CODE-SWITCHING IN AN INSTITUTIONAL SETTING: NEGOTIATING SOCIAL ROLES IN BILINGUAL ENCOUNTERS IN GUATEMALA

In this paper, I investigate the ways in which bilingual speakers utilize linguistic resources to negotiate social roles and manage social relationships at an institutional setting in Momostenango, Guatemala. Drawing from studies on code-switching and politeness, I examine the ways in which speakers utilize the available language resources—Spanish and K’iche’ on the one hand, and second-person pronouns on the other—to negotiate social roles in institutional interactions. This study contributes to our understanding of the ways in which bilinguals organize and conceptualize social roles in culturally specific and meaningful ways. The qualitative analysis of speech choi: code-switching 37 examples can adequately show facets of social life when combined with an ethnographic understanding. In addition, by focusing on actual language use and on a single institutional setting rather than multiple ones, this study brings to light the complex dynamics of using both Spanish and K’iche’ as well as formal and familiar pronouns in Guatemala.


Introduction
Goffman (I967) and Brown and Levinson (1987) showed how everyday discourse entails a multitude of face-threatening acts (FTA), such as critiques, requests, and demands.In many studies that adopted politeness and facework, the model delineated a series of communicative strategies, geared to mitigating the threat involved.One setting which is potentially rich in negotiation of roles when there are face threatening acts is the institutional setting where interlocutors have multiple roles and statuses.Although institutional setting has been of main interest to test Brown and Levinson's theory clac 60/2014, 36-61 choi: code-switching 38 (Aronsson and Rundtröm 1989;Aronsson, et.al. 1987;Curl 2008;Harris 2003;Van De Mieroop 2007), this study may differ from other studies of institutional settings in the sense that it involves bilingual language practice.The research question that I attempt to answer here is as follows: how do bilingual speakers strategically use language to perform institutional roles and manage personal relationships?To explore this question, this article provides a micro-level analysis of linguistic exchanges at a clinic in a bilingual Mayan community in Guatemala.The article focuses on the micro aspects, but also provides the sociolinguistic background necessary to make the relevant data more accessible.This article is organized in the following way: I first review the pertinent literature on code-switching prior to providing background information on bilingualism in Mayan communities in Guatemala and the participants.Starting with observations about the general pattern of language choice at the clinic, I then analyze the bilingual interaction at the clinic in the subsequent sections, focusing on how code-switching is used to achieve communicative goals.

Conversational Code-switching, Pronominal Address, and the Concept of Face
In this introduction, I briefly review the concepts and frameworks to examine the participants' code-switching behavior.Language varieties in diglossic environments commonly indicate social boundaries, expressed as group solidarity and hierarchical relationships.Scholars interested in the functions of code-switching commonly tend to assume that speakers choose one code over the other to accomplish their interactional goals (to name a few, Scotton 1988;Sebba and Wootton 1999;Urciuoli 1996).Studies on code-switching in conversation (Gumperz 1982;Heller 1988;Scotton 1988) pay a great deal of attention to its strategic use in marking social identity, role relationships, and social distance.Linguistic resources such as the choice of language or pronoun can mark differences in the social distance between the speaker and addressee.As Scotton (1988: 178) says, code-switching can be used "as negotiations of personal rights and obligations relative to those of other participants in a talk exchange." Like code switching in general, choice of pronominal address can be examined in the sense of its strategic use rather than rule-governed use.Multiple interpretations of clac 60/2014, 36-61 choi: code-switching 39 pronoun use can be analyzed only through looking at actual interaction.The two social dimensions, 'power' and 'solidarity,' identified by Brown and Gilman (1960), have traditionally seen to motivate the choice of the second-person pronoun in various European languages.The use of tu has a common definition as "the pronoun of either condescension or intimacy," whereas the use of vous carries the definition of "the pronoun of reverence or formality" (Brown and Gilman 1960:258).The shift from 'status' to 'solidarity' in the usage of the second-person pronouns reflects social changes in modern Europe, where a hierarchical order indicated by the existence of the noble class and commoners has disappeared.Although Brown and Gilman have a major influence on the studies of pronominal systems in relation to macrosocial factors expressed as power and solidarity, their limitations are mainly related to the semantic approach.Morford (1997) calls for attention to a semiotic approach rather than a semantic one to understand the multiple social meanings that the use of pronominal address can convey.She criticizes the semantic approach adopted by Brown and Gilman as follows: (Their approach) is in certain respects inadequate for understanding the process by which multiple social meanings are generated through the use of such address systems.Particularly problematic are the notions that pronominal address largely reflects a static order of relationships defined in terms of macrosociological categories; that single instances of address use are unambiguously meaningful in and of themselves.In fact, the relationship between meaning and language forms is not static, but shifted depending on the specific context (Morford 1997: 7; emphasis mine).
She further argues that "the meaning of any given usage depends on the emergent understanding of the particular contextual dimensions involved that develops in the course of social interaction" (Morford 1997:7).As she has pointed out, the importance of an 'emergent' understanding of the contextual dimension cannot be overemphasized.
However, her study does not center on how the contextual dimension determines meanings for language in use because she did not present actual data on the use of pronominal address.Although Morford's discussion of the French pronoun system illuminates interactants' understanding of what is indexed by using tu or vous, her study clac 60/2014, 36-61 choi: code-switching 40 did not show how the indexicality of the pronoun system is expressed in paired exchanges of address.As a way to understand 'the particular contextual dimensions' in social interaction, this paper also places emphasis on the ways in which two different pronoun systems are used in specific contexts.In this way, I show speakers' emergent understanding of shifts in social status and contextual negotiability.
In an attempt to situate the current study on code-switching within a broader theoretical framework, I also draw on the theory of politeness (Brown and Levinson 1987) to examine how code-switching and pronominal changes both enhance the interactional force of the interlocutors' conversational contribution.In other words, the study regards code-switching as a single, albeit highly important, language practice rather than the sole independent linguistic resource that speakers utilize to achieve their daily communicative goals.In the analysis of code-switching data, it is useful to consider Brown and Levinson's (1987) framework of power, distance, and the weight of facethreatening acts.Power and distance are wielded significantly in institutional relationships, and face-threatening strategies are used to negotiate the degree of power and distance.In both institutional and intimate relationships, giving orders or criticizing may involve face threatening.Linguistic politeness is commonly used to avoid direct conflict (Kang 2003 for a Korean case; Stewart 2003 for Spanish pronominal use; Su 2009 for a Taiwanese case).In many studies, the evidence suggests that code-switching is used as a means of negotiating conflict within the constraints of social hierarchy.
The concept of face plays a crucial role in Brown and Levinson's theory of politeness (1987).Face can be generally defined as "a person's desire to be unimpeded (negative face) and to be approved of in certain respects (positive face)" (Su 2009: 376).Certain acts, such as requests, orders, and asking favors, are intrinsically face-threatening.When speakers commit a face-threatening act, they estimate the risk of face loss and the degree of efficiency of communication.They then select a strategy from among a number of choices, thus providing more than one interpretable intention so that the actor cannot be held to have committed to a particular intent.In this study, the main interlocutor's communicative task is considered face-threatening, since she has to request or command co-workers in their own daily interactions.In response to the situation, she strategically employs various resources at hand, one of which is codeswitching.
clac 60/2014, 36-61 choi: code-switching 41 Through microanalyses of conversation, scholars have attempted to explore the ways in which speakers use code-switching to manage social relations.For example, Kang (2003) describes how bilingual Korean-Americans use Korean or English to manage and negotiate social hierarchies when interacting with Korean-American friends.For example, by using Korean terms of address while speaking in English, a speaker can reduce the likelihood of face-threatening acts within the Korean social hierarchy.
Speakers can strategically negotiate social distance and hierarchy by code-switching between Korean and English and by selecting particular terms of address.Attending to the micro aspects of interaction, such as in Kang's study, enables us to understand how speakers strategically apply the tacit norms of Korean social hierarchy in order to manage conflict.
Drawing from studies on code-switching and politeness, I examine the ways in which speakers utilize the available language resources-Spanish and K'iche' on the one hand, and formal and familiar pronouns on the other-to build and negotiate social relationships in an institutional interaction.This study focuses on how the conversation initiator, Marta, makes use of a number of available linguistic resources, especially code-switching and second-person pronouns, to achieve a communicative goal in an institutional setting.The purpose of this paper is thus to describe how speakers actively engage in managing and shifting social roles in a conversation rather than to identify a possible pattern of verbal organization among speakers.The next section provides background information to facilitate readers' understanding of the bilingual exchanges in question.

Social Context and Bilingual Resources in Momostenango, Guatemala
This study is part of a larger ethnographic study conducted over the course of an extended period fieldwork period (August 2000 to May 2001) in a bilingual Mayan community in Western Guatemala.During my field research, I lived in a Guatemalan community and volunteered once or twice a week in the clinic, which is the speech setting of this paper.In the clinic, data were gathered through participant observation, both with and without tape recordings of verbal interactions in Momostenango.In addition to participant observations, I interviewed a key informant, Marta, to obtain her clac 60/2014, 36-61 choi: code-switching 42 metalinguistic discourses regarding her language choice in the institutional setting of the clinic.I regularly assisted the nurses in this clinic on a volunteer basis.Volunteering at the clinic allowed me to observe the interactions among speakers who had assigned roles and statuses within the institution.
In Guatemala, approximately half of the population is Mayan.Diglossic distribution is common throughout the country (Garzon et al. 1998;Langan 1991), but communities like Momostenango have a greater number of indigenous Mayan people who are either Mayan monolingual or Mayan-Spanish bilingual.Spanish is rarely used in the remote aldeas (hamlets) surrounding the town center of Momostenango, but people in the town center can easily speak both languages.Spanish is primarily used in interactions with the local Ladinos 1 and in public places and situations.
Ethnolinguistically, Momostenango Indians belong to the K'iche' Maya group, which is a relatively large Mayan language group (with at least a million speakers) among the 22 distinct Mayan language groups in Guatemala.In Momostenango, 97% of the population belongs to the K'iche' Maya group, which maintains its language to various degrees.There is little use of Spanish in the aldeas (hamlets) that surround the town center of Momostenango.People in the center, however, have no trouble speaking both the languages.Spanish is primarily used in interactions with the local Ladinos and in public places and situations.Ladinos belong to non-Mayan ethnic group whose mother tongue is Spanish and they do not consist of the major part of the local population in Momostenango unlike a big city like the capital of Guatemala.The diglossic distribution is common throughout the country, but communities like Momostenango have a relatively larger Mayan indigenous population that is Maya monolingual or bilingual.
The distribution of the use of K'iche' and Spanish in different situations shows the prototypical pattern of diglossia (see Table 1).In other words, high-prestige varieties, such as the standard or official language, are used for formal purposes, while lowprestige varieties, such as vernacular languages or dialects, are used for less formal and 1 Ladinos belong to a non-Mayan ethnic group whose native language is Spanish and who generally live in an urban environment compared to Mayan ethnic groups.Although the match between a particular language and social index provides a general picture of language choice in this bilingual community, with the variants being the available indexes of varying social meanings, attempts to define particular codes and the corresponding situations or interlocutors may not always be successful.Social relationships are divided into personal and institutional, as seen in Table 1 above.In Momostenango, bilingual Mayans in a close relationship tend to use K'iche'.However, simple diglossic differentiation does not capture the complexities of the social relationships in interaction.In particular, maintaining intimate or formal relationships may require other strategies that are expressed through switching the codes available to speakers.
In addition to the distinction between K'iche' and Spanish, that of second-person pronouns can indicate formal and informal relationships.Like many European languages, both Spanish and Mayan languages are characterized by two types of second-person pronouns.It is common for the formal/polite and informal/familiar forms of second-person pronouns in K'iche' and Spanish to be used to indicate the degree of social distance and familiarity.Several dialects of K'iche', spoken by approximately one million people in the western highlands of Guatemala, have formal second-person, or honorific, pronouns.K'iche' uses a set of cross-reference markers to indicate the subject and object of verbs, the possessor of nouns, and the object of relational noun phrases.Robertson (1987) refers to the two different second-person pronouns as "polite" versus "familiar," whereas Kaufman (1977) refers to them as "formal" versus clac 60/2014, 36-61 choi: code-switching 44 "familiar."In this paper, I will use Kaufman's terms to distinguish the second-person pronouns.
Both languages not only have a distinction between formal and familiar second person pronouns but also cross-reference using morphosyntactic devices.

Second-person singular formal
Second-person singular familiar switching as a sign of transition from language maintenance to language shift or language death.For example, Barret argues that "an increase in code-switching (and in particular an increased use in the dominating language) accompanies many instances of shift and may indicate a 'turnover' to the dominating language" (2008: 284).Collins shows that those who support the Maya movement tend to code-switch in fewer contexts.While both Collins and Barret consider code-switching as evidence of a language shift and its relationship, which is often the result of language ideology, I clac 60/2014, 36-61 choi: code-switching 45 focus on the way in which language resources are utilized for various purposes as indexed in a specific context.
In order to move away from viewing code-switching as evidence of a language shift, we need to examine the implications of code-switching on social relationships and identities.Another area of interest studied by scholars of Mayan languages is the pronominal system (See Robertson 1980).In his diachronic study on the Mamean pronominal system, Robertson (1980) argues how they developed independently from Spanish second person pronominal distinction, that is, polite versus familiar.Although the Mayan pronominal system has been one of the major areas for many linguists (Bricker 1977;Robertson 1980), the analysis of actual usage in a specific context has been neglected.This study thus presents a data analysis of Mayan language in use, with an emphasis on bilingual practices and pronominal use.degree of fluency varies.For example, Marta, the chief nurse, has a full command of both K'iche' and Spanish, and mediates interactions between the monolingual Spanish doctors and monolingual K'iche' patients.While the participants can be represented in terms of their occupational positions, some are close friends in addition to being members of the institution.I became well acquainted with two workers in the clinic, Marta and Griselda, after spending time regularly with them at the clinic, which allowed me to understand their relationships better.Marta was the chief nurse who had worked at the clinic for three years at the time of my fieldwork.Griselda was the janitor, who was also bilingual, although her command of Spanish was not as good as Marta's was.
Mara had built relationships with various interlocutors with diverse positions and roles in the institution.Marta and Griselda are co-workers in the hospital, but they are also close friends who have known each other since childhood.I found it advantageous to observe the same participants interacting in different contexts in order to understand the functions and meanings of code-switching incidents.

Public Encounters at the Clinic
In this section, I describe how both language and pronoun choices are made in order to address patients at the clinic.Marta usually initiated interactions with visitors to the clinic in Spanish, including phone calls.Further, all signs in the clinic were written in Spanish, and the doctor and patient interacted in Spanish except for cases of monolingual K'iche' patients.In this case, Marta translated during the medical examination.In general, Spanish was used in the clinic.To illustrate the implicit speech norms required in the clinic, I provide the following two examples, as conventionalized exchanges between the nurse and patients.To borrow Scotton's (1988) term, the participants in the exchanges have a sense of the "script" to be used in the routine interaction.In these typical exchanges, the nurse's choice of code mainly depends on the patient's (or visitor's) sex and region, in addition to the familiarity that she perceives when interacting with them.In the following speech event, the nurse uses the formal command form (underlined) in both Spanish and K'iche' when she encounters a visiting patient at the reception desk.
clac 60/2014, 36-61 choi: code-switching 47 1) 2 Participants: J is the researcher, A is the patient, and M is the chief nurse.
A woman in traditional dress, along with a child, came into the pharmacy where Marta was sitting.I was sitting near the door through which the woman walked.

2.
A: (looking at me and M, and then to me) eh.....

4.
A: Quiero solo preguntar = (looking at me and then at Marta) I just want to ask…
I need to consult with Dr. Victor.
Ah… but he's not here.

A: k'o taj?
He's not?
Here, no, but he's down there.Yes, he will be here tomorrow The first use of the formal second-person pronoun was in Spanish, whereas the following use switched to K'iche'.By maintaining the formal second-person pronouns, it is likely that Marta switched to K'iche' upon noticing that the visitor spoke Spanish haltingly.While there is no way to know immediately whether an interlocutor has full command of both languages, Marta's responses depended on the visitor's clothing, sex, age, and other characteristics, such as the manner of speaking as in this case.When I asked her about this issue, Marta simply answered that she chose to speak K'iche' "when they look like aldeanos (people from a hamlet)" (emphasis mine), which mostly means clothing or manner of speaking.
As another interaction illustrates, however, this rule does not always apply.In one incident, only Spanish was used, although the visitors were also from aldea.Marta often spoke to visiting patients in Spanish using a formal code.Although these visitors were wearing traditional dress, Marta used the code that combines Spanish and formal second-person pronouns.The interlocutor in this case, however, was a man.
2) Participants: A is a patient who came with his wife to visit the doctor at the clinic and M is Marta, the chief nurse.
(A enters the lobby of the clinic) 1. M: (to the man) Pase adelante.
Come in, please.

A: Está el doctor?
Is the doctor in?
3. M: Sí, está, solo es que ahorita está ocupado.(Walking in) Yes, he is, but the only thing is that right now he's busy.
Sit down and wait a minute, please.
clac 60/2014, 36-61 choi: code-switching 49 As this example shows, Marta does not always communicate in K'iche' with Mayan patients who are from remote hamlets.Marta initiates the interaction with the man who accompanied his wife to the doctor's office.I later asked Marta why she used Spanish in this particular interaction.She said that she did not personally know the individuals because they were not from the center of Momostenango.In this case, unlike the first example, her use of Spanish was motivated by the degree of social distance existing between her and the interlocutor.In addition to the exclusive use of Spanish throughout the interaction, Marta uses formal second-person pronouns to address the patient to index her institutional role in this interaction.Although she did not explicitly mention this point, another reason why Marta used Spanish in this exchange is related to the addressee with whom she interacted.In Momostenango, a man is usually considered the "spokesperson" for his wife or other family members.For example, when his wife is sick, a man may ask her a question in K'iche' and then translate it into Spanish in his interaction with a nurse or doctor.In general, men in Momostenango tend to be more competent in Spanish than women are, but the degree of competence is not the only factor: it is relevant who represents the speaker's voice in the public sphere in Mayan society.
As we can see in these examples, the code commonly used for the patient in the clinic is either K'iche' or Spanish, which must be combined with formal pronominal addresses to index social distance.The institution's implicit speech norm is Spanish and the formal pronoun Usted between coworkers or K'iche' and the formal pronoun la with patients when they are from the aldea and/or are elderly.According to my interview data, bilingual Mayan workers in the clinic feel that using tú seems to indicate that the speaker does not conform to the rules of social etiquette.In addition, these workers feel that using formal pronouns in Spanish elevates the speaker's position.However, they have no problem with using familiar personal pronouns in K'iche' between friends, as will be shown in the next section.

Redefining Social Roles and Saving Face
The data analysis in this section demonstrates the common usage of K'iche' and Spanish to index intimacy and distance between two interlocutors, and how K'iche' and clac 60/2014, 36-61 choi: code-switching 50 Spanish are used in order to redefine institutional roles in a face-threatening situation.
While the above examples show Marta's use of Spanish and formal pronouns in public encounters, the example below shows her use of the symmetrical V (Usted) in Spanish and an instance of switching to T (tu) in K'iche' during the course of an interaction.
3) In the hospital, M(arta) found out that the floor of the clinic room was excessively wet, so she spoke to G(riselda) about this problem.
1. M: (entering the pharmacy room and seeing G working there) Gris, puede hacer un favor?
Gris, would you formal do me a favor?
(G leaves the room.After a short interval, G returns.)

M: (looking at G) xawilo? 3
Did you familiar see it?

Yes.
In this linguistic exchange, Marta initiates the utterance expressing her intention to make a "request" to Griselda, thereby she plays an institutional role.Marta moves from an institutional to a non-institutional role by switching codes from Spanish to K'iche'.
In addition to switching the language, Marta first addresses Griselda with the formal pronoun (Usted) in Spanish (line 1: haga, as conjugated from the infinitive verb, hacer 'do') but later expresses intimacy by combining K'iche' and a second-person familiar pronoun (line 4).Marta positions herself as Griselda's intimate friend rather than her coworker by shifting from Spanish for a polite request to using K'iche' to rebuild an intimate relationship between friends.As Goffman showed (1974), role relations among interlocutors can be overlaid in this interaction.
3 X--aw-il -o past tense 'it' inf 'you' 'to see' final marker clac 60/2014, 36-61 choi: code-switching 51 Both K'iche' and Spanish include the distinction between the formal and familiar second-person pronoun forms.The familiar second-person pronouns are generally used for persons with whom the speaker is on a first-name basis, although usage can vary with locality.In nearly all cases, the familiar form is used when addressing children, family members, and close friends.Although these pronouns are in the second person, formal pronouns in both languages take a third-person verb.In K'iche', the formal second singular pronoun is la, which comes at the end of a sentence.K'iche' also serves the function of mitigating a face-threatening situation.
Adopting Brown and Levinson's terminology (1987), we can say that Marta's request was face-threatening for both parties.Even though Marta asks Griselda for a favor politely in Spanish, she took the risk of losing face because her request involves blame of Griselda for the messy floor.Therefore, even though it is Griselda's institutional obligation to clean the floor, the fact that Griselda and Marta are close friends affects how Griselda might feel about the request.To avoid a possible authoritarian situation, Marta's question after Griselda's action, 'Did you see it?', is encoded in K'iche'.In addition to avoiding the assertion of authority, this question requests confirmation from Griselda that the condition of the floor requires her to clean it.Thus, Marta can avoid intentionality in commanding her friend to do what she wishes, because the mess itself requires Griselda to do as Marta requests.Thus, Marta's switch to K'iche' indexes outside Griselda's role as a janitor, thereby suspending her institutional roles.
According to Goffman, communication is an act that is subject to rules of conduct because "it represents a way in which selves are confirmed-both the self for which the rule is an obligation and the self for which it is an expectation" (1956: 475).In this example, Marta and Griselda situate themselves in a position of obligation and expectation, as required by the institution, but they quickly shift to a non-institutional rule of conduct by switching to K'iche' combined with familiar second-person pronouns.In dealing with such a face-threatening situation, Marta used certain rhetorical strategies at the level of actual content and contextualization cues.
In the following example, the use of Spanish and K'iche' is again combined with two second-person pronouns in each code.However, unlike example (3), where K'iche' + clac 60/2014, 36-61 choi: code-switching 52 the familiar pronoun was utilized to mitigate a face-threatening moment, Spanish + the formal pronoun assumes this role in example (4).4) G(riselda) and M(arta) are co-workers in the hospital.J(insook) is the researcher.When G comes into the room, she takes off her sweater.As G leaves the room, M grabs her skirt playfully.
It's done, it's done.Why did she go away?
8. M: Se enojó, porque yo dije, "arregle su corte!" She's mad because I said, "Fix your formal skirt!" Unlike the previous example, this case shows that the initiation of Marta's playful interaction was facilitated by switching from Spanish to K'iche'.Marta's playful criticism embarrassed Griselda because of the nature of the criticism; it concerns her appearance.In Momostenango, like many Mayan communities, the majority of Mayan women wear traditional dress of a blouse and skirt.It is important for Mayan women to have their skirt neatly tightened and their blouse tucked in.Therefore, it embarrassed Griselda that her dress did not appear neat.This excerpt is interesting not only because 4 Ch -∅ -a -b'an -a la uq IMP-3 rd sing abs-2 nd sing erg-'to do' -final marker article 'skirt' clac 60/2014, 36-61 choi: code-switching 53 of the way Marta switches between K'iche' and Spanish, but also how she reports her previous utterance to Griselda.Analytic emphasis is given in three different parts (lines 1, 3, and 8) that include command forms.I will now discuss each of the occasions of the command form separately.
First, as shown in line 1, Marta addressed Griselda with the code that indexes familiarity and intimacy, conjugating the K'iche' verb and cross-referencing the familiar secondperson singular form.As a co-worker, Marta would normally treat Griselda formally by using Spanish along with the formal second-person pronoun.However, Griselda was probably annoyed and embarrassed when Marta playfully teased her about how she was wearing her skirt.By responding in Spanish, Griselda rejected Marta's redefinition of the social relationship and the context as informal.In other words, Griselda attempted to keep their relations professional by complying with institutional speech norms, thereby attempting to make Marta recognize her social role and their institutional relationship at the clinic.
Marta responded to Griselda's effort to redefine the social relationship in line 3.By conjugating the Spanish verb mirar 'to look at' as a formal command form (mire), Marta addressed Griselda using the formal pronoun.We can compare this example with the previous one.Depending on the situation, switching to Spanish or K'iche' involves the speaker's intention to reduce face-threatening interactions.Noticing that Griselda did not respond to her playfulness but seemed rather annoyed, Marta tried to rectify the situation by switching to Spanish, the formal code between co-workers.Finally, in line 8, when Marta reported to me what she had said to Griselda, she conjugated the verb arreglar using the formal second-person pronoun form instead of translating what she actually phrased in K'iche'.This suggests that she habitually uses the familiar pronoun when she talks to Griselda in K'iche', but the use of Spanish constrains her switch from the familiar K'iche' to the formal second-person pronoun in Spanish.
This example illustrates that code-switching is "an important part of social mechanisms of negotiation and definition of social roles, networks, and boundaries" (Heller 1988: 1).
As we observe in this example, the negotiations of the rights and obligations of participants occur more visibly with changes in pronominal address.The institutional clac 60/2014, 36-61 choi: code-switching 54 setting forces speakers to enact diverse social roles, as seen in the interactions between Marta and Griselda.Speakers' social roles are determined in relation to the interlocutor in a specific context.The clinic as a speech setting provides a social context in which social roles are shifted from institutional to non-institutional and vice versa.Throughout the conversations, Marta makes a range of choices to negotiate the interpersonal distance constantly and so achieve the communicative goal.What makes her linguistic practices effective in the negotiation of distance is the multiple meanings attached to a code choice and the personal histories of Marta and Griselda.4.4.Institutional Use of K'iche' and Spanish: Code-Switching to Evoke Authority In the previous section, I analyzed examples of code-switching to illustrate how it is strategically used to modify social roles.In this section, I discuss how code-switching is used to invoke institutional identity among speakers.In the following example, Marta initiates an interaction with patient in K'iche', which indicates respect for the elderly.K'iche' is used to address an elderly patient and Spanish to address a lower-ranking coworker.In this excerpt, all of the participants have fixed institutional identities, which are displayed using both pragmatic and grammatical resources.5) An old Mayan female patient (P) is lying down in bed in a patient room of the clinic.M(arta) wants to introduce her to T(helma), a nurse's aide who plans to stay at the hospital overnight.A is the patient's son.
Hi, ma'am, I am leaving.

P: kat b'ek? Y chuweq?
You are?And tomorrow?
She is the nurse who will stay.She will stay.
When the saline solution runs out, it needs to be replaced.Maybe within three hours.No, only fruit juice or water.
In this excerpt, Marta interacts with the patient, her family member, and an assistant.
Thelma is the assistant nurse who is in training at the clinic to become a professional nurse.Like other trainees that I met, she attends classes regularly and learns in the field under Marta's supervision.I often observed that Marta and the trainees interacted mostly in K'iche', but in Spanish in the class where Marta lectured.Throughout the interaction in this example, Marta provides information and instructions to three different interlocutors.While delivering medical information, she carefully chooses each code to foster a formal relationship with each interlocutor.Her code-switching 5 Atol is a thick corn drink that is consumed mostly at breakfast, and incaparina is a mixed grain drink that is also consumed for breakfast or between meals in Guatemala.Both are thick drinks, so they are not recommended for patients who need to drink only clear fluid.
clac 60/2014, 36-61 choi: code-switching 56 evokes the authority that the information represents.Marta first begins speaking in K'iche' to express respect for the patient, an elderly indigenous woman.It is common in Mayan communities like Momostenango for formal K'iche' to be used to address an elderly woman, in this case by using wi.In the Momostenango K'iche' variety, wi commonly means 'yes', as it does in this example, but it can also be used to greet someone (in the sense of 'hi') and may be used to express respect when greeting an elderly person.
In line 6, Marta switches to Spanish when she talks to Thelma, who is also bilingual.By using Spanish instead of K'iche' in this interaction, however, Marta redefines her social relationship with Thelma, the nurse's aide.She does not only switch to Spanish from K'iche', the language that she just spoke with the patient, but she also speaks more authoritatively.Her choice of K'iche' is most likely motivated by the power dimension identified by Brown and Gilman (1960) as well as the need to display her institutional identity to the patient and her family.
Marta's code-switching to Spanish is also intended to enhance her authority when she says, 'It needs to be done'.Here, when Marta gives her instruction to the assistant, she impersonalizes both the speaker and listener by using the impersonal verb haber ('there is') in Spanish.According to Brown and Levinson (1987), impersonalization (i.e., avoiding 'I' and 'you') is a negative politeness strategy.In this case, Marta's use of impersonalization has a similar function to the use of the formal second-person pronoun in Spanish because she avoids direct commands, yet still conveys a message of authority because providing more than one interpretable intention allows the actor to avoid committing to a particular intent.In line 8, Spanish is used to express medical authority to the patient's son, but it also indicates the distance between the nurse and the patient's family as opposed to the authority directed from a higher to a lower rank.Thus, switching to Spanish can mean both distance and authority, depending on the interlocutors in this example.
In this excerpt, both K'iche' and Spanish are utilized to express formality with the interlocutors.Marta assigns actions to be taken and achieves authority through both language and grammatical choice.The use of K'iche' initially involves respect for the elderly, while the use of Spanish suggests authority expressed by a higher-ranking clac 60/2014, 36-61 choi: code-switching 57 worker to a lower-ranking one and by an authoritative member of the medical staff to the patient's family.Marta maintains a formal relationship with various people by displaying her institutional identity to each participant.The use of Spanish enhances the voice of authority as embodied by the medical staff and enables Marta to construct her status in relation to both the nurse's assistant and the patient's family.Using Spanish in this situation reflects a typical institutional norm applied to interactions with patients and co-workers.
We have seen that participants like Marta have several social roles that are specific to the interlocutor in each interaction.Drawing on the literature on interactional codeswitching and the theory of politeness, and I examined how code-switching enhances the interactional force of Marta's conversational contribution.Multiple social meanings are indexed through code-switching between Spanish and K'iche' as well as between the formal and familiar second-person pronouns in each language.In the interactions in which Marta asserts her institutional status and role, each language does not always carry the same status.The institutional roles of nurse, patient, and janitor are assigned to each interlocutor prior to their encounters, and thus they establish the authority of one over another, such as the authority of a nurse over a patient.As shown in my analysis, code-switching is one of the primary resources for the realization of social distance and hierarchy.Switching between Spanish and K'iche' is used strategically, and the secondperson pronoun is incorporated in both Spanish and K'iche' to redefine the type of relationship between interlocutors.

Summary and Conclusion
In this paper, I demonstrated the ways in which the choice of Spanish or K'iche' and formal or familiar second-person pronouns in each language can indicate both the degree of social distance between the speaker and addressee as well as the various social attributes of individual speakers, such as status and role.In addition, face saving is Marta and Griselda, such a code choice may index a lack of authority.Spanish is often associated with formality as we saw in the use of Spanish between Marta and her interlocutors.In a face-threatening situation where any imbalance between the two interlocutors may have serious consequences, code-switching and pronoun shifting work as conversational resources through which speakers can express multiple roles more effectively than they could by using a single code.
This study contributes to our understanding of daily bilingual practices among bilingual Mayas in Guatemala by examining how bilingual speakers organize and conceptualize social status in culturally specific and meaningful ways.To date, although some studies focus on bilingualism and bilingual language use, such as code-switching, little attention has been paid to the microanalysis of code-switching as a strategy in constructing social roles, as attempted in this study.Although there are some limitations in this study in the sense that it deals with a limited number of examples and a select number of interlocutors at an institution, a qualitative analysis of a limited number of speech examples can adequately show facets of social life when combined with an ethnographic understanding.In addition, by focusing on actual language use and on a single institutional setting rather than multiple ones, this study brings to light the complex dynamics of using both Spanish and K'iche' as well as formal and familiar pronouns in Guatemala.As shown in this paper, social groups organize and conceptualize social roles in culturally specific and meaningful ways.One important implication of this study is that, as Woolard (2008) rightly points out, an analysis of code-switching can and should be brought into relation with more general theoretical constructs-such as politeness (Brown and Levinson, 1987)-which go beyond the confined scope of research on language choice and code-switching.
Social Roles in the Clinic Spanish was commonly used in the clinic under examination.The clinic has eight workers, including a chief nurse, two nurse's aides, and a janitor-all of whom are Maya-as well as two ladino doctors (one part-time) and one ladino administrator.The clinic is located in a two-story building with three rooms, a doctor's office, kitchen, bathroom, and small lobby.This is considered a large clinic in the town, and it is open seven days a week.Sunday is the busiest day because it is a market day when aldeanos, people from remote hamlets, visit the town center and are thus likely to visit the doctor at the clinic.Two crucial aspects regarding interactions in the clinic need to be noted.First, the clinic was chosen as a speech setting to illustrate the participants' awareness of the implicit speech norms of the institution.The social relationships in this institution are established among co-workers who relate to each other within a stratified ranking system with the following institutional hierarchy: doctor, chief nurse, assistant nurse, and janitor.With the exception of the two doctors and administrator, who are Spanish monolinguals, all staff members are bilingual in K'iche' and Spanish, although the clac 60/2014, 36-61 choi: code-switching 46
the patient's son, who is sitting next to the bed on which the patient is lying) Ella puede tomar solo líquido claro.She can drink only clear fluid.9. A: ah... y atol, o incaparina?ah... and [Can't she have any] atol or incaparina? 5 10.M: no, solo jugo de frutas, o agua.
negotiated through the choice of second-person pronouns or impersonalized expressions.The code-switching examples in this paper illustrate how code-switching can be used as a tool to create ambiguity and invoke multiple voices in a facethreatening situation.The use of K'iche' may bring out a sense of solidarity between clac 60/2014, 36-61 choi: code-switching 58

Table 1 .
clac 60/2014, 36-61 choi: code-switching 43 more personal purposes.The detailed diglossic distribution of the two languages in Momostenango is shown below.Diglossic distribution of the two languages: Functional division of K'iche' and Spanish in different situations with different interlocutors.

Table 2
. Second PronCollins 2005)n Spanish and K'iche'As indicated in Table2, both K'iche' and Spanish include the distinction between the formal and familiar second-person pronoun forms.The pronouns known as pronouns in the familiar form are generally used with persons with whom the speaker is familiar on a first-name basis, although usage can vary with locality.In nearly all cases, it is used when addressing children, family members and close friends.Although these pronouns are in the second person, formal pronouns in both languages take a third-person verb.In K'iche', the formal second singular pronoun is 'la', which comes at the end of a sentence.Bilingual use of the Mayan language and Spanish prevails in Mayan communities in Guatemala.Research on bilingualism in Guatemala's Mayan communities has examined various topics, but little attention has been paid to bilingual interaction.Code-switching in particular has not been of main interest of scholars of Mayan languages, with the exception of two recent studies(Barret 2008;Collins 2005).The two studies treat code-