Contact Linguistics. Chap

Non-Austronesian Austronesian

Anêm (isolate) Bibling family Bariai family Whiteman family

Aria Kabana Amara (isolate)

Lamogai Kove
Mouk Lusi
Tourai

Anêm appears to be the sole survivor of a larger group of NAN languages that were once spoken throughout the area, and which have gradually lost ground to the AN languages. Today, it is spoken by only about 350 people in small mountain villages where it is in contact with AN languages like Lusi, Mouk and others. Contact among languages is the norm, so that everyone in NWNB is multilingual in three or more languages, including Tok Pisin.

Demographically and geographically, the contact situation in the pre-colonial past was somewhat different from that in Kupwar, in that contact was spread across contiguous languages spoken by different speech communities in a broad geographical area, a situation akin to a Sprachbund. In other respects, however, the situation is similar to that in Kupwar in that each village has long been characterized by intimate contact among groups who interact on a daily basis. For instance, the Anêm and Mouk speak each other's language since they have a long tradition of cultural relationships and are very much alike in social structure. The Anêm speak Mouk or Lusi as their secondary vernaculars, while speakers of the other isolate, Amara, speak either Kabana or Mouk, and in some cases Anêm, as their secondary language. The general pattern is that the Anêm and Amara (the two minority groups) typically use a Bariai language such as Lusi or Kabana for outgroup communication, since these languages act as lingue franche across the area.
Just as in Kupwar, switching between languages in NWNB involves mostly changing vocabulary while using a common grammatical and semantic structure. The following examples from Thurston (1994:687) illustrate.

(9) NAN isolate Anêm: doxa led u-tl-î aba kan abul

AN isolate. Amara: o-togou ane i-se esNei ne e-ila
BIBLING Mouk: gute ka mtex abax tan uala
BARIAI Kabana: eaba ne i-gal gaea Nan ido
BARIAI Lusi: tanta ne i-gali gaea Nani izo
man this 3sg-spear pig with spear
"This man speared the pig with a spear"

(10) Anêm: nê-b-î kaoa êbêl

Amara: pun o-goune kate
Mouk: om-mluk oulei sakam
Kabana: rau kaua mina
Lusi: u-rau kaua mina
2sg-hit dog don't
"Stop hitting the dog" or "Don't hit the dog"

The major structural characteristics that are shared among the languages of NWNB include the following (Thurston 1994:587-8):

• Fairly rigid SVO order, and similar constituent structure.

• Absence of tense as a grammatical category. Time reference is indicated optionally by temporal words.
• No verbal inflection for distinctions of voice or focus.
• Use of clause-final words to indicate aspect and negation.
• Absence of articles (though there are nominal prefixes in Amara and Malau).

There are various other similarities among the languages in causative, reciprocal, possessive and adverbial constructions, as well as in lexicon.

Most of these similarities seem to have developed as a result of changes in the AN languages under influence from Anêm or very similar NAN languages in their prehistory. A comparison of the AN languages of NWNB with their relatives outside the area makes this clear. As Thurston (1994:585-6) points out, most AN languages outside Melanesia share the following characteristics:

• The verb phrase occurs initially in neutral statements.

• TMA categories (tense, aspect, negation, etc.) are expressed within the VP either by pre-verbal particles or verbal inflections.
• Noun phrases are marked by particles which have the properties of both articles and prepositions.

The AN languages of NWNB show significant departures from the above typology, and marked similarity to NAN languages like Anêm. What processes of contact-induced change could have brought about this high degree of structural convergence? The major factor seems to have been continuing shift over the centuries from NAN to AN languages. As the influx of AN speakers grew, they merged with the indigenous NAN groups, who gradually adopted more and more of the former's culture, including their languages. It was this that led to the gradual expansion of the groups speaking the AN languages and the areas they occupy. It would appear then that the primary vehicle of convergence among the AN languages was (common) substratum influence from the NAN languages, bolstered by widespread borrowing under conditions of extensive multilingualism. The shift scenario is supported by the fact that the AN languages appear to have undergone extensive simplification and regularization of structure, in areas such as their pronominal and tense-aspect systems. Such simplification is commonplace in cases of untutored second language acquisition leading to the creation of new contact vernaculars, particularly when the new creations serve as lingue franche for groups speaking mutually unintelligible languages, as is the case in NWNB. This is particularly true of the Bariai languages, Kabana, Kove and Lusi. The isolates, Amara and Anêm, present a clear contrast to these languages, in that they seem to have largely preserved their original grammars.

In addition to the kinds of structural convergence outlined above, there has also been a great deal of diffusion of lexical items across the languages of NWNB. They differ from one another primarily in basic vocabulary, but much of their non-basic vocabulary is shared. The interior languages (especially Anêm and Amara) have borrowed extensively from the Bariai languages in domains such as fishing, canoe building and other maritime activities, while the Bariai languages have borrowed vocabulary from Anêm in areas pertaining to sound and motion, and the rainforest. The combinations of lexical and structural diffusion that have led to convergence among all these languages make it difficult to classify the outcomes strictly in terms of maintenance vs shift, since both seem to apply. Moreover, the mechanisms of convergence include both substratum influence under the agency of learners acquiring new means of communication, and borrowing under the agency of speakers maintaining their native languages. The situation in NWNB does not fall neatly into any one category, and poses a problem for Thomason & Kaufman's view that shift and borrowing situations can be clearly distinguished by the degree of structural vs lexical diffusion that takes place.
It remains for us to relate the cases of structural convergence discussed here to a typology of the social and linguistic factors that regulate these kinds of contact-induced change in maintained languages.


8. The social contexts of structural convergence.

As Thomason & Kaufman (1988:67) note, the classic situation that promotes structural borrowing involves heavy cultural pressure exerted on a subordinate population by a politically and numerically superior group. The various situations discussed here seem to bear this out, though not all cases of heavy pressure have resulted in the same degrees of structural change. The extent of structural diffusion is dependent on a broad range of social and linguistic factors, many of which are specific to each situation.

A full account of the social contexts of these contact situations is beyond the scope of this chapter. Suffice it to say that they vary according to the parameters outlined in Chapter 1, which are the basis for establishing typologies of contact settings. They include:

• The demographics of the groups involved, including numerical ratios, power and prestige relationships, etc.

• The community settings in which the contact occurred.
• The frequency and type of social interaction among the groups.
• The ideologies concerning language and its relation to social identity in each group, and, hence, attitudes toward language mixture.
• The directionality of change, and the degree of stability or shift in the contact setting.

All of the cases of extensive structural convergence we have considered here involve stable multilingualism as an integral part of the social relations among the groups in contact. But each situation has led to different kinds of outcome. In Kupwar, convergence has resulted in near isomorphism of grammars, though lexical differences are preserved. Language boundaries here have proven weak and permeable. The major reason seems to be the extensive use of code-switching and mixture in everyday interaction, and the fact that Marathi is used by everyone as a lingua franca. In NWNB, both structural and lexical convergence has occurred, leading to considerable isomorphism of structure. This seems to be due primarily to shift-induced changes resulting in common substratum influence from NAN on AN languages, though regular borrowing seems to have played some role as well. Moreover, the conservative interior languages preserve their distinctive complexity of grammar, while the AN languages which serve as lingue franche show evidence of simplificatory change. Finally, in Arnhem land, multilingual contact has resulted in a pattern of diffusion very different from that of either Kupwar or NWNB. Convergence here is to a large extent due to borrowing, both structural and lexical, with shift-induced changes playing a secondary role. There is no evidence of the kind of structural isomorphism found in Kupwar, or the simplification and regularization found in NWNB. The languages preserve distinct and rather complex morphological systems and areas of lexicon.

Clearly, then, convergence in multilingual situations can take different forms, and lead to quite different outcomes. One explanation for this may lie in the kinds of multilingual interaction practiced in each community, as described earlier. The full details of the different contact settings and patterns of social interaction that contributed to these different outcomes are yet to be fully understood. Some of these social factors will be discussed more closely in the following chapter, with particular reference to the kinds of bilingual language mixture known as code switching.


9. Linguistic constraints on structural diffusion into a maintained language.

The traditional view held by historical linguists such as Meillet, Sapir and others was that grammatical subsystems and morphology in particular were highly resistant to cross-linguistic influence. This belief has proven untenable in the face of evidence from many contact situations. But it appears that structural diffusion is for the most part the result of substratum influence under sl agentivity, rather than direct borrowing. Hence Meillet and others may well have been right, as far as the latter is concerned. In seeking constraints on structural diffusion, then, we need to be clear as to which vehicle of change is involved, since constraints that apply to one may not apply to the other. For the present, let us confine our attention to constraints on direct structural borrowing, in the strict sense. The constraints that apply to substratum influence (in cases of language shift) will be discussed in Chapter 7.

9.1. Constraints on borrowing of morphology.

There is still much disagreement on the extent and type of structural borrowing possible under contact. With regard to morphology, there seems to be consensus that derivational morphemes are more likely to be borrowed than inflectional ones. Moreover, it would appear that derivational morphology tends to be introduced via lexical borrowing, as we have seen.

There is rather less agreement as to whether inflectional morphology can be borrowed, either directly or indirectly. The cases we have examined here suggest that both are possible, under the right circumstances. Weinreich attempted to identify various structural constraints on such diffusion. According to him, the borrowing of morphemes follows these tendencies:

• Highly congruent structures allow for substitution of one morpheme by another.

• Zero morphemes tend to be replaced by overt ones, and bound morphemes by free forms.
• Allomorphic variants are likely to be leveled into a single form.
• Morphemes with opaque and complex grammatical functions are less likely to be borrowed than those with a single transparent function.
• Morphological borrowing is favored by functional factors such as gaps in the recipient language's system.

While these observations are supported by many examples in Weinreich's own data, they do not predict what will happen in all contact situations. Heath (1978:73), for instance, discusses several instances of morphosyntactic borrowing in Arnhem land which run counter to some of Weinreich's claims. He therefore suggests alternative characteristics of morphemes which seem to favor their "diffusability", as follows: (1978:105)

• Syllabicity. (Morphemes that are independently pronounceable.)

• Sharpness of boundaries. (Morphemes which are in opposition to zero)
• Unifunctionality. (Morphemes with a single function, as opposed to portmanteau morphs.)
• Categorial clarity. (Morphemes whose function is clear without having to examine the broader environment in which they occur.)
• Analogical freedom. (Morphemes that are not dependent on other morphemic systems in the same language.)

The factors or constraints proposed by Weinreich and Heath are not mutually exclusive, and indeed complement one another. They fall into three general categories:

• Constraints based on congruence of morphological structures.

• Constraints based on transparency.
• Constraints based on functional considerations.

As will become apparent, these constraints involve considerations of typological distance and markedness.

9.2. Congruence:

The typological constraint most frequently proposed for morphological borrowing is some version of Weinreich's suggestion that "the transfer of morphemes is facilitated between highly congruent structures." This may be formulated as follows: