ChapterPDF Available

Abstract and Figures

Introduction The present contribution is concerned with the areal concentration of a number of linguistic features in the Transeurasian languages and its historical motivation. The label ‘Transeurasian’ was coined by Johanson and Robbeets (2010: 1–2) with reference to a large group of geographically adjacent languages, traditionally known as ‘Altaic’, that share a significant number of linguistic properties and include up to five different linguistic families: Japonic, Koreanic, Tungusic, Mongolic and Turkic. The question whether all similarities between the Transeurasian languages should be accounted for by language contact or whether some are the residue of a common ancestor is one of the most debated issues of historical comparative linguistics (see Robbeets 2005 for an overview of the debate). Since the term ‘linguistic area’ implies that the shared properties are the result of borrowing, I will refrain from a priori attaching it to the Transeurasian region and rely on the concept of ‘areality’ instead, that is, the geographical concentration of linguistic features, independent of how these features developed historically. Only after evaluating 27 structural features shared across the Transeurasian languages will I consider how the insights from the data are relevant to historical statements about the way the languages may have come to share these features, considering diffusion, genealogical relationship or an interaction of both factors as possible explanations. In spite of the strong polarization in the Transeurasian field between so-called ‘retentionists’, who view the similarities as arising from common descent, and ‘diffusionists’, who view them as arising from areal interaction, detailed characterizations of Transeurasian as a linguistic area are surprisingly rare in the linguistic literature. Poppe (1964) analysed Altaic as a ‘language type’ on the basis of a list of structural parallels shared between Korean, Tungusic, Mongolic and Turkic languages, and Rickmeyer (1989) elaborated on this research, adding data from Japanese. Even if these contributions provide an impressive list of shared features, they do not strictly identify Transeurasian as a language area because they do not (i) delimit the language type in relation to its neighbours, (ii) list deviations from the prototypical type in the peripheries, (iii) consider the extent to which the features in question are common or rare across the world as a whole, or (iv) attempt to distinguish contact-induced from genealogically motivated features.
Content may be subject to copyright.
Robbeets -- page 1 of 38
The Transeurasian languages
Martine Robbeets
1 Introduction
The present contribution is concerned with the areal concentration of a number of
linguistic features in the Transeurasian languages and its historical motivation. The
label “Transeurasian” was coined by Johanson & Robbeets (2010: 1-2) with
reference to a large group of geographically adjacent languages, traditionally
known as “Altaic”, that share a significant number of linguistic properties and
include up to five different linguistic families: Japonic, Koreanic, Tungusic,
Mongolic, and Turkic. The question whether all similarities between the
Transeurasian languages should be accounted for by language contact or whether
some are the residue of a common ancestor is one of the most debated issues of
historical comparative linguistics (see Robbeets 2005 for an overview of the
debate). Since the term ‘linguistic area’ implies that the shared properties are the
result of borrowing, I will refrain from a priori attaching it to the Transeurasian
region and rely on the concept of “areality” instead, i.e. the geographical
concentration of linguistic features, independent of how these features developed
historically. Only after evaluating 27 structural features shared across the
Transeurasian languages, will I consider how the insights from the data are relevant
for historical statements about the way the languages may have come to share these
features, considering diffusion, genealogical relationship or an interaction of both
factors as possible explanations.
In spite of the strong polarization in the Transeurasian field between so-called
“retentionists”, who view the similarities as arising from common descent and
“diffusionists”, who view them as arising from areal interaction, detailed
characterizations of Transeurasian as a linguistic area are surprisingly rare in the
linguistic literature. Poppe (1964) analysed Altaic as a “language type” on the basis
of a list of structural parallels shared between Korean, Tungusic, Mongolic and
Turkic languages and Rickmeyer (1989) elaborated on this research, adding data
from Japanese. Even if these contributions provide an impressive list of shared
features, they do not strictly identify Transeurasian as a language area because they
do not (i) delimit the language type in relation to its neighbors, (ii) list deviations
from the prototypical type in the peripheries, (iii) consider the extent to which the
features in question are common or rare across the world as a whole or (iv) attempt
to distinguish contact-induced from genealogically motivated features.
In this chapter, I attempt a partial answer to these concerns by providing a
typological profile of selected Transeurasian languages, along with their oldest
linguistically reliable historical varieties and by comparing this profile with the
behavior of languages immediately outside the Transeurasian region. In order to
examine external boundaries, I have included adjacent languages to the east (Ainu
and Nivkh in the northeast and Rukai in the southeast), to the south (Mandarin
Robbeets -- page 2 of 38
Chinese) and to the north (Kolyma Yukaghir, Ket and Eastern Khanty).1 These
languages are taken as horizontal comparative points representative of surrounding
areas such as the Siberian area (Nivkh, Kolyma Yukaghir, Ket, Eastern Khanty) or
the Mainland Southeast Asia area (Mandarin) and neighboring families such as
Austronesian (Mantauran Rukai), Sino-Tibetan (Mandarin), Yukaghiric (Kolyma
Yukaghir), Yeniseic (Ket), Uralic/Ob-Ugric (Eastern Khanty) or Ainuic (isolate
Ainu). Although Eastern Khanty can be taken as a representative of the Uralic
languages, the main boundary to the west, I have paid less attention to additional
western boundaries, excluding sample languages from the Caucasus region or from
the Indo-European languages because of the limited space available here.
The vertical comparison points in my analysis consist of a list of 27 features,
chosen to maximize positive (+) values for Transeurasian as opposed to
neighboring languages. Although all features reflect a certain internal coherence,
about half of them (i.e. 13) display deviations from the prototypical type in the
peripheries. Where possible, I add an estimation of the degree to which the feature
under discussion is common or rare across the world’s languages, relying on the
counts in the World Atlas of Language Structures or on other typological research
to be specified below.
Given the controversy between diffusionists and retentionists, we cannot simply
amass a number of shared features among the Transeurasian languages and allow
geographical adjacency to imply the probability of diffusion, without requiring any
linguistic support for this. Therefore, historical evidence suggesting the diffusion or
the retention of traits may be particularly telling in this particular case. For
representatives of the contemporary varieties of the five families belonging to the
Transeurasian continuum, I have chosen Turkish (Turkic), Khalkha Mongolian
(Mongolic), Evenki (Tungusic), Korean (Koreanic) and Japanese (Japonic) as
horizontal comparison points.2 However, in order to allow a diachronic perspective,
their profile will be supplemented by values from the oldest linguistically reliable
historical varieties of the individual families, i.e. Old Turkic (8th-14th century),
Middle Mongolian (13th-17th century)) and/or Written Mongolian, Manchu (17th-
19th century)), Middle Korean (15th-16th century)) and Old Japanese (8th
century)).3 In case a diachronic variety does not openly or productively reflect a
certain feature, but nevertheless preserves a trace of it, indicating that the value was
1 The following sources were consulted for retrieving linguistic data underlying the feature
values in neighboring languages: Gruzdeva 1998 for Nivkh; Maslova 2003a for Kolyma
Yukaghir; Werner 1997, Vajda 2004 and Georg 2007 for Ket; Filchenko 2007 for Eastern
Khanty; Li & Thompson 1989 for Mandarin; Zeitoun 2007 for Mantauran Rukai and;
Shibatani 1990 and Tamura 2000 for Ainu.
2 The following sources were consulted for retrieving linguistic data underlying the feature
values in contemporary Transeurasian languages: Göksel & Kerslake 2005 for Turkish;
Janhunen 2012 for Khalkha Mongolian; Bulatova & Grenoble 1999 and Nedjalkov 1997
for Evenki; Martin 1992 and Sohn 1994 for Korean and; Martin 1988 and Kaiser et al.
2001 for Japanese.
3 The following sources were consulted for retrieving linguistic data underlying the feature
values in historical Transeurasian languages: Erdal 2004 for Old Turkic; Street 1957,
Weiers 1966 and Rybatzki 2003 for Middle Mongolian; Poppe 1954 for Written
Mongolian; Gorelova 2002 for Manchu; Martin 1992 and Lee & Ramsey 2011 for Middle
Korean and; Vovin 2005, 2009 and Frellesvig 2010 for Old Japanese.
Robbeets -- page 3 of 38
positive in an earlier stage of the language, the historical variety will be marked
with a plus. In this way, we can obtain a glimpse of the unrecorded typological past
of the language in question.
The organization of this chapter is as follows. In Section 2, I will set up a
typological profile of the Transeurasian languages in relation to that of the selected
languages immediately outside the continuum. The linguistic levels discussed will
include phonology, lexicon and semantics, morphology and syntax. I intend to treat
grammaticalization patterns as a distinct level of analysis because, rather than
representing a static feature value, they are concerned with a dynamic force,
leading in languages to change from a less to a more grammatical status. I will
close Section 2 with a tabular overview, summarizing the presence of the 27
examined features in the selected languages by way of plus (+) and minus (-)
values. In Section 3, I will consider how the insights from these data are relevant
for general statements about areality, paying attention to the delimitation of
areality, peripheral deviations from the prototype, changes in areality and the
distinction between diffused and inherited features. In Section 4, I will conclude
this chapter.
2 Typological profile of the Transeurasian languages
2.1 Phonology
1. Predominantly polysyllabic root structure The Transeurasian languages, together
with their historical varieties, display a preponderance of polysyllabic roots, as do
most languages in North Asia. Contemporary and Old Japanese possess a relatively
great number of monosyllabic roots, many of which are attributed to root-internal
consonant loss and subsequent vowel contraction (Whitman 1990). These
phonological reductions argue against Janhunen’s (1997) suggestion that Japanese
derives from an originally monosyllabic language. Austronesian languages such as
Rukai are typically polysyllabic as well. Mandarin is the only language in the tables
that is marked with a negative value. Similar to the languages of Mainland
Southeast Asia it is predominantly monosyllabic, but, in comparison to Classical
Chinese, it has developed a greater number of polysyllabic roots through
compounding (Norman 1988: 86). As such, Japanese and Chinese occupy an
intermediate position between the languages of North and Southeast Asia.
2. Absence of complex tonal distinctions None of the Transeurasian languages is
tonal in the sense that each syllable is characterized by a distinctive pitch pattern.
This is also true for Austronesian languages, such as Rukai. With the exception of
Ket, which has been attributed a tone system with five oppositions in recent
descriptions by Vajda (2004) and Georg (2007), the neighboring languages of
North Asia are typically non-tonal as well. However, Nivkh, Japanese and some
varieties of Korean have suprasegmental systems which can be seen as transitional
between tonal and non-tonal languages. Nivkh makes distinctive use of two types
of tones, whereas Middle Japanese and Middle Korean use a system of pitch accent
that differentiates words according to the position of one prominent syllable after
which the pitch drops. This system survives in Contemporary Japanese, but it has
Robbeets -- page 4 of 38
been lost in Contemporary Standard Korean, where it developed into a vowel
length distinction. The two-way tone distinction and the pitch accent system are
highly restricted in comparison to complex tonal systems such as in Ket and
Mandarin, where each syllable is marked with one out of five distinctive tones.
Tonal languages are not only extremely widespread throughout Southeast Asia, but
also 42% of languages in Maddieson’s (2005: 58-61) sample of 526 languages
across the world are tonal.
3. Presence of vowel harmony Vowel harmony can be defined as a phenomenon
whereby vowels within a domain agree with each other in terms of one or more
features (Ko 2012: 7). It is a characteristic feature of the Transeurasian languages,
except Japanese, but it is also present in most Uralic languages, including Khanty
and in many other languages in North Asia such as in Yukaghir, Nivkh and Ainu.
Ket lacks vowel harmony and so do Rukai and Mandarin, as such reflecting
prototypical Austronesian and Mainland Southeast Asian behavior, respectively. In
Old Japanese, however, there is a restriction on the shape of root morphemes,
whereby the vowel o2 cannot occur in a root together with the vowels u, o1 or a.
This phenomenon, known as Arisaka’s law, has been taken as a kind of vowel
harmony, but it has been rejected from comparisons with other Transeurasian
languages because it applies to roots rather than to suffixes and because it does not
reflect palatal harmony, the type of harmony which was attributed to the
Transeurasian languages until recently (e.g. Frellesvig 2010: 44). However, in
lexicalized verb stems incorporating derivational suffixes, as well as in the noun
inflectional suffixes such as the plural suffix and the genitive suffix, there are traces
of a ~ o2 vowel alternation according to the quality of the vowels in the preceding
root, e.g. OJ no2 genitive vs. OJ -na- petrified in compounds such as OJ mi1-na-
moto (< water-GEN-base) ‘source, the headwaters’ (Rickmeyer 1989: 316;
Robbeets forthcoming c).
4. Presence of tongue root vowel harmony Among the various types of vowel
harmony, the most frequently attested ones across the languages of the world are
palatal harmony, labial harmony, height harmony and tongue-root harmony. Palatal
harmony requires all vowels within a domain to be exclusively front or back. It can
be found in most Uralic languages such as in Khanty as well as in the Turkic
languages (e.g. Tk. ip-ler (rope-PL) ‘ropes’ vs. pul-lar (stamp-PL) ‘stamps’). Since
the western Mongolic languages Oirat and Kalmuck display palatal harmony as
well, it has been proposed that the original system of Mongolic harmony was
palatal (Poppe 1955, Svantesson 1985). However, Ko (2012) demonstrated that the
original vowel harmony in Mongolic was in fact based on the opposition between
the advanced vs. retracted position of the tongue root, rather than on a palatal
contrast. He argued that the tongue root retraction system in Khalka (e.g. od-o:s
(feather-ABL) vs.
ɔ
d-
ɔ
:s (star-ABL)) represents retention rather than innovation.
Furthermore, he supported the view that Tungusic vowel harmony is Retracted
Tongue Root (RTR) based, as it is in Manchu and Evenki, and that the reduced
vowel harmony in contemporary Korean derives from a tongue-root based system
in Middle Korean. As far as the harmony-like opposition between o2 and u, o1 or a
in Old Japanese is concerned, the recent reconstuction of a seven-vowel system in
proto-Japonic by Frellesvig and Whitman (2008) implies an underlying opposition
Robbeets -- page 5 of 38
between pJ *ɨ , *! and *u, *o, *a, which does not exclude an original RTR based
contrast. Whereas Vovin (1993: 50-51) and Bugaeva (2013: 26-28) reconstruct
palatal harmony in Ainu, Shibatani (1990: 15) speculates that the Ainu opposition
between o and u, a might have its origin in tongue root harmony, but here the
indications are even weaker than in the Japanese case. According to Maslova
(2003: 35), Yukaghir might be more appropriately described as having tongue root
harmony than palatal harmony. Chukchi also displays tongue-root harmony.
Although Gruzdeva (1998: 10) suggests that Nivkh leaves traces of height
harmony, Janhunen (1981) and Ko, Whitman and Joseph (2014) interpret the
system in terms of tongue root harmony. Cross-linguistically, tongue root harmony
seems to be concentrated in Niger-Congo and Nilo-Saharan languages. Outside
Africa and the Northeast Asian region the phenomenon seems to be rather rare:
only Native American languages such as Nez Perce and Coeur d’Alene Salishan are
known to have the feature (Ko 2012: 11-12). A rough estimate would be that less
than 10% of the world’s languages have a tongue root vowel harmony system.
5. Absence of initial velar nasal In most Turkic languages as well as in Mongolic
languages and Korean the velar nasal ŋ- cannot appear in word-initial position.
Japanese lacks a velar nasal phoneme. In the Tungusic languages, with exception of
Manchu, however, ŋ- can appear word-initially, but generally restricted to a specific
phonological environment, notably when it is followed by the sonorants n, r , l, m,
y, e.g. Evk. ŋene- ‘to go’, Ma. genu- ‘to go together’, Evk. ŋe:le-, Ma. gele- ‘to
fear’, etc. According to Poppe (1964: 4) the initial velar nasal in Tungusic is the
result of secondary assimilation of pTg *g-, which implies that originally *ŋ- was
absent in Tungusic as well. The assimilation was probably triggered by influence
from languages in the Siberian area, such as Nivkh, which allow initial velar nasals
(Anderson 2006). It is under the same influence that initial ŋ became allowed in
Dolgan (Turkic), e.g. ŋassa ‘pipe’. In Khanty, Ket, Kolyma Yukaghir, Ainu and
Mandarin ŋ- does not occur in word-initial position. Rukai allows an initial velar
nasal, e.g. ŋa
ɭ
ai ‘saliva’. In Anderson’s (2005: 42) sample of 468 languages, 69 %
lack an initial velar nasal. Among the languages of the world that have a velar nasal
phoneme, as is the case for most Transeurasian languages, only 35 % do not use it
in word-initial position.
6. Absence of initial r- Throughout the Transeurasian languages, the consonant r- is
not allowed to occur word-initially, except in borrowings (e.g. J rajio, K latiwo,
Even radio, Khal. radio, Tk. radyo ‘radio’). This is also true for Kolyma Yukaghir.
Ket lacks a phoneme /r/ altogether. Although initial *r- is not reconstructed for
proto-Uralic, Khanty is atypical in this sense, e.g. ra
ɣ
ta ‘to drop, slide’ and
ɣ
‘garbage’. Nivkh, Ainu, Mandarin and Rukai also have native words in initial r-.
7. Absence of initial consonant clusters None of the Transeurasian languages
tolerates initial consonant clusters, although medial clusters are tolerated in Turkic,
Mongolic, Tungusic and Korean but, on the face of it, not in Japanese. On the basis
of morphological, etymological, dialectal and textual evidence, however, it is safe
to assume that the Old Japanese obstruents OJ b, d, g, z resulted from the
rephonologization of nasal obstruent clusters pJ *np, pJ *nt, pJ *nk, pJ *ns
Robbeets -- page 6 of 38
(Robbeets 2008). Reminiscent of how the Transeurasian languages do not allow for
consonant clusters in initial position, Old Japanese did not permit word-initial
voiced obstruents except in mimetic adverbs. From the the ninth century onwards,
as loans from Chinese began to have a major impact, the restriction was relaxed and
initial voiced obstruents began to appear in borrowings and in contracted native
forms. The avoidance of consonant clusters is further characteristic of Uralic
languages, such as Khanty. Similarly, Yukaghir, Ket, Ainu and Rukai tolerate only
single consonants in word-initial position. Wordinitial clusters may comprise at
most two consonants in Nivkh, e.g. mra ‘fault’ and ksynz ‘witch’. Although
Mandarin lacks consonant clusters, there is strong evidence that in Old Chinese
(first millennium BC) a variety of consonant clusters could occur at the beginning
of the syllable as well (Norman 1988: 9-10). The simplification and eventual loss of
consonant clusters appears to be a tendency affecting most of the Mainland
Southeast Asia area. It is possible that early contacts between Chinese and
Transeurasian, that has never tolerated initial clusters, have triggered the
development along these lines.
8. Presence of voicing distinction for stops Turkic, Mongolic and Tungusic
languages share a voiced-voiceless opposition for stops. In Contemporary and
Middle Korean, stops display an opposition between lax (p), aspirated (ph) and
tensed (p’). Even if the lax stops become lightly voiced between voiced sounds,
there is no phonemic voicing distinction. The Japanese voicing distinction for stops
is a secondary development. As mentioned in the previous paragraph, voiced stops
derive from prenasalized voiceless stops, so originally, Japanese lacked a voicing
distinction. Khanty lacks a voicing distinction for stops, a feature characteristic of
proto-Uralic, although many contemporary Uralic languages have developed an
original singleton-geminate contrast into a voicing distinction. For example,
although the contrast between /p/ and /pp/ in proto-Uralic *lapa ‘flat surface; leaf’
and *tappa- ‘to stamp with feet; to hit, knock’ is maintained in the Finnish reflexes
lapa ‘shoulderblade, leaf surface’ and tappa- ‘to beat to death, kill’, it has usually
developed into a distinction between /b/ and /p/ such as in Estonian laba ‘surface’
and tapa- ‘kill, slaughter’.. Ket and Yukaghir display a voicing distinction, but
languages to the extreme northeast such as Ainu, Nivkh and Chukchi do not.
Mandarin, like Nivkh, has a distinction between aspirated and unaspirated stops,
but lacks a voiced-voiceless oppositon. Characteristic of most Austronesian
languages, Rukai also displays voice distinction for stops.
2.2 Lexicon and semantics
9. Preference for a non-verbal strategy with (extra-family) verbal borrowing As far
as the mechanisms of loan verb accomodation are concerned, most recipient
languages can be categorized into two distinct groups: borrowed verbs either arrive
as verbs, needing no formal accommodation, or, they arrive as non-verbs and need
formal accommodation. In Wohlgemuth’s (2009) terminology, the first group
represents “Direct Insertion”, while the second group represents either “Indirect
Insertion”, when the formal accommodation involves a verbalizer or else, “Light
Verb Strategy”, when the borrowed verb is integrated into a complex predicate.
Turkic, Mongolic, Korean and Japanese can be assigned to the second group
Robbeets -- page 7 of 38
because they display a clear preference for the non-verbal strategy (Wohlgemuth
(2009: 159, 161)); for instance, Tk. klik-le- and klik et- << English click; Khal. zee-
l- << Mandarin zhài ‘borrow, lend’; K coking ha-, J zyogingu suru ‘to jog’ <<
English jog; J demo-r- << English demonstrate. Whereas the northern Tungusic
languages prefer to borrow verbs through direct insertion, e.g. Evk. vypolńaj- <<
Russian vypolnja-t’ ‘to fulfill, carry out’, the southern Tungusic languages use
verbalizers, e.g. Ud. tancewa-la- << Russian tancewa-t’ ‘to dance’ and Na.
voprosa-la- << Russian voproša-t’ ‘to inquire, question’. In contrast to the
Transeurasian languages, Uralic languages such as Khanty, Austronesian languages
such as Rukai, Ainu and Mandarin show a strong preference for direct insertion
(Wohlgemuth 2009: 158, 161; Tamura 2000: 267). Yukaghir and Nivkh did not
integrate any recognizable verbal borrowings from Russian or from other foreign
languages into their lexicons (Fubito Endo p.c.; Ekaterina Gruzdeva p.c.). In
Wohlgemuth’s (2009: 157) sample, 55% of languages worldwide are found to use
direct insertion, while the remainder prefer non-verbal strategies such as indirect
insertion and the light verb strategy.
10. Presence of a two-way proximal-distal distinction in demonstrative pronouns
Although Old Turkic displays a two-way distinction in its demonstratives, i.e. OT
bo / bun- ‘this’ vs. ol / an- ‘that’, many contemporary Turkic languages such as
Turkish make a three-way distinction, e.g. Tk. bu ‘this’, şu ‘that’, o ‘that (over
there)’. Demonstrative pronouns in earlier and contemporary varieties of Mongolic
and Tungusic exhibit a proximal-distal distinction: MMo. ene ‘this’ vs. tere ‘that’,
Khal. e- ‘this’ vs. te- ‘that’, Ma. ere ‘this’ vs. tere ‘that’ and Evk. er(i) ‘this’ vs.
tar(i) ‘that’. Demonstrative pronouns in Contemporary and Middle Korean,
however, show a proximal-mesial-distal opposition: K i ‘this’, ku ‘that’, ce ‘that
over there’ and MK i ‘this’, ku ‘that’, tye ‘that over there’. This is also true for
Contemporary Japanese: J ko- ‘this’, so- ‘that’, a- ‘that over there’. In contrast to
most accounts of Old Japanese demonstratives, which posit a three-way contrast
between OJ ko2 ‘this’, so2 ‘that’ and ka ‘that over there’, Frellesvig (2010: 139-142)
argued that OJ ka was not a productive member of the demonstrative system and
that pre-Old Japanese had a simple proximal-distal distinction. While Khanty
distinguishes between proximal timi ‘this (here)’ and distal tom
ɨ
‘that (there)’,
Yukaghir, Ket and Ainu have a three-way opposition, each demonstrative pronoun
denoting a different degree of proximity: Yukagir tiŋ ‘this’ (proximal), adiŋ ~ ediŋ
‘that’ (mesial), taŋ ‘that’ (distal); Ket tu- ‘this, that’ (neutral), ki- ‘this, that’
(proximal); qa- ‘this, that’ (distal) and Ainu ta an ‘this’ (distal), ne an ‘that’
(mesial), to an okai ‘that over there’ (distal). Nivkh makes as many as five
distinctions: tyd’ ‘this’ (near and visible), hyd’ ‘this, that’ (distant), ad’ ‘that’ (more
distant and visible), aixnt ‘that’ (most distant), kud’ ‘that’ (absent).4 Rukai
distinguishes four demonstrative pronouns in terms of visibility and distance: ’ina
‘this’ (proximal), ana ‘that’ (mesial), ona ‘that over there’ (distal but visible),
dhona ‘that over there’ (distal and invisible). Mandarin has a two-way distinction
4 Note that this analysis deviates from the feature values given for distance contrasts in
demonstratives by Diessel (2005: 170-173), since he marks Ainu, Nivkh, Yukaghir and
Turkish as having a two-way contrast.
Robbeets -- page 8 of 38
between proximal zhè(ge) ‘this’ and distal nà(ge) ‘that’, which developed from a
three way-distinction in Classical Chinese between neutral, proximal and distal. In
Diessel’s (2005: 170-173) sample of 234 languages, 54% exhibit a two-way
distance contrast in demonstratives, while 38% exhibit a three-way contrast.
11. Inclusive-exclusive distinction in first person plural pronouns Among the
Turkic languages, there are no unique pronominal forms that distinguish inclusive
from exclusive person forms. Although Old Turkic and most presently spoken
varieties of Turkic distinguish between a first person plural (Tk./OT biz ‘we’) and
an augmented plural form (Tk. / OT biz-ler ‘we (as a group)’), Nevskaya (2010:
124) argues for a collective interpretation of the augmented plural, denoting “an
isolated group of people who want to oppose themselves to the others”, rather than
an inclusive interpretation as suggested by Grönbech (1936: 81). The Middle
Mongolian distinction between exclusive ba and inclusive bida is formally
preserved in the Khalka oblique paradigm in the variation between formally
exclusive man- and formally inclusive bidn-, but the functional distinction has been
lost. In the Tungusic languages, however, the inclusive-exclusive opposition is
generally well preserved, e.g. exclusive Ma. be, Evk. bu vs. inclusive Ma. muse,
Evk. mut ~ mit. Similar to the Turkic languages, Middle and Contemporary Korean
distinguish between a first person plural (K/MK wuli ‘we’) and an augmented
plural form (K wuli-tul, MK wuli-tolh ‘we (as a group)’ in which K tul, MK ·tolh is
a collective marker. Contemporary Japanese lacks an inclusive-exclusive
distinction and cannot derive an augmented plural from the first person plural
watasi-tati (I-PL) ‘we’. Old Japanese also lacks the distinction, but the stem OJ wa-
‘I, we’ can be used as a first person plural in the possessive case form, but it can
also be augmented with a collective marker -ra ~ -re to OJ ware ‘we’, a form which
in its turn has been augmented into ware-ra ‘we’ later in Japanese. As is the case
for many Uralic languages, Khanty marks a dual distinction, but not an inclusive-
exclusive distinction on its person pronouns. While Ket and Yukaghir lack the
distinction, Nivkh distinguishes between exclusive n’yŋ and inclusive mer ~ mir.
Although Ainu personal affixes on the verb have an inclusive-exclusive distinction,
the first personal pronoun aoka(i) only has a single form.5 The distinction found in
the first person plural pronouns between exclusive wǒmen and inclusive zánmen
‘we’ of Beijing and certain other northern Chinese dialects may be due
Transeurasian influence. Such a distinction was not found in Old Chinese, and it
began to appear in North China during the period of Altaic rule. It is significant in
this regard that both Middle Mongolian, spoken under the Yuan dynasty and
Manchu distinguish exclusive and inclusive forms. Rukai distinguishes exclusive -
nai ~ nai- (NOM) from inclusive -mita ~ ta- (NOM), a feature characteristic of
Austronesian languages. In Cysouw’s (2005: 166-167) sample of 200 languages,
31% distinguish between an inclusive and an exclusive with independent pronouns.
12. Property words may be verbally or nominally encoded Cross-linguistically
adjectives have no prototypical encoding strategy of their own: they will align
themselves either with verbs or with nominals (Stassen 1997: 30). Across the
5 Note that my evaluation differs from Cysouw’s (2005: 166-167) analysis, which marks
Ainu as having an inclusive/exclusive distinction with independent pronouns.
Robbeets -- page 9 of 38
Transeurasian languages, the encoding of property words appears to be mixed
because, at least in the earlier stages, both the nominal and the verbal strategy is
used (Robbeets forthcoming c). Generally, this mixed encoding is split in the sense
that most property words have only a single encoding option, with the exception of
some instances of switched encoding mentioned under feature value 13. In Old
Turkic, most property words are nominally encoded, but there seems to be a
tendency to apply the verbal strategy to words expressing time-unstable properties
such as OTk. bädü- ‘to be(come) big, great’, OTk. isi- ‘to be hot’, OTk. kat- ‘to be
hard, firm, tough’, OTk. kïz- ‘to be red’, OTk. tumlï- ‘to be cold’, OTk. us- ‘to be
thirsty’, OTk. tïgra- ‘to be tough’, etc. Contemporary Turkic languages maintain
only few reflexes of these verbal property words, e.g. Tk. büyü- ‘to be(come) large’
but in the majority of cases, the earlier verbal property word has been derived
through a deverbal noun suffix into a nominal adjective (e.g. Tk. büyük ‘big’).
Similarly, most property words are nominally encoded in Mongolic, but there is a
tendency to apply the verbal strategy to less permanent properties in Middle
Mongolian such as MMo. ayu- ‘to be(come) afraid’, MMo. čat- ‘to be ripe,
be(come) saturated’, MMo. hiče- ‘to be ashamed’, WMo. qala- ‘to be(come)
warm’, MMo. sohta- ‘to be drunk’, etc. Contemporary Mongolic languages such as
Khalkha maintain only few reflexes of these verbal property words, e.g. Khal. ayu:-
‘be afraid’. The same is true for Tungusic, where contemporary languages such as
Manchu (e.g. Ma aka- ‘to be sad’, Ma. bere- ‘to be lame’, Ma. ebi- ‘to be satiated’)
and Evenki (e.g. Evk. ukti- ‘to be hungry’, Evk. uwi- ‘to be satiated’, Evk. buli:- ‘to
be sad’) may occasionally exhibit verbal encoding.6 In Korean, there are property
words such as K kanan ha- ‘to be poor’ and phikon ha- ‘to be tired’ that consist of
a nominal root and the auxiliary ha- ‘to be in the state of’ and whose bases are
called “adjectival nouns” (Martin 1992: 189, 190; Sohn 1994: 219-220). However,
the majority of property words are inflected in essentially the same way as verbs,
e.g. K kwut-, MK kwut- ‘to be(come) hard’, K noph-, MK nwoph- ‘to be high’, etc.
Some Japanese property words, such as J sizuka, OJ siduka ‘quiet’, J/OJ tasika
‘trustworthy’ are encoded exclusively nominally, while others such as J/ OJ taka-
‘to be high’, J/OJ kata- ‘to be hard, tough’ are essentially inflected in a similar way
as verbs. In line with most Uralic languages, property words in Khanty are
exclusively nominally encoded. This is also true for Ket. In Yukaghir, Ainu and
Nivkh, however, property words are exclusively verbally encoded. As in the case
with most Transeurasian languages, Ainu property verbs express both the property
and the process leading to the property, e.g. pirka ‘to be(come) good’. In line with
Mainland Southeast Asian and Austronesian languages, Mandarin and Rukai use
verbal encodings for property words. In Stassen’s (2005: 478-481) sample of 386
languages, 27% have mixed encoding in predicative adjectives.
13. Some property words exhibit switched encoding Some property words in the
Transeurasian languages, especially in the earlier varieties, further exhibit traces of
switching, whereby the same property word can have both nominal and verbal
encoding: e.g. OT. ač ‘hungry’ / ač- ‘to be hungry’, OT. keč ‘late, slow’/ keč- ‘to
6 Note that my analysis deviates from the feature values inserted for predicative adjectives
by Stassen (2005: 480-481), in which Evenki and Manchu are marked as exclusively
verbal encoding, in line with the traditional view.
Robbeets -- page 10 of 38
be late, slow’; MMo. bulqa ‘hostile; hostility’ / bulqa- ‘to be hostile’; Ma. jalu
‘full’/ jalu- ‘to be full’, Ma. sula ‘loose, free’/ sula- ‘to be loose, be free’; MK toso-
vs. MK toso ho- ‘to be warm’; OJ taka ‘high’ / taka- ‘to be high’, OJ opo ‘big / OJ
opo- ‘to be big’, etc. None of the neighboring languages, except Tundra Yukaghir,
exhibits such behavior. There, two property words, i.e. juku ‘small’ and t’ama ‘big’
occur as noun modifiers without overt adnominalizers, e.g. t’ama-d’ohoje (big-
sword) ‘sabre’ in addition to having a verbal encoding, for example, in the deverbal
inchoative t’ama-mu- (be.big-INCH) ‘to grow, become big’ (Maslova 2003b: 14).
Logically, the proportion of languages exhibiting mixed and switched encoding will
be lower than 27%, i.e. the proportion of languages with mixed encoding in
general.
14. Partial emphatic reduplication of nominal property words Partial emphatic
reduplication is a phenomenon whereby the first consonant (if present) and vowel
of a nominal property word are repeated with the addition of another consonant to
indicate the presence of the property to the utmost degree. Whaley (2000) found
that is widespread in Turkic, Mongolic and Tungusic, e.g. Tk. bem-beyaz ‘snow
white’, OT kap-kara ‘quite black’, Kal. xob-xoldu: ‘frozen through’, WMo. ub-
ula
ɣ
an ‘completely red’, Evk. ab-aya ‘very good’. I have not been able to find
examples in Manchu, but the phenomenon is present in Sibe, a presently spoken
variety of Manchu, e.g. fak-farxun ‘extremely dark’. In Tungusic, emphatic
reduplication is restricted to Sibe, Kile-Nanai, Solon Evenki and Oroqen, i.e. the
languages spoken on Chinese soil, which have been under strong influence of the
Mongolic languages Khalka and Dagur. On the basis of this distribution and
because the greatest flexibility, in terms of both the number of reduplicated words
and the type of concepts they denote, is found in Turkic, Whaley (2000: 358)
argued for a diffusion of the feature from Turkic to Mongolic to Tungusic.
Japanese, Korean and the neighboring languages under examination do not display
partial emphatic reduplication. In Rukai, however, descriptive verbs are partially
reduplicated in comparative constructions (see feature 23).
2.3 Morphology
15. Morphology is agglutinative Agglutinative languages connect morphemes
linearly in a way that there is a one on one relationship between a morpheme and its
meaning. The Transeurasian languages belong to a North Asian and European belt
of agglutinative languages together with the Uralic languages, including Khanty
and other languages of the Siberian area such as Ket, Yukaghir, Chukchi and
Nivkh. Ainu has agglutinative morphology and so do the Austronesian languages,
including Rukai. Chinese, the only analytic language under examination, has
triggered a decrease of agglutinating features in Tungusic as one moves from north
to east and further south. Manchu is the most analytical among the Transeurasian
languages; it treats case forms, for instance, as particles rather than suffixes.
16. Inflectional morphology is predominantly suffixing Across the strongly
suffixing Transeurasian languages, prefixation is rare and restricted to derivational
morphology, such as the partial emphatic reduplication (see feature 14 above) and
some derivational prefixes in Korean (e.g. K yel- ‘young, new’ in yel-cwungi ‘a
Robbeets -- page 11 of 38
chick out of its shell’) and in Japanese (e.g. ma- intensive in ma-siro ‘snow white’).
As is the case for most Uralic languages, Khanty is strongly suffixing and so is
Yukaghir. Nivkh is considered to be weakly suffixing. In Ket, nominal inflectional
morphology is strongly suffixing, whereas verb inflection is predominantly
prefixing. In Ainu and Rukai, inflection makes use of both prefixes and suffixes.
Probably due to Transeurasian influence, Mandarin is hard to assign unequivocally
to either the isolating or weakly suffixing type, but Sinitic varieties in general tend
towards the isolating pole. In Dryer’s (2005: 110-113) sample of 894 languages,
43% are strongly suffixing.
17. Absence of obligatory numeral classifiers Although in Turkic and Mongolic
some nouns of low countability may be accompanied by a unit of measure by
means of which they can be counted, e.g. Tk. sekiz bardak su (eight glass water)
‘eight glasses of water’, OT yeti tutum talkan (seven handful parched.grain) ‘seven
handfuls of parched grain’, Khal. gourben debter nom (three volume book) ‘three
volumes of books’, etc. these languages do not make obligatory use of sortal
numeral classifiers. Similar to the use of collective suffixes for counting people in
Old Turkic (OT -(A)gU in e.g. u
̈
čägü ‘three together’) and Middle Mongolian
(MMo. -UlA ~ AlA in e.g. qoya’ula ‘two together’), the Tungusic languages use a
variety of collective suffixes following numerals from ‘two’ to ‘ten’ such as Evk. -
kt(e) and -ni for counting people (e.g. d’u-kte ‘two (people together); (we, you,
they) two’), Evk. -gdA/ -ngnA for counting objects, Evk. -llA for counting the
number of days (e.g. nada-lla ‘seven days, a week’), Evk. -nu / -pu for counting the
number of tents (e.g. ilan-nu ‘three tents’) and Evk. -musa denoting the number of
places or directions. However, only Manchu has developed about 70 sortal numeral
classifiers, which divide the inventory of count nouns into semantic classes, each of
which is associated with a different classifier, such as fesin which is used for
objects equipped with a handle, e.g. ilan fesin loho (three CLAS sword) ‘three
swords’. These words have original lexical meanings, e.g. fesin ‘haft, shaft,
handle’, but under Chinese influence they have grammaticalized into classifiers,
which are not obligatory in Manchu. Loho ilan (sword three) ‘three swords’, for
instance, is equally possible. Whereas the standard pattern in Middle Korean was to
modify a noun with a preposed numeral, e.g. twu kalh (two knife) ‘two knives’, the
most common pattern in Contemporary Korean makes use of a classifier, e.g. pus
se:k calwu (writing.brush three CLAS) in which calwu denotes long objects with
handles. However, the original pattern surfaces in expressions such as K twu nala
‘two countries’ and the use of classifiers remains optional in Korean, e.g. kalh
hana-ka issta (knife one-NOM be.present) ‘there is one knife’.7 With Chinese
influence inundating the language from Middle Korean times onwards, the
classifiers developed from native words under Chinese influence or were borrowed
as such from Chinese. Note that Middle Korean leaves traces of specialized suffixes
to count days, e.g. *-(o/u)l in saol ‘three days’, naol ‘four days’ etc. and that some
Korean dialects use a suffix -i to count persons, e.g. se:-i ‘three people’, ne:-i ‘four
people’, which recalls the use of collective suffixes in the other Transeurasian
languages. While there is an extensive list of obligatory classifiers in Contemporary
7 Note that my evaluation differs from Gil’s (2005: 228-229) interpretation that Korean has
obligatory numeral classifiers.
Robbeets -- page 12 of 38
Japanese, e.g. enpitu san-bon (pencil three-CLASS) ‘three pencils’, the system of
classifiers is much less developed in Old Japanese, where Chinese influence is
restricted to a minimum. Numerals could be used with nouns, without intervening
classifiers, e.g. OJ nana se (seven rapid) ‘seven rapids’ and the so-called
“classifiers” are restricted to roughly six suffixes, i.e. -ka for counting days starting
from the numeral ‘two’, -tu / -ti for counting objects, -ri for persons, -mo2to2 for
grassy plants, -pe1 for layers and -ka for plants. It is not unlikely that these suffixes
originate in collective suffixes. Numeral classifiers are absent in Uralic languages
such as Khanty as well as in Yukaghir and Ket. Ainu makes use of a small set of
obligatory classifiers such as -n / -iw for persons; -pe / -p for animals and things,
and rerko for counting days starting from the numeral ‘two’ (with irregular forms
tutko ‘two days’ and rerko ‘three days’). Nivkh distinguishes between 26 semantic
classes with different numeral forms for each class. The obligatory use of classifiers
is a widespread feature shared by Mandarin and the languages of Southeast Asia,
but the use of classifiers in Classical Chinese was the exception rather than the rule.
In Rukai the use of classifiers is optional in the sense that it uses a set of unaffixed
numerals without classifiers as well as a set of bound numerals which combine with
five different sortal classifiers to form verbs. In Gil’s (2005: 226-229) sample of
400 languages, 80% lack obligatory numeral classifiers.
18. Presence of mi-Ti opposition in first vs. second singular person pronouns
Nichols (2012) observes that m-T pronominal paradigms with first person labial
nasal m and second person apical or palatal obstruent t, c, s, etc. are much more
common in northern Eurasia than elsewhere in the world. Janhunen (2013: 213)
adds that there is a smaller group of mi-Ti languages extending from Uralic in the
west, to Turkic, Mongolic and Tungusic in the east to Yukaghir in the north, in
which not only the initial consonant but also the root vowel of the singular stems
shows a basic similarity, in that it contains a non-low unrounded front vowel i or e.
Although m is absent in the nominative first person singular in the Turkic,
Mongolic and Tungusic languages, e.g. Tk ben, OT ben, Khal. bii, MMo. bi, Ma.
bi, Evk. bi:, it has developed in oblique forms through assimilation to the nasal
oblique suffix -n, e.g. OT min-, Khal. min-ii (GEN), MMo. mi-nu (GEN), Ma. min-,
Evk. min-. The second person singular forms all reflect a voiceless dental T, i.e. Tk.
sen, OT sen, Khal. cii, MMo. ci, Ma. si, Evk. si:. The Korean pronouns are, among
others, first singular K/MK na and second singular K/MK ne. In Japanese, J watasi
and OJ wa are among others used in the first singular, while a variety of
contemporary pronouns and OJ na are used in the second singular. Although the
proto-Uralic first and second singular pronouns *mun and *tun reflect an m-T
distinction (Janhunen 1982: 35), Khanty is deviant in having first singular and
second singular ŋ. In Yukaghir, however, the mi-Ti opposition is present in first
singular met vs. second singular tet. In Nivkh, the distinction is absent in the
singular pronouns, first person n’i vs. second person či, but it is present in the
opposition between the first plural inclusive mir/mer and the second plural pronoun
čiŋ. The opposition is not found in Ket, Ainu, Chinese and Rukai. In Nichols and
Peterson’s (2005: 546-551) sample of 230 languages, 13% display an m-T
opposition in first vs. second person pronouns. Logically, languages reflecting a mi-
Ti opposition will represent an even smaller proportion.
Robbeets -- page 13 of 38
19. Formation of a secondary oblique stem of personal pronouns With the
exception of Korean, the Transeurasian languages share a tendency to form a
secondary oblique stem of the personal pronouns by means of a suffix which
phonologically may be identified as the dental nasal -n-. In most contemporary
Turkic languages, the nominative and oblique forms have merged, e.g. Tk. ben for
the first singular nominative and oblique, but in Old Turkic the first singular
nominative bän is distinguished from the oblique stem min-, which can be derived
from an original pTk *bi-n- (1SG-OBL-). Similarly, the Mongolic and Tungusic
languages derive oblique pronominal stems from the nominative roots through a
nasal suffix, for instance, in the first person plural pronouns MMo. ba (NOM) vs.
man- (OBL) and Khal. bid (NOM) vs. bidn- (OBL) and in the first person singular
pronouns Ma. bi (NOM) vs. min- (OBL), Evk. bi: (NOM) vs. min- (OBL). There
are no oblique pronominal stems in Contemporary Japanese, but Old Japanese
leaves traces of an oblique nasal suffix in some case forms, e.g. in the Eastern OJ
first person singular dative wa-nu-ni in alternation with Western OJ wa-ni. Vovin
(2005: 229-230) further found that an original Japonic pronominal oblique *-n- is
well supported by Northern Ryukyuan dialects where the first person pronoun uses
waa- as the nominative and genitive base and extended waN- in the oblique cases.
The erosion of the pronominal paradigm in Korean and Japanese may be due to the
gradual de-pronominalization in the recorded history of these languages, whereby
the system of personal pronouns became replaced by various terms of address and
self-reference, probably under Chinese influence. The oblique nasal suffix is an
important element in the Uralic pronominal paradigm as well, e.g. the Khanty first
person pronoun (NOM) vs. män- (OBL). Ket, Yukaghir, Ainu and Mandarin,
however, do not derive secondary oblique stems. The third person singular pronoun
in Nivkh has both regular and suppletive case forms, e.g. if-øn (3SG-NOM) vs. if-
toX ~ e-rx (3SG-DAT/ADD), but here the oblique form is not derived from the
nominative base. Rukai personal pronouns have different shapes for nominative,
topic, genitive and oblique cases, e.g. the first person singular -lrao (NOM), ilrae
(TOP), -li (GEN) vs. -iae (OBL), in which the oblique seems to be formally derived
from the nominative base by means of the same i- …-e marking as in the topic
form.
2.4 Syntax
20. SOV (Subject-Object-Verb) sentence order Syntactically, the Transeurasian
languages pattern as typical SOV languages but the sentence order is not rigid.
SOV is also among the characteristic features of the Uralic languages, here
represented by Khanty. Languages to the north such as Yukaghir and Ket or to the
northeast such as Ainu and Nivkh are almost all SOV languages, while those to the
southeast are virtually all SVO languages. Mandarin, and in fact all major varieties
of Chinese, corresponds with Southeast Asia with respect to verb-object order. Like
most Austronesian languages, Rukai tends to be verb-initial, but the word order is
non-rigid, switching freely between VSO and VOS. In Dryer’s (2005: 330-333)
sample of 1228 languages, 40% have SOV sentence order.
Robbeets -- page 14 of 38
21. GAN (Genitive-Noun / Adjective-Noun) phrase order A modifier-before-
headword word order in the sentence (SOV) is expected to correlate with a
modifier-head order within the noun phrase (GAN), whereby adjectives, genitives
and modifiers in general occur before the nouns to which they refer. This is the case
for the Transeurasian languages, the Uralic languages including Khanty and other
languages of North Asia such as Yukaghir, Ket, Ainu and Nivkh. Mandarin,
however, runs against the implicational expectation since genitives and adjectives
occur before the nouns to which they refer in spite of SVO sentence order. This
combination of feature values is absent in almost all the other languages of
Southeast Asia and has probably arisen under influence of the Transeurasian
languages. Rukai combines an adjective-noun order with a noun-genitive order
(ANG). In Dryer’s (2005: 350-357) sample of 1105 languages, 55% have Genitive-
Noun order, while out of 1213 languages, 28% have Adjective-Noun order.
22. Extensive use of converbs Converbs, also known as gerunds or adverbial
participles, can be defined as nonfinite verb forms whose main function is to mark
adverbial subordination (Haspelmath 1995: 3). Originally coined by the Altaic
scholar Ramstedt, the term converb was adopted from Transeurasian linguistics to
denote a cross-linguistic category. The Transeurasian languages are converb-
prominent languages in the sense that they use converbs rather than adverbial
subordinators as found in many European languages; see the examples below
(Bisang 1995, 1998; Johanson 1995; Nedjalkov 1995; Alpatov & Podlesskaya
1995; Sohn 2009; Malchukov 2012).
(1) Turkish
Ali gel-ince şaşır-d-ı
Ali come-CONV be.surprised-PST-3SG
‘When Ali came, he was surprised’ (Johanson 1995: 314)
(2) Khalkha
Ger-ees-ee gar-aad
house-ABL-REFL exit-PFV.CONV
deuc-en jil-iin daraa ol-d-lao
forty-ADN year-GEN after find-PASS-FIN
‘She went away from home and was found forty years later’ (Janhunen 2012:
280)
(3) Even
Dagam-mi, kunte-le d’u-v it-ti-n
approach-CONV clearing-LOC house-ACC see-PST-3SG
‘When he came nearer, he saw a house on a clearing’ (Malchukov 2012: 213)
(4) Korean
Kiho-nun nol-ko ca-ss-eyo.
Kiho-TOP play-CONV sleep-PST-POL
‘Kiho played and then slept’ (Sohn 2009: 300)
(5) Japanese
Robbeets -- page 15 of 38
Taroo-ga bangohan-o tabe-te furo-ni hai-ta.
Taroo-NOM dinner-ACC eat-CONV bath-DAT enter-PST
‘Taroo took a bath after he ate dinner’ (Alpatov & Podlesskaya 1995: 473)
Although the Uralic languages are characterized by an extensive use of converbs,
Khanty is rather atypical in this sense because it has only a single converb in -m
ɨ
n,
which is the least frequent nonfinite verb form. Yukaghir and Nivkh also use a
variety of converbs to link clauses. Ainu, however, employs subordinating
conjunctions. Ket has no converbs or serial verb constructions of any kind. In
Mandarin, verbs or verbal phrases are merely juxtaposed, the relation between the
items being largely unmarked. Rukai marks adverbial subordination through a
variety of means such as subordinating conjunctions, changes in word order and
nominalized verb forms.
23. Use of a locative existential construction to encode predicative possession The
Transeurasian languages show a clear preference to express the concept “X has Y”
on the basis of an existential sentence, whereby the possessed noun phrase
functions as the grammatical subject of the ‘exist’-predicate, while the possessor
noun phrase is in a dative-locative case form. Although locative possessive
constructions were standard in Old Turkic, Turkish uses genitive existential
sentences as well as locative existential sentences. ‘I have a book’, for instance, can
be expressed by Ben-de bir kitab var (I-LOC a book exist) or by Ben-im bir kitab-
ïm var (I-GEN a book-1SG.POSS exist). Middle Mongolian and Khalkha make use
of either a conjunctional possessive which construes the possessor noun phrase as
the grammatical subject of the copula and marks the possessed with the comitative -
tai, e.g. Khalkha Bi nom-tai bai-n’ (I book-COM be-DUR) or else, a locative
possesive, e.g. Nad-ed nom bai-n’ (I-DAT book be-DUR). As is the case for most
Tungusic languages, Manchu and Evenki employ locative existential constructions,
e.g. Evk. Min-du: kniga bisi-n (I-DAT book be-3SG); Ma. Min-de bithe bi (I-DAT
book be). Korean uses a locative existential construction, e.g. K Na-hanthey chayk-i
issta (I-LOC book-NOM exist), but the possessor can also be construed as the topic
of the noun phrase, e.g. Na-nun chayk-i issta (I-TOP book-NOM exist). This is also
true for Japanese, e.g. Watashi-ni hon-ga aru (I-DAT book-NOM exist) and
Watashi-wa hon-ga aru (I-TOP book-NOM exist). Topic possessives may have
developed under influence of Chinese, since they represent the standard strategy in
Mandarin. Among the strategies used to encode predicative possession in the Uralic
languages, we find locative possession such as in Finish and Hungarian, genitive
possession such as in Nenets and possession encoded by a transitive verb ‘to have’
such as in Khanty. Whereas Yukaghir employs a conjunctional possessive and Ainu
a ‘have’-possessive, Ket and Nivkh use locational possessives. Although many
Austronesian languages employ topic possessives, Rukai makes use of locative and
genitive possessive constructions. In Stassen’s (2005: 474-477) sample of 240
languages, 20% use a locative existential construction to encode predicative
possession.
24. Use of the ablative case form to encode predicative comparison The
Transeurasian languages all form comparative constructions in which the standard
Robbeets -- page 16 of 38
noun phrase is constructed in the ablative case form, e.g. Tk. bu araba-dan daha
büyük (this car-ABL more big) ‘bigger than this car’, OT barča-da üzä-räk
(everything-ABL high-COMP) ‘higher than anything else’, Khal. ene xun-ees iluu
(this person-ABL good) ‘better than this person’, MMo. qola-sa qola (far-ABL far)
‘farther than far’, Evk. oron-duk gugda-tmar (deer-ABL tall-COMP), Ma. ere
niyalma ci sain (this person ABL good) ‘better than this person’, OJ ware-yo1ri mo2
mantusi-ki1 pi1to2 (I-ABL PT be.poor-ADN person) ‘people poorer than me’ and J
chikyu:-yori omoi (globe-ABL be.heavy) ‘heavier than the globe’. In literary
Korean the ablative marker eyse ‘from’ can be used in comparative constructions,
e.g. K i eyse te khu-n salang (this ABL more be.big-ADN love) ‘a greater love than
this’, but it is more common to use a comparative particle pota ‘than’, e.g. K kicha
pota ppaluta (train PT be.fast) ‘faster than a train’, MK nyey pwota thak.wel hota
(past PT superior be) ‘superior to the past’. This particle has grammaticalized from
the verb MK pwo- ‘to see’ and the transferentive -·ta·ka, which signals the
interruption of an event before its completion, i.e. ‘when one looks at’. It replaced
earlier particles for comparison in Middle Korean, such as MK tukwo ‘than’ and
lawa ‘than’. The Uralic languages differ from one another with regard to
comparative constructions: languages to the west such as Finish and Hungarian use
particle comparatives as in European languages, languages to the east, such as
Nenets and Udmurt, mark the comparative standard with the ablative case ending,
as in the Transeurasian languages. In Khanty, the marker of comparison is a
postposition niŋǝ ‘since, from’, which has ablative-like semantics but differs from
the standard ablative case ending -o
ɣ
or the ablative-elative ending -i. Yukaghir and
Ket mark the comparative standard with the ablative case ending. In Nivkh, the
comparative suffix -yk is traditionally considered as a separate case form as there is
no evidence to relate it to the formally similar locative-ablative suffix -(u)
ɣ
e; -(u)x
(Gruzdeva p.c.). Ainu forms comparative constructions by means of the particle
kasuno ‘than’. In comparative constructions in Mandarin the standard noun phrase
is constructed as the direct object of a verb ‘to exceed’. In Rukai, a comparative
construction is formed through partial reduplication (CVV) of the descriptive verb
stem. In Stassen’s (2005: 490-493) sample of 167 languages, 47% use locational
comparatives, but the proportion of languages that specifically use the ablative case
form to encode predicative comparison is logically expected to be lower.
2.5 Grammaticalization
25. Direct insubordination One of the driving forces of morphosyntactic change in
the Transeurasian languages is a recurrent tendency to grammaticalize non-finite
suffixes to finite suffixes (Robbeets 2009, forthcoming a). In line with Evans
(2008: 367), I call this development “insubordination”, i.e. the conventionalized
main clause use of what appear to be formally subordinate clauses, but it can be
further specified as “direct” insubordination because non-finite suffixes are directly
reanalyzed as finite ones, without the omission of a specific matrix predicate
(Robbeets forthcoming a). Deverbal noun suffixes such as OTk -(A)r in OTk. tug-
‘to be born, to rise (of sun) (intr.)’ -> tugar ‘sunrise, east’; MMo. -m in MMo. quri-
‘to come together (intr.)’ -> qurim ‘feast’; Ma. -rA in mute- ‘to be able’ -> mutere
‘ability’; MK -(·u/o)m in yel- ‘to bear’ -> yelum and OJ -sa in naga- ‘to be long’ ->
Robbeets -- page 17 of 38
nagasa ‘length’ develop over intermediate stages of clausal nominalizers and
relativizers into finite suffixes, as illustrated in the examples (1) to (5).
(1) Old Turkic
Ölüm- oz-upan ögir-ä savin-ü yorï-r.
death-ABL escape-CONV rejoice-CONV be.happy-CONV go.on-FIN
‘Having been saved from death it happily goes on with its life.’
(Erdal 2004: 325)
(2) Middle Mongolian
udurit-basu ber ulu busire-m.
guide-COND PT NEG believe-FIN
‘Even if you guide them, they don’t believe. (Weiers 1966: 144)
(3) Manchu
si nene-me isinji-ci uthai sin-de bu-re
you be.first-CONV come-CONV at.once you-DAT give-FIN
‘If you come first, I shall give [it] to you straight away.’ (Gorelova 2002: 256)
(4) Korean
onul-un swuep-i eps-um.
today-TOP class-NOM not.exist-FIN
‘No class today.’
(5) Old Japanese
punapi1to2-wo mi1-ru-ga to2mo2si-sa
boat.people-ACC see-NML-GEN be.enviable-FIN
‘How enviable it is to see the boat-people!’ (Wrona 2008: 206)
The Uralic languages also display a recurrent tendency towards direct
insubordination. Deverbal noun suffixes such as proto-Uralic *-k, *-, *-mə and
*-śÄ are thought to have developed into finite markers for present-day (*-k, *-)
and past (*-mə, *-śÄ) tense, either in proto-Uralic or after the separation of the
daughter languages (Collinder 1965: 110-115; Janhunen 1982: 36-37). Eastern
Khanty preserves only a faint trace of this development since the finite form of the
negative verb can be marked with the perfective participle-əm, as illustrated in
example (6). However, the phenomenon is well preserved in the Mansi cognate
deverbal noun suffix -əm in uul- ‘to sleep’ -> uuləm ‘sleep’, which has developed
into the finite past, illustrated in example (7). Nikolaeva (1999b) also observes the
development in the Northern Khanty dialects.
(6) Eastern Khanty
məta waja
ɣ
lök ənt-im
some animal track NEG-FIN
‘There is not a single animal track’ (Filchenko 2007: 429)
(7) Mansi
am joht-um-m
I come-FIN-1SG
‘I have come’ (Collinder 1965: 113)
Robbeets -- page 18 of 38
In Nivkh, there is a single instance of direct insubordination, but the phenomenon
does not seem to be recurrent. It concerns the deverbal action noun and infinitive
suffix -d’ which has developed over participial use into a finite form -d’, as
illustrated in example (8).8
(8) Nivkh
if hum-d’ hyjm-d’
he live-NML know-FIN
‘He knows the living one/ (his) life.’ (Malchukov 2013: 200)
The remaining neighboring languages under discussion display strategies other than
direct insubordination in grammaticalizing non-finite suffixes to finite suffixes. In
Yukaghir and Mandarin, for instance, clausal nominalization in construction with a
copula is the main source for developing new finite constructions. Many Sinitic
languages use focus constructions consisting of a nominalizer plus a copula verb;
dropping the copula then paves the way for developing finite constructions. The
Mandarin shi de focus constructions, for instance, consist of a copula shi and a
nominalizer de, whereas the finite stance construction appears without the copula
(Yap & Matthews 2008: 20). Similar processes are found in the Siberian area, for
instance in Yukaghir (Malchukov 2013: 192-195). In Kolyma Yukaghir, the
deverbal action noun suffix -l in pala:- ‘to escape’ -> pala:l ‘(a situation of)
escaping’ has developed into a finite form in subject focus constructions, as
illustrated in (9). The intransitive subject ‘I’ takes a focus marker -ek, which is also
used to mark nominal predicates, thus pointing to its origin as a copula-like form.
As such, the example in (9) can be derived from a cleft-like construction ‘It is me
sitting’.
(9) Kolyma Yukaghir
Met-ek moda-l
I-FOC sit-FIN
‘I sit’ (Malchukov 2013: 194)
Ket displays yet another strategy to develop finite markers, namely to reduce the
matrix predicate to an affix on the former dependent verb. In example (10), for
instance, the matrix verb bimbata ‘it is audible’ is reduced to a present suffix -b
ɛ
ta
~ -bata on verbs expressing sound production (Malchukov 2013: 196-197).
(10) Ket
(10a) tam bis’
ɛ
ŋ in’ŋ
ɛ
j bi-mbata
PT what sound be.audible-FIN
‘a certain sound is audible’ (Werner 1997: 170)
(10b) p-kutəl’ej-b
ɛ
ta
1SG.POSS-whistle-FIN
‘I whistle’ (Werner 1997: 187)
8 Note that Kortlandt (2004: 4) identified the Nivkh suffix -d with the Indo-Uralic
participial suffix *-nt, considering it as evidence of a common origin.
Robbeets -- page 19 of 38
In Ainu, deverbal noun suffixes appear to be functioning as both derivational
suffixes and syntactic clausal nominalizers, but there is no indication that they have
developed into finite endings. Ainu lacks other nonfinite markers such as participial
or converb affixes that could be open to develop into finite markers. Similarly,
Rukai does not exhibit traces of direct insubordination.
The languages of the world use a variety of mechanisms for developing finite
function on formerly non-finite forms such as (1) verbalization of nominal
predicates plus finite copula with subsequent copula erosion; (2) reduction of a
finite verb to affix; (3) insubordination through ellipsis of a matrix clause and (4)
direct insubordination.
26. Grammaticalization from negative verb to verbal negator The historical
development of negation in the Transeurasian languages involves a recurrent
development of an independent negative verb into a negative auxiliary verb, which
may move from preposed to postposed position and eventually assume suffix status
(Robbeets forthcoming b). All Tungusic languages, except Manchu have preserved
evidence supporting the reconstruction of a negative verb pTg *e- ‘not to be, not to
exist’. There are some instances of independent use of the negative verb, i.e.
without a lexical verb, where it means ‘not to exist, not to live’ as in the Evenki
example in (11a). In example (11b), the negative verb acts as a finite auxiliary to
the lexical verb, which assumes an invariant adnominal form. In spite of SOV word
order, the finite negative verb is preposed to the lexical verb. In emotive sentences,
such as in example (11c), the negative auxiliary may move to a postposed position.
The Nanai example in (12) represents the final stage of the negative cycle, i.e.
fusion, whereby the auxiliary negative verb has assumed the status of derivational
suffix on the lexical verb and its phonological form is reduced to lengthening of the
stem-final vowel. Although its predecessor Jurchen preserves traces of pTg *e- ‘not
to be, not to exist’, Manchu does not, but a similar negative cycle can be
reconstructed for the verbal negator Ma. aku:.
(11) Evenki
a. esile e-dyeli-m tadu-gla
now NEG-FUT-1SG there-ENCL
‘Now I will not be (live) there. (Nedjalkov 1994: 27)
b. nungan nekun-mi e-ce-n suru-v-re.
he younger.brother-POSS.REFL NEG-PST-3SG go.away-CAUS-ADN
‘He did not lead his younger brother away.’ (Nedjalkov 1994: 11)
c. nungan songo-ro e-ce-n
he cry-ADN NEG-PST-3SG
‘He did not cry [what’s the use of crying?]. (Nedjalkov 1994: 8)
(12) Nanai
xola:-ci-si
read.NEG-PST-2SG
‘You didn’t read.’
Like Old Turkic, Turkish has a verbal negative suffix -mA- that can be derived from
an original negative verb pTk *ma- ‘not to exist’. The verbal origin is supported by
Robbeets -- page 20 of 38
the occurrence of a negative postposition mar in Chuvash, which contains a
deverbal noun suffix *-r and can take a nominal argument such as the directive case
in debtitive constructions. The Middle Mongolian negative verb stem ese- ‘not to
be, not to exist’ survived in a number of conjugated forms such as with the past
marker -be- in example (13a), but gradually the negative auxiliary became used as
an invariant form, transferring its entire inflection to the lexical verb, i.e. the past
marker -be is attached to ire- ‘to come’ in example (13b).
(13) Written Mongolian
a. ükü-be-üü ese-be-üü
die-PST-INTER NEG-PST-INTER
‘Did [he] die or did [he] not?’ (Poppe 1954: 175)
b. manu ba
ɣ
ši ese ire-be
our teacher NEG come-PST
‘Our teacher did not come.’ (Poppe 1954: 175)
The Middle and Contemporary Korean verbal negator MK a·ni, K an(i) can be
derived from an original negative verb *an- and the suffix MK -i that derives both
nouns and adverbs from verbs. Gradually, the negator ani is being replaced by an
analytic construction consisting of ani augmented by the finite auxiliary MK ho-, K
ha- ‘to do, be’, which usually contracts to anh- ‘not to do’ in Contemporary
Korean. This seems to reflect the start of the next negative cycle whereby the
grammaticalized verbal negator is replaced by a new negative construction in which
a negative verb is restored in its function as finite auxiliary to the lexical verb,
which is nominalized with the suffix -ci.
(14) Middle Korean
¨es·tyey a·ni ·wo-no-·n-ywo
why NEG come-PROC-ADN-INTER
‘Why is [he] not coming?’ (Martin 1992: 420)
(15) Korean
apenim un ka-ci anh-usy-e
father TOP go-NML NEG-HON-FIN
‘Father is not going.’ (Robbeets forthcoming c)
Old and Contemporary Japanese use an independent negative existential adjective
na- ‘to be non-existent, not to exist’, illustrated in (16a), which is thought to derive
from the same origin as the Old Japanese negative suffix -(a)n-, illustrated in (16b)
(Martin 1987: 821). As such, an original negative verb pJ *ana- ‘not to exist’ seems
to have developed into a negative suffix. Negative imperative constructions with
na- preserve a trace of the originally preposed position of the negative auxiliary.
(16) Old Japanese
a. s-uru sube1-no2 na-sa
do-ADN way-GEN NEG-NML
‘Nothing can be done.’ (Vovin 2009: 483)
b. ki1mi1-ga k-i1-mas-an-u
Robbeets -- page 21 of 38
lord-GEN come-CONV-deign-NEG-ADN
‘You did not come, [my] lord.’ (Vovin 2009: 788)
Similar to the Transeurasian languages, one of the characteristics of the Uralic
languages is the expression of negation by means of a construction, comprising a
fully inflected negative auxiliary and a largely invariant lexical verb (Comrie 1981;
Janhunen 1982: 37; Payne 1985: 215-221; Honti 1997; Suihkonen 2002: 173). The
construction may develop in ways which result in a redistribution of inflectional
categories between the negative verb and the lexical verb. Eventually, as is the case
for the negative particles Khanty
ǝ
nt
ǝ
or Estonian ei, the negative auxiliary may
become totally free of inflections and turn into an invariant verbal negator, which
recalls the situation in Mongolic in (13b). However, there are no examples in Uralic
in which the negative auxiliary ultimately becomes a suffix, as it does in Turkic,
Tungusic and Japanese. In Yukaghir, clausal negation is expressed by a proclitic el-
, which usually precedes the verb, but in spite of the formal similarity with the
proto-Uralic negative auxiliary *e-, there are no language-internal indications that it
originated in a negative verb or auxiliary (Fubito Endo p.c.). Clausal negation in
Nivkh is expressed in three ways, i.e. (i) a construction with the negative verb ķ’au-
‘not to be, not to have’ preceded by the lexical verb in an invariant dative case
form; (ii) an incorporation of the negative verb ķavr-/ģavr- ‘not to be, not to have’
incorporated into the verbal form and; (iii) a negative suffix -rla / -tla. Given the
formal similarity between the negative verb ķ’au- and the negative affix -ķavr-, it is
not unlikely that they go back to the same source (Ekaterina Gruzdeva p.c.).
Among the many negatives used in Mandarin, the most general and neutral
negation is expressed by bu, whereas the existential negative mei ‘there is not, has
not’ is used to negate the completion of an event ‘not yet’. Both mei and bu
originate as verbs (van Gelderen 2008: 225). Ainu uses a negative particle that
precedes the verb (e.g. somo ku-oman (NEG 1SG-go) ‘I do not go’). The non-
verbal status of somo is indicated by the fact that it does not take any personal affix.
There are also no indications that the Ket negative particle b
ǝ
:n or the Rukai
negative suffix -ka originate in a negative verb.
The expression of negation via negative auxiliaries is worldwide a minor type to
begin with, found in only 40 (17%) out of 240 languages in Dahl’ s (1979) sample,
which is areally biased towards Uralic and Altaic languages, in 45 (4%) out of 1011
languages in Dryer’ s (2005) sample, and in 16 (5%) out of the 297 languages in
Miestamo’s (2005) sample. As a consequence, the particular development of
negative verbs to auxiliaries to particles or suffixes is hence even rarer.
27. A morphologically simplex first person plural pronoun is complemented by the
grammaticalization of the first person pronoun augmented with a collective-plural
marker When dealing with the inclusive-exclusive distinction in first person plural
pronouns in Section 2.2. (Feature 11), it was mentioned that most Turkic languages
and Korean complement their first person plural pronoun (Tk./OT biz ‘we’; K/MK
wuli ‘we’) with an augmented collective-plural form (Tk. / OT biz-ler ‘we (as a
group); K wuli-tul, MK wuli-tolh ‘we (as a group)’). A similar tendency has been
found in the history of Japanese, where the first person singular /plural OJ wa- ‘I,
we’ coexists with the same form augmented by a collective marker OJ wa-re ‘we’,
Robbeets -- page 22 of 38
a form which in its turn was later augmented into ware-ra ‘we’ in the history of
Japanese. Etymologically, the Middle Mongolian inclusive bida, reflected in the
Khalkha formally inclusive oblique bidn-, derives from the first person singular
MMo. bi ‘I’ and a plural suffix -dA, which also occurs in the plural demonstratives
pronouns MMo. e-de ‘these’ vs. te-de ‘those’ (Doerfer 1985: 2; Domii 2006;
Nevskaya 2010: 119).9 Domii argues that originally, *ba and *bi-da complemented
each other as plural pronouns and that the distinction between exclusive and
inclusive meaning was a secondary development. The Tungusic exclusives Evk. bu
and Ma. be can be derived from the first person plural pTg *and an augmented
plural *bö-(x)e, respectively (Doerfer 1978: 81-83, 95-96; Janhunen 2013: 217),
whereas the inclusive Evk. mut ~ mit may go back to pTg *plus the collective
suffix pTg *-ti (Benzing 1955: 1020) and the inclusive Ma. muse may be an
extension of this root with the collective suffix -sA (Benzing 1955: 1017-1018).
This analysis suggests that successive cycles of plural augmentation on
morphologically simplex (or simplified) plural pronouns have triggered the
secondary development of an inclusive-exclusive distinction in Tungusic. As far as
the Uralic languages are concerned, Khanty makes a commonly found distinction
between pronouns in the first person singular ( ‘I’), dual (min ‘both you and
me’) and plural (məŋ~ m
ɨ
ŋ ‘we’), but it does not reflect any trace of plural
augmentation on the first plural pronoun. Similarly, no traces of plural
augmentation on first person plural pronouns are found in Ket or Yukaghir. In
Rukai, the first person plural inclusive -mita (NOM) is formally underivable from
the exclusive -nai ~ nai- (NOM). The personal pronouns in Ainu have all
grammaticalized from person affixes followed by any one of several existential
verbs meaning ‘exist’. The first person plural pronoun aoka(i), for instance,
consists of the first person plural inclusive transitive subject affix a- and the verb
oka ‘to exist’. In Nivkh, however, all plural personal pronouns can optionally be
augmented with a plural suffix; the first plural exclusive pronoun, for instance,
appears either as n’yņ or as n’yņ-gu (1PL-PL) ‘we’. In Mandarin, two separate
roots for the first person singular wǒ and zán pluralized, along with the suffix -men,
into the derived exclusive wǒmen and inclusive zánmen ‘we’. However since in
Classical Chinese wǒ ‘I’ could be used as a first person plural ‘we’ as well, the
exclusive wǒmen can be regarded as an instance of plural augmentation.
2.6 Overview
In the body of this chapter, I have set up a list of 27 feature labels, chosen to
maximize positive values for the Transeurasian languages. These features, inserted
as vertical comparison points in the tables below, have been examined for selected
representatives among the Transeurasian languages and their linguistic neighbors,
which are inserted as horizontal comparative points. In the tables, I summarize the
observations made above by introducing plus (+) and minus (-) values in the
corresponding cells. This then leads to a quantification of the number of plus values
in the last row.
9 An alternative analysis, deriving the inclusive MMo bida from the first singular pronoun
*bi ‘I’ plus the second plural pronoun *ta ‘you (many)’ is proposed by Janhunen (2013:
215), but the voicing of the medial dental stop would represent an irregular development.
Robbeets -- page 23 of 38
As far as the feature values for the Transeurasian languages are concerned, Table
1 shows the following tendencies. First, the typological coherence seems to be
greater for historical than for the contemporary stages of the languages
investigated. This suggests that Transeurasian areality has decreased over the last
millennium. Second, maximal coherence is found in the Mongolic and Tungusic
languages, with minor deviations from the prototype in the Turkic languages in the
west and somewhat more in the Japonic and Koreanic languages in the eastern
periphery. Third, the deviation from the prototype in the east does not reflect a
gradual loss as we proceed from Korean to Japanese, but rather an en-bloc
reduction of features or even a slight increase for Old Japanese.
Table 1: Feature values for selected Transeurasian languages along with their
historical stages
Tk.
(pre-
)OT
Khal.
(pre-)
MMo.
Ewk.
(pre-)
Ma.
(pre-)
MK
J
(pre-)
OJ
01
+
+
+
+
+
+
+
+
+
02
+
+
+
+
+
+
+
+
+
03
+
+
+
+
+
+
+
-
?
04
-
-
+
+
+
+
+
-
?
05
+
+
+
+
-
+
+
+
+
06
+
+
+
+
+
+
+
+
+
07
+
+
+
+
+
+
+
+
+
08
+
+
+
+
+
+
-
+
-
09
+
?
+
?
-
?
?
+
?
10
-
+
+
+
+
+
-
-
+
11
-
-
-
+
+
+
-
-
-
12
-
+
-
+
+
+
+
+
+
13
-
+
-
+
-
+
+
-
+
14
+
+
+
+
+
-
-
-
-
15
+
+
+
+
+
+
+
+
+
16
+
+
+
+
+
+
+
+
+
17
+
+
+
+
+
+
+
-
+
18
-
+
+
+
+
+
-
-
-
19
-
+
+
+
+
+
-
-
+
20
+
+
+
+
+
+
+
+
+
21
+
+
+
+
+
+
+
+
+
22
+
+
+
+
+
+
+
+
+
23
+
+
+
+
+
+
+
+
+
24
+
+
+
+
+
+
-
+
+
25
+
+
+
+
+
+
+
+
+
26
+
+
+
+
+
+
+
+
+
27
+
+
+
+
+
+
+
+
+
20
24
24
26
24
25
19
18
20
Robbeets -- page 24 of 38
As far as the feature values for representative neighboring languages are concerned,
Table 2 shows the following tendencies. First, the neighboring languages show
significantly stronger deviations from the prototype than do any of the investigated
Transeurasian varieties. This suggests that it is meaningful to apply the concept of
“areality” to the Transeurasian languages in the sense that they reflect a
geographical concentration of linguistic features that sets them apart from the
selected neighboring languages. Second, Kanthy and Yukaghir show more
typological similarity with the Transeurasian prototype than do other neighboring
languages. Note that for at least three of the examined features (i.e. 6, 18, 22),
Khanty yields a minus value where the Uralic prototype would yield a plus value.
This suggests that “areality” may also apply in a wider, but less coherent sense to
the belt of Transeurasian-Yukaghiric-Uralic languages. Third, the investigated
languages of North Asia have more typological features in common than those in
South East Asia, i.e. Mandarin and Rukai. This suggests a third ring of areality that
is the least uniform, involving the languages of North Asia.
Table 2: Feature values for representative neighboring languages
Khan.
Ket.
Yuk.
Niv.
Ain.
Ch.
Ruk.
01
+
+
+
+
+
-
+
02
+
-
+
+
+
-
+
03
+
-
+
+
+
-
-
04
-
-
+
?
?
-
-
05
+
+
+
-
+
+
-
06
-
+
+
-
-
-
-
07
+
+
+
-
+
+
+
08
-
+
+
-
-
-
+
09
-
-
?
-
-
-
-
10
+
-
-
-
-
+
-
11
-
-
-
+
-
+
+
12
-
-
-
-
-
-
-
13
-
-
+
-
-
-
-
14
-
-
-
-
-
-
-
15
+
+
+
+
+
-
+
16
+
-
+
-
-
-
-
17
+
+
+
-
-
-
+
18
-
-
+
-
-
-
-
19
+
-
-
-
-
-
+
20
+
+
+
+
+
-
-
21
+
+
+
+
+
+
-
22
-
-
+
+
-
-
-
23
-
+
-
+
-
-
+
24
-
+
+
-
-
-
+
25
+
-
-
+
-
-
-
Robbeets -- page 25 of 38
26
+
-
-
+
-
+
-
27
-
-
-
+
-
+
-
14
11
17
12
8
7
8
3 Interpretation of the observations
3.1 Delimitation of areality
The Transeurasian continuum has clear boundaries which delimit the language type
in relation to its neighbours both to the north (Yeniseic, Yukaghiric) and east
(Nivkh, Ainu) as well as to the south (Sinitic, Austronesian). Although the
observations above are in line with Janhunen’s (2009: 61-62) findings about a
certain internal uniformity in the larger Ural-Altaic belt, they also suggest including
Yukaghir in this larger belt and they indicate additional boundaries in areality
between the Uralic and the Transeurasian languages as such. Among the features
that enable us to delimit the Transeurasian languages in relation to their Uralic
neighbours are 4. tongue root harmony in Transeurasian (and Yukaghir) vs. palatal
harmony in Uralic; 8. voicing distinction for stops in Transeurasian (and Yukaghir)
vs. original singleton-geminate distinction in Uralic; 9. non-verbal strategy of
verbal borrowing in Transeurasian vs. direct insertion in Uralic; 11. inclusive-
exclusive distinction in Transeurasian vs. none in Uralic (and Yukaghir); 12./13
mixed and switched encoding of property words in Transeurasian (and perhaps
originally in Yukaghir) vs. nominal encoding in Uralic; 14. partial emphatic
reduplication in Transeurasian vs. none in Uralic (and Yukaghir); 18. absence of
intial m in the nominative first person singular vs. presence in Uralic (and
Yukaghir); 25. development of a negative verb into a suffix in Transeurasian vs.
none in Uralic (and Yukaghir); and 27. augmented first person plural pronoun in
Transeurasian versus none in Uralic (and Yukaghir). For some features such as 23.
and 24, Uralic makes use of a larger variety of strategies than the Transeurasian
languages, where all languages uniformly use locative possession or ablative
comparatives. It is remarkable that Yukaghir aligns with Uralic rather than with
Transeurasian in more than half of the delimiting features (i.e. 9, 11, 14, 18, 25 and
27), although it is geographically adjacent to Transeurasian languages such as
Yakut (Turkic) and some northern Tungusic languages, but not to the Uralic
languages. In my opinion, this observation is probably not coincidental, but it
might reflect the alleged genetic relatedness between Uralic and Yukaghir proposed
by among others Collinder 1965.
3.2 Deviations from the prototype
Along the margins of the Transeurasian continuum, we can observe examples of
gradual loss of Transeurasian features in the western and eastern peripheries, as
well as gradual adoption of Transeurasian features, as in the case of Mandarin.
Examples of original Transeurasian features changing in the western periphery
under Uralic influence are 4. Transeurasian tongue root harmony, which aligns with
the Uralic languages as palatal harmony in Turkic; 12. gradual loss of verbal
encoding of property words — mirroring Uralic nominal encoding — as one
Robbeets -- page 26 of 38
proceeds from older to contemporary varieties and from Tungusic in the east to
Turkic in the west and; 18. the secondary development of m-initials yielding an mi-
Ti opposition in first vs. second singular person pronouns in Turkic, Mongolic and
Tungusic.
Changes in areality in the eastern peripheries may take place under the influence
of the languages to the extreme northeast of the Siberian area or under Chinese
influence. Examples of original Transeurasian features in Tungusic and Mongolic
changing under Siberian influence are 5. the secondary assimilation of pTg *g- into
an initial velar nasal in Tungusic, in line with Nivkh and 11/ 27. the secondary
development of an inclusive-exclusive distinction on augmented plural pronouns,
mirroring the situation in Ainu and Nivkh. Examples of Korean and Japanese
features aligning with the extreme northeast Siberian area are the lack of voicing
distinction in Korean and Old Japanese, in line with Ainu, Nivkh and Chukchi and
10. the development of a mesial demonstrative distinction in Japanese and its
presence in Korean, similar to the situation in Yukaghir, Ainu and Nivkh.
Chinese features seem to have diffused into Manchu, Korean and Japanese, for
instance, in 1. the gradual increase of monosyllabic roots in Japanese; 2. the
development of simple tone systems in Japanese and Korean; 3./4. the alleged
erosion of tongue root vowel harmony in Old Japanese; 12. the relatively strong
proportion of verbally encoded property words in Japanese and Korean in
comparison to the other Transeurasian languages; 15. the increase of analytical
features in Manchu in comparison to the other Tungusic languages; 17. the increase
of sortal numeral classifiers in Manchu vis-à-vis the other Tungusic languages and
in Japanese and Korean vis-à-vis older varieties of the languages; 19. the gradual
de-pronominalization, which has taken place in the recorded history of Japanese
and Korean and 23. the development of topic possessives in Korean and Japanese.
Note that some Transeurasian languages to the center of the continuum such as
several Turkic and Mongolic languages of the Amdo Qinghai region have also lost
prototypical Transeurasian features under the influence of Chinese and other
languages of the area (Janhunen 2007).
However, the above observations support previous studies by Hashimoto (1986),
Norman (1988: 10-12, 20) and Comrie (2008), arguing that the Transeurasian
languages have also left a serious mark on the linguistic structure of Chinese. The
following developments illustrate how Chinese may have changed some of its
original Mainland Southeast Asian features under Transeurasian influence: 1. the
development of a greater number of polysyllabic roots compared to Classical
Chinese; 7. the simplification and loss of consonant clusters compared to Old
Chinese; 10. the development of a two-way distinction in demonstratives
compared to the three way-distinction in Classical Chinese; 11. the development of
an inclusive-exclusive distinction in first person plural pronouns in Beijing and
certain other northern Chinese dialects, which was not found in Old Chinese; 16.
the weak suffixing tendency of Mandarin as opposed to other Sinitic languages and;
21. the rare combination of SVO sentence order and GAN noun phrase order in
Mandarin, absent in almost all the other languages of Southeast Asia.
Geographically, Chinese is located between the Transeurasian languages and the
languages of Mainland Southeast Asia, an intermediate position, which it also
occupies from the point of view of typology.
Finally, some features in the Siberian languages to the extreme northeast seem to
Robbeets -- page 27 of 38
have diffused directly from Southeast Asia, without a Transeurasian intermediary,
e.g. 2. the occurrence of two distinctive tones in Nivkh in comparison to the
relatively simple pitch-accent systems of Japanese and Korean; 12. the exclusively
verbal encoding of property words in Yukaghir, Nivkh and Ainu, similar to
Mandarin, but different from mixed encoding in Japanese and Korean; 17. the
obligatory use of an extensive list of classifiers in the Nivkh lexicon, and a smaller
one in Ainu, recalling the widespread and archaic use of classifiers in Southeast
Asia, as opposed to their relatively late development in Japanese and Korean. This
observation may gain relevance in the light of theories that derive Ainu from the
south (e.g. Murayama 1992; Vovin 1993; Bengston & Blažek 2009).
3.3 Diffused vs. inherited features
A simplistic interpretation of the observations would be to assert that the properties
of the Transeurasian language type are universally so common that their parallel
occurrence in several adjacent language families is coincidental. This is certainly
not the case, however, because the Transeurasian continuum has clear boundaries
which delimit the language type in relation to its neighbours both to the west
(Uralic), north (Yeniseic, Yukaghiric), east (Nivkh, Ainu) and to the south (Sinitic,
Austronesian). Moreover, the relatively low frequency of some features indicates
that the shared properties are not due to mere universal principles in linguistic
structuring. Above I have provided an estimation of the frequency of 19 out of 27
features. Seven features are not very common (i.e. 5, 9, 10, 16, 20, 21, 23) in the
sense that they occur in less than half (50%) but more than a third (33%) of the
languages worldwide. Nine features are relatively uncommon in the sense that they
occur in less than a third (33%) of the languages worldwide (e.g. 4, 11, 12, 13, 18,
19, 23, 25). Phenomena that are relatively infrequent and randomly spread across
the world’s languages but frequent and geographically concentrated in a specific
group of languages provide evidence of a historical connection — be it areal or
genealogical — between the languages concerned (Croft 1990: 206-207). The
strength of the argument increases when a number of features correlate in a
particular part of the world, but not in the world as a whole.
It is important to note that the typological similarities among the Transeurasian
languages are accompanied by a significant number of correspondences in the
lexicon (see Robbeets 2005) as well as in verb morphology (see Robbeets 2007a/b,
2010, 2012) in such a way that in my own judgement these languages are
likely to be genealogically related. The most plausible family tree, representing the
overall relationships is given in Figure 1. The affiliation of the Transeurasian
languages remains debated, but even critics such as Janhunen (1996: 220) would
agree that before the first millennium BC the homelands of the individual language
families concerned were all located in a compact area in southern Manchuria, along
with the homelands of Ainuic and Nivkh speakers.
Figure 1: Family tree of the Transeurasian languages
Robbeets -- page 28 of 38
Although some of the shared features discussed above, such as 11. inclusive-
exclusive distinction in first person plural pronouns; 14. partial emphatic
reduplication of nominal property words or 18. mi-Ti opposition in first vs. second
singular person pronouns are almost certainly contact-induced, others appear to be
the residue of common ancestral features, as suggested by the following six
observations.
1. Geography: isolated position of Japanese Although the Sea of Japan and the
Tsushima Strait form a strong geographical boundary separating Japanese from the
other Transeurasian languages, Japanese is typologically closer to the
Transeurasian languages than geographically less isolated languages such as Ket,
Yukaghir, Ainu and Nivkh. Even within a prehistorical contact scenario, this
suggests that the Transeurasian characteristics in Japanese did not exclusively arise
through diffusion because Ainuic and Nivkh were also present in southern
Manchuria.
2. History: older varieties are more prototypical Transeurasian A comparison of
typological uniformity between historical and contemporary stages of the languages
investigated suggests that Transeurasian areality has decreased over the last
millennium. While influences diffusing from adjacent areas such as Mainland
Southeast Asia, Siberia and Uralic have demonstrably displaced earlier
Transeurasian features in certain contact zones, I find no evidence of Transeurasian
features having displaced earlier Chinese, Siberian or Uralic features inherent to the
continuum from Japanese to Turkic. Among the examples of displacement of
features in contact zones, for instance, we find that initial velar nasals have
developed in Tungusic under Siberian influence (5.), simple tone systems and
classifiers have developed in Japanese and Korean under Chinese influence (2.
/17.), palatal harmony has developed in Turkic under Uralic influence (4.) and that
nominal encoding of property words has increased in Turkic, Mongolic and
Tungusic under Uralic influence, while verbal encoding has increased in Japanese
and Korean under Siberian and Southeast Asian influence (12.). However, we find
no evidence of Transeurasian features entering, for instance, from the Turkic
languages and diffusing all over the Transeurasian area, while displacing original
and prototypical Sinitic features. This suggests that Transeurasian features are
inherent to these languages..
Robbeets -- page 29 of 38
3. Distribution: maximal coherence in Mongolic and Tungusic Maximal structural
uniformity is found in the Mongolic and Tungusic languages. This distributional
pattern conforms to the expectations for the Mongolic languages within a
diffusional scenario, since they constitute the center of the linguistic continuum, but
it is not what one would expect for the Tungusic languages, extending towards the
northeastern periphery. The structural coherence in Mongolic and Tungusic recalls
the separation of Mongolo-Tungusic in Figure 1 as a distinct genealogical unit.
4. Distribution: en-bloc reduction of features in Korean and Japanese Similarly,
the collective rather than gradual reduction if not slight increase of features
as we proceed from Korean to (Old) Japanese is not what we would expect within a
scenario of gradual diffusion. It is furthermore difficult to explain how some
Transeurasian features, such as 10 and 19, that show a gap in Korean, have diffused
into Japanese without a Korean intermediary.
5. Cyclicity: recurrent grammaticalization The features involving shared patterns
of grammaticalization in Section 2.5. are particularly good candidates for
genealogically motivation because they are recurrent in different forms and at
various chronological stages of the same language. Aikhenvald (2013)
characterized contact-induced grammaticalization as “change against the grain” or
atypical grammaticalization, while she regarded genealogically motivated
grammaticalization as “change that reinforces similarities” because it tends to
maintain uniformity between related languages. Given that languages tend to renew
their formal encodings in cyclic processes of grammaticalization while maintaining
their inherited grammatical categories, new forms are thus expected to
grammaticalize along shared conceptual pathways to restore old categories (Heath
1998: 729). Consequently, genealogically motivated grammaticalization is expected
to recur on different formal encodings at various points in time, while contact-
induced grammaticalization is expected to be restricted to a single formal encoding
(or to a very limited number of encodings) during a certain period of contact. The
repeated waves of grammaticalization and replacement involved in features 25 to
27 imply that the parallel patterns are genealogically motivated.
6. Isomorphism: shared features combine with formal correspondences The
observation that some structural features shared among the Transeurasian languages
combine with a formal correspondence of the marker reflecting the particular
feature is also indicative of genealogical retention. This is, for instance, the case for
9. the non-verbal strategy of verbal borrowing employing a deverbal noun suffix of
the common shape *-lA- (Tk. -lA-, Khal. -l-, Ud. -lA-, J -r(a)-) to accommodate for
verbal borrowings; 19. the formation of a secondary oblique stem of personal
pronouns through a common suffix *-n- in all Transeurasian languages, except
Korean; 25. direct insubordination involving deverbal noun and finite suffixes of
the common shape *-rA-, *-mA, *-n, *-xA ~ *-kA and *-sA (Robbeets 2009,
forthcoming a, c) across all Transeurasian languages and 26. the grammaticalization
from negative verb to verbal negator, involving common negative verbs of the
common shape *ana-, *e- and *ma- across the Transeurasian languages (Robbeets
forthcoming b). In instances like these in which isostructuralism coincides with
form-function isomorphism, the structural correspondence is likely to be
Robbeets -- page 30 of 38
genealogically motivated, especially when it concerns an instance of shared
grammaticalization (Robbeets 2013). Note that the Uralic languages also display
oblique personal pronouns in -n-, direct insubordination in *-k, *-mə and *-śÄ and
grammaticalization of negative verbs in *e-, an observation which seems to point to
remote genealogical ties between the Uralic and the Transeurasian languages.
4 Conclusion
In this chapter, I have tried to show that the Transeurasian languages form an
internally homogeneous linguistic continuum. For this purpose, I have examined
the areal concentration of 27 features in the Transeurasian languages, providing a
typological profile of some contemporary varieties in relation to historical stages of
the languages involved and to selected languages immediately outside the
continuum. Comparison with neighbors to the north (Yeniseic, Yukaghiric), south
(Sinitic, Austronesian), east (Nivkh, Ainu) and west (Uralic) makes it possible to
set up boundaries which delimit the Transeurasian proto-type. Along the margins of
the Transeurasian continuum, I have found examples of gradual loss of
Transeurasian features in the western and eastern peripheries, as well as gradual
adoption of Transeurasian features, as in the case of Mandarin. The data further
suggest that the Transeurasian continuum in its turn is part of a larger Uralic-
Yukaghiric-Transeurasian belt of languages, which again is part of a larger area of
North Asian languages. Although it is meaningful to apply the concept of “areality”
to the Transeurasian languages in the sense of a historically motivated geographical
concentration of linguistic features, I prefer avoiding the label “area” with reference
to these languages because this would imply that all shared properties are the result
of diffusion. Observations relating to geography, history, distribution, cyclicity of
grammaticalization and combined isomorphism indicate that this is not the case.
A fuller study would need to take more feature values into account and to insert
a larger variety of Transeurasian languages as comparative points. Neighboring
languages should also be more diversified and adjacent languages in the west such
as Indo-European languages or languages of the Caucasus region should be
included. One should also pay attention to structural dependencies between the
features and to considering whether particular features can be easier accounted for
by diffusion or by genealogical retention. For the latter purpose, it would be
particularly interesting to take common diachronic mechanisms, such as shared
patterns of grammaticalization into fuller account. Although this chapter perhaps
raises as many new questions as it answers, I hope to have contributed here to the
understanding of areality among the Transeurasian languages.
Abbreviations
a) Linguistic forms
ABL ablative
ACC accusative
ADD additive
ADN adnominalizer
CAUS causative
CLASS classifier
COM comitative
COMP comparative
CONV converb
COND conditional
Robbeets -- page 31 of 38
DAT dative
DUR durative
ENCL enclitic
FIN finite
FUT future
GEN genitive
HON honorific
INCH inchoative
INTER interrogative
LOC locative
NEG negative
NML nominalizer
NOM nominative
OBL oblique
PFV perfective
PL plural
POL polite
POSS possessive
PROC processive
PST past
PT particle
REFL reflexive
SG singular
TOP topic
b) Languages
Ain. Ainu
Ch. Mandarin Chinese
Evk. Evenki
J Japanese
K Korean
Khal. Khalkha
Khan. Khanty
Ket Ket
Ma. Manchu
MK Middle Korean
MMo. Middle Mongolian
Niv Nivkh
WMo. Written Mongolian
OJ Old Japanese
OT Old Turkic
pJ proto-Japonic
pK proto-Koreanic
pMo proto-Mongolic
pTg proto-Tungusic
pTk proto-Turkic
Ruk. Mantauran Rukai
Yuk. Yukaghir
References
Aikhenvald, Alexandra 2013. Areal diffusion and parallelism in drift: Shared
grammaticalization patterns. In: Robbeets, Martine & Cuyckens, Hubert (eds.)
2013. Shared Grammaticalization: with special focus on the Transeurasian
languages (Studies in Language Companion Series 132.) Amsterdam:
Benjamins, 23-42.
Alpatov, Vladimir & Podlesskaya, Vera 1995. Converbs in Japanese. In:
Haspelmath, Martin & König, Ekkehard (eds.) 1995. Converbs in cross-
linguistic perspective. Structure and meaning of adverbial verb forms: Adverbial
particles, gerunds. Berlin: Mouton de Gruyter, 465-486.
Anderson, Gregory 2005. The velar nasal (ŋ). In: Haspelmath, Martin et al. (eds.)
2005. The world atlas of language structures. Oxford: Oxford University Press,
42-45.
Anderson, Gregory 2006. Towards a typology of the Siberian linguistic area. In
Yaron Matras, April McMahon, and Nigel Vincent (eds.), Linguistic areas.
Convergence in historical and typological perspective. Basingstoke: Palgrave
Macmillan, 266-300.
Bengston, John D. & Blažek, Václav 2009. Ainu and Austric: Evidence of genetic
relationship. Journal of Language Relationship 2, 1-24.
Index of languages -- page 32 of 38
Benzing, Johannes 1955. Die tungusischen Sprachen. Versuch einer vergleichenden
Grammatik. Abhandlungen der geistes- und sozialwissenschaftlichen Klasse 11,
949-1099.
Bisang, Walter 1995. Verb serialization and converbs: Differences and similarities.
In: Haspelmath, Martin & König, Ekkehard (eds.) 1995. Converbs in cross-
linguistic perspective. Structure and meaning of adverbial verb forms: Adverbial
particles, gerunds. Berlin: Mouton de Gruyter, 137-188.
Bisang, Walter 1998. Structural similarities of clause combining in Turkic,
Mongolian, Manchu-Tungusic and Japanese - a typological alternative to the
hypothesis of a genetic relationship. In: Johanson, Lars (ed.) 1998. The Mainz
meeting. Proceedings of the Seventh International Conference on Turkish
Linguistics. Wiesbaden: Harrasowitz, 199-223.
Bulatova, Nadežda Ja. & Grenoble, Leonore A. 1999. Evenki. (Languages of the
World/Materials 141.) Munich: Lincom.
Bugaeva, Anna 2013. Causative constructions in Ainu: A typological perspective
with remarks on diachrony. [Ms.]
Collinder, Björn 1965. An introduction to the Uralic languages. Berkeley:
University of California Press.
Comrie, Bernard 1981. Negation and other verb categories in the Uralic languages.
In: Osmo Ikola (ed.) Congressus Quintis Internationalis Fenno-Ugristarum, vol.
VI. Turku: Suomen Kielen Seura, 350-355.
Comrie, Bernard 2008. The areal typology of Chinese: Between North and
Southeast Asia. In Djamouri, Redouane; Meisterernst, Barbara & Sybesma, Rint
(eds.) 2008. Chinese Linguistics in Leipzig (Collection des Cahiers de
Linguistique Asie Orientale, 12.) Paris: École des Hautes Études en Sciences
Sociales, Centre de Recherches Linguistiques sur l’Asie Orientale, 1–21.
Croft, William 1990. Typology and universals. Cambridge: Cambridge University
Press.
Cysouw, Michael 2005. Inclusive/exclusive forms for ‘we’. In: Haspelmath, Martin
et al. (eds.) 2005. The world atlas of language structures. Oxford: Oxford
University Press, 162-169.
Dahl, Östen 1979. Typology of sentence negation. Linguistics 17, 79-106.
Diessel, Holger 2005. Distance contrasts in demonstratives. In: Haspelmath, Martin
et al. (eds.) 2005. The world atlas of language structures. Oxford: Oxford
University Press, 170-173.
Doerfer, Gerhard 1978. Urtungusisch ö. In: Doerfer, Gerhard & Weiers, Michael
(eds.) 1978. Beiträge zur nordasiatischen Kulturgeschichte (Tungusica 1.)
Wiesbaden: Harassowitz, 66-116.
Doerfer, Gerhard 1985. Mongolo-Tungusica. Wiesbaden: Steiner.
Domii, Tumurtogoo 2006. The inclusive and the exclusive in Mongolian. In:
Shagdarsursen, Ts. (ed.) 2006. Mongol ulsin ix sürgüülijn. Erdem šinžilgeenij
bičig. (Acta Mongolica 6.267). Ulanbaatar: National University of Mongolia, 77-
78.
Dryer, Matthew S. 2005. Prefixing versus suffixing in inflectional morphology. In:
Haspelmath, Martin et al. (eds.) 2005. The world atlas of language structures.
Oxford: Oxford University Press, 110-113
Dryer, Matthew S. 2005. Order of subject, object, and verb. In: Haspelmath,
Index of languages -- page 33 of 38
Martin et al. (eds.) 2005. The world atlas of language structures. Oxford: Oxford
University Press, 330-333
Dryer, Matthew S. 2005. Order of genitive and noun; Order of adjective and noun.
In: Haspelmath, Martin et al. (eds.) 2005. The world atlas of language structures.
Oxford: Oxford University Press, 350-357.
Dryer, Matthew 2005. Negative morphemes. In: Haspelmath, Martin et al. (eds.)
2005. The world atlas of language structures. Oxford: Oxford University Press,
454-457.
Erdal, Marcel 2004. A grammar of Old Turkic. Leiden: Brill.
Evans, Nicholas 2008. Insubordination and its uses. In: Nikolaeva, Irina (ed.) 2008.
Finiteness. Theoretical and empirical foundations. Oxford: University Press,
366-431.
Filchenko, Andrey Yury 2007. A grammar of Eastern Khanty. Houston: Rice
University PhD. Dissertation.
Frellesvig, Bjarke 2010. A history of the Japanese language. Cambridge:
Cambridge University Press.
Frellesvig, Bjarke & Whitman, John 2008. Evidence for seven vowels in proto-
Japanese. In: Frellesvig, Bjarke & Whitman, John (eds.) 2008. Proto-Japanese:
Issues and prospects. (Amsterdam studies in the theory and history of linguistic
science. Series IV - Current issues in linguistic theory 249.) Amsterdam:
Benjamins, 15-41.
Gil, David 2005. Numeral classifiers. In: Haspelmath, Martin et al. (eds.) 2005. The
world atlas of language structures. Oxford: Oxford University Press, 226-229.
Georg, Stefan 2007. A descriptive grammar of Ket (Yenisei-Ostyak) Part 1:
Introduction, Phonology, Morphology. Folkestone: Global Oriental.
Göksel, Aslï & Kerslake, Celia 2005. Turkish. A comprehensive grammar. London:
Routledge.
Gorelova, Liliya 2002. Manchu grammar. Leiden: Brill.
Grönbech, Karl 1936. Der Türkische Sprachbau. Kopenhagen: Levin &
Munksgaard.
Gruzdeva, Ekaterina 1998. Nivkh. (Languages of the World Materials 111).
Munich: Lincom.
Hashimoto, Mantaro 1986. The Altaicization of Northern Chinese. In: McCoy,
John & Light, Timothy (eds.) 1986. Contributions to Sino-Tibetan Studies.
Leiden: Brill, 76-97.
Haspelmath, Martin 1995. The converb as a cross-linguistically valid category. In:
Haspelmath, Martin & König, Ekkehard (eds.) 1995. Converbs in cross-linguistic
perspective. Structure and meaning of adverbial verb forms: Adverbial particles,
gerunds. Berlin: Mouton de Gruyter, 1-56.
Heath, Jeffrey 1998. Hermit crabs: Formal renewal of morphology by
phonologically mediated affix substitution. Language 74, 728-759.
Honti, László. 1997. Die Negation im Uralischen I–III. Linguistica Uralica 2, 81-
96, 161-176, 241-252.
Janhunen, Juha 1981. Korean vowel system in North Asian perspective. Hangeul
172:129-146.
Janhunen, Juha 1982. On the structure of proto-Uralic. Finnisch-ugrische
Forschungen 44, 23-42.
Index of languages -- page 34 of 38
Janhunen, Juha 1996. Manchuria. An ethnic history. (Mémoires de la Société
Finno-Ougrienne 222.) Helsinki: Suomalais-Ugrilainen Seura.
Janhunen, Juha 1997. Problems of primary root structure in Pre-Proto-Japanic.
International Journal of Central Asian Studies 2.
Janhunen, Juha 2007. Typological interaction in the Qinghai Linguistic Complex.
Studia Orientalia 101, 85-103.
Janhunen, Juha 2009. Proto-Uralic—what, where, and when? Mémoires de la
Société Finno-Ougrienne 258, 57–78.
Janhunen, Juha 2012. Mongolian. (London Oriental and African Language Library
19.) Amsterdam: John Benjamins.
Janhunen, Juha 2013. Personal pronouns in Core Altaic. In: Robbeets, Martine &
Cuyckens, Hubert (eds.) 2013. Shared Grammaticalization: with special focus on
the Transeurasian languages (Studies in Language Companion Series 132.)
Amsterdam: Benjamins, 211-226.
Johanson, Lars 1995. On Turkic converb clauses. In: Haspelmath, Martin & König,
Ekkehard (eds.) 1995. Converbs in cross-linguistic perspective. Structure and
meaning of adverbial verb forms: Adverbial particles, gerunds. Berlin: Mouton
de Gruyter, 313-348.
Johanson, Lars & Robbeets, Martine 2009. Introduction. In: Johanson, Lars &
Robbeets, Martine (eds.) 2009. Transeurasian verbal morphology in a
comparative perspective: genealogy, contact, chance. (Turcologica 78.)
Wiesbaden: Harrassowitz, 1-5.
Ko, Seongyeon, Whitman, John & Joseph, Andrew 2014. Comparative
consequences of the tongue root harmony analysis for proto-Tungusic, proto-
Mongolic, and proto-Korean. In: Robbeets, Martine & Bisang, Walter (eds.)
2014. Paradigm change in the Transeurasian languages and beyond (Studies in
Language Companion Series 161.) Amsterdam: Benjamins, 141176.
Lee, Ki-Mun & Ramsey, Robert 2011. A history of the Korean Language.
Cambridge: Cambridge University Press.
Li, Charles N. & Thompson, Sandra A. 1989 Mandarin Chinese: a functional
reference grammar. Berkeley: University of California Press.
Frellesvig, Bjarke 2010. A history of the Japanese language. Cambridge:
Cambridge University Press.
Kaiser, Stefan; Ichikawa, Yasuko; Kobayashi, Noriko & Yamamoto, Hirofumi
2001. Japanese: A comprehensive grammar. London: Routledge.
Ko, Seongyeon 2012. Tongue root harmony and vowel contrast in Northeast Asian
languages. New York: Cornell University Ph.D. dissertation.
Kortlandt, Frederik 2004. Nivkh as a Uralo-Siberian language. In: Hyllested,
Adam; Jørgensen, Anders Richardt; Larsson, Jenny Helena & Olander, Thomas
(eds.) Per aspera ad asteriscos, festschrift in honour of Jens E. Rasmussen.
Innsbruck: IBS, 285-289.
Malchukov, Andrej 2012. Tungusic converbs and a typology of taxis. In:
Malchukov, Andrej & Whaley, Lindsay J. (eds.) 2012. Recent Advances in
Tungusic Linguistics. (Turcologica 89.) Wiesbaden: Harrassowitz, 213-228.
Index of languages -- page 35 of 38
Malchukov, Andrej 2013. Verbalization and insubordination in Siberian languages.
In: Robbeets, Martine & Cuyckens, Hubert (eds.) 2013. Shared
grammaticalization with special focus on the Transeurasian languages (Studies
in Language Companion Series 132.) Amsterdam: John Benjamins, 177-208.
Martin, Samuel Elmo 1988. A reference grammar of Japanese. Tokyo: Tuttle.
Martin, Samuel Elmo 1992. A reference grammar of Korean. Tokyo: Tuttle.
Maslova, Elena 2003a. A Grammar of Kolyma Yukaghir (Mouton Grammar
Library 27.) Berlin: Mouton de Gruyter
Maslova, Elena 2003b. Tundra Yukaghir (Languages of the World/Materials 372).
Munich: Lincom.
Miestamo, Matti 2005. Standard negation: The negation of declarative verbal main
clauses in a typological perspective (Empirical Approaches to Language
Typology 31.) Berlin: Mouton de Gruyter.
Murayama, Shichirō 1992. Ainugo no kigen [Origins of the Ainu Language].
Tokyo: San’ichi Shobo.
Nedjalkov, Igor V. 1994. Negation in Evenki. In: Kahrel, Peter & Van den Berg,
René (eds.) 1994. Typological studies in negation. Amsterdam: John Benjamins,
1-34.
Nedjalkov, Igor V. 1995. Converbs in Evenki. In: Haspelmath, Martin & König,
Ekkehard (eds.) 1995. Converbs in cross-linguistic perspective. Structure and
meaning of adverbial verb forms: Adverbial particles, gerunds. Berlin: Mouton
de Gruyter, 441-464.
Nedjalkov, Igor V. 1997. Evenki. Descriptive grammar. London: Routledge.
Nevskaya, Irina 2010. Inclusive and exclusive in Altaic languages. In: Johanson,
Lars & Robbeets, Martine (eds.) 2009. Transeurasian verbal morphology in a
comparative perspective: genealogy, contact, chance. (Turcologica 78.)
Wiesbaden: Harrassowitz, 115-128.
Nichols, Johanna 2012. Selection for m: T pronominals in Eurasia. In: Johanson,
Lars & Robbeets, Martine (eds.) 2012. Copies versus cognates in bound
morphology. (Brill’s Studies in Language, Cognition and Culture 2.) Leiden:
Brill, 47-70.
Nichols, Johanna & Peterson, David 2005. Personal pronouns. In: Haspelmath,
Martin et al. (eds.) 2005. The world atlas of language structures. Oxford: Oxford
University Press, 546-553.
Nikolaeva, Irina 1999. Ostyak. (Languages of the World/Materials 305.) Munich:
Lincom.
Norman, Jerry 1988. Chinese (Cambridge Language Surveys.) Cambridge:
Cambridge University Press.
Payne, John R. 1985. Negation. In: Shopen, Timothy (ed.) 1985. Language
typology and syntactic description. Vol 1 Clause structure. Cambridge:
Cambridge University Press, 197-242.
Poppe, Nicholas 1954. Grammar of Written Mongolian. Wiesbaden: Otto
Harrassowitz.
Poppe, Nicholas 1955. Introduction to Mongolian comparative studies. (Mémoires
de la société Finno-Ougrienne 110.) Helsinki: Suomalais-Ugrilainen Seura.
Poppe, Nicholas 1964. Der altaische Sprachtyp. In: Spuler, B. & al. (eds.) 1964.
Mongolistik (Handbuch der Orientalistik 5.2.) Leiden: Brill. 1-16.
Index of languages -- page 36 of 38
Rickmeyer, Jens 1989. Japanisch und der altaische Sprachtyp. Eine Synopsis
struktureller Entsprechungen. Bochumer Jahrbuch zur Ostasienforschung 12,
313-323.
Robbeets, Martine 2005. Is Japanese related to Korean, Tungusic, Mongolic and
Turkic? (Turcologica 64.) Wiesbaden: Harrassowitz.
Robbeets, Martine 2007a. How the actional suffix chain connects Japanese to
Altaic. Turkic Languages 11. 1, 3-58.
Robbeets, Martine 2007b. The causative-passive in the Trans-Eurasian languages.
Turkic Languages 11.2, 235-278.
Robbeets, Martine 2008. If Japanese is Altaic, why is it so simple? In: Lubotsky,
Alexander; Schaeken, Jos & Wiedenhof, Jeroen (eds.) 2008. Evidence and
counter-evidence: Essays in honour of Frederik Kortlandt. Volume 2: General
Linguistics. (Studies in Slavic and General Linguistics 33.) Amsterdam: Rodopi.
Robbeets, Martine 2009a. Insubordination in Altaic. Journal of Philology 31. Ural-
Altaic Studies 1, 61-79.
Robbeets, Martine 2010. Transeurasian: Can verbal morphology end the
controversy? In: Johanson, Lars & Robbeets, Martine (eds.) 2010. Transeurasian
verbal morphology in a comparative perspective: genealogy, contact, chance.
(Turcologica 78.) Wiesbaden: Harrassowitz, 81-114.
Robbeets, Martine 2012. Shared verb morphology in the Transeurasian languages:
copy or cognate? In: Johanson, Lars & Robbeets, Martine (eds.) 2012. Copies vs.
cognates in bound morphology. (Brill’s Studies in Language, Cognition and
Culture 3.) Leiden: Brill, 427-446.
Robbeets, Martine 2013. Genealogically motivated grammaticalization. In:
Robbeets, Martine & Cuyckens, Hubert (eds.) 2013. Shared
Grammaticalization: with special focus on the Transeurasian languages (Studies
in Language Companion Series 132.) Amsterdam: Benjamins, 147-175.
Robbeets, Martine forthcoming a. Insubordination and the establishment of
genealogical relationship. In: Evans, Nicholas & Watanabe, Honore (eds.)
Dynamics of insubordination. (Typological Studies in Language.) Amsterdam:
Benjamins
Robbeets, Martine forthcoming b. The development of negation in the
Transeurasian languages. In Whaley, Lindsay & Suihkonen, Pirkko (eds.)
Typology of languages of Europe and Northern and Central Asia. (Studies in
Language Companion Series) Amsterdam: Benjamins.
Robbeets, Martine forthcoming c. Diachrony of verb morphology in Japanese and
the Transeurasian languages. ( Trends in Linguistics Studies and Monographs.)
Berlin: Mouton-De Gruyter.
Rybatzki, Volker 2003. Middle Mongol. In: Janhunen, Juha (ed.) 2003. The
Mongolic languages. London: Routledge, 57-82.
Shibatani, Masayoshi 1990. The languages of Japan. (Cambridge Language
Surveys.) Cambridge: Cambridge University Press.
Sohn, Ho-min 1994. Korean. London: Routledge.
Sohn, Ho-min 2009. The semantics of clause linking in Korean. In: Dixon, R. M.
W. & Aikhenvald, Alexandra 2009. The semantics of clause linking. A cross-
linguistic typology. Oxford: Oxford University Press, 285-317.
Suihkonen, Pirkko 2002. The Uralic languages. Fennia 180.1–2, 165-176.
Index of languages -- page 37 of 38
Stassen, Leon 1997. Intransitive predication. Oxford: Clarendon Press.
Stassen, Leon 2005. Predicative possession. In: Haspelmath, Martin et al. (eds.)
2005. The world atlas of language structures. Oxford: Oxford University Press,
474-477.
Stassen, Leon 2005. Predicative adjectives. In: Haspelmath, Martin et al. (eds.)
2005. The world atlas of language structures. Oxford: Oxford University Press,
478-481.
Stassen, Leon 2005. Comparative constructions. In: Haspelmath, Martin et al. (eds.)
2005. The world atlas of language structures. Oxford: Oxford University Press,
490-493.
Street, John 1957. The Language of the Secret History of the Mongols. New Haven:
American Oriental Society.
Svantesson, Jan-Olof 1985. Vowel harmony shift in Mongolian. Lingua 67.4, 283-
327.
Tamura, Suzuko 2000. The Ainu language. (ICHEL Linguistic Studies 2.) Tokyo:
Sanseidō.
Vajda, Edward J. 2004. Ket. (Languages of the World/Materials 204.) Munich:
Lincom.
van Gelderen, Elly 2008. Negative cycles. Linguistic Typology 12, 195-243.
Vovin, Alexander 1993. A reconstruction of Proto-Ainu. (Brill’s Japanese Studies
Library 4.) Leiden: Brill.
Vovin, Alexander 2005. A descriptive and comparative grammar of Western Old
Japanese. Part 1: sources, script and phonology, lexicon, nominals. (Languages
of Asia 3.) Folkestone: Global Oriental.
Vovin, Alexander 2009. A descriptive and comparative grammar of Western Old
Japanese. Part 2: adjectives, verbs, adverbs, conjunctions, particles,
postpositions. (Languages of Asia 8.) Folkestone: Global Oriental.
Weiers, Michael 1966. Untersuchungen zu einer historischen Grammatik des
präklassischen Schriftmongolisch. Bonn: Rheinischen Friedrich-Wilhelms-
Universität Ph.D dissertation.
Werner, Heinrich. 1997. Die ketische Sprache. Wiesbaden: Harrassowitz.
Whaley, Lindsay J. & Li, Fengxiang 2000. Emphatic reduplication in Oroqen and
its Altaic context. Linguistics 38.2, 355-372.
Whitman, John Bradford. 1990. A rule of medial *-r- loss in pre-Old Japanese. In:
Baldi, Philip (ed.) 1990. Linguistic change and reconstruction methodology.
(Trends in Linguistics. Studies and Monographs 45.) Berlin: Mouton de Gruyter,
511-545.
Wohlgemuth, Jan 2009. A typology of verbal borrowings. (Trends in Linguistics.
Studies and Monographs 211.) Berlin: Mouton de Gruyter.
Wrona, Janick 2008. The nominal and adnominal forms in Old Japanese:
Consequences for a reconstruction of pre-Old Japanese syntax. In: Frellesvig,
Bjarke & Whitman, John (eds.) 2008. Proto-Japanese. Issues and prospects.
(Current Issues in Linguistic Theory 294.) Amsterdam: Benjamins, 193-215.
Yap, Foong Ha & Matthews, Stephen. 2008. The development of nominalizers in
East Asian and Tibeto-Burman languages. In María José López-Couso & Elena
Seoane (eds.), Rethinking grammaticalization: New perspectives (Typological
Studies in Linguistics 76), 309-341. Amsterdam: John Benjamins.
Index of languages -- page 38 of 38
Zeitoun, Elizabeth 2007. A grammar of Mantauran (Rukai). (Language and
Linguistics Monograph Series A4-2.) Taipei: Institute of Linguistics, Academia
Sinica.
Acknowledgement
The research presented here was realized with financial assistance of the DFG
(Deutsche Forschungsgemeinschaft), supporting a collaborative research project
“Die transeurasiatischen Sprachen: Kontakt in der Familie”, at the University of
Mainz from January 2010 to December 2013. I thank Anna Bugaeva, Fubito Endo,
Andrey Filchenko, Ekaterina Gruzdeva, Seongyeon Ko, Fukui Rei and Elisabeth
Zeitoun for their helpful feedback.
... Eberhard, Simons and Fennig (2022) list it as a Koreanic language along with Jejueo, a language variety spoken in Jeju Island. Most recently, it is frequently discussed under the geographically motivated classificatory term 'Transeurasian languages' (Johanson & Robbeets 2010, Robbeets 2013, often along with Japanese since the two languages share many structural characteristics (Narrog & Rhee 2013, Robbeets 2017. ...
... Some researchers(Choe 1982, Hur 1975, Kim 1955, Lee 1955, among others) hypothesize that sk was phonetically realized as[sk]. Transeurasian languages, however, tend to avoid word-initial consonant clusters(Deny 1924, Robbeets 2017. ...
Full-text available
Article
Nominal lexemes undergo in history extensive development in the lexical as well as grammatical domains. Their semantic change involves diverse threads of conceptualization which shows significant aspects of language development. Despite the series of intriguing changes in form and meaning that the Late Middle Korean lexeme kaz ‘edge’ has undergone, encompassing polylexicalization and polygrammaticalization that merit an in-depth analysis, it has not yet received earnest attention to date. Thus, this paper aims to fill the research gap by analyzing the lexeme’s development from Late Middle Korean to Modern Korean from the grammaticalization perspective, drawing upon data from a historical corpus and dictionaries. An extensive diachronic data analysis based on concordance and lexical searches shows that new meanings or functions are so diverse that some of them even form an antonymic relation. Such a wide-ranging semantic and functional diversity is attributable in part to the role of the participating forms in word formation processes such as derivation and compounding but, more importantly, to differential conceptualizations of the source meaning ‘edge’ e.g., entity-internal and entity-external conceptualizations; mapping the ‘edge’ onto different continua, e.g., degree, path, likelihood; and involvement of subjectification, e.g., evaluative-epistemic judgment such as counter-expectation. The innovated meanings resulting from such cognitive operations form a conceptual network. The developmental processes of the lexeme kaz ‘edge’, in general, can be explained with reference to grammaticalization mechanisms, such as desemanticization, extension, decategorialization, and erosion, and some mechanisms are also operative in lexicalization as well. However, erosion, or the reduction of phonetic volume, is not prominent in these changes, suggesting that the principles should be interpreted as tendencies rather than deterministic diagnostics. These findings contribute to a better understanding of the complex relationship between lexicalization and grammaticalization.
... I added 35 features on phonology and formal representation to increase the variability among language families. Eight of these feature formulations (TE004-TE008, TE018, TE019, TE027) are based on the feature set from Robbeets [30] and show some variation in the region. Each feature received the value 1, if the feature question could be answered with a 'yes', and the value 0, if the feature question could be answered with a 'no'. ...
Full-text available
Article
Within linguistics, there is an ongoing debate about whether some language structures remain stable over time, which structures these are and whether they can be used to uncover the relationships between languages. However, there is no consensus on the definition of the term 'stability'. I define 'stability' as a high phylogenetic signal and a low rate of change. I use metric D to measure the phylogenetic signal and Hidden Markov Model to calculate the evolutionary rate for 171 structural features coded for 12 Japonic, 2 Koreanic, 14 Mongolic, 11 Tungusic and 21 Turkic languages. To more deeply investigate the differences in evolutionary dynamics of structural features across areas of grammar, I divide the features into 4 language domains, 13 functional categories and 9 parts of speech. My results suggest that there is a correlation between the phylogenetic signal and evolutionary rate and that, overall, two-thirds of the features have a high phylogenetic signal and over a half of the features evolve at a slow rate. Specifically, argument marking (flagging and indexing), derivation and valency appear to be the most stable functional categories, pronouns and nouns the most stable parts of speech, and phonological and morphological levels the most stable language domains.
... Salar already has a partial alveolar series (/s, z/ but not /ts, dz, ʃ, tʃ/) and a full palatal series; since at least 15-25% of Salar's vo- cabulary consists of Chinese and Tibetan loans, it is unsurprising that the retroflex series has also become phonemic. 17 Data from Salar and Qinghai-Gansu Mongolic languages show that, unlike lan- guages from other families in the Eastern Himalayan affricate-rich zone, they de- veloped retroflex affricates exclusively from other affricates and/or acquired them from lexical borrowings (which is logical given the aversion of Altaic/Transeurasian languages to initial consonant clusters [Robbeets 2017]). ...
Full-text available
Article
This paper makes a contribution to phonological typology by investigating the distribution of affricate-rich languages in Eurasia. It shows that affricate-rich and affricate-dense languages cluster areally within Eurasia and have area-specific histories. In particular, the affricate-rich areas of western Eurasia – a ‘European’ area and a Caucasian area – are not the result of contact-induced sound changes or borrowing, while the two affricate-rich areas of eastern Eurasia – the Hindukush area and the Eastern Himalayan area – are the result of contact. Specifically, affricate-dense areas depend on the emergence of retroflex affricates. Moreover, languages outside these affricate-dense areas tend to lose retroflex affricates.
Full-text available
Article
Millet vs rice: an evaluation of the farming/language dispersal hypothesis in the Korean context - Volume 2 - Jangsuk Kim, Jinho Park
Article
This paper presents the results of an areal study of the elements known as (sentence-)final particles (FPs) in the languages of Asia. FPs constitute a crucial part of many languages of the region and are reported in language-particular descriptions under various labels. However, they have not been the subject of large-scale areal studies. In this paper, I discuss the morphosyntactic and functional properties typically exhibited by the FPs of Asian languages and the parameters of their variation. On the basis of a sample of 53 languages and 6 sample functional types of FPs, I explore the areal distribution of FPs of the Asian type. I demonstrate that different FP-isoglosses exhibit different geographical coverage, but the overlap of some of them allows us to speak of a structural phenomenon highly typical of a macroarea which includes East, Southeast, and Northeast Asia.
Chapter
The Study of Word Stress and Accent - edited by Rob Goedemans December 2018
Full-text available
Article
In this paper, I propose a hypothesis reconciling Austronesian influence and Transeurasian ancestry in the Japanese language, explaining the spread of the Japanic languages through farming dispersal. To this end, I identify the original speech community of the Transeurasian language family as the Neolithic Xinglongwa culture situated in the West Liao River Basin in the sixth millennium BC. I argue that the separation of the Japanic branch from the other Transeurasian languages and its spread to the Japanese Islands can be understood as occurring in connection with the dispersal of millet agriculture and its subsequent integration with rice agriculture. I further suggest that a prehistorical layer of borrowings related to rice agriculture entered Japanic from a sister language of proto-Austronesian, at a time when both language families were still situated in the Shandong-Liaodong interaction sphere.
Book
The questions as to why most languages appear to have more trouble borrowing verbs than nouns, and as to the possible mechanisms and paths by which verbs can be borrowed or the obstacles for verb borrowing, have been a topic of interest since the late 19th century. However, no truly substantial typological research had been undertaken in this field before the present study. The present work is the first in-depth cross-linguistic study on loan verbs and the morphological, syntactic and sociolinguistic aspects of loan verb accommodation. It applies current methodologies on database management, quantitative analysis and typological conventions and it is based on a broad global sample of data from over 400 languages and the typological data from the World Atlas of Language Structures (WALS). One major result of the present study is the falsification, on empirical grounds, of long-standing claims that verbs generally are more difficult to borrow than other parts of speech, or that verbs could never be borrowed as verbs and always needed a re-verbalization in the borrowing language. © 2009 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin. All rights reserved.
Book
This book offers fresh perspectives on “shared grammaticalization”, a state whereby two or more languages have the source and the target of a grammaticalization process in common. While contact-induced grammaticalization has generated great interest in recent years, far less attention has been paid to other factors that may give rise to shared grammaticalization. This book intends to put this situation right by approaching shared grammaticalization from an integrated perspective, including areal as well as genealogical and universal motivations and by searching for ways to distinguish between these factors. The volume offers a wealth of empirical facts, presented by internationally renowned specialists, on the Transeurasian languages (i.e. Japonic, Koreanic, Tungusic, Mongolic, and Turkic) — the languages in focus —as well as on various other languages. Shared Grammaticalization will appeal to scholars and advanced students concerned with linguistic reconstruction, language contact and linguistic typology, and to anyone interested in grammaticalization theory.