Contact Linguistics. Chap

Exercise: Compare some of the changes in Kupwar Urdu (Chapter 3, section 7.3) with those resulting from Russian influence in the creation of CIA. To what extent are they similar? (Advanced: Do the same for the changes in Asia Minor Greek under Turkish influence. See Thomason & Kaufman 1988, 215-222).


8. Summary.

In this chapter, we examined a variety of bilingual mixed codes that are used as in-group languages, and were created by more or less skilled bilinguals. Some of these are transient and not likely to survive, for instance the mixed Panjabi/English code used by younger generations of Sikhs in England. Others are restricted in scope of use, like Tsotsitaal, the blend of Zulu, English and Afrikaans used by younger people in South African townships. Though the creation of such mixed codes appears to be quite common, few have achieved autonomy as distinct languages. The few that have done so constitute the class of bilingual mixed or intertwined languages.

We examined four well-known exemplars of this type of contact language – Media Lengua, Michif, Ma’a and Copper Island Aleut. They display noticeable differences in their patterns of mixture and provide some sense of the diversity of these outcomes of contact. In general, intertwined languages differ from other contact varieties in that their components, whether lexical or structural, are preserved relatively intact, and can be easily traced back to the respective source languages. There is still some disagreement over the precise origins and classification of these languages. Much of it is due to the fact that distinctions made according to purely structural criteria do not always match those that are made according to socio-historical criteria.


8.1. Socio-historical background and social motivations.

Bilingual mixed languages have arisen under a variety of sociohistorical circumstances; hence it is not easy to generalize about the social motivations for their creation. However, two broad scenarios can be identified. On the one hand, there are languages like Michif, Copper Island Aleut, Chindo and others that were created by newly-emerged social groups who wanted a language of their own. For the most part, these groups are “creoles” of mixed race who see themselves as distinct from either of the cultural groups from which they descended. But in certain cases, such as Media Lengua, the group involved is not of mixed race, but is caught between competing cultures. The same motivation to express a separate social identity lies behind mixed codes like Tsotisaal and others that have not achieved autonomy.

On the other hand, there are languages like Ma’a, Anglo-Romani, Caló and others which arise when a group’s ancestral language has almost been replaced by a dominant host language. Among such groups are ex-nomads like the Roma and the Callahuaya healers, who need a secret, in-group language. Like the Ma’a and Elmodo, these groups wish to preserve some of their old traditions and culture by retaining parts of their ancestral language in their new language. This reflects their attempt to resist total assimilation to the dominant culture.


8.2. Processes and origins.

Most of the intertwined languages known to us display characteristics found in the code-switching behavior of skilled bilinguals. For instance, the pattern of mixture in Media Lengua involves an extreme form of the single content morpheme insertion found in some types of code switching. On the other hand, the mixture in Michif is reminiscent of the type of code switching known as EL island insertion, the main difference being that Michif does this across the board, replacing Cree nouns consistently with French ones. In both cases, it is possible to identify a Matrix Language – Quechua in the case of Media Lengua, and Plains Cree in the case of Michif – into which EL elements (from Spanish and French respectively) are incorporated. Similar processes lie behind the creation of most other bilingual mixed languages, though the ML is the ancestral language in some cases (like the two above), while it is a newly acquired language in others (like Ma’a and Anglo-Romani). The constraints on the mixture found in these cases are quite similar to those that operate in code-switching generally.

The case of Copper Island Aleut is quite different. The blend of Russian and Aleut structure represents a very unusual pattern of mixture. The processes involved in this case seem similar to those found in cases of convergence, particularly those involving extreme structural diffusion, as for instance in some varieties of Asia Minor Greek. The product in each case is pervasive mixture in every component of the grammar. In CIA, however, the pattern of intertwining is much more regular, displaying the neat compartmentalization of components typical of intertwined languages.
Despite the intricate pattern of mixture, the processes of intertwining in Copper Island Aleut appear to conform to the usual contraints on structural diffusion or convergence. Typological differences between Russian and Aleut block certain kinds of substitution, while others, e.g. the wholesale incorporation of Russian tense suffixes, are facilitated by (partial) congruence between verbal structures in both languages.
On the whole, therefore, while intertwined languages are relatively rare creations, the processes and constraints that were involved in their genesis are not different from those that operate in other kinds of bilingual mixture.


Introduction to Contact Linguistics.
Chapter 7 . Second Language Acquisition and Language Shift.


Section I. An overview of individual SLA.


I. 1. Introduction.

The types of cross-linguistic influence we have considered in previous chapters have all, by and large, involved changes in a maintained language under the agency of its own speakers. In a majority of these cases, recipient language (rl) speakers “borrow” or incorporate foreign (sl) features into their L1. This process of borrowing, as we have seen, is subject to quite strict constraints. In some cases, e.g., those involving structural convergence, both rl and sl agentivity seem to be involved, with active bilinguals playing a crucial role. In the next few chapters, we will be concerned with situations involving various kinds of second language acquisition, in which the target language (TL) is changed to varying degrees under the agency of learners.

Speakers attempting to communicate in a language they are acquiring may resort to various strategies to achieve success. In doing so, they create versions of the TL that differ in many ways from the varieties used by its native speakers. The strategies and changes involved in this procedure of trial and error offer fascinating insight into the creativity that is part of second language acquisition (henceforth SLA). Attempts to describe and explain this creative activity and the principles that regulate it have led to some degree of understanding, though there is also much disagreement on several issues.
Researchers traditionally make a distinction between second language acquisition and foreign language acquisition (see Ellis 1994:11-12). The former involves the learning of a language in “natural” settings, for instance by immigrants who have to acquire the dominant language of a host community. The latter refers usually to the learning of a foreign language through formal instruction, for example in the classroom. Henceforth I will refer to both types as SLA, except where it may become necessary to distinguish the two. Research has shown that both types of learning are similar in many ways, and each can provide a great deal of insight into the other. Our concern, however, is primarily with so-called “natural” SLA. This takes place in community settings where learners attempt to acquire the TL through more or less direct interaction with its speakers (or with others who have learnt it), usually without benefit of instruction. In such real life situations, as van Coetsem (1988:19) notes, “the primary aim is communication, not the acquisition of an optimal knowledge of the TL.” Hence learners may freely adapt the TL in ways that best facilitate their need to communicate.
Our ultimate concern will be with language shift or second language acquisition by groups or communities of speakers. But since individual SLA feeds group SLA, we need to understand the former in order to fully explain the latter. It won’t be possible to consider all aspects of individual SLA here; we will focus only on those that seem most relevant to our present interest. Both individual and group SLA provide a rich laboratory for investigating the changes that arise from learners’ creativity, as well as the principles and constraints that guide them. In general, such changes fall into three categories: (a) L1 or substratum influence on the learner version of the TL or “interlanguage”; (b) various kinds of simplification of TL structures (see section I.3 below); and (c) changes that are internal to the interlanguage system. These changes arise from strategies that learners employ to simplify the task of learning, and to compensate for lack of acquisition (mastery) of TL features. (See van Coetsem 1988, Chapter 3). These strategies will be discussed further below.


I. 2. L1 influence in SLA.

It has long been recognized that learners’ versions of a TL are subject to varying degrees of influence from their native or primary (dominant) language. The latter may not be the same as the former; however, I will henceforth refer to either of these as the L1. Weinreich (1953) was among the first to attempt a comprehensive account of these types of influence, which he referred to as “interference” phenomena. Though Weinreich himself did not always distinguish interference due to borrowing from that due to L1 influence, most of his illustrations in fact deal with the latter. Hence some scholars use the term ‘interference’ only in the sense of L1 influence on a TL. Most scholars of SLA, however, prefer to use the term ‘transfer’ instead. This term has been used in different senses. Lado (1957) interpreted it as the use of overt L1 elements or structures in the L2. This narrow interpretation failed to take account of all the ways in which the L1 might influence the L2. Moreover, Lado’s idea of the transfer of linguistic habits seemed somewhat too closely tied to behaviorist theories of language acquisition which later scholars challenged. Odlin (1989:27) offered a much broader interpretation of transfer as “the influence resulting from similarities and differences between the target language and any other language that has been previously (and perhaps imperfectly) acquired.” While this allows for a greater range of influence than Lado’s definition, it raises the question whether all of the possible types of L1 influence can be properly subsumed under the concept of ‘transfer.’

Part of the problem lies in the varied uses to which this term is currently put. It may refer both to the manifestations of L1 influence and to the psycholinguistic processes that bring them about. In the former sense, it refers to the final manifestations or products of certain kinds of L1 influence on an outcome of SLA. In other words, it is possible to claim, in retrospect, that certain L1 properties have been ‘transferred’ into learner versions of a TL, changing it in some way. Such an approach views learners’ innovations purely from the perspective of the original TL, rather than from the perspective of the IL grammar itself. We shall adopt the latter perspective, and use the term in the sense of a strategy or process. We will continue to use the more neutral term “L1 influence” to refer to the results of this process. L1 influence manifests itself in the form of L1 features that learners introduce into their developing version of the L2, or interlanguage (henceforth IL). Such features might be referred to as “L1 retentions”, which constitute a significant input to the restructuring of the early IL system.
As Jarvis & Odlin (2000:537) suggest:

Much of what is called cross-linguistic influence [in SLA – DW] can be viewed in terms of retentions: Whenever challenges of using or understanding a second language arise, learners may retain something from their L1 or some other language to aid in coping with the new challenges.”

Such retentions, as we shall see, can be overt or direct, as in the use of L1 sounds, words and sometimes morphemes in the IL, or it may be covert, involving more abstract L1 categories or patterns. In cases where L1 and L2 match each other closely, L1 retentions may result in relatively close approximations to the target element or structure, leading to what is called “positive transfer.” In cases of mismatch, such retentions often lead to “imperfect learning”, and represent forms of “negative transfer.” Or, as Weinreich (1968:1) puts it: “The greater the difference between the systems…..the greater is the learning problem and the potential area of interference.”

I. 2. 1. L1 influence on the TL lexicon.

In the simplest cases, learners may retain features of their L1 to compensate for their unfamiliarity with equivalent features of the TL. For instance, they may retain overt lexical items from the L1. Nemser (1991: 352-3) provides several examples of such L1 lexical retentions in the spoken and written L2 English of German-speaking Austrian students. Beginning and intermediate students use lexical items like brills ‘eyeglasses’ and function words like auBer ‘except.’ More advanced learners employ words like grammatik for ‘grammar’ and dement for ‘deny.’ L1 influence on the L2 lexicon may be more covert, leading learners to reinterpret TL words in terms of the semantics of similar L1 forms, thus creating new meanings for the former. For instance, Nemser’s students used meager to mean ‘thin’ (cf. German mager ‘thin’), and guilty to mean ‘valid’ (cf. German gültig ‘valid’) (1991:348). Appel (1996:392) refers to this as “negative transfer of word form.” Such L1-L2 word pairs are often called faux amis or ‘false friends.’ Similar examples of such pairs in English-French include demand vs demander ‘ask’; ignore vs ignorer ‘not to know’, etc. Negative transfer can occur in both directions, that is, when either language is the TL.

Just as insidious from the learner’s viewpoint are cases where an L2 form only partly matches an L1 form in meaning, but the learner assumes complete overlap between the two. For instance, German learners of English may extend the meaning of carry to include both ‘transport’ and ‘wear (clothes)’ on the model of German tragen. Note the similarity to cases of semantic extension in borrowing. Finally, learners often create loan translations employing L2 forms combined according to an L1 pattern. Nemser (1991:358) cites examples such as ill-car ‘ambulance’ (cf. German Krankenwagen lit. ‘sick-car’) and alp-dream ‘nightmare’ (cf. German Alptraum lit. ‘alp(?)-dream’). These coinages were created by learners at an intermediate stage of acquisition of L2 English.
All of these types of lexical contact phenomena recall those found in borrowing, except that the direction of influence is reversed. As we will see in Chapter 9, similar kinds of reinterpretation of superstrate lexical forms on the model of substrate semantic patterns are also common in creole formation.
It seems clear that learners rely heavily on their L1 lexical structure in creating their L2 lexicon. This assumption lies behind the Semantic Equivalence Hypothesis, which claims that “conceptual patterns and linguistic/semantic coding practices in the L1 provide the essential criteria for those in the L2” (Ijaz 1986:437). Other studies have provided support for this view, e.g., Graham & Belnap (1986), Hoeks (1985), etc. But L1 influence cannot explain all that is involved in L2 lexical acquisition. For that we still need a broader lexical theory that will explain the stages of the acquisition process and the ways in which learners solve the problems they encounter.


Exercise:
L2 learners’ uses of L1 lexical knowledge in interpreting L2 lexical items appear quite similar to certain kinds of adaptation characteristic of lexical borrowing from an L2 into an L1 (see discussion in chapter 2). What similarities and differences do you find?


I. 2. 2. L1 influence on L2 phonology.

As Weinreich (1953:14) noted, L1 influence on L2 phonology occurs when “a speaker perceives and reproduces the sounds of one language, which might be designated secondary, in terms of another, to be called primary.” In other words, speakers make “interlingual identifications” between L1 and L2 sounds, and simply replace the latter with the former. Similar kinds of reinterpretation may affect the phonotactics as well as the prosodic features of the TL (that is, the version of it that is acquired by the learners).

Phone substitution may involve the simple replacement of a target sound by a close counterpart in the L1, without further consequences for the phonemic system as a whole. Weinreich refers to this as “actual phone substitution”, where two sounds are identified as equivalent in the two languages in contact, though their actual phonetic realizations may differ somewhat. For instance, English learners of L2 French substitute their native /u/ for French /u/ without realizing that the two sounds differ acoustically, and hence do not accurately produce the former (Flege 1987).
When there is a significant mismatch between the L1 and L2 phonological categories, phone substitution can have various consequences for the learner version of the TL. In some cases, it results in loss of phonemic distinctions. For example, Spanish learners of English may substitute their native /i/ for both /iy/ and /ê/ of English, so that both ship and sheep are pronounced as /Sip/. Similarly, Japanese learners of English often fail to distinguish between /l/ and approximant /®/, substituting their only native liquid sound for both. Weinreich (1953:18) refers to this as ‘underdifferentiation of phonemes’, and identifies several other kinds of phonological reinterpretation that result from the substitution of L1 sounds or sound patterns for those in the TL.