The relationship between clauses is expressed by converbs adverbial action clauses , which consist of either simple converb suffixes, or combinations of a nominalizer and a case marker or a postposition. According to Norman 8—10 , the Sinitic languages are monosyllabic tonal languages. Their morphosyntactic structure is analytic with very little inflec- tional morphology and the grammatical relations are primarily expressed by word order or by independent grammatical particles.
The basic, unmarked word order generally appears to be Agent-Verb-Patient. However, it is important to note that sentences are characteristically organized on the basis of a topic-comment struc- ture rather than an argument structure Li and Thompson , Li and Thompson , and the grammatical relations of subject and direct object have not been grammaticalized as in the European languages LaPolla Due to the topic- comment clause structure and non-importance of a grammatical subject, Sinitic languages are not very sensitive to an active-passive distinction and usually lack true passive constructions.
In the noun phrase numerals, demonstratives and adjective attributes precede the noun. Most of the Sinitic languages have a rich system of numeral classifiers, which are used with numerals and demonstratives. Verbs are marked for aspect, but not for tense and person. Clause combining is achieved through serial verb constructions where verbs occur in chains without any morphology specifying their relationships.
The modern Tibetic languages are a group of about two hundred spoken varieties, all derived from Old Tibetan, and sharing one and the same literary! Most of them are not mutually intelligible. These languages are characterized by head-final word order and polysyllabic words derived from monosyllabic roots. Definiteness is optionally marked by demonstratives or by more specific definiteness markers. Morphosyntactically, these markers are enc- litics. There is no agreement, neither within the noun phrase, nor between the verb phrase and its complements.
In summary, Turkic and Sinitic languages show remarkable differences in terms of their structural properties and basic word order. However, due to linguistic convergence caused by Tibetan influence, Salar and Wutun have significantly diverged from their genetic relatives spoken outside the Amdo Sprachbund and they have approached a common typological profile. According to this theory, a given lin- guistic entity consists of the combination of at least four properties, repre- sented in Figure 1.
In a situation of contact between two or more languages, the transfer of linguis- tic material can concern either the four components of the linguistic entity altogether global copy , or only one or some of its components, e.
Partial copies are frequently attested in different languages. This means that they are often transferred without their combinational properties see, e. Curnow ; Matras 48— Similarly, if a given language has two different grammatical means to express the same function, linguistic contact can increase the relative frequency of the grammatical category that is similar to the neighboring languages.
Johanson also emphasizes the fact that the copy does not need to be identical with its model: linguistic copying is an active process, in which the speakers adapt and insert new linguistic material into their language.
Number of Endangered Languages by Country
An obvious exam- ple of such an adjustment is the above-mentioned example of the adaptation of foreign-copied words to the phonology of the replica language. The second theoretical model we will refer to deals with the general pro- cesses of formation of a linguistic area.
Janhunen defines four processes leading to the convergence of different languages in contact. Two of them are active processes, concerning the adoption or the loss of a linguistic feature, under the influence of neighboring languages.
The other two are passive pro- cesses, concerning the conservation of a given feature that also exists in neigh- boring languages, and the lack of development of features that are otherwise unknown in the area under consideration Table 1. Active Passive Positive Adoption of a new feature Conservation of an original feature Negative Loss of an original feature Non-adoption of a new feature This model offers the advantage of presenting a global overview of linguistic evolution, and draws attention to the fact that the effects of language contact are not limited to copying.
Examples discussed in Section 2. In Section 2. Thus, in the case of basic word order, the influence of Amdo Tibetan in the region triggers an alignment of Wutun towards Turkic and Tibetic languages. Examples 1 and 2 illustrate the basic word order in clauses governed by a bivalent verb in Amdo Tibetan and in Salar respectively, as representative of the basic word order in their language families, while examples 3 and 4 show the contrast between the word order in Wutun and Standard Mandarin.
The examples were recorded in various places in Xunhua and Hualong counties Haidong prefec- ture, Qinghai Province and reflect different sub-varieties of the Amdo Tibetan language. The discrepancies are mainly limited to the phonological system e. We thank Xun Gong for his useful comments on the phonological transcription. The remaining mistakes are our responsibility. If we examine the neutral order of the participants more precisely, this alignment of Wutun towards the Tibetic and Turkic languages is also observed for the participants in ditransitive sentences.
In the Turkic and Tibetic lan- guages, the position of the R-argument7 can be either before or after the Patient. Different factors explain this variation: morphosyntactic incorporation into the verb phrase, information structure the immediate pre-verbal position being the focus position , definiteness, etc. There exists a certain tendency for placing an animate or human R-argument Recipient before the Patient as a default syntactic position, whereas a Goal would more often appear in the position immediately before the verb. Transitive sentence Agent Patient Verb Ditransitive sentence Agent Recipient Patient Agent Patient Goal This tendency is not absolute and is easily overridden by pragmatic factors [at least in Amdo Tibetan and Salar], but is still attested in our data.
The examples 7 The term R-argument here is taken as a cover term for participants marked in the dative case in the Turkic and Tibetic languages. We will then further divide R-arguments into two sub- categories depending on their animacy: Recipient refers to an animate R-argument, while Goal refers to an inanimate R-argument. TEST This one, this child gives a thing to an old [man]. Examples 9 to 13 , on the other hand, illustrate the position of the Goal.
The Goal follows the Patient. The functional motivation for this word order involves a combination of the animacy hierarchy with the hierarchy of syntactic functions: animate participants are placed before inanimate participants, and Patient appears before Goal. This word order is also identical with the neutral word order in the Turkic languages, e. However, in ditransitive sentences in Standard Mandarin, as illustrated by the examples 14 and 15 , the Recipient usually follows the Patient, while the Goal precedes the Verb and the Patient. It involves copying the morphosyntactic properties of the participants according to their semantic characteristics, without any phonological form copied.
At the same time, as Matras notes: Contact induced change in word order is generally not common in our sample. This is understandable, given the fact that a change in the position of possessor and possessed does not affect the position of the verb, and so, it leaves the organization of the predication intact. Matras 60 Thus, this type of word order change suggests a very intensive and pro- longed cultural contact between the Wutun and Amdo Tibetan speakers at some stage of the linguistic history of the Amdo Sprachbund.
TEST 1. In examples 19a through 19d , the same morpheme is used in the future tense, regardless of whether the subject is first, second or third person. The suffixes were 1P. As in the case of basic word order, the lack of personal conjugation illus- trates a case of convergence, where one language Salar changes to become similar to Tibetan, whereas the other one Wutun retains its original character- istics.
The semantic and pragmatic functions generally assumed by the passive voice in such languages are 9 Emphasis added on the personal conjugation suffix in Turkish. That is, it is no longer productive, as shown by examples 21a through 21c. In this case, depending on the context, the Agent is either interpreted as impersonal, generic or beyond the scope of the communicative purpose, or as a zero-anaphora. In this case, it has a potential reading. Such an interpretation does not apply to the context of the example 20 : no specific Agent is present in the situation.
This loss of a productive passive voice derivation in Salar is probably linked to the loss of agreement in the verb with its subject, a development compensated by the introduction of a Tibetan-type evidential system cf. Dwyer In fact, both the lack of personal conjugation and the loss of passive voice in Salar have important ramifications for defining the category of subject. At the same time, however, given the multidirectionality of the linguistic convergence in the area, the loss of both the personal agreement and the passive voice in Salar would be best attributed to the combined influ- ence of both Sinitic and Tibetic languages.
The modification of word order in Wutun, and the loss of verbal agreement, as well as the loss of passive voice derivation in Salar, have led to convergence between Amdo Tibetan, Wutun and Salar. However, in cases where two of the three languages originally share the same characteristics, the role of Tibetan as the model language for linguistic change remains unclear. In fact, the linguistic change observed could as well be attributed to the influence of the other language family respectively Turkic and Sinitic , or to the combined effects of two language families of the area.
Only historical and sociolinguistic data can help to decide and reconstruct the most probable scenario for language change. Thus, in such cases, the role of Tibetan as a model language is clearly established, as long as the role of Mongolic languages is ruled out. In Amdo Tibetan, like in all Tibetic languages, the numeral always follows the noun, but unlike Old and Classical Tibetan, it follows the indefiniteness marker if it is present , as exemplified in This is the only word order attested in our data.
We are not aware of a contact-language explanation for this phenomenon. In Wutun, we observe in 26a that the numeral follows the noun, like in Tibetan. In some contexts, the original word order is maintained, as in example 26b. The numeral always precedes the noun. Conversely, in Salar, like in Wutun, both respective orders of the noun and the numeral are possible, as illustrated in 28a and 28b.
It does not correspond to a dialectal variation, as one and the same speaker can use both structures in turn. However, the pre-nominal position of the numeral is generally consid- ered to be more archaic and characteristic of the speech of the elder generation. The syntactic properties of the Tibetan numeral are partially copied into Salar and Wutun. This time, it involves the morphophonological form of the linguistic unit, with its syntactic properties, and only partly its semantic characteristics.
As already mentioned by Dwyer , sentence-final pragmatic particles have been transferred from the Tibetic languages into the neighboring languages.
Only Mother eats. Thus, we find a similar — albeit not strictly identical — discourse marker in Central Tibetan and in Amdo Tibetan. A: nga-mu liang-ge-de tangka jhi-ge yai 1-COLL two. The Wutun sentence-final markers ba and be are probably related to Tibetic tag questions as well, and the expression of uncertainty could have arisen as a semantic extension of the question forms. Synchronically, the borrowing of discourse particles is close to a full copy in Salar, whereas it is only a partial copy in Wutun.
Diachronic data would be necessary to decide whether the semantic features of the discourse particles were involved in the copying process in Wutun from the very beginning, or whether a language-internal evolution of the properties of these particles has led to the present situation. In this case, the copying process involves no morphophonological material. However, in the causal structure, it is difficult to analyze -nige as the interrogative morpheme, given that it would be the only converb made up of a nominalizer and an interrogative pronoun, instead of a case marker.
Salar Salar PS 87 b. The Wutun causal structure involves nominalization of the verb by means of the nominalizer -de which has cognates in all varieties of Mandarin Chinese , together with the use of the case marker -liangge, which, in Wutun, covers a wide range of functions typical for both comitatives and instrumentals. The instrumental case in Bonan is used to express instruments and causal relations, the latter use is probably due to Tibetic influence.
Therefore, the causal subordination construction in Wutun represents a replication of the Tibetan grammatical pattern, while the case marker involved in this construction is a loan calque from a Mongolic language. The case of causal subordination is interesting insofar as the influence of Tibetan manifests itself in different ways in Wutun and Salar. In Wutun, the syntactic structure itself is transferred, since Sinitic languages normally do not have the subordination structure V-NMZ-case marker.
Concerning Salar, such a syntactic structure is commonly attested in Turkic languages, but the use of the genitive case marker in this construction, with this semantic characteristic, can only be explained by contact with Tibetan. The influence of Tibetan on causal subordination in Salar and Wutun is in accordance with this hierarchy. However, due to limitations of space, we will not discuss these three examples in great detail. Like the causal subordination process, this feature is easily transferrable in Salar, since there are already similar — though not identical — constructions in Turkic languages position verbs used to express aspectual values in a subordination structure.
The introduction of this construction into Wutun may have been facilitated by the fact that it is a peripheral and non-obligatory category. Once again, the intensity of language contact between the Wutun and Tibetan speakers also certainly played a role in this process. The second example of borrowing in the domain of aspect is the progressive aspect based on the existential copula, which is found in both Salar and Wutun. This construction is isomorphic with the Amdo Tibetan progressive aspect, and its introduction into both Salar and Wutun has probably been facilitated by language contact with Amdo Tibetan.
In the other Mongolic languages, the progressive or durative value is usually assumed by a construction involving the locative copula. The main explana- tion for such a linguistic change can only be the intensity of language contact. In her cross-linguistic study on evidentiality, Aikhenvald has also noted that evidentials are easily borrowed from language to language and they are a prominent feature of several linguistic areas in the world.
In Section 3. This involves cases where only one language has aligned to a Tibetan grammatical pattern, while the other language has retained its original features that are different from Tibetan.
By comparing the features for which both Salar and Wutun diverge from their respective language families, we have also found other cases of convergence. In these cases, however, the role of Tibetan as a model language either remains unclear or can clearly be ruled out. One of the most prominent examples is the comparative construction, which is discussed in Section 3. Notably, it does not seem to be attested in any other Tibetic language outside this area: it is an independent development of the Amdo Sprachbund and its source remains unclear.
The same pattern is also attested in Amdo Tibetan, but not in other Tibetan dialects. This structure is, however, attested in different Mongolic languages see e. The semantic transparency of such a structure is a factor facilitating the linguistic transfer. In these cases, the influence of Tibetan has triggered divergence between Salar!
Our most important example presented in Section 3. We have already shown in Section 2. In Amdo Tibetan, demonstratives always follow the noun and the adjective, if there is one , as in example The two strategies are equally possible for adjectives of both Chinese and Tibetan origin. In Tibetic languages, adjective attri- butes can occur either before the noun, usually in a combination with the genitive case marker, or after the noun. The post-nominal position indicates neutral attribution, while the pre-nominal position is usually associated with restrictive usage.
The same is true for Wutun, where pre-nominal adjective attributes usually indicate restrictive usage and post-nominal adjective attri- butes express neutral usage. The position of adjective attributes in Wutun differs from Standard Mandarin, where adjective attributes always precede the noun. The ability to occur as nominalized, post-nominal attributes distinguishes Wutun adjectives from verbs. In Salar, only the position of the numeral has partially changed under the influence of Tibetan, whereas in Wutun, this change has been generalized to the position of the demonstrative and the adjective.
The diffusion of this areal feature is thus more advanced in Wutun than in Salar. Given the current sociolinguistic situa- tion, and the shift from Salar-Tibetan to Salar- Standard- Chinese bilingualism, it seems rather unlikely that Salar will continue its evolution in the direction of convergence with Tibetan for this feature. Other examples of linguistic copies that affect Wutun and Salar unequally include the modification of the functional range of the dative case marker in Salar under Tibetan influence.
Again, it suggests long and intensive language contact. In this case too, the original linguistic type of the Sinitic languages, which typically lack case markers, explains why Wutun is not affected by this kind of linguistic change: unlike Salar, Wutun does not have a dative case marker like the Tibetic and Turkic languages. The first three sections of the table present the cases where the role of Tibetan as a model language is established with a reasonable certainty. Some of them have not been discussed in this paper, namely, the parallel dative-case marking for the possessor or the partici- pant that accesses to possession, and the target of controlled perception verbs in Salar and Amdo Tibetan.
Section 3 of the table shows cases where the influence of Amdo Tibetan has triggered divergent evolution in Wutun and Salar. Section 4 outlines the two examples of areal convergence for which the model language is unclear, probably Mongolic in the case of the immediate future tense. Our findings outline the importance of sociolinguistic and historical factors in explaining the outcome of language contact.
Linguistic Biographies of Wutun Speakers in Qinghai
Explanations of linguistic con- vergence generally fall into three types. First, convergence is determined by the intensity of exposure to the contact language throughout history, and the linguistic outcome of the contact situation depends on the sociolinguistic factors Thomason and Kaufman Ch. Second, the structural similarities and differences among the languages in contact can explain the convergence, or the lack of convergence. Finally, inherent semantic-pragmatic or structural proper- ties of the linguistic categories are also frequently invoked to explain that some categories are more prone than others to contact-induced linguistic change Curnow —; Matras Regarding the inherent properties of linguistic units that can facilitate — or, on the contrary, restrain — linguistic transfer, Matras proposes several generalizations.
First, the likelihood of a given unit to be transferred by contact depends on its absolute frequency in the model language: the more frequently the unit is used, the more likely it is to be transferred. Similarly, the more semantically transparent and morphologically well delimited the category is cf. Weinreich 36—37; Heath , cited in Curnow , the more likely it will be a good candidate for linguistic copy.
For example, numeral classifiers are cross-linguistically prone to borrow- ing because of their semantic transparency Matras This map shows the number of endangered languages in each country. Total number of endangered languages in the world: Vulnerable: Definitely endangered: Severely endangered: Critically endangered: Extinct: What is endangered language? No single factor determines whether a language is endangered, but United Nations Educational, Scientific and Cultural Organisation UNESCO experts have identified nine that should be considered together: Intergenerational language transmission Absolute number of speakers Proportion of speakers within the total population Shifts in domains of language use Response to new domains and media Availability of materials for language education and literacy Governmental and institutional language attitudes and policies including official status and use Community members' attitudes toward their own language Amount and quality of documentation Degree of endangerment is classified in different categories ranging: Vulnerable - most children speak the language, but it may be restricted to certain domains e.
Category: Education. Tags: world status endangered language Extinct speakers. Filter : Loading References Embed or download.