Unitats de vegetació potencial

The view of code-switching as an interactional strategy, introduced by Gumperz (2.4.2), was adapted by Myers-Scotton in her Markedness Model (Shin and Milroy, 2000: 352, Myers-Scotton, 1998). In addition to drawing on Gumperz’s description of code-switching as a contextualisation cue, Myers-Scotton’s Markedness Model (1999: 1260) is heavily based on Rational Choice Models.

Elster (1989: 22) summarises the essence of Rational Choice or Rational Actor Models through the following claim: ‘When faced with several courses of action, people usually do what they believe is likely to have the best overall outcome’.

The theory underlying Rational Choice or Rational Actor models originates from economics and sociology. Myers-Scotton’s Markedness Model is particularly influenced by the work of the philosopher Jon Elster (1989) who claims that people’s actions are filtered by two separate processes before they occur (Myers-Scotton, 1999: 1260). During the first filter the speaker’s opportunity set is

formed. The second filter constitutes the moment in time where the speaker consciously chooses between different options. Both filters and related terms are explained in detail below.

The first filter contains so-called structural constraints which, for example, consist of social factors such as the participants’ social identity features (e.g. age, sex, socioeconomic status, ethnicity) or the characteristics of the discourse situation (e.g. topic, setting) (Myers-Scotton, 1998). According to Myers-Scotton (1999:

1260) ‘surface discourse structural features’ are a further type of structural constraints. This type of constraint encompasses ‘structural features that organize discourse, particularly those having to do with sequential organization’ (Myers-Scotton, 1998: 34). For example, whether a particular utterance represents the first or the second part of an adjacency pair can have an impact on its content or form (Myers-Scotton, 1998: 34). Many of these discourse structural constraints originate from the work carried out in conversation analysis (CA). The different types of constraints which form the basis of this first filter are ‘external’, as the speaker does not have any direct control over them (Myers-Scotton, 1998: 34).

These social and discourse structural factors influence the speaker’s ‘opportunity set’ which is defined as the speaker’s linguistic repertoire. This repertoire is composed of the various languages, dialects and styles the speaker is able to use.

Discourse strategies such as turn-taking, minimal responses and overlaps are also part of the speaker’s linguistic repertoire or opportunity set (Myers-Scotton, 2002:

207). Once the structural constraints of the first filter have operated and the speaker’s opportunity set has been formed, the speaker finally gains control over the interaction and is able to make conscious choices about which codes to use. In other words, the first filter establishes a set of possible choices from which the speaker is able to choose during the second filter.

The second filter proposed in Elster’s Rational Choice or Rational Actor models includes the mechanisms which allow a speaker to consciously choose a specific

during the first filter) (Myers-Scotton, 1998: 34). Rationality is the crucial factor in this second filter. At this stage the speaker consciously engages in a cost-benefit analysis and opts for the choice which offers the best overall outcome in terms of interpersonal relationships. Myers-Scotton (2001: 5) applies this theory to the analysis of code-switching in conversation and claims that ‘speakers are rational in the sense that their choices depend largely on assessments of possible options in terms of a cost-benefit analysis that takes account of their own subjective motivations and their objective opportunities’. This statement highlights the complex interplay of people’s prior attitudes, beliefs and values on the one hand and their temporary aims and desires in a given interaction on the other hand (Myers-Scotton, 1999: 1261). In fact, speakers contemplate their possibilities and rationally opt for the choice which allows them to fulfil their temporary desires without compromising their prior beliefs. This concept will be illustrated with an example below.

The Markedness Model (MM) is based on the assumption that both speakers and analysts are able to distinguish between marked and unmarked codes or choices.

Myers-Scotton (1998: 22) explains that all speakers have a ‘markedness evaluator’ which consists of a cognitive capacity to evaluate markedness. In order to be able to conceptualise markedness speakers need to develop two abilities:

(1) the ability to recognize that linguistic choices fall along a multidimensional continuum from more unmarked to more marked and that their ordering will vary, depending on the specific discourse type;

(2) the ability to comprehend that marked choices will receive different receptions from unmarked choices (Myers-Scotton, 1998: 22).

Speakers acquire these abilities through contact with both marked and unmarked codes (Myers-Scotton, 1998: 22). Consequently, speakers need to be exposed to the use of marked and unmarked codes in community discourse in order to learn which codes are expected to be used under which circumstances. The markedness

evaluator can be regarded as an additional filter which occurs in between the structural constraints of the first filter and the rationality of the second filter (Myers-Scotton, 1998: 34). In order to be able to decide which code will have the best overall outcome in a particular situation (second filter) a speaker must first know whether the code is marked or unmarked (markedness evaluator). Myers-Scotton (1998: 34) claims that the markedness evaluator further biases ‘the selection of alternatives from the initial, structurally determined opportunity set, this time in terms of ‘‘successes’’ or ‘‘failures’’ based on the actor’s previous factual experience, facts previously categorized in an unconscious cost-benefit analysis’.

Unmarked choices remain unnoticed in an interaction because they act in accordance with the social expectations attached to the various codes. Marked choices, on the other hand, violate these social expectations and can, therefore, be used strategically by speakers (Milroy and Gordon, 2006: 213). Quantitatively establishing which codes are marked or unmarked constitutes a vital step in an analysis of code-switching based on the Markedness Model. Simple frequency counts can be used to define the less frequently-used language as the marked code and the more-frequently used language as the unmarked choice (Myers-Scotton, 2002a: 206). The use of the marked code is based on the speaker’s rational decision to employ this code in order to fulfil a specific goal in a particular interaction (Myers-Scotton, 2002a: 218). Consequently, in order to be able to analyse the meaning of the use of different codes the researcher must, first of all, clearly define which codes s/he thinks are marked or unmarked in a given interaction.

The quantification of marked and unmarked codes can be exemplified with Myers-Scotton’s (2002a: 209) analysis of code-switching between English and Chichewa in a Malawi family living in the United States. Chichewa is identified as the parents’ unmarked code during home conversations as it constitutes the

and 7 percent of the mother’s utterances are English-only (Myers-Scotton, 2002b:

210). Myers-Scotton (2002b: 210) explains that this identification of Chichewa as the unmarked code based on simple frequency counts is supported by the parents’

language attitudes, which were assessed in a follow-up interview. The parents described their conscious efforts to speak Chichewa at home with their children in order to maintain their indigenous language. The children, on the other hand, largely speak English at home as approximately 70 percent of their utterances are English-only. The identification of English as the children’s unmarked code is not only supported by these frequency counts but also by the children’s attitudes towards English as the language awarding them independence from their parents (Myers-Scotton, 2002b: 210). However, both children switch to Chichewa at various points during family conversations in order to fulfil temporary goals. For example, during an argument between the two children one of them addresses their father in Chichewa, the father’s unmarked code and preferred language of communication, when seeking support from the father. This temporary switch to the marked code constitutes an example of a speaker abandoning their prior attitudes (i.e. preference for English) in order to achieve a temporary goal (i.e.

receiving support).

Socio-psychological factors such as language attitudes and speaker identity are incorporated in the Markedness Model. Li Wei (2005b: 377) draws attention to a shortcoming of Myers-Scotton’s analysis and claims that the Markedness Model only works if the analyst assumes that each individual will act rationally on all occasions. Moreover, the Markedness Model is based on the assumption that all speakers have an innate ‘markedness evaluator’ which allows them to evaluate which codes are marked and unmarked in any given interaction. Both Myers-Scotton’s Markedness Model and Gumperz’s distinction between ‘we codes’ and

‘they codes’ in bilingual interactions form part of what Cashman (2008: 276) terms the ‘symbolic approach’ to language alternation. Symbolic approaches employ speakers’ macro-level identities in order to explain language choices.

Macro-level identities are defined as the speaker’s ‘membership in social

categories such as sex-based, racial or ethnic groups’ (Cashman, 2008: 284).

Symbolic approaches are based on the idea that speakers exploit the social connotations attached to the various codes in order to make meaning in interaction (Cashman, 2008: 276). The following section introduces the sequential approach to switching. Researchers employing this alternative approach to code-switching refrain from basing any interpretations of multilingual language behaviour on socio-psychological factors such as attitudes and identity or on theoretical constructs such as rationality.