Accelerat ing t he world's research.

The Origins of Sinitic

Scott DeLancey

Related papers Download a PDF Pack of t he best relat ed papers 

Languages of China in t heir Sout heast Asian cont ext

Hilary CHAPPELL

Sino-T ibet an Linguist ics: Crit ical Concept s in Linguist ics

Randy J. LaPolla

St rat ificat ion in t he peopling of China How far does t he linguist ic evidence mat ch genet ics and archa…
Roger Blench
 he origins of Sinitic

Scott DeLancey
University of Oregon

A persistent problem in Sino-Tibetan linguistics is that Chinese is characterized

by a mix of lexical, phonological, and syntactic features, some of which link it
to the Tibeto-Burman languages, others to the Tai-Kadai, Hmong-Mien, and
Mon-Khmer families of Southeast Asia. It has always been recognized that this
must relect intense language contact. his paper develops a hypothesis about
the nature of that contact. he language of Shang was a highly-creolized lingua
franca based on languages of the Southeast Asian type. Sinitic is a result of the
imposition of the Sino-Tibetan language of the Zhou on a population speaking
this lingua franca, resulting in a language with substantially Sino-Tibetan lexicon
and relict morphology, but Southeast Asian basic syntax.

Keywords: Chinese; Sinitic; Sino-Tibetan; Tibeto-Burman; language contact

. he problem of Sinitic

Sino-Tibetan includes the Chinese languages and a very large number – several
hundred, if we count languages at the level of distinctness which we do in Europe
(see Tournadre 2008) – of languages which are lumped together under the label
Tibeto-Burman. A basic problem of Sino-Tibetan linguistics is the dramatic typo-
logical and lexical divergence between these two putative branches of the family.
It has long been clear than an account of the formation of Chinese must account
for its strong lexical, phonological, and grammatical connections both with the
Tibeto-Burman languages to the west and with the Southeast Asian languages
to the south, Tai-Kadai, Hmong-Mien and Mon-Khmer. While basic vocabu-
lary and some reconstructible morphology clearly link Sinitic genetically to the
Tibeto-Burman languages, its basic morphosyntactic proile is the isolating SVO
type characteristic of mainland Southeast Asia rather than the agglutinating SOV
structure characteristic of Tibeto-Burman. Benedict sums up the problem:
[T]he following facts in re Chinese and Tibeto-Burman (or Tibeto-Karen)
should be resumed: (a) Chinese shows almost no trace of the fairly elaborate TB
morphology, (b) the two stocks have only a small segment of roots in common,
 Scott DeLancey

(c) the phonological systems of the two stocks difer in many respects, and can
scarcely be reconciled at some points, (d) the tonal systems of the two stocks
appear not to be correlated. Our belief that the two stocks are genetically related
must rest, ultimately, on the fact that they have certain basic roots in common,
and that phonological generalization can be established for these roots. It might
be argued that the ST elements constitute only a superstratum in Chinese, and
that the substratum is of distinct origin. In historical terms, the Chou people
might be regarded as the bearers of a ST language, which became fused with, or
perhaps immersed in, a non-ST language spoken by the Shang people. In any
event, it is certain that the ST hypothesis illuminates only one of the many dark
recesses in the complex linguistic history of the Chinese. (Benedict 1972: 195–7)

his divergence is suicient to inspire occasional doubts about the genetic rela-
tionship of Sinitic to the Tibeto-Burman languages on the part of historians and
others (e.g. Beckwith 2002, 2006), though few if any linguists still doubt that the
history of Sinitic is of a Tibeto-Burman language whose lexicon and grammati-
cal structure was drastically reorganized in the mouths of a population speaking
Kadai, Hmong-Mien, and/or Austroasiatic, and quite possibly other, languages
(see discussion in Benedict 1976: 167).

. Sinitic and its southern neighbors

here is more diversity of opinion on the internal structure of Sino-Tibetan (and
hence on the appropriateness of that term). he prevalent view sees Sinitic and
Tibeto-Burman as the two primary branches of the family, on grounds of their
very evident structural and lexical divergence. his is problematic, however, since
the deining characteristics are all on one side. Sinitic is a small, cohesive unit of
between half a dozen and several dozen languages (again depending on the level
at which we count). In size, divergence and apparent time depth it is more compa-
rable to lower-level Tibeto-Burman branches like Bodic or Bodo-Konyak-Jinghpaw
than to “Tibeto-Burman” as a whole. And Tibeto-Burman itself is a very problem-
atic construct, since, unlike Sinitic, it cannot be deined by any shared innovations.
Assuming that Proto-Sino-Tibetan had the SOV typology still found throughout
the family except for Sinitic, Bai and Karen (see below), “Tibeto-Burman” languages
share no deining qualities, they are simply all the Sino-Tibetan languages that
aren’t Sinitic. On these and other grounds van Driem (1997, 1999, 2005, 2008, see
also Jacques to appear) suggests abandoning the Sino-Tibetan model for a view of
the family in which Sinitic is simply one subordinate unit among others; van Driem
in fact argues that it is a subbranch within Bodic. his issue does not bear directly
on the argument of this paper, but our conclusions are relevant to the question. he
conventional view of the family as Sinitic + Tibeto-Burman is based on divergent
characteristics of Sinitic which, as we will see, it shares with the Bai Yue languages
 he origins of Sinitic 

rather than with Tibeto-Burman. An account of how these traits came into Sinitic
from its southern neighbors within historic or proto-historic times would explain
the extreme divergence of Sinitic within Sino-Tibetan without requiring great time
depth for its split from the rest of the family, and thus vitiate the argument for
a bipartite Sino-Tibetan model. (Benedict’s (1976: 172f) extensive argument for a
TB-Sinitic split based on diferential retention in various languages of PST lexical
roots is essentially congruent with the broader argument that Sinitic must be dis-
tinct because it is structurally so diferent. Both arguments lose their force if we can
reconstruct a scenario of rapid language shit under intense contact).
he evidence which requires explanation falls into four broad categories: lexi-
cal correspondences among Chinese and one or more other languages or families,
morphological correspondences between Chinese and Tibeto-Burman, and the
striking similarities in both syntactic and phonological structure between Chinese
and the mainland Southeast Asian families. he diiculty is that there is signiicant
evidence linking Chinese with several diferent language groups, including Tai-
Kadai, Hmong-Mien, Austroasiatic, and Austronesian, but it cannot be genetically
related to all or even several of them.1 Most of what Chinese shares with most of
these languages must thus have resulted from language contact. he fundamental
problem of Sinitic historical linguistics is to unravel the various linguistic threads
which make up Old Chinese and its predecessors and understand how they came
to be woven together into the language which we know.
An important part of the problem is that the features which distinguish Sinitic
from Tibeto-Burman are shared with not one, but all of the southern language
groups – Hmong-Mien, Tai-Kadai, and Viet-Muong – which all share a charac-
teristic, very marked areal syntactic and phonological proile (Henderson 1965;
Matisof 1992; Enield 2003, 2011). hus the contact scenario which we need to
reconstruct must be considerably more complex than the simple imposition of a
Sino-Tibetan superstratum on a monolingual substrate population.

.. Lexical correlations

While a large part of the Chinese lexicon connects with Tibeto-Burman (Benedict
1976; Nishida 1976), there is a substantial body of vocabulary shared with one
or more of Tai-Kadai, Hmong-Mien, and Vietnamese. And in some respects the
phonological structure of Sinitic is closer to that of Viet-Muong or Tai-Kadai than
to Tibeto-Burman. In this section I will very briely review some of the reasons
why these congruences do not argue for a genetic relationship among any of these
languages. Most scholars accept Benedict’s assertion that the issue hinges on
“the amount of “core” vocabulary shared by the languages under consideration”
(1976: 168, cp. Baxter 1995), but the appropriate conclusions to be drawn from the
evidence remain subject to some debate. In the following section we will see that,
 Scott DeLancey

as usual in comparative linguistics, it is morphology which provides us with the

crucial evidence.
Since the earliest days of serious linguistic study of Chinese, scholars have
noted the substantial vocabulary shared between Chinese and neighboring lan-
guages. A great deal of this was obviously borrowed from Chinese, which through-
out historic times has been the major cultural force in East Asia. But there is also
a very substantial body of vocabulary shared with Tai-Kadai, Hmong-Mien, and/
or Vietnamese, which is much older, and it is not easy to determine whether such
shared forms are common inheritance or borrowing, and in the latter (more likely)
case, borrowing in what direction.
Aside from the Tibeto-Burman languages (Matisof 2003), Chinese has been
linked with Tai-Kadai (Wulf 1934; Nishida 1975; Li 1945, 1976; Manomaivibool
1975, 1976a, b, inter alia), Austroasiatic (Norman & Mei 1976), and Hmong-Mien
(Downer 1963, 1971; Wang 1986; Haudricourt & Strecker 1991).2 (he Austro-
nesian comparisons advanced by Sagart (1994, 1995, 1999) are now (2005) con-
sidered to relect an older connection between Austronesian and Sino-Tibetan,
rather than speciically Sinitic, which thus takes Austronesian comparisons out
of our purview). On the one hand, all of these proposals are supported by seri-
ous lexical comparisons, and some sort of historical connection with Tai-Kadai,
Hmong-Mien and Austroasiatic (and speciically Viet-Muong) is established both
by shared lexicon and by the astonishing correspondence in phonological typol-
ogy (see below). But these three groups are not evidently related, so Sinitic can
hardly be genetically related to all of them, much less to all of them and Tibeto-
Burman as well, except at some very hypothetical, very high, level. hus it has long
been clear that some of the evidence which has been adduced to argue for genetic
relations among these languages in fact relects sustained intense contact among
unrelated languages.
he interpretation of the lexical evidence has sometimes been confused by
unrealistic notions of when and how borrowing can take place, in particular
simplistic ideas that lexical borrowing only occurs from a more dominant into
a smaller and less “advanced” population, or under some kind of necessity. For
example, Manomaivibool (1975: 364), discussing shared Tai-Chinese vocabulary,
says “It seems implausible that Tai had to borrow that many items of such common
vocabulary from Chinese” (emphasis added). But under many contact scenarios
it is impossible to distinguish borrowings from cognates purely on how easily one
can imagine a motivation for borrowing a word with a particular meaning.3 I will
suggest a model of language contact which makes room for the sort of unsystem-
atic lexical mixture which we ind in Chinese.
It has been suggested (Li 1976; Manomaivibool 1975, 1976a, b; Nishida 1975,
1976) that if a Sino-Tai form can be reconstructed for Proto-Tai-Kadai, this is
 he origins of Sinitic 

evidence for genetic relationship between Sinitic and T–K, presumably on the
grounds that PTK is too old to have been contemporary with any stage of Chinese,
so that there would be no time at which borrowing could have taken place. But
there is no logic to this argument – whether we imagine the common vocabu-
lary to relect a common proto-language or to represent borrowings, in either case
PT-K or something ancestral to it, and Old Chinese or something ancestral to it,
must have been contemporaneous. Noting this fact does not constitute an argu-
ment for one hypothesis or the other. What is important is that Li and other schol-
ars consider the oldest layer of shared Tai-Chinese vocabulary (which certainly
represents loans in both directions, not only from Chinese to Tai) to be of at least
Old Chinese date, so that this common lexicon probably dates from the earliest
contact.

.. he Southeast Asian phonological proile

he most impressive correspondence between Sinitic and the Southeast Asian
Tai-Kadai, Hmong-Mien, and Viet-Muong languages is in their phonological
structure (Henderson 1965; Matisof 2001; R. Li 2005). All share the stereotypical
monosyllabic morpheme structure and elaborate tone systems. he most striking,
and puzzling, fact about this congruence is the perfect correspondence of the
tone systems (Wulf 1934; Haudricourt 1954a, b; Li 1945, 1976; Matisof 1973;
Ostapirat 2000; Ratlif 2010). Sinitic, Tai-Kadai, Hmong-Mien, and Vietnamese
all have a four-tone system, with a three-way distinction on “smooth”, i.e. open or
sonorant-inal syllables, and all “checked”, i.e. obstruent-inal, syllables manifest-
ing a distinct fourth tone. Each of the other three shares with Sinitic (and to some
extent with each other) a substantial body of shared vocabulary which shows regu-
lar correspondence in tone class. However, several generations of research have
made it clear that this tone system is a secondary development in each of the fami-
lies, not reconstructable to any common ancestral system:

Tonal similarities – even regular tonal correspondences – are not to be taken

uncritically as evidence for genetic relationship among languages. Indeed, tonal
criteria are not even suicient to establish genetic subgrouping for languages that
are already known to be genetically related. (Matisof 1973: 89)

In all of the languages tones originated out of inal laryngeal features, so that the
original correspondence is in the type of rime: obstruent coda, coda *-h (some-
times < *-s), inal *-ʔ, and “smooth” syllables with none of these (Haudricourt
1954a, b, 1961/1972; Mei 1970, 1980). he shared vocabulary which shows these
correspondences must have been borrowed at a stage when both the donor and
recipient languages still retained these inal laryngeal distinctions, and had not
yet developed phonemic tone; if we imagine that these items were borrowed with
 Scott DeLancey

phonemic tone, it becomes impossible to explain the regularity of the correspon-

dences. (For a clear exposition of this argument see Ratlif 2010: 187–93). he
languages must have still been in close contact when they underwent a shared
tonogenetic episode in which these laryngeal distinctions were reinterpreted as
tonal, as they were still centuries later when they all shared in the “Great Tone
Split” conditioned by mergers of initial consonant series.
he monosyllabic pattern is not really characteristic of Austroasiatic, or even
of Tai-Kadai, and the Sinitic developments do have parallels in the phonological
development of other Sino-Tibetan groups. So Benedict (1976) and Sagart (Sagart
1999) are probably right in attributing the original locus of monosyllabic structure
to Chinese:

he new linguistic standard of the Han dynasty … typologically characterized

by its incipient isolating morphology, and its emergent tonal and monosyllabic
phonology, gradually spread to all parts of the empire, north and south, and this
same typology further spread to all non-Chinese languages spoken in territories
under Chinese rule ater the Han: all of Miao-Yao, Viet-Muong (but not the rest
of Mon-Khmer), all of Kam-Tai, some south-eastern Tibeto-Burman languages
including Lolo-Burmese (but not Tibetan, Qiang, Gyarong, etc.).
(Sagart 1999: 8)

. Morphological evidence for Sino-Tibetan

he strongest evidence for the genetic ailiation of Sinitic with the Tibeto-Burman
languages is grammatical, speciically correspondences in personal pronouns and
in some reconstructable verb morphology. (For further discussion see Jacques to
appear). Benedict (1976), in summarizing the evidence for Sino-Tibetan, relies
very strongly on lexical correspondences, and makes a strong case that the lexical
evidence overall overwhelmingly supports the genetic connection of Sinitic with
the other Tibeto-Burman languages. But purely lexical evidence is never sui-
ciently convincing for the conservative comparativist, and, as we have seen, in the
case of Sinitic, there is suicient, and suiciently diverse, non-Sino-Tibetan lexical
material to lead many linguists to look for genetic explanations for it. herefore,
as always, the argument for genetic relationship must rely fundamentally on mor-
phological comparisons.

.. he pronouns
An important argument for the Sino-Tibetan ailiation of Chinese has always been
the correspondence of the 1st and 2nd person pronominal roots. Sagart recon-
structs the following pronominal paradigm found in Eastern Zhou (ca. 700–255
 he origins of Sinitic 

BCE) bronze inscriptions (Sagart 1995, 1999: 142–3, cp. Matisof 1995: 76–7; the
forms are reconstructed according to the system of Baxter 1992):
1st 2nd
a. 吾 *aŋa 汝 *bnaʔ
b. 我 * ŋajʔ 爾 *bnajʔ
a

Old Chinese 1st and 2nd person pronouns

he (a) forms closely match the forms reconstructed for Tibeto-Burman:
1st *ŋa, 2nd *na(ŋ) (Matisof 2003). he (b) forms, obviously derived from the
(a) forms, appear in the earliest strata as plurals, but are later attested in singu-
lar uses. Although the majority of TB units have something like the (a) forms,
which clearly are to be reconstructed for the proto-language, several languages
have pronominals which correspond better to the (b) forms, e.g. Jinghpaw ngai
‘I’, and probably Tibetan nged ‘we’. Benedict (1995) identiies the #-i extension in
both TB and Sinitic as a topic marker, but Matisof (1995, 2003: 487–8) is more
hesitant about assigning a function to it. Given its apparent pluralizing function
in Old Chinese, it may better be compared with Tibeto-Burman 1pl element #i
(van Driem 1993; LaPolla 2003a), which occurs throughout the family as both an
agreement index on the verb and as an element in pronominal forms.
However, Sagart argues, on the basis of its late appearance in Shang and
Zhou inscriptions, that the 1st person *ŋa root is a secondary development in
Chinese (1999: 142–4), and a late borrowing from there into the rest of Tibeto-
Burman (145–6). Instead of the well-established *ŋa, Sagart proposes that the
PTB 1st person pronoun was the stop-initial ka (Benedict 1991) which occurs
as the primary 1st person root in three (by Sagart’s count) geographically mar-
ginal branches of the family, northern Qiangic, Kuki-Chin, and “a few languages
of eastern Nepal and neighboring areas”, by which he must be referring to two
distinct units, the Kiranti group in Nepal and the Western Himalayan branch in
northwest India (see hurgood 1985). He suggests that this distribution relects
a spread of *ŋa, ultimately from Chinese, through the contiguous central TB
area, leaving only the few branches on the edges of the TB area untouched.
hese three (actually four, plus several other unclassiied languages in Nepal
and Arunachal Pradesh) then retain what Sagart takes to be the original Sino-
Tibetan 1st person root *ka.
his proposal is implausible on various grounds (see also Pulleyblank 1995b:
329). For example, it is hard to imagine any reason why Central Himalayan lan-
guages like Kham (1sg ŋa:) and Chepang (ŋa) would have adopted the new form,
since they historically have not been under signiicant Chinese, or for that mat-
ter even Tibetan, inluence. But in any case Sagart’s proposal cannot be correct,
 Scott DeLancey

since the nasal root is found in the 1st person agreement suix which is recon-
structable for Proto-Tibeto-Burman (Sun 1983; DeLancey 1989, 2010a; van Driem
1993; LaPolla 2003a), and thus long predates its irst appearance in the Chinese
inscriptions. Most crucially, we ind it as an agreement suix in all of the branches
which Sagart claims retain the original *ka as an independent root – notoriously
in Kiranti, but also in Qiangic (LaPolla 2003b), and Kuki-Chin (Henderson 1957;
Stern 1963; DeLancey 2013a, b) we ind the nasal agreement suix cooccurring
with the stop-initial independent pronoun:
(1) Hayu (Kiranti; Michailovsky 1988: 136)
gu nunukur pon la-ŋ
I woodpecker become go-1sg
‘I’ll turn into a woodpecker.’

(2) Sizang (Kuki-Chin, Stern 1963: 276)

-ke:i _pai: -le-ŋ:
I go if-1sg
‘If I go …’

Obviously, if the stop-initial pronouns in these languages represent inheritance of

the original TB pronoun, there is no possible source for the agreement index. But
on the established hypothesis that the PTB pronominal root was *ŋa, the source of
the agreement index is self-evident.
hus the nasal root is indisputably ascribable to PTB, and cannot be inter-
preted as a borrowing from Chinese. Since a late borrowing from a Tibeto-Burman
source into Chinese does not seem likely here either, we have to recognize this root
as dating back to their common ancestor, regardless of its relatively late appearance
in the inscriptional evidence in Chinese. But Sagart is correct that the *ka root is
also ancient; it now appears that it was a possessive or oblique form contrasting
with the nominative *ŋa (Jacques 2007; DeLancey 2011a). What we see in the
languages where this form has replaced the original nominative *ŋa, is the replace-
ment of the original inite construction with an innovative inite form based on a
nominalization, which thus takes a genitive rather than a nominative “subject”.
(hurgood’s (1985) interpretation of this root as originally a topicalizer, seconded
by Benedict (1995), does not appear to be correct (DeLancey 2011a), but that is
irrelevant to the issue here).

.. Ancient morphology

While we ind no inlectional morphology recorded in any form of Chinese, the
fossils of pre-Chinese preixes and suixes can be found in the phonological alter-
nations of semantically and graphically related words. Two morphological con-
structions which are securely reconstructible for both Tibeto-Burman and Sinitic
 he origins of Sinitic 

are a causative preix *s- (Conrady 1896; Mei 1980, 1988, 2008; Dai 2001) and a
nominalizing *-s suix (Downer 1959; Forrest 1960; Mei 1980; Mazo 2002).
he *s- causative is retained in Written Tibetan and a handful of other lan-
guages, though in many it is no longer productive:
Tibetan log ‘return (intransitive)’, slog ‘turn (transitive)’
Boro gab ‘cry’, səgab ‘make s.o. cry’
Trung ip55 ‘sleep’, səip ‘cause to sleep’

In most modern TB languages, we ind the preix relected in devoicing of the

initial consonant:
Tibetan nub ‘sink’, snub ‘destroy, abolish’
Zaiwa nop ‘sink in mud’, n�op ‘make s.t. get bogged in mud’
Boro gi ‘afraid’, si-gi ‘frighten’
Newar gya- ‘afraid’, khya- ‘frighten’

And we ind the same in Old Chinese (Mei 2008):

见 xiàn ‘be visible’ < *gians
见 jiàn ‘see’ < *kians < *s–k < *s–g
别 píe ‘leave, separate (intr.)’ < *bjät < *brjat
别 pìe ‘discriminate, distinguish’ < *pjät < *prjat < *s–p < *s–b

Dai (2001) demonstrates that this construction is ancient in Tibeto-Burman;

based on this and the abundant evidence for it in Chinese, Mei (2008) suggests
that it is a deining feature of Sino-Tibetan languages. Indeed, it is preserved, at
least in fossil form, in some branches which have lost almost all other inherited
morphology, e.g. Bodo-Garo.
he case for *s- is strong, but for it to be completely conclusive we need to ind
actual cognate pairs showing the same alternation in Sinitic and TB. For the other
classic comparison, the nominalizing *-s suix, we have the complete case. his
suix is relected in Old Chinese tonal alternations (Downer 1959; Forrest 1960);
which correspond to the suix which is preserved in Written Tibetan (Mei 1980):

Chinese Tibetan
量 liáng ‘to measure’ < *liaŋ ‘grang ‘to count’
量 liàng ‘a measure’ < *liaŋs grangs ‘a number’
织 zhī ‘to weave’ < *tjək ‘thag ‘to weave’
织 zhì ‘woven goods’ < *tjəks thags ‘web, woven stuf ’

In addition to these classic, and now deinitive, comparisons, there is a growing

body of plausible morphological comparisons between Sinitic and Tibeto-Burman
languages. Several promising suggestions are summarized in LaPolla 2003a. But
even if not all of these comparisons stand up, there is now suicient evidence to
 Scott DeLancey

establish that the connections between Sinitic and Tibeto-Burman are genetic.
Morphological comparisons like this are the sine qua non of comparative linguis-
tics, and without some strong argument discrediting these comparisons, we can
take this evidence as conclusively establishing the genetic relationship of Sinitic
with the rest of Sino-Tibetan.

. Syntactic patterns

he most immediately salient feature which distinguishes Sinitic from the Tibeto-
Burman languages is word order. In marked contrast to the SVO order which Sin-
itic shares with the Bai Yue languages, all the Tibeto-Burman languages except
Karen and Bai have SOV order, with typical SOV features such as postpositions,
clause-inal subordinators, and postverbal auxiliaries. here is no question that the
same was true of their common ancestor, as far back as Proto-Sino-Tibetan:

From the fact that we can clearly see changes in the word order of these three
languages [Sinitic, Karen, and Bai] over time, and cannot see such changes in
the Tibeto-Burman languages other than Bai and Karen, we assume that it was
Bai, Karen and Chinese that changed rather than all the other Tibeto-Burman
languages. (LaPolla 2003a: 28)

he broad shit from Sino-Tibetan to Sinitic is described by Sagart:

From a typological point of view, Old Chinese was more similar to modern East
Asian languages like Gyarong, Khmer or Atayal than to its daughter language
Middle Chinese: its morphemes were nontonal and not strictly monosyllabic; its
morphology was essentially derivational, and largely preixing; but it also made
use of inixes and suixes. At some point between Old Chinese and Middle
Chinese, and for unknown reasons, a cascade of changes caused the language to
move away from this model. Its aixing morphology began to freeze; its loosely
attached preixes were lost, while other aixes clustered with root segments and
were reinterpreted as root material. A new morphemic canon tending toward
strict monosyllabism, with a great variety of initial and inal consonant clusters,
emerged. Further shits saw the reduction of initial clusters, this resulting in a
more complex inventory of initial consonants, and in new vowel contrasts. Final
clusters were also reduced and the inventory of inal consonants restricted to
resonants and stops, this leading to the emergence of tones. hus the classical
‘Indochinese’ typology common in its major features to Middle Chinese,
Vietnamese, Miao-Yao, Tai, Burmese etc. was born. (Sagart 1999: 13)

I will suggest that rather than thinking of a linear sequence of shits borrowed
unidirectionally from Chinese to its neighbors, we should imagine several dif-
ferent linguistic systems competing with and succeeding one another, with the
inal crystallization into Old Chinese including lexical material and grammatical
 he origins of Sinitic 

structure from more than one source. But the topic of this section is the prominent
feature of Bai Yue-Sinitic typology which is not part of Sagart’s description, which
deals only morphophonological proile which is shared not only by the Bai Yue
languages and Sinitic, but also by Lolo-Burmese. hat is, of course, the SVO word
order shared by Sinitic, all the mainland Southeast Asian languages, and, second-
arily, Bai and Karen.
Mainland Southeast Asia is well-known for its striking areal linguistic typol-
ogy, characterized both by the elaborate and congruent tone systems discussed
above (not shared by most Mon-Khmer languages) and by radically isolating SVO
morphosyntax. Indeed the examples put forward to illustrate isolating typology
are always languages from this area; aside from modern European-based creole
languages, few if any other languages in the world are as resolutely free of any sort
of inlectional morphology. In this respect Sinitic, at least roughly, sorts with the
Southeast Asian rather than the Tibeto-Burman languages, which are characteris-
tically agglutinative, SOV, and oten morphological very complex. But the South-
east Asian typological proile is much more complex than its simple monosyllabic
SVO stereotype (see Enield 2003), and the degree to which the details of Sinitic
syntax conform to it is a topic on which more research is badly needed (see Bisang
1996, 2008).
Sinitic basic word order, at least, is a secondary feature acquired somehow
from the southern languages. Most scholars, from Terrien de la Couperie on, see
the shit in Sinitic as due to inluence from neighboring languages to the south;
Egerod (1976: 59) points out that since SVO order is inherited in hai, “Chinese
was largely a recipient rather than a donor in the early times … it is Chinese
which borrows a new word order” (see also Benedict 1972, 1994; van Driem
2008). Indeed, all of the Southeast Asian groups have SVO syntax as far back as
we can trace. And there are ample traces of earlier SOV patterning in Old Chinese
(Cheng 1983). For a summary of the case for SOV syntax in pre-Old Chinese see
the last chapter of LaPolla 1990.
To take only one striking example of SOV patterning in Old Chinese, consider
the sentential particle yě 也. his occurs frequently in equational sentences with no
overt copula (Example 3), exactly parallel to the behavior of similar sentence-inal
particles in many Tibeto-Burman languages, e.g. Classical Tibetan (Example 4):
(3) 彼丈夫也，我丈夫也
bǐ zhàng.fū yě, wǒ zhàng.fū yě
dem man final, I man final
‘hey [were] men, I [am] a man.’ (Mencius 3.1)
(4) bram=ze de dbul=po zhig go
Brahmin dem pauper a final
‘hat Brahmin [was] a pauper.’
 Scott DeLancey

In Tibeto-Burman these sentence-inal particles derive from old copulas

(DeLancey 2011b), which is why they oten continue to function as such in equa-
tional sentences. heir inal position is thus directly attributable to the SOV of the
TB languages. A parallel history for Old Chinese yě 也 is not the only imaginable
explanation for its position and behavior, but is certainly the most economical.
For that matter, we even ind clear evidence for an SOV past in synchronic
Chinese syntax (Dryer 2003; D. Xu 2006; F. Wu in this volume). Particularly strik-
ing is the fact that despite their SVO clause order, Sinitic languages all place rela-
tive clauses before the head noun, a characteristically SOV pattern. Dryer and Wu
argue that these phenomena in Sinitic languages are better explained as Altaic
inluence than as Tibeto-Burman residue. his is certainly not impossible, but not
all of the evidence is as clear as these authors suggest. It is true that both pre- and
post-head relative clause constructions occur in Tibeto-Burman languages, oten
in the same language. But the occurrence of pre- vs. post-head relative clauses in
TB is not random. here are three basic relative constructions which occur in the
family: nominalized clauses serving as genitive modiiers to the head noun, nomi-
nalized clauses serving as appositive modiiers, and, in a few languages, relative
pronoun constructions calqued from neighboring Indic languages. It is the irst
of these which is structurally comparable to Sinitic relative clause constructions,
and in all Tibeto-Burman languages, as in Sinitic, genitive-marked relative clauses
are invariably prenominal. Post-head relative constructions in TB languages are
always of one of the other two types.
A more crucial point is that, if this is the explanation for the SOV features in
modern Chinese languages, we are presumably looking at northern inluence at
the Proto-Sinitic level, since most of the SOV-like features which these authors
mention are as true in Yue as in Mandarin:
[I]t must be stressed that the proportio of N-Mod structures in Yue, Hakka, etc.
is extremely small, and that Mod-N is the overwhelmingly dominant order in all
dialects at both morphological and syntactic levels.” (Bennett 1979: 94)

here is ample evidence for Altaic inluence speciically on northern Chinese

over the last two millennia (Hashimoto 1976a, b, 1986; Wadley 1996, inter alia),
but this can explain only those features which distinguish Mandarin from lan-
guages like Min and Yue. It is clear that the southern Sinitic languages, and
especially Yue, manifest the strongest and deepest Bai Yue, and especially Dai,
inluence (A. Hashimoto 1976; Bennett 1979; R. Li 2005). his certainly relects
the non-Sinitic substratum on which Chinese culture was imposed as it spread
south, in other words, many of the substratum features which distinguish the
southern from the northern Sinitic languages must have been acquired since
 he origins of Sinitic 

Proto-Sinitic times. M. Hashimoto (1976a, b) and others trace putative Altaic

features, both syntactic and phonological, back only as far as the collapse of the
Han Dynasty and ensuing invasions from the north. Although population move-
ments certainly brought many refugees from the north to the south of China, it
is implausible that this could bring about signiicant word order shits in the
local languages, but this is a necessary part of any hypothesis in which southern
word order features are to be explained in terms of post-Proto-Sinitic inluence
from Altaic or other northern languages. hus any of Altaic or more generally
northern inluence must date back to before the southern expansion of Chinese,
that is, to Proto-Sinitic.
A simple model in which Proto-Sinitic is a typical Tibeto-Burman language
which gradually adapted certain southern linguistic features implies that the mix
of SOV and SVO constructions in Chinese should show a gradual and unidirec-
tional tendency toward more SVO and less SOV over time. An important body of
research (Peyraube 1997a, b; Djamouri 2001) shows that in fact we ind that Shang
era materials show more consistent SVO patterning than later Zhou inscriptions;
we will return to the implications of this in Section 3.

. he sources of Sinitic

here is no question that the formation of Sinitic involved contact with neighbor-
ing languages, including Austroasiatic, Tai-Kadai, and Hmong-Mien, and very
possibly others which have completely disappeared. here has been a certain ten-
sion on the question of what sorts of contact might be involved. Traditionally
there seem to be two basic possibilities: contact between adjacent languages, i.e.
imagining Proto-Sinitic, Proto-Hmong-Mien, etc. as spoken in adjacent states,
or super-substratum inluence, i.e. an “elite dominance” model in which Proto-
Sinitic formed in a state consisting of immigrant Tibeto-Burman conquerors
interacting with indigenous Proto-Tai-Kadai or Proto-Hmong-Mien subjects.
Both of these have important deiciencies; in this section I will develop an alter-
native model which is better suited to explain the kinds of data we have been
considering.

. Typology and diachrony: he “creoloid” pattern

Our problem is to imagine a context for the formation of Sinitic which provides
for a broadly mixed lexicon, and some basic Sino-Tibetan morphological struc-
ture with Southeast Asian creoloid syntax. Obviously we are looking at a language
contact or language mixture situation, as scholars have noted for over a century
 Scott DeLancey

(Terrien de la Couperie 1887; Matisof 1973, 1992; LaPolla 2001, 2010, inter alia),
but these general descriptions encompass a wide variety of phenomena, oten
with quite diferent outcomes (homason & Kaufman 1988). In homason and
Kaufman’s terms we are looking at both borrowing and language shit. he syn-
tactic and phonological convergence between Sinitic and the Bai Yue languages
is far too deep to represent simple borrowing between neighboring languages;
we have to imagine a situation with extensive long-term bi- or multilingualism.
Sinitic is, in homason and Kaufman’s terms, a language “in which a number of
structural interference features are to be attributed to the efects of language shit,
but in which enough inherited grammatical patterns remain that genetic continu-
ity has clearly not been disrupted” (homason & Kaufman 1988: 129). Typically
this involves the replacement of morphological categories by syntactic expressions
(1988: 129), which is precisely the essential diference between PST and Sinitic
morphosyntax.
When we compare the conservative, highly morphologized TB languages
(especially the rGyalrongic, Kiranti, and Nungish groups) with the transparent
agglutinative pattern found in the larger and better-known branches (modern
Tibetan, Lolo-Burmese, Bodo-Garo), we see the kinds of “simplifying” efects
which are known to occur in situations of intense language contact (homason &
Kaufman 1988; Ansaldo & Matthews 2001; Dahl 2004; McWhorter 2007; Trudgill
2009, 2011, inter alia), and when we look into the history and prehistory of the
regions where these branches emerged, we see evidence for just the kinds of
situation of intense contact which we know produces such linguistic changes
(DeLancey 2010b, 2012, to appear). In Sinitic the efects are even more dramatic,
involving a shit to SVO constituent order. But they are of the same kind, and must
be explained the same way.
he similarity of the isolating Sinitic-Southeast Asian morphosyntactic proile
to creole languages has been noted for some time.4 But history gives us no reason
to imagine that Sinitic at any stage was ever a true creole, in the traditional sense
of a language which develops from a grammarless pidgin. Ansaldo and Matthews
(2001, 2007) consider it a “creoloid” language, a pattern which arises in “heavy
contact situations involving typologically distant varieties” (2001: 311). his is a
common diachronic phenomenon resulting from “non-normal” (homason &
Kaufman 1988), “suboptimal” (Dahl 2004) or “interrupted” (McWhorter 2007)
transmission, i.e. one or more historical episodes in which a signiicant number
of adult members of the speech community were non-native speakers using the
language as a lingua franca or supralect:
Language contact has this consequence [reduction in complexity] because of
pidginization. he most extreme outcome of pidginization is the development of
a pidgin language, but this is a very rare occurrence. It is only pidiginization at its
 he origins of Sinitic 

most extreme, together with a number of other unusual factors, which combine
to lead to the development of pidgin and, even more rarely, creole languages.
Pidignization can be said to occur whenever adults and post-adolescents learn a
new language. (Trudgill 2009: 99, emphasis original)

“Pidginization” is probably not the most apt term for this phenomenon, since we
are dealing with languages which show the constellation of typological features
traditionally associated with creole languages but do not have the typical history
of a creole:

Perhaps a more constructive way to see the “prototypical creole” traits is that
languages which have been subject to intensive contact involving several
typologically distant varieties will tend to show some combinations (or subset) of
these features. (Ansaldo & Matthews 2001: 317)

Such languages arise in conditions of intense contact, when for whatever reason
some signiicant portion of the language community are second-language rather
than native speakers (McWhorter 2007). his kind of development has occurred,
and continues to occur, repeatedly in Sino-Tibetan, and it is clear that Sinitic has
the same kind of history.

. he linguistic context

he territory where Sinitic languages are spoken was an area of substantial lin-
guistic diversity from prehistoric times (Terrien de la Couperie 1887; Pulleyblank
1983, 1995a; Ballard 1984; Luo 1990; LaPolla 2001; Blench 2008, ms., inter alia).
We are particularly concerned with the languages of the southern peoples who
the Chinese referred to as Bai Yue 百越 (“Hundred Yue”), which seem to have
been of mixed provenance, including Austroasiatic, Tai-Kadai and Hmong-Mien
languages (JZ Li 1994; Meacham 1996; LaPolla 2001). he precise nature of the
Austroasiatic element is in some dispute. Strong connections between Sinitic and
Vietnamese are clear, but the existence of a deeper Austroasiatic stratum, as sug-
gested by Norman and Mei (1976), is not universally accepted; the debate partly
hinges on the question of the center of disperal for Austroasiatic (see Diloth
2011; Sidwell & Blench 2011; Sagart 2011; van Driem 2011).
he Yue people and kingdom to the south are a long-term presence in Chinese
history, but the irst explicit reference to the “Hundred” Yue is in the Qin era
Annals of Lü Buwei:

For the most part, there are no rulers to the south of the Yang and Han rivers,
in the confederation of the Hundred Yue tribes [百越之际], in the territories of
Bikaizhu, Fufeng, and Yumi, and in the states of Fulou, Yangyu, and Huandou.
(Knoblock & Riegel 2000: 112/Book 20/1.3)
 Scott DeLancey

his term is important because it makes clear that the reference of Yue is multieth-
nic (Luo 1990: 268). he question is, how does Chinese come to share large bodies
of vocabulary, and characteristic phonological and morphosyntactic typological
proiles, with these languages?
Benedict’s and Nishida’s suggestion that the language of the Shang dynasty was
of non-Sino-Tibetan provenance, and that Old Chinese represents the outcome of
the imposition of the Sino-Tibetan speech of the Zhou conquerors on a Shang
substrate, provides a possible explanation for the southern features in Sinitic –
assuming that the language of Shang was of Bai Yue stock, which is certainly likely.
But it doesn’t directly account for the distribution of the Southeast Asian mor-
phophonological proile, and the widely shared lexicon, both of which are shared
among Sinitic, Viet-Muong, Hmong-Mien and Tai-Kadai. On the simplest version
of Benedict’s hypothesis, we would rather expect to ind extensive sharing between
Sinitic and whichever of these families was represented by the language of Shang.
he fact that four quite distinct stocks share the same phonological-syntactic pro-
ile and common lexical stock suggests that the seedbed of Sinitic was broader, not
conined simply to the interaction of Shang and Zhou.
Still, this is a place to start. Nishida (1976) insists that, while the language of
Zhou “must have been very close to the Tibeto-Burman languages”, that of Shang
was non-Sino-Tibetan. He demurs from speculating about the ailiation of the
language of Shang, but presents one example which suggests a Hmong-Mien aili-
ation. Regarding Benedict’s hypothesis, he says:
his view is possible, of course, though it could be based on a somewhat diferent
assumption. he Shang language belonged to some unknown family and was of
the SVO type in its word order. his SVO word order was already evident in the
records of oracle bones and in bronze inscriptions. On the other hand, it can
be suspected that the language of early Chou originally had SOV word order,
which is a very distinctive feature of Tibeto-Burman. I supposed that the early
Chou tribe, formerly having no writing system of its own, had borrowed a writing
system from the Shang language, as a member of the Shang cultural area, as a
result, the Chou language changed from the SOV type to the SVO order under
the strong literary inluence of the Shang writing system. (Nishida 1976: 36)

We cannot accept the whole of Nishida’s suggestion; it is hardly imaginable that

the writing system as such had much to do with fundamental systemic changes
on the morphosyntactic structure of the language, especially in a context where
only a miniscule portion of the population could have been literate. But the
scenario suggested here, involving some assimilation of the language of Zhou
toward that of Shang prior to the dynastic shit, ofers the possibility of a more
nuanced history than implied by Benedict’s simple model of Zhou superim-
posed on Shang.
 he origins of Sinitic 

. he origins of Sinitic

Any account of the origins of Sinitic must conform to the essential picture of a
contact situation between western invaders speaking a TB tongue and locals
speaking languages ailiated with one or more of the attested mainland Southeast
Asian stocks. But it is not enough to simply say “contact” and pretend that we have
explained anything. In this view of Sinitic we have a very speciic outcome, with
Sino-Tibetan lexical and grammatical core, heavy Bai Yue lexical inluence, creo-
loid syntax based more on Bai Yue than on Sino-Tibetan patterns, and innovative
phonological structure. his did not come about through people overhearing each
other’s languages on market day, or learning a few phrases for doing business; we
have to imagine a situation of widespread bi- or multilingualism. his would be
the case in a scenario in which Chinese or pre-Chinese speakers conquered a Bai
Yue population, as happened as the kingdoms of Chu 楚 and then Yue 越 were
incorporated into Qin China. But this does not automatically explain the extent
of the inluence which we ind on the whole language. Ballard’s (1984) “Mother
Soup” metaphor captures the problem but doesn’t solve it. More importantly, the
most important contact phenomena predate the assimilation of the southern king-
doms into imperial China. he southern kingdoms became part of China in the
course of the Qin imperial expansion, by which time the essential features of Old
Chinese had been established for centuries. Once again, with Nishida, we have to
be concerned with contact beginning prior to the Zhou conquest of Shang.
Following the suggestions of Benedict and Nishida, we focus particularly on
the time of the replacement of the Shang Dynasty by Zhou, formerly a western
vassal state of presumably Tibeto-Burman origin. But in this context it is a mistake
to suppose that the deep and pervasive areal phenomena which we see could be
the simple result of a single dramatic historical event. Ater the establishment of
the Zhou Dynasty we can certainly expect the language of Zhou to have had some
inluence on surrounding languages. But prior to that, when Zhou was subordi-
nate to Shang, there would presumably have been inluence in the other direc-
tion. And, since there must have been a history, over at least a century or two, of
increasing Zhou strength and inluence ultimately leading to the dynastic shit,
the language of Zhou could well have been inluential in the region for some time
prior to the fall of Shang.
Although the historicity of the Xia Dynasty is not considered to be established
archaeologically, we can hardly imagine Shang to have emerged ex nihilo, and
there is ample archaeological evidence for urbanization and early state formation
well prior to the traditional dates of Shang, and even Xia. Major imperial states
like Shang are the result of centuries of conquest and consolidation of smaller city-
states. Chang (2005: 126) cites a Qing dynasty historian Gu Zuyu 顾祖舆 who in
 Scott DeLancey

the Du Shi Fang Yu Ji Yao estimated that there were 10,000 states (guo 国) at the
beginning of the Xia dynasty (2100 BCE), but that 500 years later at the founda-
tion of the Shang dynasty these had been consolidated to 3,000, and by 1,000 BCE
at the beginning of the Zhou dynasty there were only 1,800. hus already in the
third millennium BCE we have a picture of imperial expansion which would have
involved substantial armies being raised, and marched of into foreign lands, and
substantial populations being subjugated or enslaved. And it is this process that
I propose is at the root of the Southeast Asian typological proile, and the birth of
Sinitic.
So we have a set of linguistic phenomena requiring a fairly complex model of
language contact and interaction, at a time and place where we have ample evi-
dence that the linguistic situation was indeed extremely complex. he history of
Sinitic involves more than simply the contact of two languages in a conquest situ-
ation. What I propose is that the features which so dramatically distinguish Sinitic
from other Tibeto-Burman branches, and connect it with each and all of the Bai
Yue languages, relect the use of Proto-Sinitic as a lingua franca, used widely by
non-Chinese (by whatever deinition) outside of the actual administrative control
of the Chinese state. In the multilingual context of early China and its neighbors,
we can imagine the utility of a vehicular lingua franca even without reference to
the Chinese state and its inluence. By the time the Chinese state is present on the
historical stage, some version of its language would be a likely candidate for this
role, but it is very plausible, indeed more likely than not, that there would already
have been a widely-used vehicular language in the region.
Let us hypothesize, following Nishida’s implicit suggestion, that the language
of Shang was of Bai Yue stock. here are non-linguistic reasons to suppose that
the predominant element was Hmong-Mien, which more and more seems to have
been the language spoken by the irst cultivators of rice in the Yangtze basin (van
Driem 2011), and thus likely to have been a dominant language in the region from
3–4 millennia ago. Genetic evidence further singles out Hmong-Mien as a major
player in the formation of the Han ethnicity. Most research connecting genetic and
linguistic distributions is primarily concerned with the origins and higher-order
connections of the major families, and thus focuses on earlier eras then we are
interested in. (See van Driem 2005; Chu 2005; Poloni et al. 2005 for discussion of
some of the issues). Rather than any deinite correlation with linguistic groups, the
major division in East Asia is between a northern and a southern population, with
both Sinitic and Tibeto-Burman-speaking groups split across the north-south
divide (Chu et al. 1998; Bo, Xie et al. 2004; Bo, Li et al. 2004).
One robust inding is the considerable genetic diversity of southern Han
populations, strongly implying that Chinese speakers include substantial popula-
tions that once spoke some other language. he genetic picture is consistent with
 he origins of Sinitic 

the implications of linguistic analysis that Sinitic was born of an intrusive Tibeto-
Burman language meeting an entrenched language of Bai Yue origin:
[T]he interaction between Chinese and other southern populations occurred
ater the divergence of the Chinese and Tibeto-Burmans, and a limited gene low
occurred between them ater the divergence. (Bing et al. 2000: 585)

Many researchers report evidence of long-term contact among the various south-
ern groups (Chu et al. 1998; Yao et al. 2002; Bo, Li et al. 2004; van Driem 2011),
with later introduction of Sino-Tibetan elements. Bing et al. single out Hmong-
Mien among the southern populations as particularly involved in the origins of
the Han, and thus presumably of Sinitic, and deduce a history of:

… strong interactions between the Han and Hmong-Mien peoples that have
lasted for several thousand years, as is conirmed in history literature, although a
possible shared ancestry can not be ruled out. (Bing et al. 2000: 585)

We ind a disproportionately male contribution of Sino-Tibetan-associated

genetic markers in the southern Han population (Bo, Xie et al. 2004; van Driem
2011), evoking van Driem’s “father tongue” scenario, in which “mothers passed
on the language of their spouses to their ofspring” (2012: 198), which would go
a long way toward explaining the massive and fundamental Bai Yue efect on the
structure of Chinese. But in historical context we do not necessarily need the
explanatory power of the “father tongue” scenario, as we have a perfect context for
creolization in any case.
As the Shang state grew in power and inluence, its language must have been
widely used among the neighboring groups, both the Bai Yue to the south and
Sino-Tibetan neighbors such as Zhou to the west, as a lingua franca. Here we have
a major motive force in the development of the Southeast Asian typological pro-
ile. As a lingua franca, this language would accommodate vocabulary from vari-
ous language communities, and this is the origin of the substantial body of lexical
material which is shared across all four of the SEA-type stocks. As the Zhou state
increases in power and inluence within the sphere of Shang, the representation of
its Sino-Tibetan language in the lingua franca grows. By the time of the establish-
ment of the Zhou Dynasty, the lingua franca is widely spoken in a version with
signiicant Sino-Tibetan vocabulary, but still with syntax based on the language
of Shang.
When Zhou takes over the empire, there is, as on Benedict’s model, a tempo-
rary diglossic situation, in which genuine Zhou speech is, for a while, retained
in the ruling class, but among the formerly Shang population, Shang speech is
gradually replaced not by “pure” Sino-Tibetan Zhou, but by a heavily Tibeto-
Burman inluenced version of the lingua franca. In similar situations, certainly
 Scott DeLancey

in all subsequent instances of the institution in China of a foreign dynasty, we

see the vehicular language, rather than the speech of the foreign invaders, end-
ing up as the language of administration. We do see, in Old Chinese, SOV as
well as SVO syntactic patterns, such that Cheng (1983) speaks of “two sub-
languages coexisting in early archaic Chinese”, an earlier SOV stratum and an
innovative SVO syntax, and Xú (2004) of a “typologically mixed” language.
his would, essentially, be “pure” Sino-Tibetan Zhou with SOV syntax, and the
Zhou-inluenced lingua franca spoken with the SVO pattern of the Bai Yue
languages.
In this context the observation noted in Section 1.3 that Pre-Archaic Shang
oracle bone inscriptions are more consistently SVO than Zhou era material, which
show noticeably more SOV constructions, makes complete sense: as Benedict and
Nishida suggest, the language of Zhou represents a Tibeto-Burmanization of a
previously substantially Bai Yue, and thus presumably SVO, language.
Sinitic as it emerges into history is then the result of a subsequent process of
assimilation of the lingua franca of the subject population toward the Sino-Tibetan
speech of their rulers. hus we see overwhelmingly Sino-Tibetan vocabulary, but
with extensive, but unsystematic, lexical remnants of earlier versions of the lin-
gua franca. he Sino-Tibetan inheritance extends to pronouns and derivational
morphology, but, as is typical in language replacement, the complex Sino-Tibetan
inlectional morphology disappeared. he stubborn retention of the SVO word
order template can probably be suiciently explained, as Benedict suggests, by the
fact that the vast majority of the population spoke an SVO language or languages
and always had.

Notes

. If all of these languages should be related in something like Sagart’s (1994a) or Starosta’s
(2005) East Asian phylum, it would be at greater time depth than we are considering here, and
the relevant evidence is of a different and more obscure sort.

. I ignore here suggested connections to Uralic, Indo-European, North Caucasian, and

Na-Dene.
. Consider the following French borrowings into English: family, dinner, supper, soup, easy,
difficult, quiet, silent, noisy, lake, river, mountain, valley, forest, marsh, village, city, language,
story, color, attack, defend, argue, agree, beautiful, flower, stupid, count, real, false, very, front. It
would be hard to argue that any of these, or thousands of similar items, “had” to be borrowed
for cultural reasons.

. I don’t know who first made this observation; I first heard it in the 1970’s from David
Strecker and Brenda Johns.
 he origins of Sinitic 

References

Ansaldo, Umberto, and Stephen Matthews. 2001. “Typical creoles and simple languages: he
case of Sinitic.” Linguistic Typology 5 (2–3): 311–326.
Ballard, W.L. 1984. “he Mother Soup: A South China recipe for tonometamorphogenesis.”
Computational Analysis of African and Asian Languages 22: 43–70.
Baxter, William. 1992. A Handbook of Old Chinese Phonology. Berlin and New York: Mouton
de Gruyter.
____. 1995. “ ‘A stronger ainity … than could have been produced by accident’: A probabilistic
comparison of Old Chinese and Tibeto-Burman.” In Wang (ed.), 1–39.
Beckwith, Christopher. 2002. “he Sino-Tibetan problem.” In Medieval Tibeto-Burman
Languages, ed. by Christopher Beckwith,113–57. Leiden: Brill.
____. 2006. “Introduction: Toward a Tibeto-Burman theory.” In Medieval Tibeto-Burman
Languages II, ed. by Christopher Beckwith, 1–38. Leiden: Brill.
Benedict, Paul. 1972. Sino-Tibetan: A Conspectus. Cambridge: Cambridge University Press.
____. 1976. “Sino-Tibetan: Another look.” Journal of the American Oriental Society 96 (2):
167–197.
____. 1991. “A note on PST-level morphosyntax.” Linguistics of the Tibeto-Burman Area 14 (1):
137–41.
____. 1994. “On VO vs. OV in Southeast Asia.” Linguistics of the Tibeto-Burman Area 17 (1):
173–4.
____. 1995. “Proto-Tibeto-Burman/Proto-Sino-Tibetan *-i suix.” In Nishi et al. (eds.) 31–4.
Bennett, Paul. 1979. “A critique of the Altaicization hypothesis.” Cahiers de Linguistique: Asie
Orientale 6: 99–104.
Bing Su, Chunjie Xiao, Ranjan Deka, Mark T. Seielstad, Daoroong Kangwanpong,· Junhua Xiao,
Daru Lu, Peter Underhill, Luca Cavalli-Sforza, Ranajit Chakraborty, and Li Jin. 2000. “Y
chromosome haplotypes reveal prehistorical migrations to the Himalayas.” Human Genet-
ics 107: 582–590.
Bisang, Walter. 1996. “Areal typology and grammaticalization: Processes of grammaticalization
based on nouns and verbs in East and mainland South East Asian languages.” Studies in
Language 20: 519–97.
____. 2008. “Underspeciication and the noun/verb distinction: Late Archaic Chinese and
Mon-Khmer.” In he Discourse Potential of Underspeciied Structures, ed. by Anita Steube,
55–81. Berlin and New York: de Gruyter.
Blench, Roger. 2008. “Stratiication in the peopling of China: How far does the linguistic
evidence match genetics and archaeology?” In Sanchez-Mazas et al., 105–32.
____. Ms. “Why we don’t need Austric or any other macrophyla in SE Asia: the Southern
Yunnan Interaction Sphere.”
Bo Wen, Xuanhua Xie, Song Gao, Hui Li, Hong Shi Xiufeng Song, Tingzhi Qian, Chunjie Xiao,
Jianzhong Jin, Bing Su, Daru Lu, Ranajit Chakraborty, and Li Jin. 2004. “Analyses of Tibeto-
Burman populations reveals sex-based admixture in southern Tibeto-Burmans.” American
Journal of Human Genetics 74: 856–865.
Bo Wen, Hui Li, Daru Lu, Xiufeng Song, Feng Zhang, Yungang He, Feng Li, Yang Gao, Xianyun
Mao, Liang Zhang, Ji Qian, Jingze Tan, Jianzhong Jin, Wei Huang, Ranjan Deka, Bing Su,
Ranajit Chakraborty and Li Jin. 2004. “Genetic evidence supports demic expansion of Han
culture.” Nature 431: 302–305.
 Scott DeLancey

Chang, Kwang-chih. 2005. “he rise of kings and the formation of city-states.” In he Forma-
tion of Chinese Civilization: An Archaeological Perspective, ed. by Sarah Allan, 125–39. New
Haven and London: Yale University Press.
Cheng, Tsai Fa. 1983. “SOV sentence patterns in Early Archaic Chinese.“16th International
Conference on Sino-Tibetan Languages and Linguistics, Seattle.
Chu, Jiayou. 2005. “Genetic analysis of minority populations in China and its implications for
multiregional evolution.” In Sagart et al. (eds.), 248–251.
____, W. Huang, S.Q. Kuang, J.M. Wang, J.J. Xu, Z.T. Chu, Z.Q. Yang, K.Q. Li, P. Li, M. Wu, Z.C.
Geng, C.C. Tan, R.F. Du, and J. Li. 1998. “Genetic relationships of populations in China.”
Proceedings of the National Academy of Science 95 (20): 11763–8.
Conrady, August. 1896. Eine indochinesische causative-denominativ-Bildung und ihr zusammen-
hang mit den Tonaccenten. Leipzig: Harrasowitz.
Dahl, Östen. 2004. he Growth and Maintenance of Linguistic Complexity. Amsterdam and
Philadelphia: Benjamins.
Dai, Qingxia. 2001. “Zang-Mian yu zu yuyan shidong fanchou de lishi bianhua [he historical
development of the causative category in Tibeto-Burmese languages].” Journal of Chinese
Linguistics 29 (1): 1–10.
DeLancey, Scott. 1989. “Verb agreement in Proto-Tibeto-Burman.” Bulletin of the School of
Oriental and African Studies 52 (2): 315–33.
____. 2010a. “Towards a history of verb agreement in Tibeto-Burman.” Himalayan Linguistics
9 (1): 1–38.
____. 2010b. “Language replacement and the spread of Tibeto-Burman.” Journal of the Southeast
Asian Linguistics Society 3 (1): 40–55.
____. 2011a. “Agreement preixes in Tibeto-Burman.” Himalayan Linguistics 10 (1): 1–35.
____. 2011b. “Finite structures from clausal nominalization in Tibeto-Burman languages.” In
Nominalization in Asian Languages: Diachronic and Typological Perspectives, ed. by Foong
Ha Yap, K. Grunow-Hårsta & J. Wrona, 343–362. Amsterdam and Philadelphia: Benjamins.
____. 2012. “On the origins of Bodo-Garo.” In North East Indian Linguistics 4, ed. by Gwendo-
lyn Hyslop, Stephen Morey, and Mark Post. New Delhi: Foundation/Cambridge University
Press India.
____. 2013a. “he history of postverbal agreement in Kuki-Chin.” Journal of the Southeast Asian
Linguistics Society 6: 1–18.
____. 2013b. “Verb agreement suixes in Mizo-Kuki-Chin.” In: North East Indian Linguistics 5,
ed. by Gwendolyn Hyslop, Stephen Morey, and Mark Post, 138–150. New Delhi: Founda-
tion/Cambridge University Press India.
____. to appear. “Creolization in the divergence of Tibeto-Burman.” In TransHimalayan Linguis-
tics, ed. by Nathan Hill and Tom Owen-Smith. Berlin: Mouton.
Diloth, Gerard. 2011. “Austroasiatic word histories: boat, husked rice and taro.” In Enield
(ed.), 295–313.
Djamouri, Redouane. 2001. “Markers of predication in Shang bone inscriptions.” In Sinitic
Grammar: Synchronic and Diachronic Perspectives, ed. by Hilary Chappell, 143–171.
Oxford: Oxford University Press.
Downer, Gordon. 1959. “Derivation by Tone-Change in Classical Chinese.” Bulletin of the School
of Oriental and African Studies 22 (1/3): 258–290.
____. 1963. “Chinese, hai, and Miao-Yao.” In Linguistic Comparison in South East Asia and the
Paciic, ed. by H.L. Shorto, 133–139. London: SOAS.
 he origins of Sinitic 

____. 1971. “he further relationships of the Miao-Yao languages.” 4th International Confer-
ence on Sino-Tibetan Languages and Linguistics, Bloomington, IN.
Driem, George van. 1993. “he Proto-Tibeto-Burman verbal agreement system.” Bulletin of the
School of Oriental and African Studies 56 (2): 292–334.
____. 1997. “Sino-Bodic.” Bulletin of the School of Oriental and African Studies 60 (3): 455–488.
____. 1999. “A new theory on the origin of Chinese.” Indo-Paciic Prehistory Association Bulletin
18: 43–58.
____. 2005. “Tibeto-Burman vs. Indo-Chinese: Implications fvor populations geneticists,
archaeologists, and prehistorians.” In Sagart et al. (eds.), 81–106.
____. 2008. “To which language family does Chinese belong, or what’s in a name.” In Sanchez-
Mazas et al. (eds.), 219–53.
____. 2011. “Rice and the Austroasiatic and Hmong-Mien homelands.” In Enield (ed.), 361–390.
____. 2012. “Glimpses of the ethnolinguistic prehistory of Northeastern India.” In Origins
and Migrations in the Extended Eastern Himalayas, ed. by Toni Huber, and Stuart Black-
burn,187–211. Leiden: Brill.
Dryer, Matthew. 2003. “Word order in Sino-Tibetan languages from a typological and geograph-
ical perspective.” In hurgood and LaPolla (eds.) 43–55.
Egerod, Søren. 1976. “Benedict’s Austro-Tai hypothesis and the traditional view on Sino-hai
relationship.” Computational Analysis of African and Asian Languages 6: 51–59.
Enield, N.J. 2003. Linguistic Epidemiology: Semantic and Grammar of Language Contact in
Mainland Southeast Asia. London: Routledge.
____. 2011. “Linguistic diversity in mainland Southeast Asia.” In Enield (ed.), 63–79.
____. (ed.). 2011. Dynamics of Human Diversity: he Case of Mainland Southeast Asia. [Paciic
Linguistics 627]. Canberra: Australian National University.
Forrest, R.A.D. 1960. “Les occlusives inales en chinois archaïque.” Bulletin de la Société Linguis-
tique de Paris 55: 228–239.
Hashimoto, Anne Oi-Kan Y. 1976. “Substratum in Southern Chinese – the Tai connection.”
Computational Analysis of African and Asian Languages 6: 1–9.
Hashimoto, Mantaro. 1976a. “he agrarian and pastoral difusion of languages.” In Hashimoto
(ed.), 1–14.
____. 1976b. “Language difusion on the Asian continent.” Computational Analysis of African
and Asian Languages 3: 49–64.
____. (ed.). 1976. Genetic Relationship, Difusion and Typological Similarities of East & South-
east Asian Languages: Papers for the 1st Japan-US Joint Seminar on East & Southeast Asian
Linguistics. Tokyo: Japan Society for the Promotion of Science.
____. 1986. “he altaicization of Northern Chinese.” In Contributions to Sino-Tibetan Studies,
ed. by John McCoy, and Timothy Light, 76–97. Leiden: Brill.
Haudricourt, André. 1954a. “De l’origine des tons en vietnamien.” Journal Asiatique 242: 69–82.
____. 1954b. “Comment reconstruire le chinois archaïque.” Word 10 (2–3): 351–64.
____. 1961. “Bipartition et tripartion des systèmes de tons dans quelques langues d’Extr̂me-
Orient.” Bulletin de la Société de Linguistique de Paris 56: 163–80.
____. 1972. “Two-way and three-way splitting of tonal systems in some Far Eastern languages.”
In Tai phonetics and phonology, ed. by Jimmy Harris and Richard Noss, 58–86. Bangkok:
Central Institute of English Language. (English translation of Haudricourt 1961).
____, and David Strecker. 1991. “Hmong-Mien (Miao-Yao) loans in Chinese.” T’ oung Pao n.s.
77 (4–5): 335–41.
 Scott DeLancey

Henderson, Eugenie. 1957. “Colloquial Chin as a pronominalized language.” Bulletin of the

School of Oriental and African Studies 20: 323–7.
____. 1965. “he topography of certain phonetic and morphological characteristics of Southeast
Asian languages.” Lingua 15: 400–434.
Jacques, Guillaume. 2007. “A shared suppletive pattern in the pronominal systems of Chang
Naga and Southern Qiang.” Cahiers de linguistique Asie Orientale 36 (1): 61–78.
____. to appear. “he genetic position of Chinese.” In Encyclopedia of Chinese Languages and
Linguistics, ed. by Rint Sybesma, Wolfgang Behr, Yueguo Gu, Zev Handel, and C.-T. James
Huang. Leiden: Brill.
Knoblock, John and Jefrey Riegel (tr.). 2000. he Annals of Lü Buwei: A Complete Translation
and Study. Stanford: Stanford University Press.
LaPolla, Randy. 1990. Grammatical Relations in Chinese: Synchronic and Diachronic Consider-
ations. Ph.D. dissertation, University of California (Berkeley).
____. 2001. “he role of migration and language contact in the development of the Sino-Tibetan
language family.” Areal Difusion and Genetic Inheritance, ed. by Alexandra Aikhenvald and
R.M.W. Dixon. Oxford: Oxford University Press.
____. 2003a. “Overview of Sino-Tibetan morphosyntax.” In hurgood and LaPolla (eds.), 22–42.
____. 2003b. A grammar of Qiang, with annotated texts and glossary. Berlin; New York: Mouton
de Gruyter.
____. 2010. “Language contact and language change in the history of the Sinitic languages.”
Procedia Social and Behavioral Sciences 2 (2010): 6858–6868. http://www.sciencedirect.
com/science/article/pii/S1877042810011705
Li, Fang-kuei. 1945. “Some old Chinese loan words in the Tai languages.” Harvard Journal of
Asiatic Studies 8: 333–42.
____. 1971. “Shang gu yin yánjiū [Studies on Old Chinese].” Tsing Hua Journal of Chinese Studies
9 (1–2): 1–61.
____. 1976. “Sino-Tai.” In Hashimoto (ed.), 230–39.
Lǐ, Jingzhong. 1994. Yuyan yanbian lun. Guangzhou: Guangzhou Chubanshe.
Lǐ, Rúlóng. 2005. “Gwuānyú xinan fanyan de “dǐcéng” yánjiū [On “substratum” in Southeast
Chinese dialects].” Minzu Yuwen 2005 (5): 1–15.
Luo, Xianglin. 1990. Zhongguo Minzu Shi [History of Minorities of China]. Taibei: Zhonghua
Wenhua Chuban Shi Ye She.
Manomaivibool, Prapin. 1975. A Study of Sino-hai Lexical Correspondences. Ph.D. dissertation,
University of Washington.
____. 1976a. “hai and Chinese – are they genetically related?” Computational Analysis of
African and Asian Languages 6: 11–31.
____. 1976b. “Layers of Chinese loanwords in hai.” In Tai linguistics in honor of Fang-Kuei Li,
ed. by homas Gething, J. Harris, and P. Kullavanijaya, 179–184. Bangkok: Chulalongkorn
University Press.
Matisof, James. 1973. “Tonogenesis in Southeast Asia.” In Consonant Types and Tone, ed. by
Larry Hyman, 73–95. (Southern California Occasional Papers in Linguistics 1). Los Ange-
les: Linguistics Program, University of California Los Angeles.
____. 1992. “Southeast Asian languages.” In International Encyclopedia of Linguistics, Vol IV, ed.
by William Bright and Bernard Comrie, 44–48. New York and Oxford: Oxford University
Press.
____. 1995. “Sino-Tibetan palatal suixes revisited.” In Nishi, et al. (eds.) 35–91.
 he origins of Sinitic 

____. 2001. “Genetic vs. contact relationship: prosodic difusibility in South-East Asian lan-
guages.” In  Areal Difusion and Genetic Inheritance: Problems in Comparative Linguis-
tics,  ed. by Alexandra Aikhenvald and R.M.W. Dixon, 291–327. Oxford and New York:
Oxford University Press.
____. 2003. Handbook of Tibeto-Burman: System and philosophy of Sino-Tibetan reconstruc-
tion. [University of California Publications in Linguistics 135]. Berkeley and Los Angeles:
University of California Press.
Mazo, Olga. 2002. “Relexes of the suix *-s in Classical Tibetan and Old Chinese.” Tibet, Past
and Present, ed. by Henk Blezer, 455–472. Leiden: Brill.
McWhorter, John. 2007. Language Interrupted: Signs of Non-Native Acquisition in Standard
Language Grammars. Oxford: Oxford University Press.
Meacham, William. 1996. “Deining the Hundred Yue.” Bulletin of the Indo-Paciic Prehistory
Association 15: 93–100.
Mei, Tsu-lin. 1970. “Tones and prosody in Middle Chinese and the origin of the rising tone.”
Harvard Journal of Asiatic Studies 30: 86–110.
____. 1980. “Sisheng bieyi zhong de shijian cengci.” Zhōngguó Yǔwén 1980 (6): 427–43.
____. 1988. “Neibu nigou Han yu san li [hree examples of internal reconstruction in Chinese].”
Zhongguo Yuwen 1988 (3): 169–81.
____. 2008. “Shanggu Hanyu dongci zhuoqing bieyi de laiyuan [he original diferences between
verbs with voiced and voiceless initials in Ancient Chinese].” Minzu Yuwen 2008 (3): 3–19.
Michailovsky, Boyd. 1988. La langue hayu. Paris: CNRS.
Nishi, Yoshio, James Matisof, and Yasuhiko Nagano (eds.) 1995. New Horizons in Tibeto-
Burman Morphosyntax. (Senri Ethnological Studies no. 41). Osaka: National Museum of
Ethnology.
Nishida, Tatsuo. 1975. “Common Tai and Archaic Chinese.” Studia Phonologica 9: 1–12.
____. 1976. “Some problems of the morpheme stock in Sino-Tibetan: A preliminary observa-
tion.” In Hashimoto (ed.), 30–38.
Norman, Jerry, and Mei Tsulin. 1976. “he Austroasiatics in ancient south China: Some lexical
evidence.” Monumenta Serica 32: 274–301.
Ostapirat, Weera. 2000. “Proto-Kra.” Linguistics of the Tibeto-Burman Area 23 (1): 1–251.
Peyraube, Alain. 1997a. “On word order in Archaic Chinese.” Cahiers de linguistique – Asie
orientale 26 (1): 3–20.
____. 1997b. “On word order and word order change in Pre-Archaic Chinese.” Zhungguo Jingnei
Yuyan ji Yuyanxue 4: 105–124.
Pulleyblank, E.G. 1983. “he Chinese and their neighbors in prehistoric and early historic
times.” In Keightley (ed.), 411–466.
____. 1995a. “Historical and prehistorical relationships of Chinese.” In Wang (ed.), 145–94.
____. 1995b. “Comments from Pulleyblank.” In Wang (ed.), 325–35.
Ratlif, Martha. 2010. Hmong-Mien Language History. (Studies in Language Change 8).
Canberra: Paciic Linguistics.
Sagart, Laurent. 1994a. “Proto-Austronesian and Old Chinese evidence for Sino-Austronesian.”
Oceanic Linguistics 33 (2): 271–308.
____. 1994b. “Questions of method in Chinese-Tibeto-Burman comparison.” 27th International
Conference on Sino-Tibetan Languages and Linguistics, Paris.
____. 1995. “Some remarks on the ancestry of Chinese.” In Wang (ed.), 195–223.
____. 1999. he Roots of Old Chinese. Philadelphia: Benjamins.
 Scott DeLancey

____. 2005. “Sino-Tibetan-Austronesian: an updated and improved argument.” In Sagart et al.

(eds.), 163–78.
____. 2011. “he Austroasiatics: East to west or west to east?” In Enield (ed.), 345–359.
____, Roger Blench, and Alicia Sanchez-Mazas (eds.) 2005. he Peopling of East Asia: Putting
Together Archaeology, Linguistics and Genetics. London: Routledge Curzon.
Sanchez-Mazas, Alicia, et al., (eds.). 2008. Past Human Migrations in East Asia: Matching
Archaeology, Linguistics and Genetics. London and New York: Routledge.
Sidwell, Paul, and Roger Blench. 2011. “he Austroasiatic Urheimat: the Southeastern Riverine
hypothesis.” In Enield (ed.) 315–343.
Starosta, Stanley. 2005. “Proto-East-Asian and the origin and dispersal of languages of East and
Southeast Asia and the Paciic.” In Sagart et al. (eds.), 82–197.
Stern, heodore. 1963. “A provisional sketch of Sizang (Siyin) Chin.” Asia Major (n.s.) 10:
222–78.
Sun, Hongkai. 1983. “Woguo Zang-Mian yu dongci de rencheng fanchou [he personal category
of verbs in Tibeto-Burman languages spoken in China].” Minzu Yuwen 1983 (2): 17–29.
Terrien de la Couperie. 1887. he Languages of China Before the Chinese. London: David Nutt.
homason, Sarah, and Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic
Linguistics. Berkeley and Los Angeles: University of California Press.
hurgood, Graham. 1985. “Pronouns, verb agreement systems, and the subgrouping of Tibeto-
Burman.” In Linguistics of the Sino-Tibetan area: he state of the art, ed. by Graham
hurgood, James Matisof, and David Bradley, 376–400. Canberra: Department of Lin-
guistics, Australian National University. [Paciic Linguistics Series C, no. 87].
____, and Randy LaPolla (eds.) 2003. he Sino-Tibetan languages. London and New York:
Routledge.
Tournadre, Nicolas. 2008. “Arguments against the concept of ‘conjunct’/’disjunct’ in Tibetan.”
In Chomolangma, Demawend und Kasbek: Festschrit für Roland Bielmeier zu seinem 65.
Geburtstag, ed. by Brigitte Huber, M. Volkart, and P. Widmer, 281–308. Halle: International
Institute for Tibetan and Buddhist Studies GmbH.
Trudgill, Peter. 2009. “Sociolinguistic typology and complexiication.” In Language Complexity
as an Evolving Variable, ed. by Geofrey Sampson, David Gil, and Peter Trudgill, 98–109.
Oxford and New York: Oxford University Press.
____. 2011. “Social structure, language contact, and language change.” In he SAGE Handbook
of Sociolinguistics, ed. by Ruth Wodak, Ruth, Barbara Johnstone, and Paul Kersill, Los
Angeles: SAGE.
Wadley, Stephen. 1996. “Altaic inluences on Beijing dialect: the Manchu case.” Journal of the
American Oriental Society 116 (1): 99–104.
Wang, William S–Y. (ed.) 1995. The Ancestry of the Chinese Language. [Journal of Chinese
Linguistics Monograph Series #8]. Berkeley: Project on Linguistic Analysis, University of
California.
Wang, Fushi. 1986. “A preliminary investigation of the genetic ailiation of the Miao-Yao lan-
guages.” International Symposium on the Minority Nationalities of China, Santa Barbara.
Wu, Anqi. 1996. “Han-Zang yu shidong he wancheng ti qianzhui de cancun yu tongyuan de
dongci cigen [Surviving preixes of causative and perfective aspect and cognative verb roots
of Sino-Tibetan languages].” Minzu Yuwen 1997 (6): 21–32.
____. 2008. “Shiqian huanan diqu de yuyan anchu [Language contact in Southern China in
prehistoric ages].” Minzu Yuwen 2008 (3): 21–35.
Wu, Fuxiang. “Characteristic syntactic patterns of Mandarin Chinese.” In this volume.
 he origins of Sinitic 

Wulf, Kurt. 1934. Chinesisch und Tai: Sprachvergleichende Untersuchungen. (DanskeVidenska-

bernes Selskab, Historisk-ilologiske Meddelelser 20.3). København: Levin & Munksgaard.
Yao, Yong-Gang, Long Nie, Henry Harpending, Yun-Xin Fu, Zhi-Gang Yuan, and Ya-Ping
Zhang. 2002. “Genetic relationship of Chinese ethnic populations revealed by mtDNA
sequence diversity.” American Journal of Physical Anthropology 118: 63–76.
Xu, Dan. 2006. Typological Change in Chinese Syntax. Oxford and New York: Oxford University
Press.