Contact Linguistics. Chap

10. Summary.

We’ve seen that lexical changes due to contact involve not just direct importation of words, but a variety of other processes leading to innovations in the lexicon of the recipient language. Lexical borrowings are integrated in varying degrees into the phonology, morphology and syntax of the borrowing language. They are also subject to different kinds of semantic change, as they vie for a place in the lexicon with native items that may already occupy similar semantic space. Once incorporated, they become fair game for both derivational and inflectional processes internal to the recipient language. Many of the products of these processes – loan blends, creations and the like – are not strictly speaking borrowings, but innovations that have no counterparts in the source language.

Lexical borrowing must therefore be seen as just one aspect of a creative process of lexical change under contact, which builds on both native and foreign resources. This process not only adds to the lexicon, but also enriches its potential for further expansion. Foreign additions to the lexicon may also trigger phonological and morphological innovations. These structural changes appear to be subject to strict constraints, not least of which is the extent of the lexical borrowing involved. The higher the latter, the more likely it is that some structural features will be transferred. But since these structural innovations have to fit into the existing phonology and morphology of the recipient language, they are also subject to typologically-based constraints as well. Hence they tend to be rare even in cases of substantial lexical borrowing. The conditions under which structural features diffuse to more significant degrees across language boundaries will be the subject of the next chapter.
An Introduction to Contact Linguistics.
Chapter 3. Structural diffusion in situations of language maintenance.


1..Introduction.

We saw in the last chapter that languages can import large numbers of foreign lexical items while still retaining their basic grammar, and indeed their genetic affiliation to earlier forms of the language. But do languages also import structure from external sources, and if so, under what conditions? What kinds of agency are involved in the diffusion of structural features? And what limits are there on it?

We have already seen that significant lexical borrowing can introduce new derivational morphemes and processes, as well as new phonemic distinctions (though not necessarily new phones) to a language. Does this mean that all structural borrowing is mediated by lexical borrowing? Or can structure be borrowed separately, in its own right?
Some scholars, e.g., Thomason & Kaufman (1988), distinguish between “lexical” and “structural” borrowing as though the two could in fact proceed independently of each other. Indeed, Thomason & Kaufman give many examples of what they call “structural borrowing” and even go so far as to claim that, under conditions of intense contact, massive structural borrowing can result in entire grammatical systems being replaced. But despite the many claims concerning the “borrowing” of phonology, morphology and even syntax, there has been no convincing demonstration that such structural change occurs without mediation by some other medium or process.
In fact, when we examine cases of language maintenance in which the maintained language is subject to extensive lexical borrowing from an external language, we find that in many if not most cases, little structural innovation occurs. Moreover, such structural change is practically always mediated by lexical transfer. This tends to support the traditional wisdom that (maintained) languages put up stiff resistance to importation of foreign structural elements. This constraint appears quite valid for structural borrowing in the strict sense of that term. It would be very strange indeed for speakers of a maintained language to single out structural features from a foreign source and install them as part of their native grammar. In the first place, why would they? Secondly, how could they identify and access such features for importation? Speakers do not have access to a list or a complete inventory of structural features from which they can select at will the ones they wish to import. For instance, when Middle English speakers “borrowed” derivational morphemes like –tion, -able, etc. from French, it certainly wasn’t because they isolated them in the relevant French words and imported them independently of the stems to which they were attached. The process of borrowing – which, recall, involves the agency of recipient language speakers – would therefore appear to be subject to very strong constraints indeed, when it applies to structure. Such constraints may be weaker in the case of derivational morphology (which is more lexical in nature anyhow), but are more robust in the case of phonology, inflectional morphology and syntactic structure.
This is not to say that such structural features can never be diffused from one language into another. Indeed, it was suggested in Chapter 1 that there is in principle no limit to what can be transferred across languages, given the right circumstances. But when structural features are transferred, it is rarely the result of direct borrowing. Rather, such transfer is either mediated by lexical borrowing (as noted) or introduced under the agency of speakers of the external source language, whose innovations are then adopted by other speakers of the recipient language. In most if not all cases, as we shall see, the speakers who initiate such structural change are bilinguals.
The distinction being made here between direct borrowing and structural diffusion that is mediated by source language (sl) agentivity is similar to the distinction that Heath (1984:367) makes between “direct transfer of forms from the other language” and “structural convergence (using native morphological material).” For Heath, “direct transfer” includes borrowing and code-switching, while “structural convergence, also called pattern transfer or calque, is the rearrangement of inherited material because of diffusional interference” (ibid.).⁷ As Heath notes, such “interference” can cause recipient language morphemes, words or phrases to undergo rearrangements that make the recipient language (rl) structures more similar to those of the source language. This would suggest that abstract structural patterns may be transferred at the morphological and syntactic levels. It isn’t clear, however, what mechanisms underlie this type of convergence. It will be argued later that these mechanisms involve L1 or substratum influence under the agency of bilinguals who impose sl features on the rl.
To summarize briefly at this point, we can distinguish several types of contact-induced structural changes in cases of language maintenance. They include direct borrowing of structural elements (quite rare), indirect structural diffusion via lexical borrowing (fairly common) and indirect structural transfer via sl agentivity in situations of bilingualism and mutual accomodation between the linguistic groups. Strictly speaking, all of these types of transfer are vehicles of convergence, since they all lead to greater structural similarity between the languages involved. In general, then, two languages can be said to have converged structurally when previous differences in grammar between them are reduced or eliminated either because one adopts structural features from the other as a replacement for its own, or both adopt an identical compromise between their conflicting structures. The vehicles of convergence may include direct borrowing, substratum influence and other mechanisms of levelling such as simplification. As we shall see, it is not always easy to determine which of these applies in a given instance. In the following sections, we will explore various instances of structural convergence and examine the sociolinguistic as well as structural factors which brought about the diffusion of linguistic features.
Let us first examine cases where it has been claimed that structural features have been directly borrowed.


2. Is there direct borrowing of structural features?

As we saw in Chapter 2, there is good evidence that certain function words such as conjunctions, prepositions and even complementizers can be borrowed directly. For example, various Amerindian languages in Latin America have borrowed Spanish conjunctions like pero ‘but’ and como “as, like” and so on. However, such borrowings have had little or no impact on the grammar of the recipient languages.⁸

It has been claimed that direct borrowing of inflectional morphology can occur in cases of close typological fit between the languages in contact. Thus Weinreich (1953:32) informs us that Meglenite Rumanian speakers have replaced the native verb inflections -u (1st per. indicative) and -i with Bulgarian inflections -um and -is* respectively. Such importation of inflections appears to be generally rare in situations of language maintenance, though it does occur if there is sufficient congruence between the inflections involved. Another type of structural borrowing that has been claimed to occur involves changes in the semantics and/or grammatical function of native morphemes under foreign influence. An example is the extension of the function of American Yiddish ver from interrogative pronoun to both interrogative and relative pronoun on the model of English who (Weinreich 1953:30). Cases like these are the morphosyntactic equivalent of lexical calquing, or “loanshifts.”
The so-called “borrowing” of morphology may take various other forms, such as the introduction of new morphological categories or the loss of existing categories. An example of a newly-introduced morphological category is the development in Balkan languages of a distinction between two complementizers, one used in factives after verbs like say, think, etc. (e.g., Rumanian ca&), the other after desideratives like want, demand, etc. (e.g., Rumanian sa& - (Weinreich 1953:40). This distinction corresponds to the Middle Greek distinction between óti and ná (ibid.). Examples of the loss of morphological categories include the demise of the usual Dravidian distinction between exclusive and inclusive "we" and the old Dravidian gender system in Brahui under Balochi influence (Thomason & Kaufman 1988:93, citing Emeneau 1962:56). Many other examples of these kinds of contact-induced change can be found in Weinreich (1953:39-42) and Thomason & Kaufman (1988:78-97).
The problem is that all of these cases are cited without any explanation of the processes by which the changes occurred, or of the agentivity involved. Many of the situations in which these changes took place involved bilingualism and/or ongoing language shift. Hence the mechanisms involved might have included heavy lexical borrowing or code-mixing on the part of recipient language speakers, as well as substratum influence by bilingual source language speakers whose innovations were imitated by rl speakers. In short, the case for direct borrowing of structure in any of these situations has yet to be convincingly made.
In order to find explanations for contact-induced changes in structure, we need to consider specific cases of contact and examine their social settings and the dynamics of language use by the groups involved.


3. Factors affecting structural convergence.

The extent and nature of structural convergence depend, as usual, on a range of social, historical and linguistic factors. In the first place, there are differences in the socio-historical circumstances that gave rise to each situation. In many cases, it is difficult to reconstruct the earlier history of the contact, and one must rely more on speculation (aided by linguistic evidence) to establish the sources and directionality of change. Second, the social contexts of the contact vary significantly from one situation to another. Some situations involve shift from a minority language to a dominant one with one-way bilingualism within the minority group, while others involve a high degree of bi- or multi-lingualism across groups. To complicate matters, there may be differences in degrees of bilingualism between different groups in the same community. Some situations involve diffusion of features across geographical and linguistic boundaries, while others involve relatively close contact within the same community. Finally, there are cases where diffusion has taken place gradually over long periods without extensive multilingualism, as opposed to those where diffusion has been swift as a result of intense contact and pervasive communal multilingualism. It is clear also that many of these convergence situations overlap considerably with those in which we find contact phenomena such as code-switching. As usual, the task is to determine what combination of historical, demographic, socio-cultural and other factors is responsible for these different outcomes. In some cases, this may prove impossible.

We also need to account for convergence in terms of linguistic processes and constraints. The first problem is how to distinguish internally-motivated changes from those due to diffusion from external sources. Second, we must identify the precise external source, as well as the agents of change - a task made all the more difficult by the absence of relevant socio-historical data for many contact situations in the distant past. Finally, in cases where we have the necessary data, we must explain the mechanisms of change - whether borrowing, substratum influence or other causes - as well as the constraints that affect their spread.
4. Structural convergence in stable bilingual situations.

Let us first examine those situations involving more or less stable bilingualism within the community. These settings correspond to some of those examined in the previous chapter. We saw that, in general, such situations tend to promote varying degrees of lexical borrowing with only marginal diffusion of structural features. Indeed, even minority languages that are under heavy cultural pressure from a dominant host language resist importation of structure, with the exception of derivational morphology and some function words, as already noted. For instance, Pennsylvania German (PG) appears to have undergone very little structural change as a direct result of contact with English (Louden 1997). Fuller (1996) claims that PG is converging toward English in structure. But the changes she documents are relatively few in number, and most of them involve no new (morpho)syntactic patterns in PG. In fact, several of the changes involve minor adjustments in the subcategorization properties of verbs, similar to those discussed in the following section for Los Angeles Spanish. Finally, Fuller (p. 511) notes that “few overt English system morphemes are being brought into PG.”

Similarly, the French spoken in the Ottawa-Hull region of Ontario, Canada shows very little evidence of structural influence from English. Concerning this contact situation, Poplack (1996:286) informs us that, despite extensive code switching and lexical borrowing among bilingual speakers, there is no evidence that French grammatical structure has been replaced by more “English-like mechanisms.” Similar conclusions apply to many other minority languages in bilingual situations, for example Alsatian and Flemish, both in contact with French, or Spanish in many Hispanic communities in the United States. Let’s look briefly at the last of these situations.


4.1. Spanish in L.A.

Silva-Corvalán (1994) presents many examples of apparent structural changes in Los Angeles Spanish that have been attributed to English influence. Such changes are particularly characteristic of the varieties of Spanish spoken by persons bilingual in Spanish and English. But Silva-Corvalán argues that such changes do not in fact represent anything new to Spanish syntax. For instance, an apparent change in the “affected dative” construction yields structures like (1a) in LA, whereas Spanish normally has (1b)

(1) a. LA Span. y quebraron mi, mi jaw.

And they-broke my, my jaw

b. Gen. Span. y me quebraron la mandibula

and to-me they-broke the jaw
“And they broke my jaw”

Yet Spanish does allow structures like (1a) under different pragmatic conditions, e.g. when the affected object is not an inalienable possession, as in (2).

(2) Tiraron una piedra y quebraron mi portalápices.

“They threw a stone and broke my pencilholder” (1994:139)

All that is involved in (1a), then, is a relaxation of the pragmatic constraint on the use of the “affected dative.” No new syntactic structure is “borrowed” from English. Other apparent syntactic changes involve extensions of the meanings or functions of individual lexical items. For instance, in (3a), atrás, which normally means “behind”, has acquired the English sense of “back.” (2000:14)

(3) a. LA Span. Se lo dió p’atrás (

to-him it she-gave back
b. Gen. Span. Se lo volvió
to-him it she returned
“She gave it back to him”

By way of final example, in (4a) jugar “to play” (games) has had its meaning extended to include the playing of musical instruments, a sense normally conveyed in General Spanish by tocar. The change also involves modification of the subcategorization frame of jugar, as in the following example from Otheguy (1995, ex. 2).

(4) a. LA Span. Cuando no tengo nada que hacer, juego música.

b. Gen. Span. Cuando no tengo nada que hacer, toco música.
when NEG I-have nothing what to-do I-play music
“When I have nothing to do, I play music.”

Silva-Corvalán (1998) discusses several other examples like these, and argues that the changes involved are not the outcome of direct borrowing of syntactic structures or rules from one language into another (1998:225).Rather, what is involved is a kind of “lexico-syntactic calquing” (ibid.) triggered by partial congruence between Spanish and English words. This results in Spanish words assuming the semantic and/or subcategorization properties of the apparent English equivalents (faux amis). Her findings lead her to the conclusion that “what is borrowed across languages is not syntax, but lexicon and pragmatics” (1998:226). This claim certainly seems to hold for cases of language maintenance with bilingualism such as that in LA. Clearly, intense contact and heavy cultural pressure are insufficient to bring about significant structural change in many minority languages. Under what conditions, then, do such changes occur? To answer this, let us examine certain bilingual situations in which structural diffusion has taken place.