/ of -ze in the
formation of the diminutive.
5

As I discussed in §7.1, although synchronically the word for pig is væ, historically it was probably vɛ.
This is the form on which veze and other compounds in which ‘pig’ is a component are based.

133

(137) [vɮɛ-ze]
vɮɛ-ze
tongue-dim
‘uvula’
Despite surface appearances, the nonconformity of vɮɛze constitutes a legitimate
exception because there is a restriction against e after ɮ, either specifically in the Mazi
dialect, or synchronically in Stau. In Huang (1992)’s data on Daofu, she has ten occurrences
of ɮe. Where words that coincide with Huang’s ɮe words exist in my data, the vowel in
those words is i, as (138) displays.
(138) Mazi Stau
vɮi
ɮi

Daofu (Huang 1992)
vɮe
ɮe

Gloss
‘neck’
‘wheat’

Furthermore, crucially, no ɮe sequences occur in my database. Thus, vɮɛze is not a
counterexample to the rule that /ɛ/ assimilates to the /e/ in the diminutive suffix.
The third vowel assimilation observed in my data is that the assimilation of the mid
front vowel /ɛ/ when it precedes the high front vowel i that forms the genitive particle. In
(139), /ɛ/ of scɛ ‘tool’ takes on the tenseness of the genitive i and is realized as /e/.
(139) [ndərjæ-sce-i-lu]
ndərjæ-scɛ-i-lu
sweep-tool-gen-pole
‘hoe’

7.4 Reduplication
Reduplication is a common process in Stau among nouns, verbs, and especially
adjectives. In verbs, reduplication indexes plurality (Gates, forthcoming). For adjectives,
reduplication is a sign of their adjectival status; no particular semantic nuance is conveyed
through reduplication in adjectives. Among nouns, the semantic function of reduplication

134

is less obvious. It may express smallness (‘indent’, ‘button’), endearment (‘friend’), or that
a noun consists of many of the same small thing (‘arm hair’, ‘grass’).
Reduplicated words are derived through total reduplication; the whole root is selected as the base for reduplication. In my database, no root except vdʑə ‘friend’ occurs
as an independent word; the others were only found in reduplicated form. Many roots are
reduplicated with no change, such as the following:
(140) sɲæsɲæ
‘bitter’
ʁərʁər
‘circular, round’
vdʑəvdʑə
‘friend’
zuzu
‘button’
rgurgu
‘to burp’
ndzəndzə-fɕə ‘to whisper’ (hide-speak)
In other cases, the vowel of the reduplicant changes from the original vowel of the
base. For example, the word for ‘grass’ is rŋərŋæ. Its base is rŋæ. The reduplicant copies
this sequence, and then changes its vowel to /ə/ before being prefixed to the base.
rŋərŋæ belongs to a group of reduplicated words that change their reduplicant vowel
to /ə/. These words have either /æ/ or /e/ as the original vowel of their base. All the examples
in my dataset are listed in (141):
(141) ɕəɕæ ‘to wipe’
vəve ‘to do’
vəvæ ‘to repair, build’
ɮəɮæ ‘to plow’
rŋerŋæ ‘to stirfry’
spəspæ ‘arm hair’
rŋərŋæ ‘grass’
The second type of vowel change happens to reduplicated words whose original
base vowel is /ə/ or /o/. In the process of reduplication, these vowels change to /ɛ/. For
sxɛsxo ‘to shake’, the reduplicant copies the base sxo, changes the /o/ nucleus for /ɛ/, and

135

then is prefixed onto the base. Three examples of this type of reduplication occur in the
data:
(142) sxɛsxo ‘to shake’
ɬɛɬə ‘wet’
vzɛvzə ‘to scratch’
Two reduplicated words occur that could be seen as counterexamples to the reduplicant vowel change rules I proposed. The first potential counterexample is vrɛvræ ‘to
thresh’. Since the vowel of the base is /æ/, the reduplicant vowel should change to /ə/, not
/ɛ/ as is the case.
An alternate explanation is that vrɛvræ is historically vrɛvrɛ, a case of reduplication
without changes, like those in (140). The difference between the vowels came about through
the vowel shift of /ɛ/ > /æ/ (introduced in §7.1). It fits the tentative pattern that the vowel
shift happens most frequently in word-final position. This explanation is further supported
by the fact that, in sentence context, the word began to sound like vrævræ.
The second potential counterexample is rtʰærtʰə ‘right’ (opposite of ‘left’). It also
does not fit the proposed rules. With /ə/ as its base vowel, the reduplicant vowel should
come out as /ɛ/, not /æ/. Here again the vowel shift presents a possible explanation. If we
hypothesize that the first syllable’s /æ/ is historically /ɛ/, then the underlying rtʰɛrtʰə would
fit well into the second type of reduplicated words with vowel change.

Chapter 8
Pitch phenomena

8.1 Pitch in related languages
Among the languages related to Stau, some have no tone system, such as Japhug
Rgyalrong (Jacques 2004:203), while others do. In Zhuokeji, Lin Y.J. (2012) finds a twoway contrast between a privative falling tone and toneless. A number of other varieties
have been analyzed as having pitch-accent systems, e.g. Caodeng Rgyalrong (J. Sun 2008),
Showu Rgyalrong (J. Sun 2004), Puxi Shangzhai (J. Sun 2000b). These languages use
tonal alternations to mark tense, aspect, and modality on verbs. In addition, in Puxi, tone
is lexically contrastive in some cases. Prins (2011) also describes Jiaomuzu Rgyalrong as
having a simple pitch-accent system, in which accent is predictably assigned to the final
syllable of a word.
Concerning “pitch-accent” systems, Hyman (2009:219) argues that pitch-accent is
not a coherent notion, and that no language should be analyzed in this way. Often, pitchaccent is spoken of as the intermediate on a continuum between stress and tone. Hyman
emphasizes that this is not the case. Stress and tone systems do not exist on a continuum;
they are separate phenomena that each have a prototype. Systems that have been called
pitch-accent should be reanalyzed as mixed systems that pick and choose properties from
the tone and stress prototypes (2009:213).

136

137

8.2 Pitch patterns in Stau
The pitch patterns found on citation forms of monosyllabic and disyllabic words
in Stau fall into several distinct patterns. Pitch patterns on words of three or more syllables do not fall into such clear categories, which is not surprising since all these words are
morphologically complex.
On one-syllable words, the most frequent pitch pattern―in 272 of 280 or 97% of
the words―is a high-falling pitch, as represented in the pitch track below:

Figure 21. Pitch track for pi ‘ball of tsampa’
There are eight monosyllabic words that are exceptions, having high level pitch
(143).

138

(143) ɕwæ ‘night (after midnight)’
ndʑəp ‘to suck’
toŋ ‘hole’
xɕɛt ‘to whip’
ʁa
‘door’
wur ‘pillow’
wut ‘light’
wo ‘again’
Figure 22 shows a pitch track of one of these monosyllabic words, toŋ ‘hole’.

Figure 22. Pitch track for toŋ ‘hole’
The latter four words in (143) do not necessarily sound like they have level pitch
because their consonant onsets are very sonorant and are able carry pitch. These begin at a
low pitch during the consonant segment, and rise to high, as exemplified by Figure 23.

139

Figure 23. Pitch track for ʁɑ ‘door’
As for disyllabic words, two main pitch patterns emerge: high followed by low pitch
(HL), and high followed by high-falling (H-HF). The latter is much more prevalent. 558 out
of 627 disyllabic words, or 89%, have a H-HF pattern, as for example, does ɮɛki ‘bracelet’,
shown in Figure 24.

140

Figure 24. Pitch track for ɮɛki ‘bracelet’
The HL pitch pattern is found on 59 disyllabic words in my dataset. A pitch track of
the HL pattern is displayed in Figure 25. At first glance, it may look like the second syllable
has a falling pitch. It does, but this due to utterance-final falling, a common intonational
phenomenon cross-linguistically. More salient is the fact that the pitch contour of the second
syllable starts much lower than that of the first syllable.

141

Figure 25. Pitch track for vədæ ‘wife’
Though not all the contents of the HL category are predictable, some types of words
are predictably found with HL pitch. These are: most kinship terms and most two-syllable
words in which the second syllable is the constative suffix -rə. Kinship terms, both those
that begin with the prefix æ- and those that have no kinship affix, almost without fail have
a HL pitch pattern. The examples in (144) follow this pattern.
(144) vədæ ‘wife’
dzəvə ‘husband’
mæmæ ‘grandmother’
æpe ‘grandfather’
There are some exceptions to this rule, kinship terms which have the H-HF pattern
instead of the HL one. The following words are all that occur in the data:
(145) lɛjɛ ‘daughter-in-law’
pətʰoŋ ‘son-in-law’
pərɟi ‘grandchild’

142

pərɟə ‘grandchild’
Interestingly, this pattern even extends to smi-ze. That same string of segments has
two different meanings depending on what pitch pattern it is pronounced with, one being a
kinship term and the other not. When smi-ze is pronounced with the HL pattern, it means
‘daughter’; it is said with H-HF pitch, it means ‘girl’. Witness pitch contours of the two
side-by-side:

Figure 26. Pitch track for smi-ze ‘daughter’ (HL)

143

Figure 27. Pitch track for smi-ze ‘girl’ (H-HF)
To the casual eye, the two pitch contours may look very similar, but they do show
a significant difference. The contours in Figures 26 and 27 are both relatively flat on the
first syllable; this is the H pitch they have in common. On the second syllable of Figure
27 the pitch contour starts at the same height as its previous H pitch, and steadily drops,
not flattening out until the very end; this represents a HF pitch. The contour of Figure 26’s
second syllable starts lower, and flattens out mid-fall; this is its L pitch.
Besides kinship terms, words in which the second and last syllable is -rə are also
predictably pronounced with HL pitch. A few examples of words with these suffixes follow:
(146) cʰɛ-rə ‘idle’
ɣʑæ-rə ‘to make a sound’
tʰo-rə ‘slow’
kon-rə ‘to understand’
rgan-rə ‘to want’

144

There are also two exceptional categories within disyllabic words. First, in (147)
there are four words with level pitch pattern, in which the two syllables have the same level
pitch. The first three measured in the range of 240-250 Hz, the last with higher pitch in a
270-280 Hz range. The first two words are Tibetan loanwords, but this likely has no bearing
on their pitch behaviour, since many borrowings take the H-HF pattern.
(147) tsʰoŋkoŋ ‘shop’
kɛdʑoŋ ‘straight’
qoqo
‘indent’
kæɕæ ‘morning’
The last word kæɕæ may actually be part of the second exceptional category, that of
greetings, leave-takings and times of day. They are a category that behaves very unusually.
In my recordings, times of day, greetings, and leave-takings have a particular intonation
pattern that is due to their pitch patterns being descending melodic intervals. Greetings and
leave-takings are pronounced with a major third and times of day with a minor third (the
pitch of a doorbell). Since producing a musical interval necessitates holding one’s pitch
steady, these words are also characterized by fairly consistent fundamental frequency on
each syllable, in contrast to other words that typically fall utterance-finally. The figures
below display the pitch tracks, respectively, of a leave-taking and a time of day.

145

Figure 28. Pitch track for gaca ‘goodbye (evening)’; represents a major third

Figure 29. Pitch track for sɲəɣɟi ‘afternoon’; represents a minor third

146

Although kæɕæ ‘morning’ in its citation form does not fit the pattern for times of
day, as it has a HH pitch, in sentence context it behaves the same as all other time of day
words, giving reason to class it with them.
All time-of-day words were recorded in the same sentence frame: pəsɲi ___ rku-rə
‘it’s cold this ___’, and the intonation pattern also carries over into these sentences, but the
minor third is not produced on the time-of-day word. Instead rku-rə receives the interval.
The syllables preceding rku-rə, including the time-of-day, are all pronounced with the pitch
of the first tone of the interval, which falls on rku. This can be witnessed in the example
below, where kəzə ‘night’ is presented in the frame.

Figure 30. Pitch track of pəsɲi kəzə rku-rə ‘it’s cold this night’ (kəzə ‘night’ between
vertical lines)
In Figure 30, the fundamental frequency of both syllables of kəzə is essentially equal
to that of rku; the pitch tracks are on the same level horizontally. The same goes for sɲi in
pəsɲi. The first syllable of pəsɲi has a little lower frequency, which is understandable as
it also has less amplitude and length than the other syllables. The final syllable of the

147

utterance, -rə, has a lower pitch than rku; three semi-tones lower than it, the two forming
together a minor third.
The non-triviality of the minor third pattern is even clearer if we look at another
example of a time-of-day word in context. When mdʑudʑu ‘late morning’ was recorded in
the frame sentence, an extra morpheme tɕʰæ ‘on’ was added, so that the sentence became:
(148) pəsɲi mdʑudʑu
tɕʰæ rku-rə
today late.morning on cold-const
‘It’s cold this morning’
The addition changed the pitch pattern of the sentence: the intonation pattern persisted; in fact, the minor third was repeated twice. Figure 31 below shows, as in the previous
example, rku-rə being pronounced with a minor third. The other minor third precedes it,
formed by the last syllable of mdʑudʑu and tɕʰæ (the syllable following the second vertical
line). One can see in Figure 31 that dʑu is assigned the same pitch as rku, and tɕʰæ as -rə.

Figure 31. Pitch track of pəsɲi mdʑudʑu tɕʰæ rku-rə ‘it’s cold this morning’ (mdʑudʑu ‘late
morning’ between vertical lines)

148

This phenomena of using a specific intonation pattern for a particular set of words
is not without parallel in English. We English speakers sometimes use minor thirds when
we greet or take our leave (e.g. “Morning!” or “See ya!”), or call someone’s name.

8.3 Interpretation of pitch patterns
Stau is not a tone language, based on my observations in §8.2. Monosyllabic words
are only found with one pitch pattern; disyllabic words for the most part also occur with
one pitch pattern. Huang (1991) also comes to this conclusion. In her study of Daofu,
she did not find tonal contrast used to distinguish meaning either, in a database of more
than 3000 words. She did find, as I did, that each word has a habitual pitch pronunciation
(1991:14). Her findings differ a little from my own, in that on monosyllabic words she finds
two common pitch patterns. Falling pitch is one, and occurs frequently on syllables in which
the initial consonant or consonant cluster is voiceless. The other pitch pattern is rising, and
that occurs mostly on syllables that begin with voiced consonants. As for disyllabic words,
Huang finds, as I did, that they usually use what she transcribes as 33-53, and what I call
H-HF.
Based on the observations made in §8.2, it is possible to hypothesize that Stau has a
fixed stress system. In general, the pitch patterns of Stau words are predictable. Almost all
monosyllabic words have the same falling pitch pattern, which is congruous with a stress
system. In a tone system, we would expect at least two pitch patterns on monosyllables.
By contrast, in a stress system, isolated monosyllables by default receive stress and are
pronounced with the same pitch pattern, as we find in Stau.
The pitch patterns on disyllabic words also show evidence of a fixed stress system.
Most disyllabic words have H-HF pitch. This could be interpreted as stress on the second syllable. Where disyllabic words do not follow this pattern, the exception is usually
justified. Disyllabic words that have HL pitch are complex, formed by attaching affixes

149

to roots. We would interpret these words as having stress on the first syllable. Greetings, leave-takings, and times-of-day are also predictable exceptions to the rule of secondsyllable stress. Although these words do not have second-syllable stress, they obtain their
pitch from an utterance-level pitch pattern, which trumps stress applied at the lexical level.
Thus, these words do not count as true counterexamples.

Chapter 9
Conclusion
In conclusion, I will briefly summarize my findings and suggest areas for further
study.
In this description of the phonology of Stau, I have found that Stau has a large
consonant inventory of 42 consonants, and a moderately large vowel inventory of eight
vowels. My findings coincide well with those of Huang (1991), who studied Stau before
me. She lists more consonants in her inventory, but the discrepancy arises because she is
more concerned to describe all the sounds that occur phonetically, while in my inventory
I only include those that are phonemic. As to vowels, we both describe the same eight
vowels.
Stau has a large syllable canon with six licit syllable types: CV, CCV, CCCV, CVC,
CCVC, and CCCVC. An additional syllable type, V, appears in prefixes; only the low
vowels /æ/ and /ɑ/ are allowed as single V syllables. In consonant clusters formed with a
preinitial consonant followed by an initial consonant, Ci position licenses stops, fricatives,
affricates (voiced, voiceless, and if applicable voiceless aspirated of the foregoing), and
nasals. Cp licenses fricatives, the rhotic, and nasals. Ci Cm clusters allow stops, fricatives,
the rhotic, and /m/ in initial position. In medial position they allow the rhotic, glides, /l/, and
/v/. Clusters of three consonants are more restricted than any of the two-consonant clusters.
Ci in these clusters licenses only stops; Cp licenses alveolar fricatives, the rhotic, and nasals,
while Cm allows glides, /v/, and the rhotic. Final position licenses ten consonants: /p, /t/;
/x/, /ɣ/, /χ/; /m/, /n/, /ŋ/; /r/ and /l/.
There are fewer phonotactic constraints on vowels. There are a few [vowel + coda]
co-occurrence restrictions: velar and uvular fricative codas only follow non-high vowels;
150

151

/ŋ/ only appears after /o/ and /ɑ/; and /l/ only occurs after /ə/ and /ɛ/. In addition, the contrast
between /e/ and /ɛ/ is neutralized in favour of the latter before word-final codas. In the same
context, the high front vowel /i/ is realized as its allophone [ɪ].
Four phonological processes were identified in Stau: vowel assimilation, voice assimilation, nasalization, and vowel changes in reduplication.
Fairly consistent pitch patterns occur on mono- and disyllabic words in Stau that can
be interpreted as a fixed stress system. Monosyllabic words predominantly are pronounced
with a HF pitch pattern, which is analyzed as stress. Disyllabic words usually have stress
on their second syllable; this is realized as a H-HF pitch pattern. Some complex disyllabic
words, such as kinship terms and words ending with -rə, are realized with HL pitch; these
have stress on the first syllable. Greetings, leave-takings, and times of day are pronounced
with an utterance-level intonation pattern; the latter with a minor third interval, and the two
former with a major third.
The two topics just mentioned are the ones most in need of further study. Much more
time could be given to investigating the phonological processes that arise in the interaction
of morphemes in derivation and inflection. Stau has quite complex verb morphology; this is
likely to be a promising area of study, as Jacques et al. (2013) has already shown. Additional
research is also needed to study the pitch patterns of Stau, particularly of words of three or
more syllables and of words as they are placed in the context of utterances.

Appendix A
Minimal pairs
This appendix provides minimal pairs for similar phones as evidence for their phonemic status.

A.1 Consonant minimal pairs
Stops
(149) pi
‘ball of tsampa’ pʰi
‘to run away’
pubæ ‘Tibetan’
pʰupʰa ‘male pig’
xopi ‘table’
ræpʰi ‘mahjong’
tɕʰæpæ ‘to punish’
nəpʰæ ‘to split’
(150) popo ‘pocket’ bobo ‘zhaoji’
(151) toŋbe ‘stem’
tʰoŋbe ‘pot’
tutu ‘basket carried on back’ tʰutu ‘mixed together’
bəti ‘cheek’
ɣətʰi ‘to drink’
ɣrətoŋ ‘well’ (n.)
pətʰoŋ ‘son-in-law’
(152) gaca ‘goodbye (evening)’ necʰa ‘good morning’
vcə ‘to harvest’
vcʰə ‘to weigh’
(153) kæɕæ ‘morning’ kʰæɕjæ ‘lips’
krə
‘boat’
kʰrə ‘to shake s.t.’
ʁnɛkuku ‘dark’
pukʰu ‘mosquito’

152

153

(154) qoqo ‘indent’ qʰosto ‘back’
qoqo ‘indent’ ŋaŋeqʰo ‘myself’
Affricates
(155) tʂæ ‘to cut with scissors’ tɕitɕæ ‘skin’
tʂəma ‘body dirt’
tɕədə ‘book’
(156) mdʐu
‘thunder’
mdʑudʑu ‘midday’
ndʐændʐæ ‘same’
ndʑæ
‘rainbow’
ndʐə
‘time, instance’ ndʑəv ‘to suck’
(157) tʂʰatʂoŋ ‘mug, cup’ tɕʰatsoŋ ‘all’
(158) tsoŋma ‘clean’ tsʰoŋkoŋ ‘shop’
tsəgə ‘clothing’ tsʰə
‘salt’
rtsɛ
‘deer’
rtsʰe
‘lung’
(159) tɕæ ‘tea’
tɕʰæ ‘on’
ætɕɛ ‘together’ ætɕʰə ‘what’
Fricatives
(160) səli ‘to roll’
zəli ‘to fall’
səqə ‘small piece of machinery’ zəbə ‘body’
məsi ‘stick of wood’
qʰəzi ‘bowl’
pʰjəsu ‘outside’
zuzu ‘button’
(161) ɕu
‘strength’ ʑu
‘yogurt’
ɕə
‘teeth’
ʑəvə ‘village’
ɕuɕu ‘behind’ ʑuʑə ‘can, able to’
kʰæɕjæ ‘lips’
zʁæʑjæ ‘comb’

154

(162) xə ‘yak-bull crossbreed’ ɣəzə ‘bird’
maxe ‘water buffalo’
maɣər ‘a type of tree’
(163) χodzupare ‘cloth worn on head’ ʁoɲu ‘back’
χjə
‘to destroy’
ʁjə ‘fish’
kəχo
‘bark (of tree)’
jəʁo ‘upstairs’
(164) xotsʰɛv ‘pepper’
χodzupare ‘cloth worn on head’
ʂcaxpa ‘excrement’ tɕaχpa
‘to steal’
Nasals
(165) mə ‘younger brother’ ne ‘to rest’
mdʑu ‘thunder’
ndʑə ‘time’
smi ‘young women’ sni ‘nose’
(166) nə
‘to rest’
ɲə
‘ear’
nətso ‘sun’
nɛrdʑə ‘fingernail’ ɲɛndʑə ‘sunflower seed’
noŋ ‘inside’
ɲoŋpæ ‘old’
(167) nərbə ‘treasure’
ŋəmɛ ‘cow’
æneze ‘paternal younger aunt’ æŋæze ‘baby’
(168) ɲəlɛm ‘dream’ (n.)
ŋəmɛ ‘cow’
ʁoɲu ‘back’ (n.)
səŋun ‘who’
sɲu ‘beans’ (for pigfeed) sŋuscæ ‘blue’
Laterals
(169) ləpʰu ‘tree’
ɬə ‘cow’s milk’ ɮə ‘field’
lærɲə ‘celtuce’
ɬæ ‘god’
ɮæ ‘hand’
lɛvdɛr ‘peeler’
ɬɛ ‘to get wet’ ɮɛ ‘to come’
vlɛ
‘to put down’ vɬɛ ‘ashes’
vɮɛ ‘tongue’

155

Approximants
(170) vo ‘stomach’ wo ‘again’

A.2 Vowel minimal pairs
(171) pʰi ‘to run away’ pʰe ‘to vomit’
ʑele ‘turnip’
zəli ‘to fall’
qʰre ‘to pull down’ kʰri ‘chair’
(172) pʰe ‘to vomit’ pʰɛ ‘to dig’
mɬe ‘to braid’ fɬɛ ‘ashes’
nene ‘breast’ nɛmɛ ‘finger’
(173) ɕwæ ‘night’
ʂqwa ‘Adam’s apple’
mæmæ ‘grandmother’ tʂəma ‘body dirt’
nəpʰæ ‘to split’
mupʰa ‘sow’
(174) spu ‘incense stick’ spo ‘grassland’
turu ‘to find it’
ro ‘1’
ʂtɕʰu ‘wine bottle’ xtɕʰo ‘6’
(175) ɕi ‘highland barley’ ɕə ‘teeth’
ɮi ‘wheat’
ɮə ‘field’
mi ‘mole’
mə ‘younger brother’
(176) rŋe ‘to hear’ rŋə ‘knee’
ʂtsʰe ‘lung’ ʂtsʰə ‘cough’
ɣne ‘2’
sɬəɣnə ‘moon’
(177) skɛ ‘language’ skə ‘chives’
ɣmɛ ‘wound’ ɣmɛ ‘fire’
ɮɛ ‘to come’ ɮə ‘field’

156

(178) rŋæ ‘face’ rŋə ‘knee’
ɮæ ‘hand’ ɮə ‘field’
ɲæ ‘fish’ ɲə ‘ear’
(179) ʁa ‘door’
ʁə ‘head’
ʁja ‘male yak’
ʁjə ‘fish’
na ‘post for building house’ nə ‘to rest’
(180) mo ‘eye’
mə ‘younger brother’
spo ‘grassland’ spə ‘pus’
ʂko ‘root’
ʂkə ‘foot’
(181) ɕu ‘strength’ ɕə ‘teeth’
ndʑu ‘to sit’
ndʑə ‘to hide s.t.’
mkʰu ‘cowshed’ mkʰə ‘smoke’

Appendix B
Lexicon
The following lexicon contains much of the Stau data that I used in analysis for this
thesis. I hope its inclusion will be helpful to fellow Rgyalrongic scholars.
Num
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.

Lexical

Phonetic

English Gloss

su-ndʐə
vrə
æpæ
æmə
æti
ædæ
mə
sqʰi
æpe
mæmæ
ækʰə-kɛcʰɛ
ækʰə-ze
æne-kɛcʰɛ
æne-ze
ækʰə
æne
æʑu-kɛcʰɛ
æʑu-ze
æjæ-kɛcʰɛ
æjæ-ze
æjæ
məbærmæ
sno
kʰædʑu
ʁzaʁzav
təvdəze
ʑələze
smi-ze

sundʐə
vʐə
æpæ
æmə
æti
ædæ
mə
sqʰi
æpe
mæmæ
ækʰəkɛcʰɛ
ækʰəze
ænekɛcʰɛ
æneze
ækʰə
æne
æʑukɛcʰɛ
æʑuze
æjækɛcʰɛ
æjæze
æjæ
məbærmæ
sno
kʰædʑu
ʁzɑʁzɑv
təvdəze
ʑələze
smeze

third time
to pour
dad
mom
older brother
older sister
younger brother
younger sister
grandfather
grandmother
paternal older uncle
paternal younger uncle
paternal older aunt
paternal younger aunt
paternal uncle
paternal aunt
maternal older uncle
maternal younger uncle
maternal older aunt
maternal younger aunt
maternal aunt
second younger brother
younger sister male speaker
thank you
thank you
son
son
daughter

157

Category
V
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N

N
N
N

158

Num
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.

Lexical

Phonetic

English Gloss

Category

dzəvə
vədæ
jovə
kæɕæ
necʰɑ
gɑcɑ
mdʑudʑu
sɲəɣɟi
lɛɲu
ɕwæ
kəzə
tə-skrə-sə
kɛdərə
xɑvdu
ŋəkʰæ
ɣədʑelə
jɛʑoŋbu
pəsɲi
ævəsɲi
qʰəsji
bəlætəsɲi
æ-sɲi
ro
ɣne
xsu
ɣɮə
ɴɢwɛ
xtɕʰo
zɲe
rjɛ
ŋgə
zʁɑ
ʁɑvro
ʁɑmne
ʁɑvsu
ʁɑvɮə
ʁɑɴɢwɛ
ʁɑvtɕʰo
ʁɑzɲe
ʁɑrjɛ
ʁɑŋgə

dzəvə
vədæ
jovə
kæɕæ
necʰɑ
gɑcɑ
mdʑudʑu
sɲəɣɟi
lɛɲu
ɕwæ
kəzə
təskrəsə
kɛdəʐə
hɑvdu
ŋəkʰæ
ɣədʑelə
jɛʑoŋbu
pəsɲi
ævəsɲi
qʰəsji
bəlætəsɲi
æsɲi
ʐo
ɣne
xsu
ɣɮə
ɴɢwɛ
xtɕʰo
zɲe
ʐjɛ
ŋgə
zʁɑ
ʁɑvʐo
ʁɑmne
ʁɑfsu
ʁɑvɮə
ʁɑɴɢwɛ
ʁɑftɕʰo
ʁɑzɲe
ʁɑrjɛ
ʁɑŋgə

husband
wife
wife
morning
hello
goodbye
late morning
afternoon
evening
night (after midnight)
night
late
early
now
before
after
often always
today
yesterday
tomorrow
everyday
day
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

N
N
N
N

N
N
N
N
N
ADJ
ADJ

N
N
NUM
NUM
NUM
NUM
NUM
NUM
NUM
NUM
NUM
NUM
NUM
NUM
NUM
NUM
NUM
NUM
NUM
NUM
NUM

159

Num
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.

Lexical

Phonetic

English Gloss

Category

nə-sqʰɑ
rtsudʑu
ndʐə
æ-ndʐə
nə-ndʐə
səmbə
tɕʰɑtsoŋ
kɛ-nɮə
kɛ-ɣji
kɛ-ɣri
ŋæ
ɲi
tʰɛ
ŋæji
ɲiɲi
tʰiɲi
vdʑəvdʑə
ŋɑ-ɣne
ɲi-ɣne
tʰi-ɣne
ŋɑŋeqʰo
ɑdɛ
tʰɛ
sə
ætɕʰə
ludɛ
səŋun
ʁə
kɑpəlɑ
mormi
mojo
mospə
mo
bəti
sni
jæ
ɣmur
ɕə
vɮɛ
mæŋgɛ
ɲə

nəsqʰɑ
ʂtsudʑu
ndʐə
ændʐə
nəndʐə
səmbə
tɕʰɑtsoŋ
kɛnɮə
kɛɣji
kɛɣʐi
ŋæ
ɲi
tʰɛ
ŋæji
ɲiɲi
tʰiɲi
vdʑəvdʑə
ŋɑɣne
ɲiɣne
tʰiɣne
ŋɑŋeqʰo
ɑdɛ
tʰɛ
sə
ætɕʰə
ludɛ
səŋun
ʁə
kɑpəlɑ
mormi
mojo
mospə
mo
bəti
sni
jæ
ɣmu˞
ɕə
vɮɛ
mæŋgɛ
ɲə

20
number
time
first time
second time
all
all
heavy
light
many
1 sg
2 sg
3 sg
1 pl
2 pl
3 pl
friend
1 dual
2 dual
3 dual
myself
this
that
who
what
which
who
head
forehead
eyebrow
eyelid
eyelashes
eye
cheek
nose
mouth
mouth
teeth
tongue
chin
ear

NUM
N
N

ADJ
ADJ

N

N
N
N
N
N
N
N
N
N
N
N
N
N

160

Num

Lexical

Phonetic

English Gloss

Category

111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
143.
144.
145.
146.
147.
148.
149.
150.
151.

sneˤdoŋ
rŋæ
vɮi
rqwɑ
xpoŋ
ɮɛrko
ɮæ
nɛmɛ
nɛrdzə
qʰosto
tɕo
noptɕʰo
xpʰə
plɛ
rŋə
rdə
rko
rkə
ʁɑlo
vo
wocʰi
rtsʰə
xæcʰo
zʁoræ
zəpə
tɕo
wocæ
spəspæ
ʁəptɕæ
tɕitɕæ
xɕi
ʑæʁɟæ
rŋe
scici
no
qʰæ
zjuræ
skəri
tsəgə
rə-gi
tʰɛ

sneˤdõ
ʐŋæ
vɮi
ʂqwɑ
xpõ
ɮɛʂko
ɮæ
nɛmɛ
nɛrdzə
qʰosto
tɕo
noptɕʰo
xpʰə
plɛ
ʐŋə
ʐdə
ʂko
ʂkə
ʁɑlo
vo
wocʰi
ʂtsʰə
hæcʰo
zʁoʐæ
zəpə
tɕo
wocæ
spəspæ
ʁəptɕæ
tɕitɕæ
xɕi
ʑæʁɟæ
ʐŋe
scici
no
qʰæ
zjuræ
skəri
tsəgə
ʐəgi
tʰɛ

bridge of nose
face
neck
Adams apple
shoulder
arm
hand
finger
fingernail
back
lower back
side
butt
thigh
knee
lower leg
ankle
foot
chest
stomach
lower abdomen
cough
sneeze
yawn
body
waist
navel
hair on arm
hair on head
skin
sweat
limp
to hear
to look at, see
to smell
to laugh
to cry
to shout, call
clothing
to clothe
to take off clothes

N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
V
V
V
N
N
N
N
N
N
N
N
V
V
V
V
V
V
N
V
V

161

Num
152.
153.
154.
155.
156.
157.
158.
159.
160.
161.
162.
163.
164.
165.
166.
167.
168.
169.
170.
171.
172.
173.
174.
175.
176.
177.
178.
179.
180.
181.
182.
183.
184.
185.
186.
187.
188.
189.
190.
191.
192.

Lexical

Phonetic

English Gloss

Category

kəgəmæ
ʁɑv
reskə
nə-ŋgi
nə-ndjælæ
tɕʰɑxtɕə
ɣə-tʰi
pʰe
mdʑu-rə
spæ-rə
ɮo-rə
jo
ʁɑ
ɲɟɛm
ʁnɛmju
ɣmə
tə-rjɛ
mobrɛ
ʑɛrvæ
mbjɛ-mə
rʁu
kʰæɕjæ
vɟə
rwu
skɛ
ɟɛuræ
ɬə
zjær
noŋtɕʰə
si
spə
tʂəmɑ
ʁərjɑ
ɕu
ŋgələ
rɬə
ndzure
sciskɛr
ægəjo
tə-rkæ-sə
ŋo-rə

kəgəmæ
ʁɑv
reskə
nəŋgi
nəndjælæ
tɕʰɑxtɕə
ɣətʰi
pʰe
mdʑuʐə
spæʐə
ɮoʐə
jo
ʁɑ
ɲɟɛm
ʁnɛmju
ɣmə
tərjɛ
mobʐɛ
ʑɛrvæ
mbjɛmə
ʐʁu
kʰæɕjæ
vɟə
ʐwu
skɛ
ɟɛuræ
ɬə
zjæʂ
nõtɕʰə
si
spə
tʂəmɑ
ʁərjɑ
ɕu
ŋgələ
ʂɬə
ndzure
sciskɛʂ
ægəjo
təʂkæsə
ŋoʐə

naked
needle
thread
to eat
lick
chew
drink
to vomit
hungry
thirsty
tasty
house
door
wall
roof
fire
to stand up
tears
blind
deaf person
dumb
lips
saliva
breath
voice
facial hair
cows milk
heart
guts
liver
pus
body dirt
bone
strength
food stuff
wheat flour
to live
to be born
to grow up
to get tired
illness

ADJ
N
N
V
V
V
V
V
ADJ
ADJ
ADJ
N
N
N
N
N
V
N
ADJ
N
ADJ
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
V
V
V
V
N

162

Num
193.
194.
195.
196.
197.
198.
199.
200.
201.
202.
203.
204.
205.
206.
207.
208.
209.
210.
211.
212.
213.
214.
215.
216.
217.
218.
219.
220.
221.
222.
223.
224.
225.
226.
227.
228.
229.
230.
231.
232.
233.

Lexical

Phonetic

English Gloss

Category

ɣmɛ
ŋo-rə
zæntɕʰæ
vzɛvzə
smənvəgəɟi
vsɛ
tə-sɛ-sə
ɬæ
ɲcʰəcə
jækʰæ ɲcʰɛɲcə
pʰi
scʰɛco
skɛ
nene
nə-qe
xsəxsə
tə-vsɛ-sə
skɛsnɑ
ʁlə-və
ʑɛdʑuntɕʰɛm
χtɛ
rɲi
pjɛno
ɕontʰo
mdʑəsnæ
zgəŋæ
tsʰə
wɛrtsʰi
ɣrə
zʁe
ɣələm-sə
ndʑəv
ɣməlɛv
mkʰə
vɬɛ
nə-sqʰu
ndzu
kə-xcʰi
ɣə-vi
zɑmɑsotɕi
pərzi

ɣmɛ
ŋoʐə
zæntɕʰæ
vzɛvzə
smənvəgəɟi
fsɛ
təsɛsə
ɬæ
ɲ̥cʰəcə
jækʰæ ɲcʰɛɲcə
pʰi
scʰɛco
skɛ
nene
nəqe
xsəxsə
təfsɛsə
skɛsnɑ
ʁləvə
ʑɛdʑuntɕʰɛm
χtɛ
rɲi
pjɛno
ɕõntʰo
mdʑəsnæ
zgəŋæ
tsʰə
wɛʂtsʰi
ɣʐə
zʁe
ɣələmsə
ndʑəp
ɣməlɛp
mkʰə
fɬɛ
nəsqʰu
ndzu
kəxcʰi
ɣəvi
zɑmɑsotɕi
pərzi

to be hurt
to be painful
to feel itchy
to scratch
to make medicine
to kill
to die
god
to fight
to argue
to run away
to chase
language
mother’s milk
to throw out s.t.
to be alive
to be killed
languages
to sing
to dance
to return s.t. to s.o.
to wait
meat
fruit
seed
egg
salt
lard
water
cook by boiling
ripe
to suck
fire
smoke
ashes
to extinguish
to sit down
to open
to do
pot
knife

V
V
V
V
V
V
V
N
V
V
V
V
N
N
V
V
V
V
V
V
V
N
N
N
N
N
N
N
V
ADJ
V
N
N
N
V
V
V
V
N
N

163

Num
234.
235.
236.
237.
238.
239.
240.
241.
242.
243.
244.
245.
246.
247.
248.
249.
250.
251.
252.
253.
254.
255.
256.
257.
258.
259.
260.
261.
262.
263.
264.
265.
266.
267.
268.
269.
270.
271.
272.
273.
274.

Lexical

Phonetic

English Gloss

Category

mbjo
ɕəɕæ
tʰəɣə
ɕækɛcʰɛ
rʁəbə
xtɕærə
mdæ
ʁʑə
leskæ-və
tɕitɕæɣʑɛ
ɕɑŋgu
ɮɛ
ŋkʰərvæ
rtin
pɛrspo
ɲɟəræ
mɟo-rə
tʰo-rə
rdzulu
tɕɛ
dzo
ŋkʰurlu
krə
vʑo
pʰɛ
zəli
ɬɛɬə
ɣroɣro
tə-rmu-sə
tʰətʰə
sɲæsɲæ
tə-tsə-sə
vəvɑŋgu
tə-scʰusə
ndjɛv
ɲəlɛm
məsi
vdzi
pəŋæ
smi
kɛdi

mbjo
ɕəɕæ
tʰəɣə
ɕækɛcʰɛ
ʐʁəbə
xtɕæʐə
mdæ
ʁʑə
leskævə
tɕitɕæɣʑɛ
ɕɑŋgu
ɮɛ
kəŋkʰərvæ
ʂtɪn
pɛrspo
ɲɟəræ
mɟoʐə
tʰoʐə
rdzulu
tɕɛ
dzo
ŋkʰu˞lu
kʐə
vʑo
pʰɛ
zəli
ɬɛɬə
ɣʐoɣʐo
tərmusə
tʰətʰə
sɲæsɲæ
tətsəsə
vəvɑŋgu
təscʰusə
ndjɛv
ɲəlɛm
məsi
vdzi
pəŋæ
smi
kɛdi

cutting edge
to wipe
rope
fat
thin
sword
arrow
bow
to work
to peel
to go
to come
to turn, grind
to stop
to walk
to run
fast
slow
to crawl
road
bridge
wheel
boat
to plane
to dig
to fall
wet
to dry
to forget
sweet
bitter
rotten
to build
to burn
to sleep
dream n
stick of wood
human
man, male
woman, female
child

N
V
N
ADJ
ADJ
N
N
N
V
V
V
V
V
V
V
V
ADJ
ADJ
V
N
N
N
N
V
V
V
V
V
V
ADJ
ADJ
ADJ
V
V
V
N
N
N
N
N
N

164

Num
275.
276.
277.
278.
279.
280.
281.
282.
283.
284.
285.
286.
287.
288.
289.
290.
291.
292.
293.
294.
295.
296.
297.
298.
299.
300.
301.
302.
303.
304.
305.
306.
307.
308.
309.
310.
311.
312.
313.
314.
315.

Lexical

Phonetic

English Gloss

Category

lutɕʰoŋ
ʑəvə
vqe
rkəmə
nə
kə-rjɛn
rmə
ɲcʰæræ
kə-dʑə-sə
ɲcʰə
rcʰɛ
vzu
tə-kʰri
ɣe
nə-xsɛrni
læʁju
tə-rko-sə
rə-ɲɟu-sə
rtsʰu
ntʰvæ
ndzə
ɕjæ
kə-scʰici
rəle
rə-vlɛ-sə
leskæ
nə-vəve
χjə
tə-pətɕo-sə
vəvæ
spjæ
nəpʰæ
vku
ɲtɕʰɑχpɑ
rʁe
zgozgo
tɕʰətɕʰə
pəcæ
lu
ndərjæ-sce-i-lu
tɕɑzjɛ

lutɕʰõ
ʑəvə
fqe
ʂkəmə
nə
kərjɛn
ʐmə
ɲcʰæræ
kədʑəsə
ɲ̥cʰə
ʂcʰɛ
vzu
təkʰʂi
ɣe
nəxsɛrni
læˠju
təʂkosə
ʐəɲɟusə
ʂtsʰu
n̥tʰfæ
ndzə
ɕjæ
kəscʰici
ʐəle
ʐəvlɛsə
leskæ
nəvəve
χjə
təpətɕosə
vəvæ
spjæ
nəpʰæ
fku
ɲ̥tɕʰɑχpɑ
ʐʁe
zgozgo
tɕʰətɕʰə
pəcæ
lu
ndərjæsceilu
tɕɑzjɛ

young
village
to shoot gun
to steal (hidden)
to rest
to ask
name
to play
to meet, run into
to beat, hit; to thresh
to bite
to take
to seize
to touch
to rub hands
wave
to push
to carry on back
to kick
to tread on it
to hide it
to seek it
to show it
to put it
to put it
work
to do
to destroy it
to wreck, tear
to repair s.t.
to split
to split
to bend it
to fold
to wash it
sour
salty
stick of wood
pole of tool
hoe
rake

ADJ
N
V
V
V
N
V
V
V
V
V
V
V
V
N
V
V
V
V
V
V
V
V
V
N
V
V
V
V
V
ADJ
V
V
V
ADJ
ADJ
N
N
N
N

165

Num
316.
317.
318.
319.
320.
321.
322.
323.
324.
325.
326.
327.
328.
329.
330.
331.
332.
333.
334.
335.
336.
337.
338.
339.
340.
341.
342.
343.
344.
345.
346.
347.
348.
349.
350.
351.
352.
353.
354.
355.
356.

Lexical

Phonetic

English Gloss

Category

tɕɑχpɑ
noŋ
pʰjəsu
pʰɛ
rə-lə-sə
rəmæləsə
bjæ
ʁərʁər
kə-rtsʰi
kə-xsi
ɣə-pʰri
tə-lu-sə
nəvi
kʰælɛv
ro
ro
æ-ndə
æ-xtɕæ
χtsɑ
nə-vtso
ɮərji
rkuvcæ
zdo
vqo
zdomə
məqʰi
ɬɛvsə
mdʐu
ndʑæ
kʰɑwɑ
rwo
kə-rwo-sə
tə-ʑə
nətso
sɬəɣnə
zgri
wut
ɣræmjæ
xpurju
ɣmɛ
cʰu-rə

tɕɑχpɑ
noŋ
pʰjəsu
pʰɛ
ʐələsə
ʐəmæləsə
bjæ
ʁərʁər
kəʂtsʰi
kəxsi
ɣəpʰʂi
təlusə
nəvi
kʰæləv
ʐo
ʐo
ændə
æxtɕæ
χtsɑ
nəftso
ɮərji
ʂkufcæ
zdo
fqo
zdomə
məqʰi
ɬɛpsə
mdʐu
ndʑæ
kʰɑwɑ
ʐwo
kərwosə
təʑə
nətso
sɬəɣnə
zgʐi
wut̚
ɣʐæmjæ
xpurju
ɣmɛ
cʰuʐə

to steal visible
inside
outside
to throw out
boiling water
not boiling
to split by itself
circular
to tie more loosely
to tie tightly
to untie, volitional
to untie, unvolitional
to cover it up
cover, lid
to swell
swollen spot
a stab
a cut
to cut oneself
to cut
to mix
to carve
cloud
sky
fog
rain
lightning
thunder
rainbow
snow
ice
to freeze
to melt
sun
moon
star
light
shadow
wind
to blow
hot of weather

V

V

V
ADJ
V
V
V
V
V
N
V
N
V
V
V
V
V
V
N
N
N
N
N
N
N
N
N
V
V
N
N
N
N
N
N
V
ADJ

166

Num
357.
358.
359.
360.
361.
362.
363.
364.
365.
366.
367.
368.
369.
370.
371.
372.
373.
374.
375.
376.
377.
378.
379.
380.
381.
382.
383.
384.
385.
386.
387.
388.
389.
390.
391.
392.
393.
394.
395.
396.
397.

Lexical

Phonetic

English Gloss

Category

tsʰeke
rku-rə
ɣrə-rkʰu
rkutsɛzɛndæ
ʑiŋgu
loŋbæ
nɑ
zbrə
mɑɣər
tæsi
zjo
lupʰu
xtʰo
spo
mtsʰu
ɣrə
rtɕʰæmbəqolu
bjolæ
kɛdʑɛdʑi
rərɮɛv
rgəmɛ
pjəmæ
tsi
ləpʰu
rqo
nɛmɛ
kəχo
qɑvlɑ
bɑrlu
doŋbə
mito
rko
kɛcʰɛ-xi-rə
ɣro
rədɑsnærŋæ
ɣəzə
bjɛrgə
pʰɛ
zjɛ
roŋqʰuræ
rtsʰulæ

tsʰeke
ʂkuʐə
ɣʐəʂkʰu
ʂkutsɛzɛndæ
ʑiŋgu
lõbæ
nɑ
zbʐə
mɑɣə˞
tæsi
zjo
lupʰu
xtʰo
spo
mtsʰu
ɣʐə
ʂtɕʰæmbəqolu
bjolæ
kɛdʑɛdʑi
ʐəʐɮɛv
ʐgəmɛ
pjəmæ
tsi
ləpʰu
ʂqo
nɛmɛ
kəχo
qɑvlɑ
bɑ˞lu
doŋbə
mito
ʂko
kɛcʰɛxiʐə
ɣro
ʐədɑsnærŋæ
ɣəzə
bjɛrgə
pʰɛ
zjɛ
ʐoŋqʰuræ
ʂtsʰulæ

hot
cold for weather
cold
warm
mountain
valley
post for building hse
tree type
tree type
tree type
tree type
tree type
ground (n.)
grassland
lake
river
bubble
to float
shore
wave
stone for building hse
sand
grass
tree
trunk
toe
bark
branch
leaf
stem
flower
roots
to grow
to wither
animal
bird
pheasant
to dig
to rake
to hoe
to move

ADJ
ADJ
ADJ
ADJ
N
N
N
N
N
N
N
N
N
N
N
N
N
V
N
N
N
N
N
N
N
N
N
N
N
N
N
V
V
N
N
N
V
V
V
V

167

Num
398.
399.
400.
401.
402.
403.
404.
405.
406.
407.
408.
409.
410.
411.
412.
413.
414.
415.
416.
417.
418.
419.
420.
421.
422.
423.
424.
425.
426.
427.
428.
429.
430.
431.
432.
433.
434.
435.
436.
437.
438.

Lexical

Phonetic

English Gloss

Category

ŋkʰərvæ
sxɛsxo
ɲuɲɟu
rtʰærtʰə
tsʰəsji
bobo
xɑkon-rə
popo
jə
scɑn-rə
tə-scɑŋ
tə-rgɑŋ
rgɑ-ndʑə
qʰætəsɬoŋ
səm
svo
ʁnɛʁnɛ
ʁnɛ-kuku
kɛ-svo
ɕoŋsnæ
ŋkʰu
tə-nəv-sə
xɕɛr
xtsʰə
tsʰə
xeʑi
ɣzəɣzə
æχe
ætsʰe
xævʑi
xɑji
mjɛmmjɛm
xovɛxovɛ
ɲæ
ʁjə
pæbə
ɕɑndʑu
kʰətæ
skʰro
pukʰu
bæɴɢe

ŋkʰərvæ
sxɛsxo
ɲuɲɟu
ʂtʰæʂtʰə
tsʰəsji
bobo
xɑkõdʑə
popo
jə
scɑndʑə
təscɑ̃
tərgɑ̃
ʐgɑ̃dʑə
qʰætəsɬõ
səm
sfo
ʁnɛʁnɛ
ʁnɛkuku
kəsfo
ɕõsnæ
ŋkʰu
tənəvsə
xɕɛʂ
xtsʰə
tsʰə
heʑi
ɣzəɣzə
æχe
ætsʰe
hævʑi
hɑji
mjɛmmjɛm
xovɛxovɛ
ɲæ
ʁjə
pæbə
ɕɑ̃dʑu
kʰətæ
skʰʂo
pukʰu
bæɴɢe

to twirl
to move things
left
right
to think
zhaoji
to know
pocket
to say
afraid
afraid
to like
to want
glad
mood
bright
dark
dark
bright
forest
owl
to sink
to float, flow
earth soil
salt
how many
some
a little
a little
yet still
yet still
smooth
a minute ago
fish
fish
insect
worm
dog
ant
mosquito
fly n

V
V
ADJ
V
V

ADJ
ADJ
V
V
ADJ
N
ADJ
ADJ
ADJ
ADJ
N
N
V
V
N
N

ADJ
N
N
N
N
N
N
N
N

168

Num
439.
440.
441.
442.
443.
444.
445.
446.
447.
448.
449.
450.
451.
452.
453.
454.
455.
456.
457.
458.
459.
460.
461.
462.
463.
464.
465.
466.
467.
468.
469.
470.
471.
472.
473.
474.
475.
476.
477.
478.
479.

Lexical

Phonetic

English Gloss

Category

mpʰri
krəmbə
zdɛrmoŋ
rŋæmɑ
ɣmur
vɕoxpæ
vɕævɟo
zgəŋæ
bjo
dʑuæ
kɛ-cʰɛ
kɛ-de
kɛ-dʑi
rɣirɣi
kɛ-ɣɮæ
xtsʰoxtsʰo
pəpə
tsʰukʰæ
ndjindji
rŋərŋə
rɲərɲə
sŋuscæ
pʰrupʰru
ɲæɲæ
ɕu-kɛcʰɛ
ɕu-kede
kɛ-ndʑu
tʰɑvtɕɑ
kɛ-ndʑɛm
kɛ-rgi
ɲoŋpæ
xsærpæ
kɛ-ji
xtsoŋma
tʂəmɑ
ŋosti
ʁoɲu
rcuqu
tɕʰæ
kɛdʑɛdʑi
və

mpʰʂi
kʐəmbə
zdɛrmõ
ʐŋæmɑ
ɣmuʂ
fɕoxpæ
fɕævɟo
zgəŋæ
bjo
dʑuæ
kɛcʰɛ
kede
kɛdʑi
ʐɣiʐɣi
kɑɣɮæ
xtsʰoxtsʰo
pəpə
tsʰukʰæ
ndjindji
ʐŋəʐŋə
rɲərɲə
sŋuscæ
pʰʂupʰʂu
ɲæɲæ
ɕukɛcʰɛ
ɕukede
kɛndʑu
tʰɑftɕɑ
kɛndʑɛm
kɛrgi
ɲoŋpæ
xsærpæ
kaji
xtsoŋma
tʂəmɑ
ŋosti
ʁoɲu
ʂcuqu
tɕʰæ
kɛdʑɛdʑi
və

snake
horn
claw
tail
beak, mouth
wing
feather
nest
to fly
to swim
big
small
long
short
thin, flat
thin like thread
thin and flat
colour
red
green
yellow
blue
white
black
strong
weak
good
bad
soft
hard
old
new
pretty
clean
dirty
front
back
between
on
far
under

N
N
N
N
N
N
N
N
V
V
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
N
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
N

169

Num
480.
481.
482.
483.
484.
485.
486.
487.
488.
489.
490.
491.
492.
493.
494.
495.
496.
497.
498.
499.
500.
501.
502.
503.
504.
505.
506.
507.
508.
509.
510.
511.
512.
513.
514.
515.
516.
517.
518.
519.
520.

Lexical

Phonetic

English Gloss

tʰɑdʑi
tʰɑni
kɛ-mtʰu
ʁmæʁmæ
kɛ-nəv
pəpə
rɟækɛcʰɛ
toχtoχ
ætɕɛ
æsɛ
stoŋbæ
ɟirə
mɲæ-rə
ndʐændʐæ
ʑuʑə
mɲə
tʰɛ
ʁərʁər
kɛ-skvɛ
me-skvɛ-rə
ʁɟi
toŋ
qoqo
kɛdʑoŋ
nə-norə
tə-ŋe-sə
rkæ-rə
ʁje-rə
xopi
sɛpʰjo
rəkɛ
xɑji
xævʑi
ɣdæmbæ
tɕʰəkɛ
kʰɛsɛr
kudʑən
toŋskə
tɑdʐu
ndʑætsə
tɕʰəpæ

tʰɑdʑi
tʰɑni
kɛmtʰu
ʁmæʁmæ
kɛnəv
pəpə
rɟækɛcʰɛ
toχtoχ
ætɕɛ
æsɛ
stoŋbæ
ɟiʐə
mɲæʐə
ndʐændʐæ
ʑuʑə
mɲə
tʰɛ
ʁərʁər
kɛskvɛ
meskvɛʐə
ʁɟi
toŋ
qoqo
kɛdʑoŋ
nənoʐə
təŋesə
ʂkæʐə
ajeʐə
xopi
sɛpʰjo
ʐəkɛ
xɑji
xævʑi
ɣdæmbæ
tɕʰəkɛ
kʰɛsɛʂ
kudʑən
toŋskə
tɑdʐu
ndʑætsə
tɕʰəpæ

far
near
high
low
deep
shallow
wide
narrow
together
full
empty
in
not
same
can, able to
can, able to
have
round
sharp
blunt
hole
hole
indent
straight
smell
correct
beautiful
pretty
table
direction
and
also
also
because
but
cloth
satin
thread
silk
woollen cloth
clothing

Category

ADJ
ADJ
ADJ
ADJ
ADJ
A
ADJ
ADJ

ADJ
V
V
V
ADJ
ADJ
ADJ
N
N
N
ADJ
N
ADJ
ADJ
N
N

N
N
N
N
N
N

170

Num
521.
522.
523.
524.
525.
526.
527.
528.
529.
530.
531.
532.
533.
534.
535.
536.
537.
538.
539.
540.
541.
542.
543.
544.
545.
546.
547.
548.
549.
550.
551.
552.
553.
554.
555.
556.
557.
558.
559.
560.
561.

Lexical

Phonetic

English Gloss

Category

tsəgə
kʰɛrcoŋ
kuvæ
vɮæzæ
zuzu
tʰoŋkʰu
smɛntɕʰɛv
χodzu-pɑre
tɕi
tʰoŋkʰuskur
ndərci
ntʰævæ
skoŋskrə
skoŋski
rkombjo
ɣzi
ndʑərtæ
zʁəʑjæ
nərbə
pjərə
loŋtʰu
pʰjɑmdə
ŋgəjɑ
ɮɛki
coɣdən
ɲəʁzɛn
pʰru
wur
ʁɮəstɛn
mbɛr
tɕeʁoŋ
rtso
jəʁo
ɲəvə
mkʰu
vɛko
rɣi-ru-re
tsʰɛko
ɣəræ-ru-re
zdi
coŋ

tsəgə
kʰɛʂcõ
kuvæ
vɮæzæ
zuzu
tʰoŋkʰu
smɛntɕʰɛv
χodzupɑre
tɕi
tʰoŋkʰuskuʂ
ndəʂci
n̥tʰævæ
skõskrə
skõski
ʂkombjo
ɣzi
ndʑəʂtæ
zʁəʑjæ
nərbə
pjəʐə
loŋtʰu
pʰjɑmdə
ŋgəjɑ
ɮɛki
coˠdən
ɲəʁzɛn
pʰʂu
wuʂ
ʁɮəstɛn
mbɛʂ
tɕeʁõ
ʂtso
jəʁo
ɲəvə
mkʰu
vɛko
ʐɣiʐuʐe
tsʰɛko
ɣəræʐuʐe
zdi
coŋ

clothing
clothing
collar
sleeve
button
pants
skirt
cloth worn on head
hat
belt for pants
girdle for coat
decorative apron
puttee
puttee
sock
shoe
boot
comb
treasure
coral
earring
necklace
ring
bracelet
wool blanket
wool for wearing
wool cloth
pillow
sleeping mat
cushion
kitchen
floor
upstairs
downstairs
cowshed
pigsty
stable
sheepfold
chicken pen
stone wall
clay wall

N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N

171

Num
562.
563.
564.
565.
566.
567.
568.
569.
570.
571.
572.
573.
574.
575.
576.
577.
578.
579.
580.
581.
582.
583.
584.
585.
586.
587.
588.
589.
590.
591.
592.
593.
594.
595.
596.
597.
598.
599.
600.
601.
602.

Lexical

Phonetic

English Gloss

Category

rɟæ
pubæ
æŋæ-ze
rgɛvʑə
vədæ-rgɛvʑə
pəŋæ
smi
pʰusɛr
musɛr
ʑələze
təvdəze
smi-ze
lɛjɛ
pətʰoŋ
pərɟi
pərɟə
ækʰə-i-vədæ
æne-i-dzəvə
æʑu-i-vədæ
æjæ-i-dzəvə
pəqe
mɑxe
ʁjɑ
mdʐu
xə
ʁjɑ-ze
tɕʰɛʁloŋ
ŋəmɛ
ʁɲæ
rcɑxpɑ
mupʰɑ
rji
rekwe
rɛmɛmæzə
rtɛpu
rɛmɛ
zivæ
tom
tom-ɲæɲæ
xsæ
rŋæma

ʐɟæ
pubæ
æŋæze
ʐgɛvʑə
vədærgɛvʑə
pəŋæ
smi
pʰusɛ˞
musɛʂ
ʑələze
təvdəze
smeze
lɛjɛ
pətʰõ
pərɟi
pərɟə
ækʰəivədæ
æneidzəvə
æʑuivədæ
æjæidzəvə
pəqe
mɑxe
ʁjɑ
mdʐu
xə
jɑze
tɕʰɛʁlõ
ŋəmɛ
ʁɲæ
ʂcɑxpɑ
mupʰɑ
rji
ʐekwe
ʐɛmɛmæzə
ʂtɛpu
ʐɛmɛ
zivæ
tom
tomɲæɲæ
xsæ
ʐŋæma

Chinese
Tibetan
baby
old man
old woman
young man
young woman
young man
young woman
boy
boy
girl
daughter-in-law
son-in-law
grandchild
grandchild
paternal uncle’s wife
paternal aunt’s husband
maternal uncle’s wife
maternal aunt’s husband
huang niu
water buffalo
male yak
wild yak
pian niu
calf
bull
cow female
cow dung
excrement
hoof
horse
foal
mother and foal
stallion
mare
mane
panda
black bear
snow leopard
horsetail

N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N

172

Num
603.
604.
605.
606.
607.
608.
609.
610.
611.
612.
613.
614.
615.
616.
617.
618.
619.
620.
621.
622.
623.
624.
625.
626.
627.
628.
629.
630.
631.
632.
633.
634.
635.
636.
637.
638.
639.
640.
641.
642.
643.

Lexical

Phonetic

English Gloss

Category

rzo
stɑ
ʑəzɛv
rbu
pæbəle
ndzɛʁloŋ
lærɲə
sɲulu
tɕʰərtsʰɛl
ɟɛzo
læχɑpe
rtɕɑtsʰɛl
rdɑŋrgoŋ
bətsʰɛl
ʁɮərko
lɛpu-rɲərɲə
pɛlrtsʰæ
rmo
tsoŋ
skə
rdokə
ndæsɲu
æɕəm
tsʰæsnəχ
tsʰɑxkɛr-kede
snəmtsʰɛl
tsoŋrəl
joskə
tɕɛlɛv-dʑoŋ
tɕɛləv-ɲu
sɲurbu
moxkɛr
kov-ɲu
tɕɑzgæ
ʑəjoŋ
ʑoŋər
doləmæ
ɬokəv
kəxkɛr
xotsʰɛv-rŋərŋə
xotsʰɛv-ndindi

ʐzo
stɑ
ʑəzɛv
ʐbu
pæbəle
ndzɛˤlõ
lærɲə
sɲulu
tɕʰərtsʰɛl
ɟɛzo
læχɑpe
ʂtɕɑtsʰɛl
ʐdɑ̃ʐgõ
bətsʰɛl
ʁɮəʂko
lɛpurɲərɲə
pɛlʂtsʰæ
ʐmo
tsoŋ
skə
ʐdokə
ndæsɲu
æɕəm
tsʰæsnəχ
tsʰɑxkɛrkede
snəmtsʰɛl
tsõʐəl
joskə
tɕɛlɛvdʑõ
tɕɛləfɲu
sɲuʐbu
moxkɛʂ
kofɲu
tɕɑzgæ
ʑəjõ
ʑoŋə˞
doləmæ
ɬokəv
kəxkɛʂ
xotsɛvʐŋəʐŋə
xotsʰɛvndindi

leopard
tiger
turtle
bee
butterfly
earth
asparagus lettuce
pea
cauliflower
potato
cabbage
celery
tomato
water spinach
bamboo shoot
carrot
lotus root
oyster mushroom
scallion
chives
cucumber
asparagus
corn
spinach
small chinese cabbage
cole
onion
garlic
mung bean sprout
bean sprout
green pea
white fungus
garlic shoot
ginger
yam
sweet potato
eggplant
pumpkin
white gourd
green pepper
chili pepper

N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N

173

Num
644.
645.
646.
647.
648.
649.
650.
651.
652.
653.
654.
655.
656.
657.
658.
659.
660.
661.
662.
663.
664.
665.
666.
667.
668.
669.
670.
671.
672.
673.
674.
675.
676.
677.
678.
679.
680.
681.
682.
683.
684.

Lexical

Phonetic

English Gloss

Category

kʰækəv
ʑele
moroɣ
ɕi
ɮi
mbre
ɣvə
tətkɛr
vɛtɛm
sɲu
rɟætɕæ
ɑrɑ
ɲəmdʑə
kə-məsə
ædʑæpædʑæ
rtsʰɛv
xoscæ
scərpə
spəvji
ɣmɛ
ɣmɛɕu
pətɛr
vɮɛ-ze
jæɴqjo
mi
rgɛrtsʰo
ɮɛvɮi
ɮɛɲu
ɮæbjænoŋ
rtsʰe
nene
væ
vɛ-ze
ɣəræ
pʰuɣər
muɣər
ɣəræ-ze
pʰupʰɑ
mupʰɑ
mdʐu
spjoŋkʰə

kʰækəv
ʑele
moʐoɣ
ɕi
ɮi
mbʐe
ɣvə
tətkɛʂ
vɛtɛm
sɲu
rɟætɕæ
ɑʐɑ
ɲəmdʑə
kəməsə
ædʑæpædʑæ
ʂtsʰɛv
xoscæ
scəʂpə
spəvji
ɣmɛ
ɣmɛɕu
pətɛʂ
vɮɛze
jæɴqjo
mi
ʐgɛ˞tsʰo
ɮɛvɮi
ɮɛɲu
ɮæbjænõ
ʂtsʰe
nene
væ
veze
ɣəʐæ
pʰuɣə˞
muɣəʂ
ɣəræze
pʰupʰɑ
mupʰɑ
mdʐu
spjoŋkʰə

bitter gourd
turnip
black fungus
highland barley
wheat
rice
oats
sesame
peanuts
beans for pigfeed
looseleaf tea
wine
sunflower seeds
cooked rice
sandals
huajiao
hot pepper sauce
soy sauce
sores
wound
scar
rash measles
uvula
palate
mole
spine
wristbone
back of hand
palm
lung
breast
pig
piglet
chicken
cock
hen
chick
male pig
sow
dragon
wolf

N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N

174

Num
685.
686.
687.
688.
689.
690.
691.
692.
693.
694.
695.
696.
697.
698.
699.
700.
701.
702.
703.
704.
705.
706.
707.
708.
709.
710.
711.
712.
713.
714.
715.
716.
717.
718.
719.
720.
721.
722.
723.
724.
725.

Lexical

Phonetic

English Gloss

Category

rtsɛ
rærgo
bæɴɢe
ɑtsɑwɑtsɑ
loŋbutɕe
səŋgi
sqevcə
rŋæmoŋ
spəɲcʰɛr
weqe
tsələ
vcə
ɣzəmdə
tʰoŋkæ
skərvæ
mtɕʰurtin
ŋkʰurlu
toŋkʰur
pʰjarɟɛv
zɮæ
tærtʰɑχ
mdəʁoŋ
mkʰɛr
ɮəɮæ
loŋcʰə
zjɛ
ɬækoŋ
ɟɛzo
ɣrə
rdzærə
tʂʰeɕɑŋ
ɣrəkɛnəv
ʁɑ
ʁzəmkʰri
ɲɛkʰri
ŋkʰvo
rdəqu
ɣɮəli
ntʰətæ
ɕɑɣrə
spru-rə

ʂtsɛ
ʐæʐgo
bæɴɢe
ɑtsɑwɑtsɑ
loŋbutɕe
səŋgi
sqefcə
ʐŋæmõ
spəɲcʰɛʂ
weqe
tsələ
fcə
ɣzəmdə
tʰoŋkæ
skərvæ
mtɕʰuʂtɪn
ŋkʰu˞lu
toŋkʰuʂ
pʰjarɟɛv
zɮæ
tæʂtʰɑχ
mdəʁõ
mkʰɛʂ
ɮəɮæ
lõcʰə
zjɛ
ɬækõ
ɟɛzo
ɣʐə
ʐdzæʐə
tʂʰeɕɑŋ
ɣʐəkɛnəv
ʁɑ
ʁzəmkʰʂi
ɲɛkʰʂi
ŋkʰfo
ʐdəqu
ɣɮəli
n̥tʰətæ
ɕɑɣʐə
spʐuʐə

deer
Tibetan antelope
spider
locust
elephant
lion
squirrel
camel
frog
rabbit
cat
rat mouse
monkey
thangka
turn prayer wheel
tower
prayer wheel
push prayer wheel
prayer position
to say mantras
prayer flags
monastery
rGyalrong stone tower
to plow
to plow
to hoe
lama’s house
potatoes
waterfall
peak of mtn
wagon
flood
window
bed for lamas
bed
key
mortar bowl
pestle
to grind
small sickle
butter churn

N
N
N
N
N
N
N
N
N
N
N
N
N
N
V
N
N
V
V
N
N
N
V
V
V
N
N
N
N
N
N
N
N
N
N
N
V
N
N

175

Num
726.
727.
728.
729.
730.
731.
732.
733.
734.
735.
736.
737.
738.
739.
740.
741.
742.
743.
744.
745.
746.
747.
748.
749.
750.
751.
752.
753.
754.
755.
756.
757.
758.
759.
760.
761.
762.
763.
764.
765.
766.

Lexical

Phonetic

English Gloss

Category

spru
zəsqræ
ndərjæ
rzæqə
tutu
pʰəru
pərzi
tsʰɛtɕə
kʰɑmbo
spu
koŋ-kɛcʰɛ
koŋ-kede
zuqʰi
mi-ʁjerə
kærkæ
kɛ-jɛ
mə-cʰɛcʰɛ
cʰɛ-rə
tsʰosusu
kɛ-ntɕʰi-rə
mdzemdze
qʰæsɬo-ndʑə
səmsci-rə
rgɑn-rə
me-rgɑn-rə
qʰæ-me-sɬurə
səm-me-scirə
stʰvæ
ʁo
ɣwæ
vkə
mɬe
ʁəpræ
ʁəmɬe
tsəntʰɛv
tʂæ
təvæ-vtʰi
xɕɛt
zbəcæ
xcʰi
zbəqe

spʐu
zəsqræ
ndərjæ
ʐzæqə
tutu
pʰəʐu
pərzi
tsʰɛtɕə
kʰɑmbo
spu
koŋkɛcʰɛ
koŋkede
zuqʰi
miˤjeʐə
kæʂkæ
kɛjɛ
məcʰɛcʰɛ
cʰɛʐə
tsʰosusu
kɛntɕʰiʐə
mdzɛmdze
qʰæsɬõdʑə
səmsciʐə
rgɑndʑə
meʐgɑndʑə
qʰæmesɬuʐə
səmmesciʐə
stʰvæ
ʁo
ɣwæ
fkə
mɬe
ʁəpræ
ʁəmɬe
tsəntʰɛv
tʂæ
təvæftʰi
xɕɛt̚
zbəcæ
xcʰi
zbəqe

to churn
broom
to sweep
basket carried on back
basket carried on back
baset carried on back
small knife
cleaver
bag
incense stick
expensive
cheap
ugly
ugly
difficult
easy
busy
idle
idle
good-sounding (haoting)
polite
happy
happy
happy
sad
sad
sad
to press down
to help
to hug
to have eaten ones fill
to braid
braid
braid
scissors
to cut with scissors
to smoke
to whip
to hit
to puncture
to urge

V
N
V
N
N
N
N
N
N
N
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
ADJ
V
V
V
V
V
N
N
N
V
V
V
V
V
V

176

Num
767.
768.
769.
770.
771.
772.
773.
774.
775.
776.
777.
778.
779.
780.
781.
782.
783.
784.
785.
786.
787.
788.
789.
790.
791.
792.
793.
794.
795.
796.
797.
798.
799.
800.
801.
802.
803.
804.
805.
806.
807.

Lexical

Phonetic

English Gloss

Category

xtɕəlæ
ʑæ
ndjɛvji
xorvæ
qur
rgurgu
rɲi
ʁɲæ
ʁəvdæ
zŋo
stʰi
gon-rə
mə-gon-rə
ku-rə
mu-ku-rə
zɮæ
vræ
sɲiʁə
χjə
pʰrɛ
rkurɟɛv
jələ
rə-rŋutsu
səli
ʁɮi
vdʐər
vge
tɕəsco
scævæ
rə-mu
kedi-zu-rə
æŋæze-zu-rə
ɮɛ
ɕə
tsʰu
qrə
zu
ɣʑi
ndʑi
vrə
ɣʑæ-rə

xtɕəlæ
ʑæ
ndjɛvji
horvæ
quʂ
ʐguʐgu
rɲi
ʁɲæ
ʁəvdæ
zŋo
stʰi
gõndʑə
məgõndʑə
kuʐə
mukuʐə
zɮæ
vʐæ
sɲiʁə
χjə
pʰʂɛ
ʂkurɟɛv
jələ
ʐəʐŋutsu
səli
ʁɮi
vdʐə˞
vge
tɕəsco
scævæ
ʐəmu
kedizuʐə
æŋæzezuʐə
ɮɛ
ɕə
tsʰu
qʐə
zu
ɣʑi
ndʑi
vʐə
ɣʑæʐə

to twist
to limp
to doze
to snore
to snore
to burp
to wait
to wait
to nod
to hang
to hang
to understand
not understand
to understand
not understand
to read
to write
pen
to snap stick
to snap thread
stamp
saying, expression
to kneel
to roll
to roll
to roll up
to cross (bridge)
to paddle
paddle n
to draw
to be pregnant
to be pregnant
to return
to leave
to milk cow
female yak
female pian niu
to teach
to learn
to pour
to make sound

V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
N
V
V
N
N
V
V
V
V
V
V
N
V
V
V
V
V
V
N
N
V
V
V
V

177

Num
808.
809.
810.
811.
812.
813.
814.
815.
816.
817.
818.
819.
820.
821.
822.
823.
824.
825.
826.
827.
828.
829.
830.
831.
832.
833.
834.
835.
836.
837.
838.
839.
840.
841.
842.
843.
844.
845.
846.
847.
848.

Lexical

Phonetic

ŋɑrɟɛr
ŋɑrɟɛʂ
ŋəzu
ŋəzu
prilærə
pʐilærə
mdʑu-rə
mdʑuʐə
metʰɛv
metʰɛv
tʰævkæ
tʰæfkæ
zoŋroŋ
zõʐõ
zdɛr
zdɛʂ
qʰəzi
qʰəzi
qʰəzi-kede
qʰəzikede
tʂʰɑtʂoŋ
tʂʰɑtʂõ
ɕɛr-qʰəzi
ɕɛrqʰəzi
mkʰə-rjɛ
mkʰərjɛ
zɑmɑsotɕi
zɑmɑsotɕi
tsʰædəm
tsʰædəm
tɕæ
tɕæ
tɕændɛvlə
tɕændɛvlə
zætər-kʰəre
zætəʂkʰəre
zætər
zætəʂ
tɕoji
tɕoji
tærqæ
tæʂqæ
vivəx
vivəx
mbre-zʁejo
mbʐezʁejo
lɛvdɛr
lɛvdɛʂ
rnæɲi
rnæɲi
ʁmɛʑi
ʁmɛʑi
zoŋqo
zõqo
tʰoŋbe
tʰoŋbe
ɕɛr-tæmbə
ɕɛrtæmbə
ɕɛr
ɕɛʂ
tɕutɕæ
tɕutɕæ
xtɕər-scɛ
xtɕərscɛ
rʁɛrʁæ-scɛ-pɑre rʁɛrʁæscɛpɑre
tɕʰəvkɑ
tɕʰəfkɑ
kindækindæ
kindækindæ
sco
sco
rtɕʰu
ʂtɕʰu
tæmbə
tæmbə
ndəʁərtse
ndəʁərtse
snəm-rʁɛ-scɛ
snəmrʁɛscɛ
kə-xcʰi
kəxcʰi

English Gloss

Category

to roar
to howl
to whinny
to moo
stove
stove
basin
plate
bowl
small bowl
cup, mug
glass
pipe
strainer
pitcher, thermos
tea
tea leaves
container for chopsticks
chopsticks
spoon
ladle
pressure cooker
rice cooker
peeler
pot w handles
pot without handles
decorated pot
pot
glass bottle
glass material
metal
clip, pin
cloth for wiping
tap
dripping
scoop for water
bottle for making wine
bottle
laundry detergent
dish detergent
to turn on

N
V
V
V
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
V
N
N
N
N
N
V

178

Num
849.
850.
851.
852.
853.
854.
855.
856.
857.
858.
859.
860.
861.
862.
863.
864.
865.
866.
867.
868.
869.
870.
871.
872.
873.
874.
875.
876.
877.
878.
879.
880.
881.
882.
883.
884.
885.
886.
887.
888.
889.

Lexical

Phonetic

English Gloss

Category

nə-sqʰi
kæntɕʰæ
ʁɮi
tɕæmtsʰæ
ʁdu
ɕədʑutʰoŋtʰoŋ
ɕə-rʁɛ-scɛ
rŋæ-rʁɛ-scɛ
ʁə-rʁɛ-scɛ
lɑɕəv
ɕɛtʰipɑku
xɑjoŋbɛtɕo
vtɕækʰæ-zŋo-re
mɑrnɑ
ɕovə-rgɛm
ŋæmbæ
ɣrə-rgɛm
tɕʰəkəv
ɲəmdʑə
ɲəmæmɛto
tɕitɕæ
rtu
rtɕɑrtɑ
ʁzɛn
ɕɛntʰəv
rŋɑ-rʁe-pɑre
tɕʰutsʰu
rdirdæ
vɬɛvɬɛpʰɛre
ɲɟɛrdo
tə-ɲɟɛrə-sə
æ-mpʰæɬinə-re
nə-mpʰæɬinə-re
nə-mbre
qʰre
ɲɛr
xsɛv
xsər
rɟæmæ
vcʰə
stʰjæ

nəsqʰi
kæntɕʰæ
ʁɮi
tɕæmtsʰæ
ʁdu
ɕədʑutʰõtʰõ
ɕərʁɛscɛ
ʐŋærʁɛscɛ
ʁərʁɛscɛ
lɑɕəv
ɕɛtʰipɑku
xɑjõbɛtɕo
ftɕækʰæzŋore
mɑrnɑ
ɕovəʐgɛm
ŋæmbæ
ɣʐəʐgɛm
tɕʰəkəv
ɲəmdʑə
ɲəmæmɛto
tɕitɕæ
ʂtu
ʂtɕɑʂtɑ
ʁzɛn
ɕɛntʰəv
rŋɑrʁepɑre
tɕʰutsʰu
rdirdæ
fɬɛfɬɛpʰɛre
ɲɟɛrdo
təɲɟɛrəsə
æmpʰæɬinəʐe
nəmpʰæɬinəʐe
nəmbʐe
qʰʂe
ɲɛʐ
xsɛv
xsəʂ
rɟæmæ
fcʰə
stʰjæ

to turn off
rolling pin
to roll out
tea strainer
wood pail
pail
toothbrush
face wash
shampoo
rubber gloves
large plastic bag
kettle
rack hanging things
oil
cardboard
wood block
sink
watermelon
sunflower seed
sunflowers
shell, rind
to cut hair
bike
lama’s clothes
other's clothes
wash cloth
watch
garbage
garbage can
change
to change
insideout
insideout
to pull down
to pull down
to taste
to repay
to stirfry
scale
to weigh
to support prop up

V
N
V
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
V
N
N
N
N
N
V
N
N
V

V
V
V
V
V
N
V
V

179

Num
890.
891.
892.
893.
894.
895.
896.
897.
898.
899.
900.
901.
902.
903.
904.
905.
906.
907.
908.
909.
910.
911.
912.
913.
914.
915.
916.
917.
918.
919.
920.
921.
922.
923.
924.
925.
926.
927.
928.
929.
930.

Lexical

Phonetic

English Gloss

Category

scor
kə-xæ
tə-xæ
ndzəndzə-vɕɛ
vdʑə
ŋe-rə
mɑ-ŋe-rə
ɣrə ndʑədʑə
ɣrətoŋ
zɮæræ
mtsʰi
rɛvræ
vtsʰu
rcʰɛ
tə-rcʰɛ
tʰəvæ
tɕonzɛr
rdæ
tə-pʰo
ŋkʰrɛ
ŋkʰrɑtɕɑ
zjærkʰu
səmnɛ
tɕʰæpæ
vsu
mpʰi
vlɛ
nə-ri
rʁɑ
rʁɑmə
tsə
xtʰəxtʰə
ɕuɕu
pʰrə
sɬə
lɑndʑə
xtsʰə
ɲɟɑbɑ
ndzure
tsəgə-tʰi-re
tɕuʁɑ

scoʂ
kəxæ
təxæ
ndzəndzəfɕɛ
vdʑə
ŋeʐə
mɑŋeʐə
ɣʐə ndʑədʑə
ɣʐətoŋ
zɮæræ
mtsʰi
ʐɛvʐæ
ftsʰu
ʂcʰɛ
təʂcʰɛ
tʰəvæ
tɕonzɛʐ
rdæ
təpʰo
ŋkʰʂɛ
ŋkʰʂɑtɕɑ
zjæʂkʰu
səmnɛ
tɕʰæpæ
fsu
mpʰi
vlɛ
nəʐi
rʁɑ
ʐʁɑmə
tsə
xtʰəxtʰə
ɕuɕu
pʰʂə
sɬə
lɑndʑə
xtsʰə
ɲɟɑbɑ
ndzure
tsəgətʰire
tɕuʁɑ

to support
to come out
to come out
to whisper
friend
to be okay
to not be okay
to draw water
well (n.)
to winnow
to lead guide
to thresh
to take out of water
to bite
to have been bitten
hammer
nail
to hammer
to lose
to shake something
to shiver
to hurt emotionally
worry
to punish
to spin (wool)
to card
to put s.t.
to add
to be crazy
crazy person
to rot
behind
behind
tangled
stairs
railing
dirt
mud
chair
clothesline
gate

V
V
V
V
N
V
V
N
V
V
V
V
V
V
N
N
V
V
V
V
V
N
V
V
V
V
V
V
N
V

N
N
N
N
N
N
N

180

Num
931.
932.
933.
934.
935.
936.
937.
938.
939.
940.
941.
942.
943.
944.
945.
946.
947.
948.
949.
950.
951.
952.
953.
954.
955.
956.
957.
958.
959.
960.
961.
962.
963.
964.
965.
966.
967.
968.
969.
970.
971.

Lexical

Phonetic

English Gloss

Category

ŋkʰvu-mɛ
ŋkʰvu-cʰi-scɛ
ŋkʰvo
tsʰoŋkoŋ
zɛkʰoŋ
ræpʰi
tɕækoŋ
smən-zjəre
mɛto-zjəre
pjozdɛr
ɣrəɲcʰɛr
ʁə-vəvæ-re
ʁə-rʁɛ-re
sɬopræ
tɕi-zjəre
sɬomɑ
rgɛrgən
xseskə-zjəre
xseskə
xændʑe
snopdʑə
mtɕʰutɕu
mpʰrivæ
zɣɑrtɕɑ
tɕədəzjəre
tɕədə
zdɛr
mdɛrə
ɕoɣdu
pʰjɛr
nə-pʰiv
ɕɛmnu-zjəre
ɕɛmnu
skərjo
kʰɛdɛr
spusnəre
pʰərwæ
tʂərvə
sqədi
ʁzə
jo

ŋkʰfumɛ
ŋkʰfucʰiscɛ
ŋkʰfo
tsʰõkõ
zɛkʰõ
ʐæpʰi
tɕækõ
smənzjəre
mɛtozjəre
pjozdɛʂ
ɣʐəɲcʰɛʐ
ʁəvəvæʐe
ʁərʁɛre
sɬopræ
tɕizjəre
sɬomɑ
rgɛrgən
xseskəzjəre
xseskə
xændʑe
snopdʑə
mtɕʰutɕu
mpʰʂivæ
zɣɑʂtɕɑ
tɕədəzjəre
tɕədə
zdɛʂ
mdɛʐə
ɕoˠdu
pʰjɛʂ
nəpʰɪv
ɕɛmnuzjəre
ɕɛmnu
skərjo
kʰɛdɛʂ
spusnəre
pʰərwæ
tʂə˞və
sqədi
ʁzə
jo

lock
key
lock and key
shop
restaurant
mahjong
sign
pharmacy
flower shop
drip outside
puddle
hair salon
hair salon
university
hat shop
student
teacher
buddhist shop
buddha
scarf
hanging
tassel
prayer beads
whip
bookshop
book
drip in house
drum
umbrella
to open unfurl
to close
glasses shop
glasses
decorative blanket
scarf
incense lighter
decorative dagger
bell
to ring
bead with religious significance
turquoise

N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
V
V
N
N
N
N
N
V
N
N

181

Num

Lexical

Phonetic

English Gloss

972.
973.
974.
975.
976.
977.
978.
979.
980.
981.
982.
983.
984.
985.
986.
987.
988.
989.
990.
991.
992.
993.
994.
995.
996.
997.
998.
999.
1000.
1001.
1002.
1003.
1004.
1005.
1006.
1007.
1008.
1009.
1010.
1011.
1012.

pjərə
mtɕʰurdzæ
ɣzi-zjəre
ʁdu
rqwarzo
rmævɕædʑo
zondo
təvæ-ɑrɑ-zjəre
pɛr
skəpɛr
kʰri
rzelo
zbræ
kɛ-rjɛ
qʰegəɟirə
sɬəkʰro
æ-bɛr
nə-bɛr
xsə
rɑtɑ
mbo
tɕʰæɣwə
vəɣvə
səqə
sərʁə
rtsæmbræ
rŋəbo
ʁdʐu
kʰətsʰi
xtɕærʁæ
cʰəpʰɛv
wərdzə
rtɛvrtɛv
qʰræqʰræ
rɑtɑdʑugəcɑŋ
rɟæræ
ɕoŋkʰæspusnəre
zoŋkʰæspusnəre
rɑkʰæspusnəre
ɮəɮæ
zuŋu

pjəʐə
coral
mtɕʰurdzæ
teapot
ɣzizjəre
shoe shop
ʁdu
circular prayer flag
ʂqwarzo
necklace
rmæfɕædʑo
peacock feather
zondo
horn
təvæɑrɑzjəre
tobacco and wine shop
pɛʂ
photo
skəpɛʂ
photo of lama
kʰʂi
chair
ʐzelo
with lifted skirts
zbʐæ
tent
kɛrjɛ
ceramic
qʰegəɟiʐə
raining
sɬəkʰʂo
step (n.)
æbɛʂ
first step
nəbɛʂ
second step
xsə
to tighten
ʐɑtɑ
mill
mbo
leather bowl
tɕʰæɣwə
mill stone
vəɣvə
stationary mill stone
səqə
small piece machinery
sərʁə
rotary paddle
ʂtsæmbræ
bowl catches tsampa
ʐŋəbo
roasted barley
ʁdʐu
tsampa
kʰətsʰi
water channel
xtɕærʁæ
wood gate at mill
cʰəpʰɛv
stick on millstone
wərdzə
yak tail
ʂtɛfʂtɛv
fine
qʰʂæqʰʂæ
coarse
rɑtɑdʑugəcɑŋ
miller
rɟæræ
first floor
ɕoŋkʰæspusnəʐe wood incense box
zoŋkʰæspusnəre copper incense box
ʐɑkʰæspusnəre
bronze incense box
ɮəɮæ
to plow
zuŋu
at first

Category
N
N
N
N
N
N
N
N
N
N
N
ADV?
N
N
N
N
N
V
N
N
N
N
N
N
N
N
N
N
N
N
N
ADJ
ADJ
N
N
N
N
N
V

182

Num

Lexical

Phonetic

English Gloss

Category

1013.
1014.
1015.
1016.
1017.
1018.
1019.
1020.
1021.
1022.
1023.
1024.
1025.
1026.
1027.
1028.
1029.
1030.
1031.
1032.
1033.
1034.
1035.
1036.
1037.
1038.
1039.
1040.
1041.
1042.
1043.
1044.
1045.
1046.
1047.
1048.
1049.
1050.
1051.
1052.
1053.

ɮə
pʰutə
vɕi
vɬə-ndəv
wo
ɕirgə
tsʰərɲi
sənəm
tə-tsʰɛr
kʰɛgɛ
rə-rmi-sə
vɛrvi
rŋə
rə-tɕe
æ-sɬə
nə-sɬə
ntsʰɛm
mbjæ
ɴɢwi
tɕegə
jɛlə-gə
kilu
mbo
rŋərŋæ
tʰɛvɕi
vɕi
ɣvə
ɕɛv
tɕedə
rtsʰɛv
vcə
stʰo
rə-zjæ
nə-zjæ
mgrə
ndzæ
vlɛ
pʰæ
ʑigə
cʰɛgə
zdæmæ

ɮə
pʰutə
fɕi
fɬəndəv
wo
ɕirgə
tsʰərɲi
sənəm
tətsʰɛʂ
kʰɛgɛ
ʐəʐmisə
vɛrvi
ʐŋə
ʐətɕe
æsɬə
nəsɬə
n̥tsʰɛm
mbjæ
ɴɢwi
tɕegə
jɛləgə
kilu
mbo
ʐŋəʐŋæ
tʰɛfɕi
fɕi
ɣvə
ɕɛv
tɕedə
ʂtsʰɛv
fcə
stʰo
ʐəzjæ
nəzjæ
mgʐə
ndzæ
vlɛ
pʰæ
ʑigə
cʰɛgə
zdæmæ

field
to cover
to need
to spread seed
again
to harrow
to pull
farming
finished
after
sprout
slowly
to become green
to come up
one month
two months
between, around
to hoe
hoe
something
so-called
each, some
box for grain
bad grass
to pull out
to want
type of weed
to pick up
time
autumn
to harvest
to leave s.t.
to bring up
to bring down
wall for drying barley against
third floor room
to put, leave
half
around
to thresh by machine
tool for threshing by hand

N
V
V
V
V
V
N

V
ADJ
V
V
N
N
V
N

N
V
V
N
V
N
N
V
V
V
V
N
N
V

V
N

183

Num

Lexical

Phonetic

English Gloss

Category

1054.
1055.
1056.
1057.
1058.
1059.
1060.
1061.
1062.
1063.
1064.
1065.
1066.
1067.
1068.
1069.
1070.
1071.
1072.
1073.
1074.
1075.
1076.

zdɑvɮə
tʰɛrʁe
ɣzo
ɣroɣro
xtsoŋmɑ
vətə
mbe
ʁdʐu
pɛrvə
rəro
tə-pɛ
mɛji
pi
ʑu
pʰɑjo
və
rni
tʰutu
ætɕʰə
rə-zu
zu
pʰɛv
ɲɟɛr

zdɑvɮə
tʰɛrʁe
ɣzo
ɣʐoɣʐo
xtsoŋmɑ
vətə
mbe
ʁdʐu
pɛrvə
ʐəʐo
təpɛ
mɛji
pi
ʑu
pʰɑjo
və
ʐni
tʰutu
ætɕʰə
ʐəzu
zu
pʰɛv
ɲɟɛʂ

flapping part on threshing tool
part of hand thresher
big wood storage box barley
dry
clean
to make
to bring
to mill
container
up
to take out
butter
ball of tsampa
yogurt
together
to make
to mix
mixed together
with
to hold
to sew
to close
to change

N
N
N
ADJ
ADJ
V
V
V
N
V
N
N
N
V
V

V
V
V
V

References
Anderson, Gregory D.S. 2013. The velar nasal. In Matthew Dryer & Martin
HaspelmathWorld atlas of language structures online, , eds. Leipzig: Max Planck
Institute of Evolutionary Anthropology. Online: http://wals.info. (Accessed 21 April
2014.) Online: http://wals.info/chapter/9. (Accessed 5 September 2014.)
Chang, Betty Shafts & Kun Chang. 1974. Gyarong historical phonology. Academia
Sinica: Bulletin of the Institute of History and Philology 46.391-524.
Chirkova, Katia. 2012. The Qiangic subgroup from an areal perspective: A case study of
languages of Muli. Language and Linguistics 13.133-70.
Clements, George N. 1990. The role of the sonority cycle in core syllabification. Papers
in Laboratory Phonology, ed. by John Kingston & Mary E. Beckman, 1, 283-333.
Cambridge: Cambridge University Press.
Duanmu, San. 2008. Syllable structure: The limits of variation. Oxford & New York:
Oxford University Press.
Duō'ěrjı́. 1995. Chuanxibei zangqu geshizahua yinxi fenxi [An analysis of the sound
system of the Geshiza speech in the Tibetan area of northwestern Sichuan]. Yunnan
Minzu Yuwen 1.34-44.
Duō'ěrjı́. 1998. Daofuyu geshizahua yanjiu [A study on the Geshiza variety of the Daofu
language]. Beijing: Zhongguo Zangxue (China Tibetology) Press.
Ebert, Karen H. 1990. On the evidence for the relationship Kiranti-Rung. Linguistics of
the Tibeto-Burman Area 13.57-78.
Gatehouse, David. 2014. People and languages of Sichuan's Ethnic Corridor. Online:
http://sichuanzoulang.com/en.html. (Accessed 18 February 2014.)
Gates, Jesse P. 2012. Situ in situ: Towards a dialectology of Jiarong (rGyalrong). Langley,
BC: Trinity Western University master's thesis.
Gates, Jesse P. 2013. Intelligibility, identity, and structure in Western rGyalrongic. 3rd
Workshop on Sino-Tibetan Languages of Sichuan. Paris.
Gates, Jesse P. 2014. Review of Pelkey, Jamin R., Dialectology as Dialectic (2011). SIL
Electronic Book Reviews 003.1-15. Online: http://tinyurl.com/oe45c47. (Accessed 14
January 2015.)
184

185

Hayes, Bruce & Tanya Stivers. 2000. Postnasal voicing. UCLA, ms. Online:
http://tinyurl.com/mc2h6te. (Accessed 19 January 2015.)
Hodgson, Brian Houghton. 1853. Sifán and Hórsók vocabularies, with another special
exposition in the wide range of Mongolidan affinities and remarks on the lingual and
physical characteristics of the family. Journal of the Asiatic Society of Bengal
22.121-51. Online: http://tinyurl.com/pybjrxr.
Hodgson, Brian Houghton. 1874. Essays on the Languages, Literature, and Religion of
Nepāl and Tibet: Together with Further Papers on the Geography, Ethnology, and
Commerce of Those Countries. London: Trübner & Company. Online:
http://tinyurl.com/nszodlf.
Huang, Bufan. 1990. Daofuyu yuyin he dongci bianhua [Phonology and verb conjugation
in Daofu]. Minzu Yuwen 5.23-30.
Huang, Bufan. 1991. Daofuyu [The Daofu language]. In Qingxia Dai, Huang Bufan, Fu
Ailan, Wangmu Renzeng & Liu JuZangmianyu shiwu zhong [Fifteen Tibeto-Burman
languages], , eds. 1-45. Beijing: Yanshan Press.
Huang, Bufan. 1992. Daofu. In Qingxia Dai, Huang Bufan, Fu Ailan, Wangmu Renzeng
& Liu JuA Tibeto-Burman lexicon, , eds. Beijing: Central Institute of Minorities.
Online: http://stedt.berkeley.edu/~stedt-cgi/rootcanal.pl/source/TBL. (Accessed 5
August 2014.)
Huang, Bufan. 2003. Lawurongyu [The Lavrung language]. Minzu Yuwen 3.60-80.
Huang, Xing. 2000. Zhongguo shaoshu minzu yuyan huoli yanjiu [Studies on minority
language vitality in China]. Beijing: Central University for Nationalities Press.
Hyman, Larry. 2009. How (not) to do phonological typology: the case of pitch-accent.
Language Sciences 31.213-38.
Jacques, Guillaume. 2004. Phonologie et morphologie du Japhug (rGyalrong).
University of Paris Ph.D. dissertation. Online: http://tinyurl.com/nszodlf. (Accessed
29 April 2014.)
Jacques, Guillaume. 2014. Denominal affixes as sources of antipassive markers in Japhug
Rgyalrong. Lingua 138.1-22.
Jacques, Guillaume, Anton Antonov, Yunfan Lai & Lobsang Nima. 2013. Person marking
in Rtau. Online: http://tinyurl.com/q8qcx6w. (Accessed 14 January 2014.)
Jacques, Guillaume & Alexis Michaud. 2011. Appendix: Approaching the historical
phonology of three highly eroded Sino-Tibetan languages: Naxi, Na and Laze.
Diachronica 28.1-25. Online: http://dx.doi.org/10.1075/dia.28.4.02jac.additional.

186

Ladefoged, Peter. 2003. Phonetic data analysis: An introduction to fieldwork and
instrumental techniques. Oxford: Blackwell Publishing.
Lai, Yunfan. 2013a. La morphologie affixale du lavrung wobzi. Paris: Université
Sorbonne Nouvelle master's thesis. Online: http://tinyurl.com/q8qcx6w. (Accessed 26
February 2014.)
Lai, Yunfan. 2013b. The person agreement system of Wobzi Lavrung. Online:
http://tinyurl.com/nlfo7h3. (Accessed 21 January 2014.)
Lai, Yunfan. 2013c. The templatic morphology of Wobzi Lavrung. Online:
http://tinyurl.com/opgmf8w. (Accessed 21 January 2014.)
Lai, Yunfan. 2013d. What lies behind Wobzi consonant clusters? Online:
http://tinyurl.com/ld9sfyn. (Accessed 21 January 2014.)
LaPolla, Randy J. 2003. Overview of Sino-Tibetan morphosyntax. In Randy J. LaPolla &
Graham ThurgoodThe Sino-Tibetan languages, , eds. 22-42. London & New York:
Routledge.
LaPolla, Randy J. 2006. Sino-Tibetan languages. Encyclopedia of Languages and
Linguistics, ed. by Keith Brown, 393-6. Elsevier. Online: http://tinyurl.com/mpe5ezc.
(Accessed 29 April 2014.)
LaPolla, Randy J. 2013. Subgrouping in Tibeto-Burman: Can an individual-identifying
standard be developed? How do we factor in the history of migrations and language
contact? Language typology and historical contingency: In honor of Johanna Nichols,
ed. by Balthasar Bickel, Lenore A. Grenoble, David A. Peterson & Alan Timberlake,
465-74. Amsterdam: John Benjamins.
LaPolla, Randy J. & Chenglong Huang. 1996. A grammar of Qiang [pre-pub]. City
University of Hong Kong, ms.
Lewis, M. Paul, Gary F. Simons & Charles D. Fennig, eds. 2013. Ethnologue: Languages
of the world. 17 edn Edition. Dallas, Texas: SIL International. Online:
http://www.ethnologue.com. (Accessed 20 March 2014.)
Lin, You-Jing. 2003. Tense and aspect morphology in the Zhuokeji rGyalrong verb.
Cahiers linguistique - Asie orientale 32.245-86.
Lin, You-Jing. 2012. By no means marginal: Privative tone in Zhuokeji Rgyalrong.
Language and Linguistics 13.625-62.
Lin, Xiangrong. 1993. Jiarongyu yanjiu [A grammar of rGyalrong]. Chengdu: Sichuan
Nationality Press.

187

Liu, Huiqiang. 1989. Ergongyu de fufuyin han shengdiao [Tones and consonant clusters
in the Ergong language]. Minzu Luncong 7.196-201.
Maddieson, Ian. 2013a. Consonant inventories. In Matthew Dryer & Martin
HaspelmathWorld atlas of language structures online, , eds. Leipzig: Max Planck
Institute of Evolutionary Anthropology. Online: http://wals.info. (Accessed 21 April
2014.) Online: http://wals.info/chapter/1. (Accessed 21 April 2014.)
Maddieson, Ian. 2013b. Lateral consonants. In Matthew Dryer & Martin
HaspelmathWorld atlas of language structures online, , eds. Leipzig: Max Planck
Institute of Evolutionary Anthropology. Online: http://wals.info. (Accessed 21 April
2014.) Online: http://wals.info/chapter/8. (Accessed 4 August 2014.)
Maddieson, Ian. 2013c. Syllable structure. In Matthew Dryer & Martin HaspelmathWorld
atlas of language structures online, , eds. Leipzig: Max Planck Institute of
Evolutionary Anthropology. Online: http://wals.info. (Accessed 21 April 2014.)
Online: http://wals.info/chapter/12. (Accessed 4 August 2014.)
Maddieson, Ian. 2013d. Vowel quality inventories. In Matthew Dryer & Martin
HaspelmathWorld atlas of language structures online, , eds. Leipzig: Max Planck
Institute of Evolutionary Anthropology. Online: http://wals.info. (Accessed 21 April
2014.) Online: http://wals.info/chapter/2. (Accessed 21 April 2014.)
Mansier, Patrick. 1983. Lexique et phonology du Gyarong de Tsenla [A lexicon and
phonology of bTsanlha rGyalrong]. Paris: École des Hautes Études en Sciences
Sociales Ph.D. dissertation.
Marshall, Steven D. & Suzette Cooke. 1997. Dawu. In Steven D. Marshall & Suzette
CookeTibet outisde the TAR, Hong Kong: Alliance for Research in Tibet. Online:
http://tinyurl.com/kfqwlxs. (Accessed 20 March 2014.)
Matisoff, James A. 1990. Bulging monosyllables: Areal tendencies in Southeast Asian
diachrony. Proceedings of the Sixteenth Annual Meeting of the Berkeley Linguistics
Society volume.543-59.
Matisoff, James A. 1991. Sino-Tibetan linguistics: Present state and future prospects.
Annual Review of Anthropology 20.469-504.
Matisoff, James A. 1996. Languages and dialects of Tibeto-Burman. Berkeley: Center for
Southeast Asia Studies.
Matisoff, James A. 2003. Handbook of Proto-Tibeto-Burman: System and philosophy of
Sino-Tibetan reconstruction. Berkeley: University of California Press. Online:
https://escholarship.org/uc/item/19d79619#page-1.

188

Matisoff, James A. 2004. “Brightening” and the place of Xixia (Tangut) in the Qiangic
branch of Tibeto-Burman. In Ying-Jin Lin et al.Studies on Sino-Tibetan languages:
Papers in honor of Professor Hwang-cherng Gong on his seventieth birthday, , eds.
Taipei: Institute of Linguistics, Academia Sinica.
Nagano, Yasuhiko. 2003. Cogtse Gyarong. In Randy J. LaPolla & Graham ThurgoodThe
Sino-Tibetan languages, , eds. 469-89. London & New York: Routledge.
Nagano, Yasuhiko & Marielle Prins, eds. 2013. rGyalrongic languages database. Online:
http://htq.minpaku.ac.jp/databases/rGyalrong/. (Accessed 21 March 2014.)
Nichols, Johanna. 1996. The comparative method as heuristic. The comparative method
reviewed: Regularity and irregularity in language change, ed. by Mark Durie &
Malcolm Ross, 39-71. New York & Oxford: Oxford University Press.
Pelkey, Jamin R. 2011. Dialectology as dialectic: Interpreting Phula variation. Berlin &
New York: De Gruyter Mouton.
Prins, Marielle. 2011. A web of relations: A grammar of rGyalrong. University of Leiden
Ph.D. dissertation.
Qu, Aitang. 1984. Jiarongyu gaikuang [Outline of the rGyalrong language]. Minzu Yuwen
2.67-80.
Qu, Aitang. 1990. Jiarongyu de fangyan: fangyan huafen he yuyan shibie [rGyalrong
dialects: Issues in dialect subclassification and language recognition]. Minzu Yuwen
4:1-8, 5.37-44.
Sun, Hongkai. 1962. Qiangyu gaikuang [An outline of the Qiang language]. Zhongguo
Yuwen 12.561-71.
Sun, Hongkai. 1983. Liujiang liuyu de minzu yuyan ji qi xishu fenlei [Minority languages
of the Six River Region and their genetic classification]. Minzu Xuebao 3.99-274.
Sun, Hongkai. 1987. A brief account of my research work. Linguistics of the
Tibeto-Burman Area 10.117-25.
Sun, Hongkai. 1990. Languages of the Ethnic Corridor in Western Sichuan. Linguistics of
the Tibeto-Burman Area 13.1-31.
Sun, Hongkai. 2004. Jiarongyu zai Zangmianyuzu yuyan zhongde lishi diwei [The
position of rGyalrong among languages of the Tibeto-Burman family]. In Ying-Jin Lin
et al.Studies on Sino-Tibetan languages: Papers in honor of Professor Hwang-cherng
Gong on his seventieth birthday, , eds. Taipei: Institute of Linguistics, Academia
Sinica.

189

Sun, Jackson T.-S. 1994. Caodeng rGyalrong phonology: A first look. Linguistics of the
Tibeto-Burman Area 17.29-47.
Sun, Jackson T.-S. 2000a. Parallelisms in the verb morphology of Sidaba rGyalrong and
Lavrung in rGyalrongic. Language and Linguistics 1.161-90.
Sun, Jackson T.-S. 2000b. Stem alternations in Puxi verb inflection: Toward validating
the rGyalrongic subgroup in Qiangic. Language and Linguistics 1.211-32.
Sun, Jackson T.-S. 2003. Caodeng rGyalrong. In Randy J. LaPolla & Graham
ThurgoodThe Sino-Tibetan languages, , eds. 490-502. London & New York:
Routledge.
Sun, Jackson T.-S. 2004. Verb-stem variations in Showu rGyalrong. Studies on
Sino-Tibetan Languages: Papers in honor of Professor Hwang-cherng Gong on his
seventieth birthday, 269-96.
Sun, Jackson T.-S. 2005. Linguistic coding of generic human arguments in rGyalrongic
languages. 11th Himalayan Languages Symposium. Chulalongkorn University,
Bangkok, Thailand.
Sun, Jackson T.-S. 2007. Morphological causative formation in Shangzhai Horpa.
Bulletin of Chinese Linguistics 2.211-31.
Sun, Jackson T.-S. 2008. Tonality in Caodeng rGyalrong. Chomolangma, Demawend und
Kasbek, Festschrift für Roland Bielmeier zu seinem 65. Geburtsag, ed. by Brigitte
Huber, Marianne Volkart & Paul Widmer, 257-80. Halle (Saale): International
Institute for Tibetan and Buddhist Studies.
Sun, Jackson T.-S. & Qianzi Tian. 2013. Verb agreement in Gexi Horpa. Bulletin of
Chinese Linguistics 7.73-113.
Thomas, Frederick William. 1948. Nam: An ancient language of the Sino-Tibetan
borderland. London: Oxford University Press.
Thurgood, Graham. 1985. The 'Rung' languages: Notes on their proto-morphosyntax and
subgrouping. Acta Orientalia 46.79-99.
Thurgood, Graham. 2003. A subgrouping of the Sino-Tibetan languages: The interaction
between language contact, change and inheritance. In Randy J. LaPolla & Graham
ThurgoodThe Sino-Tibetan languages, , eds. 3-21. London & New York: Routledge.
Tournadre, Nicolas. 2008. Arguments against the concept of ‘conjunct’/‘disjunct’ in
Tibetan. Chomolangma, Demawend und Kasbek: Festschrift für Roland Bielmeier zu

190

seinem 65 Geburtstag, ed. by Brigitte Huber, Marianne Volkart & Paul Widmer,
281-308. Halle (Saale): International Institute for Tibetan and Buddhist Studies.
Wang, Stephen S. 1970. Consonantal clusters of Tibetan loanwords in sTau. Monumenta
Serica 29.631-58.
Winkler, Daniel. 2008. Yartsa gunbu (Cordyceps sinensis) and the fungal
commodification of Tibet's rural economy. Economic Botany 62.291-305.
Wolfenden, Stuart N. 1936. Notes on the Jyarung dialect of Eastern Tibet. T'oung Pao
32.167-204.
Zhou, Minglang. 2003. Multilingualism in China: The politics of writing reforms for
minority languages 1949-2002. Berlin & New York: Mouton de Gruyter.