Language shift and maintenance can ultimately be seen as “the long-term collective results of language choice” (Fasold 1984:213, cf. Fishman 1965:73, Gal 1979:13). Accordingly, analyzing underlying patterns in individual language choice is essential for assessing language vitality.
In a bilingual community, although communication in the two languages would be possible, language choice is by no means met arbitrarily, but rather it is a social behavior subject to norms. Sociolinguistic studies of language choice have been significantly shaped by the notion of diglossia (see chapter 2.1.2.1), which demonstrated that language choice reflects social norms concerning distinct functions of the two languages in contact (Li 2007:27). The study of language choice drawing on this insight is guided by the famous question posed by Fishman: “Who speaks what language to whom and when?” (1965). In order to approach the
question, Fishman (1965) suggests the concept of “domains of language choice”. For Fishman (1965:75), a domain is “a socio-cultural construct abstracted from topics of communication, relationships between communicators, and locales of communication, in accord with the institutions of a society and the spheres of activity of a culture”. It is considered “a relatively uniform but flexible analytic scheme” (Fishman 1965:86) because relevant domains of language choice are likely to differ between settings. Therefore, it is a task for researchers to designate and define domains of language choice based on their insight into the particular multilingual settings (Fishman 1965:74, 1972:451).
Greenfield’s study on language choice among young Puerto Rican bilinguals (Spanish and English) in New York City (1972) illustrates the concept of domain in concrete research practice. Assuming that “domains are a higher order generalization derived from congruent
situations”, Greenfield defined five relevant domains for language choice in the community,
based on preceding participant observation, focused interviews and discussion: family, friendship, religion, education and employment (1972:21).44 It was hypothesized that in the Puerto Rican community in New York, Spanish – associated with intimacy and solidarity – would be used primarily in domains such as family and friendship, whereas the use of English – associated with status differentiation – would be mainly reserved for domains such as religion, education and employment (1972:20). Based on the results of two experiments45, Greenfield concludes that the language choice of Puerto Rican young bilinguals in New York differs according to the domain of interaction: with Spanish preferred in the domain of family,
44
In order to collect self-report data on normative language choice, he selected a typical situation of each domain comprising three seemingly-congruent components, namely situational interlocutor, situational place and
situational topic (Greenfield 1972).
45 Greenfield’s study (1972) comprised two experiments. The first experiment was mainly conducted to validate
his domain construct. In the first experiment, Greenfield presented two of the three components (interlocutor, place and topic) and requested test persons to complete the situation by selecting a suitable third component and indicate the amount of Spanish and English that they would use in this situation. The second experiment was conducted to observe the independent effect of each of the components, interlocutor, place and topic on language preference. In the second experiment, three components (interlocutor, place and topic) were provided to the test persons and they were requested to indicate how much of English or Spanish they would use in a given situation comprising the three components.
followed by the domains of friendship and religion and least in the domains of education and employment (1972:33). In view of the association of Spanish with intimacy and English with status, Greenfield sees bilingualism in the community confirming the model of normative diglossia (1972:27).
Fishman (1965:80, 86) sees strong potential in domain analysis as a conceptual and methodological tool enabling valid generalizations concerning different kinds of bilingualism as well as the language maintenance or shift consequences of particular configurations. Moreover, the concept of domain reveals interrelationships between individual language choice and widespread socio-cultural norms and expectations, successfully linking micro-sociolinguistics with macro-sociolinguistics (Fishman 1965:73, 1972:449, 453).
Domain analysis has also been applied to investigate patterns of language choice between Yucatec Maya and Spanish – for example – by Pfeiler (1988, 2012) and Otto (2009).
Following the concept of domain by Fishman, Pfeiler (1988, 2012) investigated language use in two communities in the state of Yucatan, Cantamayec and Chabihau in the 1980s. The two communities differ in main socioeconomic activities. Cantamayec in the maize cultivation zone is characterized by the traditional milpa agriculture, whereas Chabihau lies in the coastal region, where fishing represents the main socioeconomic activity. Based upon her analysis, Pfeiler (2012) identified different patterns of language use in the two communities. In Cantamayec, Yucatec Maya was predominantly used in the private and semi-public domains, whereas the use of Spanish was reserved for public domains. Accordingly, she classified bilingualism in Cantamayec as stable diglossia, albeit not excluding the possibility of change. By contrast, in Chabihau, the use of Yucatec Maya was often restricted to one’s own family at home. Pfeiler attributes the diminished role of Yucatec Maya in Chabihau to its fishing economy, which has attracted immigrants from other communities and encouraged communication with Spanish-speaking traders (2012:112f.)
investigated language choice in the municipality of Felipe Carrillo Puerto, Quintana Roo. By means of interviews with bilinguals in Yucatec Maya and Spanish, he collected self-report data on normative language choice in 34 hypothetical situations comprising different interlocutors, places and topics, which were attributed to seven domains46. The results obtained from the domain analysis demonstrate that Yucatec Maya is mostly used in the family domain, while the use of Spanish predominates in all other domains (Otto 2009:170)47. Moreover, in the family domain the use of Yucatec Maya is dependent on the age of participants, with the younger generation preferring the use of Spanish except for communication with the older generation (Otto 2009:61f.). As an intergenerational language transmission is an important prerequisite for language maintenance, he sees this age-dependent difference in language choice as an indicator for the ongoing language shift in Felipe Carrillo Puerto (Otto 2009:170).
The domain analysis has shown that language choice is usually met in accordance with norms concerning the association of languages with different social institutions. Language choice has also been studied within social psychology (Fasold 1984:183). While domain analysis emphasizes the role of social institutions on language choice, the social psychological approach focuses on the psychological processes of individual language choice in interactions (Fasold 1984:187). Giles’s theory of speech accommodation approaches change in people’s speech style (not necessarily language choice)48 based on the postulate that “people are motivated to adjust their speech styles, or accommodate, as a means of expressing values, attitudes and intentions towards others” (Giles, Bourhis and Taylor 1977:321f.). According to Giles, individuals communicate social approval by means of shifting their speech style towards that of
46 These domains were: family, friends and acquaintances, work and colleagues, authority, strangers, religious
activities and other daily situations.
47 However, there are some situations outside the family domain in which the use of Maya prevails. Otto mentions
the following situations: communication with the healer (jmeen), talking with the taxi driver (combi), forcing out their pet, communication in the traditional church and speaking with a barefooted stranger or a lady wearing traditional clothing (íipil) (Otto 2009:170).
48 The accommodation of speech styles includes convergence in “languages, accents, speech rates, pause and
the other (convergence)49. By contrast, shifting away from the other’s style of speech (divergence) is a sign of dissociation from the other, serving as a symbolic strategy for maintaining distinctiveness (Giles, Bourhis and Taylor 1977:321-324).
The approaches to language choice presented thus far have indicated that language choice in a particular situation is made in accordance with the ideas that speakers have about the languages in contact. Accordingly, the following part is concerned with the language attitudes underlying language choice.