Part III – Interactional Analysis
Chapter 5 – Code Switching, EMCA and Disputes
The previous part focused on the macro picture in two institutions in order to provide a description of the participants’ social and cultural alignments and identities. The ethnographic perspective enabled me to deal with various identity markers related to the participants’
environments and to provide a context for the current part, which will be concerned with Code Switching, EMCA and disputes. This part will zoom in into the participants’ interactions which are constructed through fights and disputes. The choice to focus on disputes in interaction is determined by the findings of the ethnographic study, in that the data have revealed that adversative talk is the most prominent form of interaction among the participants. Moreover, disputes can be considered one of the major problems that inhibit the bilinguals of this study, especially when they are engaged in collaborative learning designs.
Several studies have documented the phenomenon of students’ resistance to academic tasks and academic identity. A great deal of this research is concerned with university students (e.g., Felder and Brent 1996) who documented a phenomenon concerned with resistance that takes the form of challenging the teacher, joking and doing the minimum amount of work necessary. Their research is based on interviews. Phoenix and Frosh (2001) explored masculinities in the early teenage years (11-14 year-old-boys) in London schools, and they used individual and group interviews. They discussed in this study a host of issues concerned with the boys’ aspirations and anxieties, violence, and the discourse of boys’ underachievement in school. The researchers here concluded that “1-boys must maintain their difference from girls and so avoid doing anything that is seen as the kind of thing girls do. 2- Popular masculinity involves “hardness”, sporting prowess, “coolness”, casual treatment of schoolwork and being adept at “cussing”. 3- Some boys are “more masculine” than others. However, since the matter is concerned with interviews, we do not see how students “do resistance” in their interaction, nor we can fully understand what “casual treatment of schoolwork”
mean and entail. An ethnomethodological approach can be more comprehensive in this regard, like the studies conducted by Stokoe et al (2013) and Benwell and Stokoe (2005) where they investigate the same phenomenon of resistance among university students. In these studies, the researchers discuss two patterns, the first is concerned with resistance to academic tasks where students a – invoke their lack of preparation in order to avoid engaging or to appear disengaged, b- challenge the parameters of the task itself. The second pattern is concerned with the students’ reluctance to
display their academic knowledge by policing each other’s contributions, and the way students themselves downgrade their own academic achievement as being a student seems to necessitate being “average” and not standing out. The phenomena which will be reported in this study partially intersect with this research which is concerned with “resistance to academic tasks”. However, the patterns of disputes during group work are more concerned with deviating from the school task and disengagement with group work than with “doing assessments” or “downgrading each other’s contributions to a task”.
EMCA as an analytical method not only has the potential of uncovering the various disputes that ensue between participants, but also can explain the members’ methods in doing disputes, in addition to the role of language(s) and code switching in the various forms of disputes. In what
follows, I will present an overview of how the study of CS has evolved from the domain of ethnography to the domain of interaction, and I will provide definitions of the basic terminology which were used by the major researchers who investigated CS. The chapter is meant to be an introduction to the issues which will be explored in the next chapters. The excerpts which will be used in this chapter are exemplifying rather than analytic.
Code Switching:
Early studies which dealt with language contact tackled the issue of code-switching by
assuming that the phenomenon was a product of poor parenting (for example, Weinreich 1953), or even a phenomenon which has nothing to do with linguistics, (for example, Nilep (2006) quotes Vogt (1954: 368) saying “code-switching in itself is perhaps not a linguistic phenomenon, but rather a psychological one, and its causes are obviously extralinguistic.” These early considerations can be seen as reflections of the negative attitudes towards bilingualism that were dominant at that time.
Such attitudes correspond with the negative attitudes towards bilingualism in very few recent studies (for example Gasper de Alba 1995), who introduces such concepts as “cultural clash”,
“cultural schizophrenia”, and bilingualism as a “mental deficiency”, in an attempt to highlight the consequences of “bi-culturalism”. Mainstream code-switching research, on the other hand, is often dated from Blom and Gumperz’s (1972) “Social meaning in linguistic structures” where the terms situational and metaphorical switching were introduced. Situational code-switching is akin to the concept of diglossia which was introduced by Ferguson (1959), where he spoke about language alternation between varieties of the same language in different domains, and used Arabic to explain his concept; Fishman (1967) applied the concept of diglossia on alternation between unrelated languages, and explicated the concept of “domain”. With diglossia, we do not see alternation or code-switching in the same conversation, rather we see different varieties used in different settings and domains (for example, the use of informal dialect at home and with friends; and the use of the formal variety in the mosque and at school), and this functional division between the two varieties was referred to as situational code-switching (Blom and Gumperz 1972).
Gumperz’s Contextualization36: Blom and Gumperz undertook a study of verbal behavior in Hemnesberget, a small village in Northern Norway. Gumperz (1964b) compared the use of two dialects, standard literary Bokmål (B) and local Ranamål (R) to the use of standard and local
dialects of Hindi in northern India. In each population, the local dialect appeared more frequently in interaction with neighbors, while the standard dialect was reserved for communication across “ritual barriers” (caste, class, village groupings in India, religious setting in Norway). Distinct repertoires are identified in terms of participants, setting, and topic. Blom and Gumperz (1972) described (B) and (R) as distinct codes, though not distinct languages. The separation of the two codes as Blom and Gumperz argued, was maintained because each is “conditioned by social factors”, in the sense that each has somewhat distinct social functions, comparable to the specialization of function for the high variety and the low variety in diglossia (Furgeson 1959). They argued that social events (participants, setting, and topic) restrict the selection of linguistic variables. They applied the notion
36 Much of Gumperz’s early work was carried out in northern India (Gumperz 1958, 1961, 1964a), focused on Hindi and its range of dialects. Gumperz 1958 describes three levels – village dialects, regional dialects and standard Hindi – each of which may be comprised of numerous varieties, and which serve different functions. Gumperz writes “most male residents, especially those who travel considerably, speak both the village and the regional dialect. The former is used at home and with other local residents; the latter is employed with people from the outside” (1958: 669). This gave rise to the idea that linguistic form is affected by setting and participants as well as topic, but which is also comparable in many ways to Ferguson’s diglossia (1959).
of situational switching when the switching was determined by such variables. In their definition of situational code-switching, Blom and Gumperz (1972) assumed “a direct relationship between language and the social situation. The linguistic forms employed are critical features of the event in the sense that any violation of selection rules changes members’ perception of the event. P: 223”, and these changes involve clear “changes in the participants’ definition of each other’s rights and obligation”. Blom and Gumperz stressed the point that sociolinguistic variables must be
investigated empirically in order to reach for an understanding of using one language and not the other. They provided the example that, if “two locals having a heart-to-heart talk will presumably speak in (R). If instead they are found speaking in (B), we conclude either that they do not identify with the values of the local team or that they are not having a heart-to-heart talk” (P: 223). Even in situations where the two dialects are used, no change in the participants’ rights and obligation occurs, because personal issues (greetings, asking about family, etc.) were carried in one dialect, while the business part is carried on in another dialect.
Metaphorical code-switching, on the other hand, is defined as “when (R) phrases are inserted metaphorically into a (B) conversation, this may ….add a special social meaning of confidentiality or privateness to the conversation. (P: 224)”. In other words, metaphorical code-switching bears the notion of using two varieties in the same conversation (for example when changing the topic of conversation from personal issues to matters of official business, marking the transition from informal to formal conversation.) Many researchers and commentators (e.g. Jaffe (2007); Jørgensen (2008); Garrett (2007)) have pointed out that code-switching that involves a minority language and a majority language is viewed as something abnormal and negative by the majority layman, and thus situational code-switching which is akin to diglossia is seen as the ideal type of bilingualism, as the division between the languages is normally maintained and based on the setting.
However, Gumperz himself seems to have noticed the imperfection of the division of code-switching into situational and metaphorical, and he used in 1982 the new terminology
conversational code switching, and pointed to the difficulties of distinguishing the two types of switching in that native speakers generally have no recognition of their own conversational code switches. In other words, the association between language use and domains (setting, topics,
participants, activities) is highly variable, except in the case of diglossia. Gumperz (1982) suggested that the function of code-switching cannot be determined by relying on Macro social observations, and that an understanding of the functions of code-switching should come from the analysis of brief spoken exchanges. Situating the study of code-switching within the realm of conversation analysis, Gumperz noted that code switching must be treated as a contextualization cue, which in itself and in association with other cues (prosody, gesture, tone of voice, etc.) can signal and identify its
function.
Goffman’s footing: Erving Goffman (1979, 1981) introduced the concept of footing as a process in interaction similar to some functional descriptions of code-switching. It is a situation applicable to metaphorical code-switching, and can be defined as the stance or positioning that an individual takes within an interaction. Within a single interaction – even within a short span of talk – an individual can highlight any number of different roles. Changes in purpose, context, and participant role are common in interaction, and footing is a useful tool that can highlight the multiple positions taken by parties to talk in interaction. Footing attempts to show the linguistic markers that are usually associated with alternation between languages. In Goffman’s (1981: 128) words, “a change in footing implies a change in alignment we take up to ourselves and the others present as expressed in the way we manage the production or reception of an utterance. A change in our footing is
another way of talking about a change in our frame for events.” The crucial point to consider is that speaker and hearer are influenced not only by the sound of speech, but rather, there are different markers associated with the sound (sight, gestures, touch, facial features, etc.). This means that
code-switching is determined by switching from one stance or alignment to
another…..When we change voice, whether to speak for another aspect of ourselves or for someone else, or to lighten our discourse with a darted enactment of some alien interaction arrangement – we are not so much terminating the prior alignment as holding it in abeyance with the understanding that it will almost immediately be reengaged. So, too, when we give up the floor in a conversation, thereby taking up the footing of a recipient (addressed or otherwise), we can be warranted in expecting to reenter the speaker role on the same footing from which we left it….In truth, in talk it seems routine that, while firmly standing on two feet, we jump up and down on another. (Goffman 1981: 155).
Goffman makes his point clear that footing is not about code-switching per se, as it is a
phenomenon which engulfs a speaker’s way of switching and alternating between different stances, and language switching is one of them. In a way, Goffman (1981) problematizes the concept of
“hearer” and “speaker”, identifying a range of other forms of participation in interaction. A speaker may produce talk as an animator, merely delivering words formulated by some other person, the author of the utterance; and these two may act on behalf of a third party, the principal of the utterance whose position is established by the words that are spoken. These are the speaker’s footings or production formats. A speaker is always an animator, but he may or may not exploit the other two formats. The phenomenon becomes more complex when we know that listeners can be divided into several types as long as they are within the hearing distance. Thus, Goffman mentions ratified recepients (official), and nonratified (unofficial). In multiparty talk, some ratified recipients may be addressed by the speaker, while others remain unaddressed. Nonratified recipients can be overhearers (accidently) or eavesdroppers (deliberately). Goffman suggests that such participants should be labeled bystanders, who may – or may not – join the conversation, or even influence the speaker’s stance if he/she notices bystanders. The importance of footing is that it situates the motivation of switching to another stance, code or language within the interactional turn by turn context, and without a need to allude to broader contextual frames governed by ideological
understandings of switching or by assigning a social or psychological value to each language. Both, Goffman’s footing and Gumperz’s contextualization cues situated the study of code-switching within the realm of talk-in-interaction.
Code-Switching and the Sociocultural perspective:
There are two approaches attempting to investigate the meaning of code switching from a sociocultural perspective. One approach is ethnographic, and can be divided into two branches, one led by Myers-Scotton and her theory of markedness, and the other is led by Monica Heller and Ben Rampton. The second approach is the Ethnomethodological Conversation Analytic approach which attempts to situate the meaning and function of code-switching within conversation.
Markedness theory: is a socio-psychological approach explicated by Myers-Scotton (1993:
Social motivation for codeswitching: evidence from Africa). This theory depends on the assumption that languages in multilingual communities are associated with particular social roles (rights and obligations) sets, and participants have knowledge and understanding of the social meanings of the rights and obligations associated with each code, and this can be, thus, considered an explication of
Gumperz’ early concept of situational code-switching. She constitutes that in the absence of such knowledge, participants will not be able to discern the significance of particular code choices. This theory was criticized by Auer (1998), Wei (1998), and Gafaranga (2007) for relying on external knowledge, as they stick to Gumperz’ notion of code-switching as a contextualization cue. Nilep (2006) criticized this approach on the grounds that “(it) requires the analyst to make assumptions about each individual speaker’s knowledge and understanding of the speech situation. Code
switching is then explained on the basis of the analyst’s assumptions about speakers’ internal states (including shared judgments about rights and obligations) rather than its effects on the conversation at hand.”
Ethnographic approaches and identity discourses: like the markedness model, ethnographic approaches to code-switching require external knowledge about the value attached to each
language. Monica Heller, for example, tackles issues of power and authority in her research. Heller (1992, 1995, 2006) argues that dominant groups rely on norms of language choice to maintain symbolic domination, while subordinate groups may use code switching to resist or redefine the value of symbolic resources in the linguistic marketplace. While Heller associated code switching with identity, Ben Rampton (2005) focused on ethnicity and ethnic issues in crossing. The term
“crossing”, has gained currency in sociolinguistics, as it deals with phenomena related to language use in multilingual communities, because it can be used as an analytical tool that conflates different levels of analyzing multilingual communities. The increase of immigration led to the emergence of urban communities or neighborhoods where several speech communities coexist and intermingle.
Linguists used such terms such as “borrowing” and “interference” to describe alternations between languages, in which a member of a certain speech community uses the language or “code” of another speech community. However, these terms are famously unstable and each linguist defines them and uses them differently and sometimes interchangeably. Moreover, Rampton argues that the use of terms like code-switching and code-mixing in sociolinguistics is no more satisfactory in that these terms are usually concerned with purely linguistic issues and do not extend to the realms of sociology or to the situated social meaning when one uses more than one language. Code-switching is considered to be dependent on a range of factors or “domains”, the most important of which are interlocutor, topic, and setting, without giving much credit to the role of social meaning – or to the existing discourses in the society that are related to race, ethnicity, dominance and power.
According to Rampton, it is adequate to deal with code-switching when bilinguals are members of the same speech community and speak the same languages. “Crossing, in contrast, focuses on code-alternation by people who aren’t accepted members of the group associated with the second
language they employ. It is concerned with switching into languages that aren’t generally thought to belong to you” Rampton (2005: 270-1). Rampton’s study was concerned with stylized Asian
English, Panjabi and Creole in England’s urban settings, where he describes “young people transgressing the conventional equation of language and ethnicity prescribed for them in ethnic absolutism, it’s relevant to wider public discussion about new ethnicities and cultural hybridity,…(
and he sees this hybridity)… as evidence of cultural innovation in globalized urban spaces”
Rampton (2005: 5). Rampton also stresses the point that crossing serves a variety of purposes: it could be an end in itself and it could emphasize disdain or respect and it occurs both when people are among co-ethnics, and also when they are with members of ethnic out-groups. He sees crossing as a natural event that “occurs at the boundaries of interactional enclosure, in the vicinity of delicts and transgressions, in self-talk and response cries, in games, cross-sex interaction and in the context of performing art” Rampton (2005: 271). Seen from this perspective, crossing is concerned with creativity and the continuous reconstruction of identity, and it allows the examination of macro and micro social processes. Markedness theory and ethnographic approaches to code-switching rely
heavily on the analyst’s macro social observations, and might thus sacrifice the pragmatic function of code-switching in interaction by imposing extralinguistic socio-cultural meanings for the
motivation of switching.
EMCA:
To understand the premises of the second approach, i.e., micro-analysis, and how it differs and diverges from the two branches mentioned above which take into account macro observations, we need to return to some of the premises of ethnomethodological conversation analytic approach (EMCA). Ethnomethodology (Garfinkel 1967, 1968) considers that the object of sociology is to account for the social order, and proposes accounts of the orderliness of practical social action. The accounts are meant to come from a “member’s perspective” as Garfinkel suggests, or to have an
“emic orientation” (ten Have, 1999). In other words, order is the major characteristic of social action, and the practical impossibility of disorderly action was established by Garfinkel in his
“breaching experiments”, where students were asked to “suspend” the natural order (e.g. not to return greetings) and see what happens. The outcome of these experiments was that the students either could not do it or their behavior was interpreted as strange. An understanding of the notion of order in Ethnomethodology depends to a great extent on the assumption that social action happens in a normative framework. Garfinkel (1967) argues that three types of acts can be identified with reference to a particular norm (e.g. if greeted, return the greeting). An act can be 1- a direct application of the norm (normative action) if the greeting is returned; 2- an instance of deviance from the norm if the greeting is not returned, and which can be either a- an instance of repairable deviance (when members assume that he/she didn’t hear my greeting, and so attempt to make themselves heard), or b- an instance of functional deviance (when members assume that the hearing has taken place, and attempt to interpret the non-return of the greeting as a deviation from the norm). In other words, the norm is merely a scheme of interpretation, and return of the greeting is judged as normal and normative action, while non-return indicates a deviant action which can be taken as further evidence that the norm exists.
The conversation analytic approach which is based on talk-in-interaction (Schegloff, 1968) and Sacks’ Lectures on Conversation Analysis (2000) highlighted the importance of social action in ordinary talk. The aim of CA is to describe the orderliness of talk-in-interaction. Studies within this field have demonstrated that talk-in-interaction is highly orderly, e.g., at the level of turn-taking (Sacks 2000), or entry into conversation, or at the level of actions which have been demonstrated to come in pairs (adjacency pairs), etc. Preference in CA is akin to norm in ethnomethodology, in that
“each act always comes either before or after another act” (Gafaranga 2007: 118), for example, a greeting act or a question may or may not be returned or answered, and such options do not have the same status, as one is “expected and normative” and is construed as action which is not marked, and the other “unexpected and deviant” and thus considered as physically marked. The two are
structurally different in that “dispreferred seconds are typically delivered: a) after some significant delay, b) with some preface marking their dispreferred status; c) with some account of why the preferred second cannot be performed” (Levinson: 1983: 307; Church 2009; Kangasharju 2009).
Auer’s (1984) conception of code-switching is established on the principle of what he calls
“preference for same language talk”37. According to this principle, “bilinguals monitor each other’s
37 Gafaranga (2007) proposes that this principle should be modified into “preference for same medium”, in that “the need for a base language is felt because the base language, if clearly identified, works as such a point of reference against which instances of language choice are identified as deserving the researcher’s attention (…) the base language works as a norm, a scheme of interpretation” (P: 138-9). As to how to identify the base language, it can be
turn constructional units (TCU) to be able to take turns in an orderly manner and to choose
language in an orderly manner, in that “code-switching …is conceptualized as divergence from the language of the prior turn or TCU” (Auer 2000: 137). Preference for same language talk entails that the preferred choice is to keep talk going in the same language, and language alternation is a
dispreferred occurrence. Being dispreferred, code-switching must be treated as “unexpected and deviant” and thus physically marked and accompanied by other dispreferrence markers or contextualization cues. In other words, code-switching must be treated and analyzed as a contextualization cue (prosody, gestures, etc) because it works in many ways like other contextualization cues (Auer 1984; Gumperz 1982).
Auer identifies two types of language alternation, discourse-related and participant-related. The former generates meaning regarding the organization of talk and turn-taking system; the latter generates meanings about participants, the way they negotiate what language to use, implicitly or explicitly (See Auer 1984: 32-54). In Gafaranga’s (2007: 133) words, “Auer’s model is a case of functional deviance, for it always tells something either about the organization of talk or about the speaker”. Proponents of this model argue that “the meanings of code-switching must be interpreted with reference to the language choices in the preceding and following turns by the participants themselves, rather than by correlating language choice with some externally determined values”
(Wei 2000: 164) See also (Gafaranga 2007; Wei 1998, 2002 ; Auer 1998). The turns preceding and following the code-switching constitute the “context”. Auer (1998) and Wei (2002; 1998) argue that a great deal of research related to code-switching is deviating from the CA method and premises, simply because they “tend to analyze the meanings of CS in terms of power relations with the speech community, the symbolic values of different languages, and/or the socio-psychological motivations of speakers.” The multiple studies related to the Køge project, for example, addressed the issue of language attitudes regarding minority languages (Turkish), and imposed later on the results of their findings on the code-switching in interaction among the young Turkish-Danish speakers. In other words, the meaning of code-switching was not interpreted with reference to the language choices in the preceding and following turns by the participants themselves, rather by correlating language choice with externally determined values (see also Jørgensen (2008) V1 and V2; Jørgensen 1998). Wei (1998) considers that this approach relies on the premises of markedness theory of code-switching which “places its emphasis on the analyst’s interpretation of bilingual conversation participants’ intention and explicitly rejects the idea of local creation of meaning of linguistic choices” (P: 157). Conversation Analysis, on the other hand, focuses on the members’
procedures of arriving at local meaning of language alternation. It is, in other words, a matter of emic vs. etic perspectives.
Proponents of the sociolinguistic approach do not claim that they are using “pure CA”.
Jørgensen, for example, states that his method is “inspired by conversation analysis, with important modifications”, and he dwells on the notion of context in CA, which he perceives as “narrow” in that “it relates to the concept of turn and usually covers only the turns adjacent to the specific turn being analyzed”, and discards important sociolinguistic variables like the speakers’ socioeconomic status, gender, age, ethnicity, etc. He plausibly adds “An analysis which deals with conversations between interlocutors who know each other well before the conversation, and who have known each other for years, must accept that the speakers will also bring shared histories into the
conversation” (Jørgensen 2008: 325-330). Stroud (1998: 321 ) argues in defense of this approach, done through observation, surveys, interviews, questionnaires, etc. Gafaranga suggests that the analysis must target the language which is least represented. In the case of the current study, the base language is “Danish”, while the language which is least represented is “Arabic”.
by echoing Gumperz’ premises and views on the subject, and he sees that “language in a bilingual environment is inevitably expressing meanings of either solidarity, informality and compassion (the in-group or we-code), or formality, stiffness and distance (the out-group or they-code)” (Gumperz 1982: 66). For Stroud (1998), switching to another code is to juxtapose the “we-code” and the
“they-code”, and “the code-switches serve to index the associations or identities linked to each code. By knowing the details of the local we-they situation, the intention and meanings of the switches can be extrapolated by listeners and researchers perceiving the switch.”
Auer (1998), Wei (1998; 2002), and Gafaranga (2007) particularly criticize the practice of many analysts who take issue with “we code” vs. “they code” dichotomy. This dichotomy refers to an ethnic/minority language (e.g., we – Arabic in relation to the broader mainstream language they - Danish), and the switch from (we) to (they) is perceived by researchers like Heller, Stroud,
Jørgensen, etc. as marked and symbolizes social distance or authority (for example, Power Wielding in Jørgensen’s terms). Against this approach, Wei argues that the boundaries of communities and languages are not always clear-cut, and the we- and they-codes are often hard to establish
empirically. The distinction between the CA approach and the other sociological inquiries, is that the former attempts to uncover the methods through which ordered activity is generated, while the latter considers language as a medium for the expression of intentions, motives, or interests (Wei 2002: 163), or in Gafaranga’s (2007) and Wei’s (1998; 2002) terms, the technical concept of
“preference” in CA has often been wrongly equated with the attitudinal notions of liking, acts of compliance, or the grammatical construction of affirmatives.
To sum up, there are two approaches that attempt to uncover the meanings of code-switching within the sociocultural tradition. One depends on the assumption that speech communities attach different rights, identities and obligations to each of their languages (Jørgensen, Heller, Stroud, Myers-Scotton, Rampton, to name a few), and “speakers who code-switch are seen as appealing to the rights, obligations and identities associated with each language” (Stroud 1998: 321); the other (Auer, Wei, Gafaranga) on the other hand, considers that the “definition of the codes used in code-switching may be an interactional achievement which is not prior to the conversation but subject to negotiation between participants. If anything, it is not the existence of certain codes which takes priority, but the function of a certain transition in conversation” (Auer 1998: 9).
In the following chapters, following Gumperz, Goffman and on a more overall level, Sack’s notion of “order at all points”, meaning that in principle, anything in interaction, also
codeswitching, is orderly, systematic and may therefore be analyzed like that, I analyze in detail sequences in which codeswitching is taking place in disputes. It is important to point out that
sometimes the participants accomplish their dispute in one language only (either Danish, or Arabic), and at other times it seems that the two languages represent the common code for peer interaction.
Using one language and not the other, or alternating between the two languages is to be determined by the context of the interaction, rather than by the assumption that the participants switch in order to wield more power. The language which is least used in the school domain is Arabic, and in this regard, switching to Arabic is to be seen as marked, since the norm and what the school requires is to use Danish. Moreover, in the study of code-switching, the CA methodology has already been used by Auer, Gafaranga, Cromdal, Wei (to name a few), and these researchers have demonstrated that bilingual speakers use code-switching to structure the conversational activity in which they are engaged, and bilingual talk is, thus, orderly and systematic.