University of Texas at El Paso

As the Quincentennary celebration of Columbus's first voyage to the new world approaches, nothing could be more appropriate than an assessment of the state of the art in investigations of language contact and bilingualism involving Spanish and Portuguese. For the arrival of Europeans in the Americas signals nothing so much as the contact of Spanish and Portuguese with myriad languages -indigenous languages, African languages, and other European languages, including with each other. This contact has, in many cases, not only endured, but has produced extensive bilingualism.

Investigators of bilingualism and language contact face three principal tasks. One is to understand the nature of bilingual societies; a second is to understand bilingual individuals; the third is to understand what happens to grammatical systems when they are in prolonged contact with other systems.

The study of bilingual societies involves the attempt to understand how they come into existence, how they are maintained, and how they (in some cases) become monolingual. The study of bilingual individuals attempts to understand how the persons become bilingual, how they remain bilingual, how they use their two languages to function in bilingual societies, and how the neurolinguistic apparatus of bilingual speakers may or may not differ from that of monolinguals. Understanding of the effects of bilingualism and contact on grammatical systems depends to a certain degree on knowledge of the bilingualism of societies and individuals who use the grammatical systems, as well as on substantive and equivalent descriptions of the grammars involved in their monolingual forms. The principal question is the relationship between grammars -between the grammars underlying the pre-contact and post-contact forms of the language, or between the grammars underlying non-contact or contact varieties.

Obviously, the specific questions pertaining to the general areas of inquiry mentioned above could be investigated in any of several situations of contact and bilingualism. But the fact is that contact situations involving Spanish and Portuguese are particularly appropriate for such research. They are accessible, extensive, and varied. They have provided opportunity and material for understanding of contact and bilingualism from the very beginning, and continue to do so. Haugen 1956, of course, marked out the territory, bibliographically and conceptually, for others to follow. In the middle years since then, two of the massive landmark accomplishments in the field, Rubin 1968 and Fishman 1968, are concerned with Spanish, in one instance in contact with an indigenous language, in the other in contact with English.

What I shall try to accomplish here is an assessment of the substantive progress in the field in recent years. By that I mean not just knowledge of specific cases of language contact and bilingualism, but those ways in which we have advanced our understanding of the fundamental nature of contact and bilingualism, including its possibilities and prohibitions².

Still, «state of the art» essays are always problematic. Should they be more than grouped and annotated bibliographic entries? How much more? Clearly, one must give a reasonably thorough idea of the range of activity in the field. But I do not pretend to bibliographic completeness, and readers are advised that this is not a bibliographic essay. As

––––––––   811   ––––––––

Haugen noted in wrestling with the same problem some 35 years ago, the topic had already attracted the attention not just of linguists, but also of sociologists, anthropologists, psychologists, educators, politicians, and others. The relevant literature was therefore «mountainous». If the climb necessary for an overview was daunting in 1956, today it is probably not possible.

This is not to say that we can do nothing, however. While the literature has exploded, many of the issues remain the same, and the intrepid explorer must traverse many of the same rocks, glaciers, and crevasses Haugen encountered. In some cases the progress of the intervening years has amounted to the unearthing of new examples of well-known phenomena. In others, genuinely new answers to new problems have emerged. In yet others, little progress at all is evident. The problem is to distinguish one from the other.

Until recently, the area most resistant to substantive advance has been understanding relationships among grammars. It is true that traditional approaches to historical linguistics have always included language contact as a cause of change, and have attributed a contact-based origin -substratum- to the different dialects of Spanish. Mexican Spanish was said to be different from Peruvian (Ecuadorean, Bolivian, etc.) because of its Nahuatl (Quechuan) substratum. But precise understanding of how, beyond lexical borrowing, contact-induced change occurs has proven elusive. The arguments have tended toward the cataloging of individual items, with little emphasis on the underlying grammars.

Progress in understanding contact-induced grammatical change has come only with difficulty -which should not be surprising, from one point of view, for the understanding of grammatical (non-phonological) change in general has not come easily. On the other hand, it is strange that there has been so little progress -indeed so little attention paid- to understanding grammatical change under contact, for at least some very tight constraints were formulated, and further investigation invited, well over a century ago.

Several late nineteenth-century linguists were interested in the issue of language mixing and possible changes in languages which were in contact with others. Max Muller (cited in Whitney 1881) maintained that while mixture in the sense of lexical and phonological borrowing occurred, mixing (i. e., borrowing of basic grammatical rules) is impossible. Wilham Dwight Whitney (1881) refined Muller's notions a great deal, arguing that structural influence was possible, but only through the mechanism of a secondary process, whereby a precursor already existed in the borrowing language. The precursor itself could have already existed in the language, or it could have come from another language, through the borrowing of a sufficient number of forms which, at a time much later than the original borrowing, might, as now nativized forms, suggest an extension, or analogical process. The essential role of the precursor is a linguistic equivalent to preadaptation in biological evolution (Gould 1977). Whitney thus maintained Muller's general claim that basic word order change is prohibited.

Explicit testing of this and other hypotheses concerning the influence of one grammatical system on another has not been plentiful. It has begun to appear, however, and despite its paucity, is extraordinarily important.

I will begin with the one aspect of contact-induced change which has always seemed to be the best understood -lexical borrowing. Any situation of extensive contact produces significant and immediately noticeable borrowing. Traditional accounts of lexical borrowing have tended to emphasize either the reasons for borrowing (new words for new things, the necessity for euphemism, etc.), the cultural matrix of borrowing (the direction from subordinate to superordinate or the reverse, the fact that such pairs as sheep/mutton, deer/venison, cow/beef, pig/pork reflect social conditions and relationships in Norman England), or cultural attitudes toward borrowing (some linguo/cultural groups borrow readily while others do not).

What has received less attention has been linguistic aspects, particularly constraints on borrowing. Sobin 1976 has shown conclusively that cutting across all the socio-cultural aspects of borrowing are a set of (probably universal) grammatical constraints on borrowing. These constraints take the form of a hierarchical continuum which predicts the order of grammatical categories in borrowing. Nouns are the first category borrowed, followed by

––––––––   812   ––––––––

verbs, followed by adjectives. Two characteristics of this continuum are particularly important. One is that it is established primarily on the basis of data from Spanish in South Texas, an area in which lexical borrowing is rampant, and often considered to be unconstrained. Even so, however, there is only one attested example of a borrowed adjective, tofudo «tough». Without the search for grammatical conditions on borrowing, such a fact is unexplainable. The other is that the hierarchy is based on abstract linguistic features. Borrowed verbs, for example, do not form just one undifferentiated category. Those verbs which are borrowed first (and perhaps the only ones) are those which possess the feature [+do so], a feature which in English defines a class of verbs similar to but somewhat narrower than the class of [active] verbs³. What is interesting is that a very similar structure exists in Spanish as well and appears to be relevant in the analysis of Spanish loans in indigenous languages.

The second aspect of grammatical influence I will discuss is influence in lexico-syntactic structure. It has long been acknowledged that the lexical structure of one language may change under the influence of another without actually incorporating a borrowed word. Examples would be loan translations and shifts, such as asistir/atender used in their English rather than Standard Spanish senses («to assist» and «to attend» respectively), or calques, such as llamar pa'tras «to call back». But it has now been established that such influence may be much more subtle, and that it, too, can affect abstract grammatical features.

Silva-Corvalán 1986 has demonstrated that the Spanish of Spanish-English bilinguals in Los Angeles, CA is affected by English in proportion to their degree of bilingualism (which is to say, the relative amount of use of the two languages and their relative proficiency). The specific aspect of Spanish examined was the use of the verbs ser and estar, which, as every first semester student knows, are distributed according to an extensive set of rules. In Los Angeles Spanish Silva-Corvalán found a tendency for estar to encroach on the territory of ser, that is, to be used where the dialects of monolinguals would use ser, such as in

(1) Está inteligente «He is intelligent»

where the use of estar with a permanent characteristic is innovative.

In examining the innovation, a number of questions become important, questions which can be answered only through the application of newer methods of linguistic investigation. These questions concern 1) the process by which the extension of estar proceeds; 2) the meaning to be postulated for estar in order to account for its extension; and 3) the role of language contact on the actuation of the change. The methodology used to answer such questions is that developed within the general area of variation study. Instead of the citation of individual uses of estar, it requires the collection of a representative sample of naturalistic speech, the analysis of each occurrence of ser or estar, and the search for both categorical (absolute) and statistical patterns of occurrence, correlated with other linguistic and social factors.

The answers to these questions appear to be as follows. The extension of estar proceeds through a series of steps involving adjective types in the predicate (innovation is more likely with an adjective of size than with one of sensory character), or involving semantic transparency of the ser/estar choice (innovation is more likely in contexts where the standard language choice of ser/estar marks subtle differences in the relationship between the subject and the attribute than where the use of ser/estar marks a clear semantic difference in the interpretation of the adjective). Second, the extension of estar is associated with lower Spanish proficiency (though it is hardly prohibited by higher Spanish proficiency), and with earlier acquisition of English. Third, the extension of estar is encouraged, but not caused by language contact. The history of Spanish shows continual extension of estar into contexts previously allocated to ser. The contact of English and Spanish and the resulting widespread bilingualism have accelerated the already existing change. Thus, direct influence from English is not always easy to identify, even when it seems intuitively likely.

The questions raised by Silva-Corvalán's results both mirror and extend long-standing puzzles. Her results illuminate to a certain extent the process of change, and perhaps something which could be considered a type of mixing, but do not show a change in the basic grammar of Spanish. The fact that the possibility for the innovative use of estar already

––––––––   813   ––––––––

existed in Spanish provides an example of Whitney's claim.

A similar case is discussed in Klein 1980. The structure under examination is the use of the simple present and the progressive present in Spanish and English, which are similar in form and function, but only partly so in function. In both cases , the progressive present refers to an action taking place at the moment of speech. In Spanish, the simple present can also refer to such an action, but in English it cannot. This is illustrated in (2) and (3).

(2) Mira, está saliendo el sol

«Look, the sun is coming out»

(3) Mira, sale el sol

*«Look, the sun comes out»

Klein hypothesized that the partial similarity in the two grammatical systems would create an opportunity for influence, which would be reflected in the speech of bilingual speakers, and that the direction would be from English to Spanish rather than the reverse. After an investigation which used quantitative methodology much like that used by Silva-Corvalán, Klein determined that indeed, bilingual speakers did tend to reduce their use of the simple present for a moment-of-speech event and to extend their use of the progressive present.

One of the prime factors which Klein identifies as contributing to the changes is the fact that in this area of the grammar, the Spanish is more «ambiguous» than the English. That is, English reveals a more direct, one-to-one correspondence between form and function, while Spanish uses one form, the simple present, for more than one function, one of which is also expressed by another form, the progressive present. Spanish, therefore, presents an opportunity for a change which English does not present, and the fact that English offers a «solution» to this ambiguity suggests to speakers a logical path of change, which explains why the bilingual speakers' Spanish shows a change and not their English. That is, the change is predicted largely by linguistic factors, and not by such psycho-social factors as dominance.

The innovation in Spanish is not, of course, completely new, but involves extending an already existing tendency, which we would expect from Whitney's constraint. Both Silva-Corvalán's and Klein's investigations examine lexico-syntactic-semantic variables and appear to suggest a further prediction that contact-influenced changes are more likely to be in the direction of leveling of oppositions than creation of new oppositions, i. e., simplification. This general notion is not new, of course, but is now being investigated in terms of the two major influencing factors: the social matrix of the society in which the two languages are spoken in terms of dominance and role relationships, and the grammars of the two languages involved⁴. Because theory-driven investigation of change under contact is in its infancy, it is as yet unclear whether this further prediction is correct or not, but it is the type of prediction which now requires investigation.

Another similarity in Silva-Corvalán's and Klein's work is methodological. Both use quantitative methods of the sort pioneered by Labov. The reasons for use of such methods are several, but they include the fact that, as noted by Klein, contact (like other phenomena) may instigate changes that only quantitative methodology can detect. That is, incipient changes and changes in progress exist only in naturalistic speech and do not occur overnight. They begin in restricted linguistic environments and in small social groups, and then radiate outward in both linguistic and social space. They can be documented and analyzed only through the recording of naturalistic speech and the accountable reporting of every token of the structure under investigation. In order to analyze contact-influenced changes exhaustively without waiting an unacceptable time period for the changes to move toward completion, quantitative methodology is essential.

Neither of the two studies mentioned above deals with basic syntax, the area mentioned by Whitney as being immune to influence. Other situations provide data which may bear on the answer, particularly those in which Spanish is in extensive contact with a language which is typologically very different. Luján, Minaya, and Sankoff (1984) discuss the acquisition and use of three phrase orders in Peruvian Spanish by bilingual (Spanish-Quechua) children in terms of two issues. One is whether bilingual speakers tend to show the typical Quechua syntactic orders SOV, GN (genitive-nominative), AN (adjective-noun) as opposed to the more normal Spanish orders SOV, NG, NA, and the other is whether the possible paths of change from SOV, GN, and

––––––––   814   ––––––––

AN to SOV, NG, and NA confirm one or another of two general hypotheses concerning possible paths of syntactic change.

In this work the level of discussion is raised considerably in relating questions of language acquisition, linguistic variation, language universals, and the process of linguistic change. In acquisitional terms, what the Quechua-speaking child must do in learning Spanish is to change from OV to VO order in the verb phrase, and in the noun phrase to change from GN to NG and from AN to NA. The questions concern how these changes occur: 1) whether the basic verb phrase change occurs before the noun phrase changes, after them, or simultaneously; 2) whether the changes can be tracked and analyzed in normal speech; and 3) what aspects of linguistic theory might constrain and predict the changes.

To deal with the issues in a slightly different order, it was found that the variant orders could be tracked in quantitative terms across three age groups, 5, 7, and 9 year old speakers. While speakers in all three groups showed some incidence of variant orders in all three patterns, the variation showed a dear progression across the three age groups. The larger (verb) phrasal change must occur before the changes in the noun phrase, so that, for example, the speech of 5-year olds shows a rise in the incidence of SVO order while still maintaining predominantly GN and AN order in the noun phrase. By the age of 9, speakers have by and large changed to SVO, NG, and NA, though they still produce some verb final, genitive-noun, and adjective-noun orders even as adults. This state of affairs is found to be consistent with the Universal Consistency Hypothesis, which predicts that syntactic changes remain consistent with linguistic universals at every step, rather than with the Trigger Chain Theory, which holds that change begins with a violation of some linguistic universal, which then provokes further changes in order to eliminate the first stage incongruency.

While the results show the gradually diminishing (with age and, presumably, greater exposure to Spanish) influence of Quechua on Spanish, they do bear on the question of mixture, for, as Luján, Minaya, and Sankoff note, «features characterizing the various acquisition stages in children's speech are often preserved as variable styles in their adult speech, along with the standard norm assimilated in the acquisition process» (370). That is, adult speakers of Peruvian Spanish then have available to them patterns, such as OV, GN, and AN, to use among their other stylistic markers. This is noted in various works on Andean Spanish (such as Muysken 1981, for Ecuador, and also Luján, Sankoff, and Bordelois 1982 and Pozzi-Escot 1972 for Peru), and is the sort of feature attributed correctly, though anecdotally and imprecisely, to the «Quechua substratum».

If OV persists in Andean Spanish, it might be considered something of a counterexample to the Whitney hypothesis, insofar as basic word order syntax, as opposed to morphosyntax, is affected. On the other hand, one could easily argue that the precursor condition is easily met, since Spanish word order is flexible to the point of having variant OV order, as well as AN (GN is somewhat more problematic). The resulting question would be why OV apparently persists longer than AN in adult speech.

The point I would emphasize is not so much the specific claims of this work, but the more general point of view so well stated by LMS: «While universal principles [of change-JA] may receive relative empirical validity through the historical changes spanning several generations of monolingual speakers, the data afforded by speakers in areas of language contact provide ideal laboratory conditions for thoroughly validating them by means of straightforward quantitative analysis. Language contact studies may thus bear crucially on theories of language acquisition and change» (369).

The three investigations summarized above document contact influenced changes in Spanish. In two of them, Spanish is the socially subordinate language (to English). In the other, Spanish is the superordinate language (to Quechua). Clearly, influence can go in either direction. What is needed now is a greatly increased level of the same sort of investigation we've seen above, focused on specific hypotheses concerning language change as well as on particular sociolinguistic situations: Spanish as superordinate language, Spanish as subordinate; Spanish in contact with a relatively similar language (English), Spanish in contact with typologically different languages of varying sorts. The relevant situations all exist. It is simply a matter of identifying the appropriate sociolinguistic

––––––––   815   ––––––––

situation for the linguistic hypothesis to be investigated.

An additional interesting investigation of contact-induced grammatical change is Guy 1981. Speculation concerning a possible creole source for many characteristics of Caribbean Spanish has existed for many years. Guy's work, however, on Brasilian Portuguese has permitted interpretations that could not be accomplished before. Investigation of s-aspiration and deletion in Caribbean Spanish makes that process relatively well understood. s is more likely to be deleted when the following word begins with a consonant, when plurality is marked elsewhere in the noun phrase, when the s-bearing word is not the first in the noun phrase, when the following syllable is unstressed-in short, under a complex set of phonological, morphological, and semantic conditions, all of which are variable⁵. Now, Guy demonstrates convincingly that an astoundingly similar set of conditions exists in Brasilian Portuguese, despite the fact that the specific marker(s) are only partially similar -i. e., the specific linguistic process is not only s-deletion in the noun phrase, but also other indications of subject verb agreement. The similarity, therefore, is in a highly abstract grammatical rule.

As Guy notes, there are four possibilities for accounting for such linguistic similarities: 1) borrowing 2) parallel [but independent] historical developments; 3) common genetic origin; 4) linguistic universals; 5) common socio-historical phenomena which explain the common origin of the specific linguistic processes. In this case (though I will not take the space to detail the reasons), the first four can safely be discounted. Additionally, it is easy to document the operative socio-historical factors which would have created the two parallel processes -the slave trade and the resulting socio-economic situation in the colonies.

The parallel variable processes in Caribbean Spanish and Brasilian Portuguese exist in the context of a number of other common characteristics which also point to creole origins, so the argument does not rest solely on highly abstract variable rules. To my mind, however, the coincidence between such variable rules is far too great to be due to anything but a common source.

In drawing his conclusions, Guy attributes the similarity to contact with «creole», without entering the debate about the nature of creole -i. e., African substratum, universal grammar, fossilized acquisition error, etc. Guy has mentioned to me (p. c.), however, that at least one of the constraints in both Caribbean Spanish and Brasilian Portuguese is highly reminiscent of a common West African pattern. That is the morphological tendency to mark plural only on the first element of the noun phrase. Whether this turns out to be the case or not remains to be seen, but may provide an area for research in the future⁶.

Perhaps the paradigm case of linguistic influence is code-switching, especially the intra-sentential variety⁷. Examination of codeswitching has revealed a number of its properties. Both Pfaff 1979 and Poplaçk 1980 agree that intra-sentential switching obeys some version of the equivalence constraint, which essentially ensures that switched elements do not violate patterns of either language. Thus, (English-Spanish) switches between such major constituents as subject noun and verb are grammatical, as are switches between determiner and noun (el perro chewed him up (from Pfaff 1979), el sheetrock. Switches within the noun phrase are often not grammatical, however, because of the basic grammatical difference between the two languages -English is Adj N while Spanish is N Adj. The effect of this analysis is to suggest that code-switching is not produced by a separate grammar, but by the overlapping portions of the grammars of Spanish and English.

Joshi 1985 argues that the equivalence constraint alone is insufficient, however, but that one language or the other must be considered the matrix, with the other the embedded language. The matrix language is the language of the finite verb. According to him, such an approach captures more generally restrictions on switching bound forms and other closed class items than the simple equivalence constraint.

Blake 1987 notes an asymmetrical pattern of acceptability in switched complement clauses. He argues that INFL, clearly an example of a closed class item, must be taken into account if one is to explain why (4) and (6) are grammatical but (5) is not.

(4) I am going to decidir (from Timm 1975)

(5) *Quieren to come (Timm 1975)

(6) No van a bring it up at the meeting (Pfaff 1979)

According to Blake, what accounts for this pattern of acceptability is the Projection Principle

––––––––   816   ––––––––

(Chomsky 1986) and other aspects of abstract syntax.

One thing that the work to date on code-switching has made abundantly clear is that much more research is needed on switching in languages that are typologically more different than English and Spanish. Obviously, the equivalence constraint will seem to account for most switches when the two languages are typologically similar to begin with. Spanish is spoken in extensive contact with enough typologically diverse languages so that it should offer an ideal laboratory for the investigation of these questions. To my knowledge, however, that work remains to be done. (Of course, if intra-sentential switching tends not to occur between typologically different languages, that also corroborates the hypothesized constraint, in one form or another).

Much of the early work on the nature of bilingual communities focused on macro-sociolinguistic⁸ patterns, such as linguistic attitudes, factors which predispose the use of one or the other language (as Fishman has said, «who uses what language to whom and when?»), language shift and maintenance, diglossia, and the like. After the publication of Ferguson 1959, which is not so oriented (see Hudson-Edwards 1984 for a review of much of the recent thinking on diglossia), investigation of these issues has largely centered on survey and questionnaire approaches. The vast amount of such research has added much to our understanding of the specifics of many bilingual situations throughout the hispanic world as well as to our understanding of the macro-dynamics of bilingual communities. These include such matters as the past and future of a bilingual situation: Who is bilingual? Who is monolingual? Are more speakers becoming bilingual? Or fewer? Under what conditions do more speakers become bilingual? Under what conditions do more become monolingual? Is the community diglossic? What are the salient attitudes concerning the languages of the community? Is language shift evident, or language maintenance? What conditions favor each of these possible results?

Aside from investigation of such overt socio-cultural influences as lexical borrowing, and, to a certain extent, code-switching, detailed analysis of the ways in which the social situation, the use of language, and the structure of language mutually interact have only recently been treated in depth.

Much of the interesting work in this area in recent years has examined the notion of styles, or repertoires, and has attempted to relate the analysis of stylistic variation to a number of other factors -the nature of the speech community, acquisition, individual competence/proficiency, and socio-historical factors.

The attempt to disentangle various aspects of these phenomena and weave them back into a coherent picture go back a number of years. Sawyer (1970, 1973), in a comprehensive study of Spanish-English bilinguals in San Antonio, Texas, concluded that their English did not constitute a dialect of English, but rather an imperfect stage in the mastery of English. The reason was that they had not assimilated to the local (Texas) norms of the surrounding monolingual population. By «assimilating to local norms» Sawyer had in mind a very common-sensical interpretation: San Antonio Mexican-Americans simply did not for the most part sound like stereotypical Texans.

Ma and Herasimchuk (1971), working within the context of another large scale investigation of bilingualism, came to similar conclusions, though they used a much more sophisticated methodology. One of their aims was to apply the variationist methodology developed by Labov to a bilingual speech community. A significant portion of the variationist view concerns the nature of a speech community as a community which shares certain norms of communication and evaluation (Labov 1972). These norms may be very subtle, and may consist of orderly patterns of behavior which are revealed only statistically in what appears superficially to be disparate behavior by different sub-groups of the community.

Even using a framework which emphasized both heterogeneity and pattern, Ma and Herasimchuk concluded that the Puerto Rican Spanish/English bilinguals studied did not constitute a «unitary English speech community» (although they did constitute such a community in Spanish). The reason was, again, that they did not use the same norms as the local monolingual English community. For Ma and Herasimchuk, however, the concept of norm is different than for Sawyer. It involves phonological style-shifting. The Puerto Rican speakers marked style shifting within Spanish

––––––––   817   ––––––––

in the same way as did monolinguals by varying their use of certain phonological elements, most notably the variants of s, which in Puerto Rican Spanish are [s], [h], and [Ø]. The use of [s] marks a discourse as more formal, and the use of [h] and [Ø] mark it as progressively less formal. In English, however, the bilinguals did not use the stylistic markers of the local monolingual community appropriately. In most cases, they simply tended to use only one of the variants of a variable known to be a style marker in the monolingual English community⁹.

Later work has dealt with these issues in slightly different ways, attempting to arrive at a coherent view of the bilingual speech community which includes its relationship to the surrounding monolingual community[ies], the linguistic characteristics of individual speakers, and the nature of the linguistic varieties used. In Amastae 1981 the assumption that bilinguals would necessarily adopt the norms of the surrounding monolingual speakers is questioned. Perhaps more importantly, the fact that bilingual communities typically reveal a continuum of speakers, with incipient bilinguals at one end, very fluent native bilinguals at the other, and all intermediate positions also occupied, is noted. Finally, the role of phonological style shifting is examined. The solution to the related problems of the range of speakers and the matter of style shifting is suggested to be a continuum, along which individual speakers can be located in acquisitional terms. In addition, the continuum provides the framework for the selection and use of particular variants for stylistic marking.

The role of style shifting in bilingual situations has received further attention. Lavandera 1979 proposes that bilinguals often do not acquire the subtle, non-referential aspects of language use such as appropriate manipulation of style marking variants in their second language. Thus, in a manner almost completely parallel to that exhibited by the New York Puerto Ricans, Italian immigrants Argentina who are speakers of cocoliche, the inmigrant variety of Spanish characterized by interference from Italian and other learner features, often do not acquire the appropriate rules for the use of the variants of (s), for example (and similarly to the Puerto Rican norms): [s] for formal occasions, [h] and [Ø] for informal speech. Instead, they simply adopt one or the other variant, and use it categorically.

This is not to say that the cocoliche speakers are completely monostylistic speakers, however. As Lavandera notes, insofar as many bilingual speakers interact to a great extent with other bilinguals like themselves, they tend to switch languages for stylistic purposes, or to incorporate Italian variants into their Spanish. It is only when the situation requires them to use exclusively Spanish that they may be limited to a repertoire in which they cannot manipulate all of the socio-stylistic possibilities presented by the larger monolingual community.

Gal 1984 describes a similar situation. In a Hungarian-speaking village in eastern Austria the younger generation of Hungarian speakers does not control the non-referential, stylistic and expressive variants of Hungarian, even though Hungarian is for most the (temporally) first or «native» language. As Gal notes, «it is the political and economic developments of the postwar years that have deprived many young speakers of the social contexts in which the phonological variants of Hungarian are meaningfully distinguished» (299).

Since in the New York and Argentine situation the language which is stylistically reduced is the (temporally) second language and in the Hungarian case it is the (temporally) first language, it is clear that the operative factor is not order of acquisition, but social and role relationships among speakers and groups of speakers in the bilingual situations. Where bilinguals interact primarily with each other, their total range of repertoire includes both languages, and different varieties of each language fulfill different, but probably nonoverlapping, functions. When they must limit their language use to one language only, they may appear to monolinguals to be of limited proficiency because switching to the other language is not an option for fulfilling a particular stylistic function when dealing with monolinguals. The language that is spoken in such a reduced form is the language which the total social environment allocates to limited social uses and role relationships.

While the nature and function of style shifting in bilingual situations is now better understood, defining overlapping speech communities still appears problematic. A part of the problem may stem from an overly restrictive interpretation of Labov's definition of «speech community». Obviously, at one level

––––––––   818   ––––––––

any two speakers who speak the same language and can communicate at all form some sort of speech community. In order to frame a definition which can encompass communities of the sort Labov has described -extremely tightly-knit communities in modern, industrialized urban settings- as well as others such as those found in newly formed, non-urban, multi-lingual, perhaps ex-colonial settings, Le Page and Tabouret-Keller (1985) have introduced a basic distinction. A focused speech community has clear norms of linguistic production and evaluation, and is tied together by those norms. It is the type of community described in classical Labovian investigations, whether it is the Lower East Side in New York or Panama City (Cedergren 1973). A diffuse speech community, on the other hand, is one which does not have such clearly focused norms of production and evaluation. The reason is largely socio-historical, insofar as diffuse communities are typically composed of disparate groups rather recently formed into a single group. Diffuse communities are, in fact, in the process of forming their identities as communities, which means that they are creating their norms. Naturally, not all members of the society will adopt those norms at the same rate or in the same way. A proto-typical diffuse community is Belize, where (English) creole, English, Spanish, and indigenous languages are all used to varying degrees. Almost by definition, any recent bilingual community will be a diffuse speech community in some respect. It is precisely the process of norm formation and definition that is interesting, and is now being investigated.

While the question of community role relationships, language use, attitudes, and norms tends to be put in non-individual terms, the individual speaker cannot be ignored. As Teschner 1981 notes, psychometric (=quantitative) approaches have their limits, and the natural alternative is a psycho-historical approach. The goal of such an approach is to explain differences in individual behavior (including attitudes) that can be difficult to detect through survey and questionnaire methodologies. The approach involves the development of an extensive personal history of each speaker in order to isolate and explain differences in behavior which are obscured in the «normal» statistical variability of a group. Teschner discusses the cases of two Spanish-English bilinguals which would very likely be classed alike in a questionnaire investigation of sociolinguistic phenomena. While very similar in socio-economic and demographic characteristics, the two display very different linguistic attitudes and resulting patterns of language choice in dealing with other bilinguals. The explanation of the differences is found in their individual family backgrounds and upbringing.

The linguistic variables Teschner deals with are attitudes and language choice, not actual language elements. However, the situation is reminiscent of the debate over the match of individual language with group patterns, within the framework of variability. The logical extension of the approach is a detailed examination of the use of linguistic forms, seen against the backdrop of the individual speaker's psycho-history. Such an approach might well be the only method of solving one critical problem in contact studies. That is the tracking of a nascent variety. One example of this is Chicano English, which conceptually presents little problem. Chicano English is simply a variety, analogous to Indian (East) English, which has grown out of the contact between English and Spanish and the resulting bilingualism of many speakers. Because many of its speakers are now monolingual English speakers, Chicano English cannot be identified completely with interference or other learner phenomena, though it clearly shows instances of them. The problem has been to sort out speakers of one type from speakers of the other, since the linguistic phenomena may overlap (see Amastae 1981 for another approach to this problem). A psycho-historical approach to individuals may offer the key to this difficulty¹⁰.

One of the more newly developed areas of linguistics is in ethnographic analysis of speech events and discourse analysis. A number of investigators have made significant progress in the analysis of a variety of speech events. While these investigations frequently focus on one linguo-cultural group, ethnographic research is inherently, if only implicitly, comparative. Thus, there is opportunity for not only the investigation of events in one language in contact with another but also of specifically bilingual events. Sánchez 1983 offers an overview of all of the discourse uses in the bilingual Chicano community (see also

––––––––   819   ––––––––

Duran 1981). Limón 1983 treats in some detail the bilingual, including inter- and intro-sentential code-switching, communicative strategies used in a meeting of a student organization, showing that those participants who most fluently manipulate the bilingual resources have very clearly delineated communicative strategies and structures. Valdés 1983 deals with the discourse reasons for intro-sentential switches. Hart-González 1979 examines an artistic event in terms of language.

An important extension of the focus on bilingual discourse involves an investigation which is currently underway by Elías-Olivares (p. c.). She is applying ethnographic methods to the investigation of (discourse) argument structure in a Spanish-English bilingual community. While it is too early to predict the results, it can nevertheless safely be predicted that the investigation of discourse structures in bilingual communities is just coming into its own and will constitute an important area of research in the future.

In closing, I want to reiterate a point made earlier. During the last five centuries Spanish and Portuguese have been, and continue to be, in contact with a wide variety of languages, with a wide range of social factors. They provide enormous possibilities for the empirical investigation of language contact and bilingualism, and perhaps ultimately for the development of a true theory of language contact. My survey has been anything but complete, and perhaps suffers from idiosyncrasy. Many, many excellent works have been omitted, not because of disregard, but simply because of space limitations. What I have tried to do has been to identify what I see as the significant newer trends in research on LusoHispanic contact and bilingualism, and to illustrate those trends with some of the relevant literature. Of necessity, omissions have had to be made. I hope, however, that I have been successful enough so that interested readers will appreciate the substance of recent work, and will be able to follow the bibliography I have given to more extensive listings.

Amastae, Jon. 1981. «Learner Continuums and Speech Communities». Papers in Linguistics 14.2: 155-96.

——. and Lucía Elías-Olivares. 1982. Spanish in the U. S.: Sociolinguistic Aspects. Cambridge: Cambridge University Press.

Blake, Robert. 1987. «Is to/a the Head of INFL?: I Don t Want to decidir la cuestión, but I'm Going To». In Variation in Language: NWAV-XV at Stanford, ed. by K. Denning et al, 22-34. Stanford: Stanford University.

Cassano, Paul V. 1982. «Language Influence Theory Exemplified by Quechua and Maya». Word 33. 1-2; 127-41 (Chang-Rodríguez 1982).

Cedergren, Henrietta. 1973. «The Interplay of Social and Linguistic Factors in Panama». Unpublished dissertation, Cornell University.

Chang-Rodríguez, Eugenio. 1982. «Spanish in the Western Hemisphere». Word 33: 1-2.

Chomsky, Noam. 1986. Knowledge of Language: Its Nature, Origins, and Use. New York: Praeger,

Clyne, Michael G. 1984. «the Decade Past, the Decade to Come: Some Thoughts on Language Contact Research». International Journal of the Sociology of Language (IJSL) 45: 9-20.

Dorian, Nancy. 1973. «Grammatical Change in a Dying Dialect». Language 49: 414-38.

——. 1978. «The Fate of Morphological Complexity in Language Death». Language 54: 590- 609.

——. 1981. Language Death. Philadelphia:

––––––––   820   ––––––––

University of Pennsylvania Press.

Dressier, Wolfgang. 1972. «On the Phonology of Language Death». Proceedings of the Chicago Linguistic Society 8: 448-5Z

Durán, Richard P. 1981. Latino Language and Communicative Behavior. Norwood, NJ: Ablex.

Ferguson, Charles. 1959. «Diglossia» Word 15: 325-40.

Fishman, Joshua, and R. Cooper, 1968. Bilingualism in the Barrio. Bloomington: Indiana University Press.

Haugen, Einar, 1956. Bilingualism in the Americas. University, Alabama: University of Alabama Press. (Publication of the American Dialect Society, No. 26).

Gal, Susan. 1984. «Phonological Style in Bilingualism: the Interaction of Structure and Use». In Meaning, Form, and Use in Context, ed. by Deborah Schriffrin, 290-302 (GURT 1984). Washington, D. C.: Georgetown University Press.

Gilmer, Paul. 1986. «Judeo-Spanish in Turkey». Unpublished Ph. D. dissertation. University of Texas at Austin.

Gould, Stephen Jay. 1977. Ever Since Darwin. New York: Norton.

Guy, Gregory R. 1981. «Parallel Variation in American Dialects of Spanish and Portuguese». In Variation Omnibus, ed. by D. Sankoff and H. Cedergren, 85-96. Edmonton: Linguistic Research.

Hart-González, Lucinda. 1979. «The huayñito: a Multi-lingual, Multiple Channel Event». In Speaking Singing, and Teaching: a Multi-disciplinary Approach to Language Variation, 109-17, ed. by F. Barkin, et al. Tempe: Arizona State University. (ASU Anthropological Research papers #20).

Hudson-Edwards, Alan. 1984. «Rediscovering Diglossia» Southwest Journal of Linguistics 7. 1: 5-15.

Joshi, Arvind. 1985. «Processing of Sentences with Intrasentential Code-switching». Natural Language Parsing: Psychological, Computational, and Theoretical Perspectives, ed. by David Dowty, Lauri Karttunen, and Arnold Zwicky, 190-205. Cambridge: Cambridge University Press.

Klein, Flora. 1980. «A Quantitative Study of Syntactic and Pragmatic Indications of Change in the Spanish of Bilinguals in the U. S.». In Locating Language in Time and Space, ed. by William Labov, 69-82. New York: Academic Press.

Labov, William. 1972. Sociolinguistic Patterns. Philadelphia: PA: University of Pennsylvania Press.

Lavandera, Beatriz. 1979. «The Variable Component in Bilingual Performance». In International Dimensions of Bilingual Education, ed. by James E. Alatis, 391-409. Washington, D. C.: Georgetown University Press.

Le Page, Robert and Andree Tabouret-Keller. 1985. Acts of Identity: Creole-based Approaches to Language and Identity. Cambridge: Cambridge University Press.

Limón, José. 1982. «El Meeting». In Spanish in the U. S.: Sociolinguistic Aspects, ed. by J. Amastae and L. Elías-Olivares, 301-32. Cambridge: Cambridge University Press.

Luján, Marta, D. Sankoff, and I. Bordelois. 1982. «El principio de consistencia universal en el español andino del Ecuador». El español de America (Actas del I Congreso Internacional), ed. by H. López Morales and M. Vaquero. Río Piedras: Universidad de Puerto Rico.

Luján, Marta, L. Minaya, and D. Sankoff. 1984. «The Universal Consistency Hypothesis and the Prediction of Word Order Acquisition Stages in the Speech of Bilingual Children». Language 60.2: 343-71.

Ma, Roxanna and E. Herasimchuk. 1971. «Linguistic Dimensions of a Bilingual Neighborhood». In Bilingualism in the Barrio, ed. by J. Fishman, et al, 347-464. Bloomington , IN: University of Indiana Press.

Muysken, Pieter. 1981. «The Spanish that Quechua Speakers Learn: L2 Learning as Norm-governed Behavior». To appear in the Proceedings of the First North American-European Workshop on Second Language Learning, ed. by Roger W. Andersen.

Pfaff, Carol. 1979. «Constraints on Language Mixing: Intrasentential Code-switching and Borrowing in Spanish/English». Language 55.2: 291-318. (Also in Amastae and Elías-Olivares 1982).

Poplack, Shana. 1980. «Sometimes I'll Start a Sentence in Spanish y termino en español: Toward a Typology of Code-switching». Linguistics 18.7-8: 581-618. (Also in Amastae and Elías-Olivares 1982).

Pozzi-Escot, Inés. 1972. «El castellano en el Perú: Norma culta versus norma regional». El reto del multilingüismo en el Perú, ed. by Alberto Escobar, 123-42. Instituto de Estudios Peruanos.

Rubin, Joan. 1968. National Bilingualism in Paraguay. The Hague: Mouton.

Sawyer, Janet. 1970. «Spanish-English Bilingualism in San Antonio, Texas». In Texas Studies in Bilingualism, ed. by G. G. Gilbert, 18-41. Berlin: Walter de Gruyter.

——. 1973. «Social Aspects of Bilingualism in San Antonio, Texas». In Varieties of Present Day English, ed. by R. Bailey and J. Robinson, 226-32. New York: Macmillan.

Sánchez, Rosaura. 1983. Chicano Discourse. Rowley, MA: Newbury House.

Silva-Corvalán, Carmen. 1986. «Bilingualism and Language Change». Language 62.3: 587-608.

Sobin, Nicholas. 1976. «Texas Spanish and Lexical Borrowing». Papers in Linguistics 9. 1-2: 15-47. (Also in Amastae and Elías-Olivares 1982).

Teschner, Richard V. 1981. «Historical-Psychological Investigations as Complements to Sociolinguistic Studies in Relational Bilingualism: Two Mexican-American Cases». The Bilingual Review/La revista bilingüe. 8.1: 42-55.

Timm, Leonora A. 1975. «Spanish-English Code-switching: El por qué and how-not-to». Romance Philology 28: 473-81.

Valdés, Guadalupe. 1983. «Social Interaction and Codeswitching Patterns: A Case Study of Spanish/English Alternation». In Bilingualism in the Bicentennial and Beyond, ed. by G. Keller, R. Teschner, and S. Viera, 53-85. New York: Bilingual Review Press. (Also in Amastae and Elías-Olivares 1982).

Whitney, William Dwight. 1981. «On Mixture in Language». Transactions of the American Philological Association. 5-26.

Hispania [Publicaciones periódicas]. Volume 72, Number 4, December 1989