Thresholds leading to shift: Spanish/English codeswitching and convergence in Georgia, U.S.A.

Abstract

In the speech of members of a Georgia (U.S.A.) community of immigrants from Latin America, as their bilingual sentences of Spanish/English reach a certain threshold of English content, Spanish morphosyntactic patterns begin to converge toward those of English. Data from naturally-occurring conversations by 56 children and adults of both sexes are analyzed within Myers-Scotton's (1993 [1997], 2002) Matrix Language Frame model. Eight language types were identified, including monolingual Spanish and English turns, codeswitched turns, and turns showing convergence (morpheme strings from one language with some grammatical structure from the other). An instance of each language type per turn was counted as a token of that language type. Each sentence of a multisentence turn was counted as a separate token. Tokens of each type were counted per informant. A rank ordering of the data by percentage of monolingual Spanish allows observation of how certain thresholds signal changes in the types of language mixing. Analysis reveals that percentages of monolingual and codeswitched utterances pattern in relation to percentages of utterances showing convergence, indicating that informants' Spanish does not begin to converge toward English until fewer than 70% of their utterances are monolingual Spanish. The data thus show that both codeswitching and convergence are mechanisms of language shift and change from dominance in one language to another. Social factors are also shown to be associated with the linguistic patterns.

Key words

codeswitching convergence language contact

I Introduction

Bilingual language contact over time may lead to language change. Linguistically, this change involves the contact of different lexical and grammatical systems. This study focuses primarily on lexical and grammatical patterns in language change with additional reference to related social factors including gender, age, place of origin, time spent in contact with a second language, time spent in school, type of employment, preferred language, and which language is used where, when and with whom. The study shows three steps or stages of language change in a language contact community: (Stage 1) mostly monolingual sentences in the community's first or original dominant language with slight lexical and even less grammatical influence from the community's new or second language; (Stage 2) increased lexical influence from the new language, and the two languages affect each other grammatically; (Stage 3) the new language has become the dominant language but with continuing grammatical and lexical influence from the first language. Between stages one and two and between stages two and three there are apparent thresholds of change: (between stages 1 and 2) at around 70% monolingual first language utterances; and (between stages 2 and 3) at around 40% monolingual first language utterances.

Unlike most other language contact studies in the literature that either examine well established contact situations (e.g., Thomason & Kaufman, 1988) or languages that have been in contact at least several decades (e.g., Silva-Corvalan's (1994) Los Angeles study), this study treats a very new language contact situation in a Hispanic speech community in Northeast Georgia (U.S.A.). Established in the late 1980s in Habersham County, Georgia, and parts of two bordering counties (Banks and Stephens), this community has flourished and is a good site to observe early language contact phenomena between English and Spanish.

Codeswitching (CS), the alternation between two different languages, (1) is a frequent phenomenon of language contact. Convergence (2) is the use of grammatical patterns from more than one language in the same sentence, a possible result from language contact. Other possible outcomes of contact include language maintenance or language shift. The current study examines the frequency of CS, convergence, and monolingual speech with reference to the three stages of language change and language maintenance and shift. Shift here does not necessarily mean that a speaker or speakers have shifted from the community's first to the second language so completely that they never utter anything in any other language in any context, though it could mean that. It is more realistic to assume that speakers who have shifted speak with most of their interlocutors, especially peers, and in most contexts, almost entirely in what was at least at one time the community's second language.

This investigation examines CS and convergence across a community including both genders and both children and adults. Analysis of informant speech samples reveals that percentages of monolingual and codeswitched utterances pattern in relation to percentages of utterances showing convergence. Spanish begins to converge toward English grammatical patterns when fewer than 70% of informants' utterances are monolingual Spanish. When fewer than 40% of their utterances are monolingual Spanish, monolingual English is more frequently used than monolingual Spanish, and many of their other English morpheme utterances show convergence with Spanish grammatical patterns. These data lead to predictions about what new migrants might do in their language patterns. Whether individual speakers will converge to English (in their Spanish) and or shift to English can be predicted on the basis of the patterns of bilingual usage across a range of informants, with patterns ranging from complete monolingualism in Spanish to complete monolingualism in English and varying degrees and types of bilingualism in between.

The subjects studied are all members of the Hispanic community of the previously mentioned Georgia counties. They are a coherent group to this extent: they are all from Latin America, mostly Mexico; local schools, churches, and other social organizations treat them for the most part as a unit and either refer to them as Hispanics or Mexicans; interactions among members of this community are generally in Spanish, and for children whose main language is already English, older members of their families still interact with other members of the community in Spanish, even if their home language is mainly Q'anjob'al Maya; (3) shared physical, ethnic, cultural, and linguistic characteristics set all members of the Hispanic community apart from any other non-Hispanic groups in the area.

2 Codeswitching and convergence

This study focuses on the morphology and syntax of CS, specifically the points at which switches are made within and between sentences, and the morphology and syntax of convergence. From Myers-Scotton's (1993 [1997], 2002) Matrix Language Frame (MLF) model of CS it is assumed that every clause has a matrix language. The MLF model is based on two major hierarchies, (1) the Matrix Language (ML) versus Embedded Language (EL) hierarchy and (2) the system versus content morpheme hierarchy (Myers-Scotton, 1993 [1997], 2002). The first hierarchy assumes that every utterance has a ML grammatical frame into which morphemes may be inserted or embedded, whether they derive from one or more than one language. The ML frame is from one language or is a composite of more than one language. The ML frame dictates word order. There is therefore an inherent asymmetry between the two languages or the ML and the EL, the ML taking precedence.

The notion of Matrix Language and asymmetry are also important attributes of a bilingual community's use of its two languages. Spanish is the ML of the majority of the speech samples of the majority of the speakers in the community. English is the ML, however, of the majority of the speech samples of some of the speakers, and in time, depending on social factors, this minority of English ML speakers may indeed grow, causing a shift from a community-wide Spanish ML to a community-wide English ML. Thus the asymmetry between the ML and the EL in a single utterance is replicated in the speech of an entire community in which the community ML is at first Spanish and the community EL is English, and later the former EL, English, becomes the new community ML.

This asymmetry between the two languages is further illustrated by the second hierarchy which assumes that all morphemes are either content or system morphemes. Content morphemes assign and receive thematic roles while system morphemes do not. Content morphemes usually include nouns, verb stems, and descriptive adjectives, while system morphemes usually include inflections and determiners. Single morphemes from the EL embedded into the ML will be content morphemes and some types of system morphemes. Under the 4-M model, discussed in Myers-Scotton (2002), the system morphemes that must come from the ML only (under the System Morpheme Principle introduced in Myers-Scotton (1993)) are called "'outsider' late system morphemes" (2002, p. 75). (4) Distinguishing the different types of system morphemes and analysis of composite ML frames is beyond the scope of this paper. Myers-Scotton (2002, pp. 53-107) is an excellent reference for these analyses. Example (1) illustrates the MLF model. Note that the ML is Spanish and that content morphemes, not system morphemes, from English are inserted into the ML Spanish frame.

(1) An utterance of a 13-year-old Mexican male (Informant 33, cf. Table 1)

Maestro, y a onde vamos air al SWIMMIN' onde onde ... ?

maestro y a onde vamos a ir

teacher and to where we go[1PL/PRES] (5) O to go

al swimmin' onde onde

to the swimming where where

'Teacher, and where are we going to go swimming where where ...?'

A structural change, including word order or borrowing or deletion of some system morphemes, may indicate a composite ML frame and structural convergence of one language toward the other, and/or ML turnover (Myers-Scotton, 2002) as a last step. ML turnover is the replacement of one language as the ML with the other in a bilingual context. This may be in the speech of one individual or of an entire community. Convergence is the process resulting in a composite ML. A composite ML is the structure of an utterance which takes abstract grammatical structure, with surface consequences, from more than one language. Convergence is defined as the use of grammatical structure from at least two languages in the same utterance even in cases in which all the morphemes are from only one of the languages (Myers-Scotton, 2002). Convergence has therefore occurred if word order is altered and/or system morphemes are deleted in a sentence of morphemes from one of the languages by analogy with the other language. What is sometimes referred to in the literature as transfer or interference (6) can be explained in terms of structural convergence, as shown in examples (2-5) from our data.

Example (2) illustrates all Spanish morphemes with some abstract grammatical structure from English. Spanish distinguishes gane (simple past 'won') from ganado (past participle in the construction he ganado 'I have won'). English uses won for both the simple past and the past participle. The speaker in example (2) used the simple past, incorrectly for standard Spanish, instead of the past participle, most likely because there is no distinction between the two verb forms in English.

(2) An utterance of an eight year old Peruvian male (Informant 39, cf. Table 1)

Yo he gang.

yo h-e gan-e

I[1S/PAST] have-1S/PRES win-1S/PAST

'I have won.'

(standard native Spanish: Yo he ganado)

Example (3) is another example of all Spanish morphemes with abstract grammatical structure from English. The speaker in example (3) used yo 'I' with me gusta, literally 'to me it is pleasing' in which the verb gusta assigns the experiencer the dative case me in Spanish. Because the English equivalent is I like in which the verb like assigns the experiencer nominative case, I, the speaker has apparently added the Spanish nominative yo to the sentence.

(3) An utterance from a nine year old Peruvian male (Informant 48, cf. Table 1)

YO ME GUSTA ESA CASA MAMI ...

yo me gust-a esa casa mami

I to me[DAT] please-3S/PRES that house mommie

'I like that house, mommie.'

(standard native Spanish: Me gusta esa casa mami)

Examples (4) and (5) illustrate all English morphemes with abstract grammatical structure from Spanish. Spanish descriptive adjectives, certainly color words, follow the nouns they modify. In English, descriptive adjectives precede the nouns they modify. The speaker of example (4) has opted for the Spanish word order, while using all English morphemes.

(4) An utterance from an 11-year-old Mexican female (Informant 51, cr. Table 1)

SEE, I PUT SOME DOTS RED.

(standard native English: See, I put some red dots)

(Spanish: circulos rojos, literally dots red 'red dots')

In example (5) the speaker also uses Spanish word order and even changes the surface morpheme from not to no because this more closely matches Spanish word order and morphology, hence, I no can see rather than the standard English I cannot see or I can't see.

(5) An utterance from a 10-year-old Honduran female (Informant 26, cf. Table 1)

I NO CAN SEE.

(standard native English: I can't see)

(Spanish: no puedo, literally no can I 'I can't')

The two types of convergence (examples (2-3), Spanish morphemes with some English grammatical structure; examples (4-5), English morphemes with some Spanish grammatical structure), are essentially opposites of each other. A Spanish morpheme string with a composite ML from Spanish and English is Spanish converging toward English. An English morpheme string with a composite ML from Spanish and English is English converging toward Spanish. Convergence does not imply categorically that one language will eventually turn completely into the other but only that one language is influenced grammatically by the other.

3 Simplification or not?

One could argue that yo he gane (example (2)) is not convergence of Spanish and English grammar, but merely an example of child language. I contend, however, that this example illustrates Spanish/English grammatical convergence because of the following. The informant, age eight, who said the utterance, also said several similar utterances; his older brother and two other boys (not of the same national origin or family and not in the data corpus proper), 9-10 years old, also said similar utterances. All four of these children were more relatively isolated from the Hispanic community at large than any of the other informants in our study. All four exhibited very similar language patterns, showing disregard for morphosyntactic distinctions between English and Spanish (e.g., English won = Spanish gane / ganado). These patterns were very different from those of all other children informants, ages eight and younger, who demonstrated utterances that conformed more closely to standard Spanish morphosyntactic patterns.

The grammatical convergence shown in examples (2-5) cannot be brushed aside as instances of simplification or of child language. All the children who uttered examples (2-5) are old enough (ages 8-11) to be past the stages at which …