"Code Switching" in Sociocultural Linguistics

“Code Switching” in Sociocultural Linguistics
Chad Nilep
1. Introduction
The term code switching (or, as it is sometimes written, code-switching or
codeswitching)1 is broadly discussed and used in linguistics and a variety of
related fields. Benson (2001) cited nearly 1,300 articles on the subject in the
Linguistics and Language Behavior Abstracts database published in virtually
every branch of linguistics between 1990 and 2000. A search in 2005 swells that
number to more than 1,800. However, despite this ubiquity – or perhaps in part
because of it – scholars do not seem to share a definition of the term. This is
perhaps inevitable, given the different concerns of formal linguists,
psycholinguists, sociolinguists, philosophers, anthropologists, etc. This paper will
attempt to survey the use of term code switching in sociocultural linguistics and
suggest useful definitions for sociocultural work.
Since code switching is studied from so many perspectives, this paper will
necessarily seem to omit important elements of the literature. Much of the work
labeled “code switching” is interested in syntactic or morphosyntactic constraints
on language alternation (e.g. Poplack 1980; Sankoff and Poplack 1981; Joshi
1
My personal preference is to spell code switching as two words, with white space between them,
a practice I will generally follow throughout this paper. Original spelling will be preserved in
quotations and when paraphrasing scholars who routinely use an alternate form.
Chad Nilep – University of Colorado Boulder
1985; Di Sciullo and Williams 1987; Belazi et al. 1994; Halmari 1997 inter alia).
Alternately, studies of language acquisition, second language acquisition, and
language learning use the term code switching to describe either bilingual
speakers’ or language learners’ cognitive linguistic abilities, or to describe
classroom or learner practices involving the use of more than one language (e.g.
Romaine 1989; Cenoz and Genesee 2001; Fotos 2001, inter alia). These and other
studies seem to use code as a synonym for language variety. Alvarez-Cáccamo
(2000) argues that this equation may obscure certain interactional functions of
such alternation.
Praticamente todo o trabalho sobre o “code-switching”, ou

comutação de códigos, tem-se baseado numa identificação estrita
entre as noções de “código” e “variedade linguística”, seja esta
uma língua, um dialecto, um estilo, ou mesmo um registo
prosódico. Porém, esta focagem estrutural deixa sem explicar
convincentemente certos fenómenos conversacionais relativos à
relevância ou significatividade (ou ausência de tal relevância) das
alternâncias entre variedades aparentemente em contraste.
Practically all work on “code-switching,” or changing codes, has

been based on a strict identification between the notions of “code”
and “linguistic variety,” be that a language, dialect, style, or
prosodic register. However, this structural focus fails to
convincingly explain certain conversational phenomena relative to
the relevance or significance (or lack of relevance) of alternations
between contrasting varieties. [Alvarez-Cáccamo 2000:112; my
translation]
Certainly, the study of language alternation has been fruitful over the past several
decades. The identification of various constraints, though controversial, has
inspired a great deal of work in syntax, morphology, and phonology. A structural
focus has been similarly constructive for production models (e.g. Azuma 1991,
2
1996) or as evidence for grammatical theory (e.g. MacSwann 2000; Jake, Myers-
Scotton and Gross 2002). By ignoring questions of function or meaning, though,
this structural focus fails to answer basic questions of why code switching occurs.2
Auer (1984) warns, “Grammatical restrictions on code-switching are but
necessary conditions” (2); they are not sufficient to describe the reason for or
effect of a particular alternation. If linguists regard code switching simply as a
product of a grammatical system, and not as a practice of individual speakers,
they may produce esoteric analyses that have little importance outside the study of
linguistics per se, what Sapir called “a tradition that threatens to become
scholastic when not vitalized by interests which lie beyond the formal interest in
language itself” (1929:213).
This paper is thus positioned within the discipline of sociocultural
linguistics, an emerging (or one might say, revitalized) interdisciplinary approach
to linguistics that looks beyond formal interests, to the social and cultural
functions and meanings of language use.
Periodically over the last century, linguists have proposed to bring their
own studies closer to other fields of social inquiry. In 1929, Edward Sapir urged
linguists to move beyond diachronic and formal analyses for their own sake and
2
Woolard (2004) suggests that the basic question should be not why speakers make use of the
various forms available to them, but why speakers would not make use of all available forms.
Thus she suggests, “It could be argued that linguists, with their focus on constraints against rather
than motivations for codeswitching, do ask this alternative question” (91).
3
to “become aware of what their science may mean for the interpretation of human
conduct in general” (1929:207). He suggested that anthropology, sociology,
psychology, philosophy and social science generally would be enriched by
drawing on the methodologies as well as the findings of linguistic research. He
also exhorted linguists to consider language within its broader social setting.
It is peculiarly important that linguists, who are often accused, and

accused justly, of failure to look beyond the pretty patterns of their
subject matter, should become aware of what their science may
mean for the interpretation of human conduct in general. Whether
they like it or not, they must become increasingly concerned with
the many anthropological, sociological, and psychological
problems which invade the field of language. [Sapir 1929:214]
Sapir was not alone in his hopes for a more socially engaged linguistics. Indeed
the development during the 1930s-1950s of what has come to be known as
sociolinguistics and psycholinguistics suggests that, at least for some linguists,
social interaction and human cognition were as important as the forms and
structures of language itself.
Nonetheless, by the 1960s some scholars once again felt the need to argue
for a more socially engaged linguistics. In his introduction to a special issue of
American Anthropologist dedicated to the Ethnography of Communication
(Hymes and Gumperz 1964), Dell Hymes (1964) lamented that the socially
integrated linguistics Sapir had called for was disappearing. Hymes and others
worried that new formal approaches, as well as the push for linguistics as an
autonomous field, threatened to once again isolate linguists. At the same time,
4
though, the growth of ethnolinguistics and sociolinguistics offered a venue for the
socially engaged linguistics Sapir had called for four decades earlier.
Four more decades have passed, and once again scholars are calling for a
revitalization of socially and culturally oriented linguistic analysis. Bucholtz and
Hall (2005) position their own work on language and identity as what they call
sociocultural linguistics, “the broad interdisciplinary field concerned with the
intersection of language, culture, and society” (5). Just as Hymes (1964) worried
that linguistics had been bleached of its association with the study of human
interaction in the wake of formalist studies, Bucholtz and Hall point out that
sociolinguistics has in turn been narrowed to denote only specific types of study.
Sociocultural linguistics is thus suggested as a broader term, to include
sociolinguistics, linguistic anthropology, discourse analysis, conversation analysis
and sociology of language, as well as certain streams of social psychology,
folklore studies, media studies, literary theory, and the philosophy of language.
What follows is a look at some foundational studies of code switching in
linguistic anthropology, sociolinguistics, and sociology, as well as a brief survey
of recent work on the topic within sociocultural linguistics. Finally (section 4), I
will suggest an operative definition for the term code switching that I hope will
serve as a basis and context for sociocultural discussions of the contextualizing
functions of language alternation and modulation.
5
2. Foundational Studies
The history of code switching research in sociocultural linguistics is often
dated from Blom and Gumperz’s (1972) “Social meaning in linguistic structures”
(e.g. Myers-Scotton 1993b; Rampton 1995; Benson 2001). This work is certainly
important and influential, not least for introducing the terms situational and
metaphorical switching (see below). However, by 1972 the term code switching
was well attested in the literature, and several studies in linguistic anthropology
and sociolinguistics prefigured later code switching research in sociocultural
linguistics. Section 2.1 briefly reviews some of the important “pre-history” of
code switching research; that is, early studies of the use of multiple grammatical
systems in discourse. Section 2.2 considers sociologist Erving Goffman’s notion
of footing, a theoretical construct closely related to (and influenced by) code
switching. Section 2.3 returns to Gumperz’s work on the subject, including both
work with Jan-Petter Blom, and later work on code switching and
contextualization.
2.1. Pre-history of code switching research

One of the earliest American studies3 in linguistic anthropology to deal
with issues of language choice and code switching was George Barker’s (1947)
3
Benson (2001) discusses work by Espinosa (1980 [1911]), a folklorist and Romance
dialectologist, as the earliest code switching research in the United States. Espinosa, who retired
from Stanford University in 1946, does not appear to have had a great influence on later code
switching research.
6
description of language use among Mexican Americans in Tucson, Arizona. In
addition to his analysis of the economic relations, social networks, and social
geography of Tucson residents, Barker sought to answer the question, “How does
it happen, for example, that among bilinguals, the ancestral language will be used
on one occasion and English on another, and that on certain occasions bilinguals
will alternate, without apparent cause, from one language to another?” (Barker
1947:185-86)
Barker suggested a continuum between intimate family relationships,
extending through increasingly public situations, to formal interaction with
English speaking neighbors. He found that interactions among family members
were most likely to be conducted in Spanish, while formal talk with Anglo-
Americans was most likely to use the medium of English (even when all parties in
the interaction were able to understand Spanish). At intermediate points along the
scale, language choice was less fixed, and elements from each language could
occur. Furthermore, Barker proposed that younger people were more apt to use
multiple languages in a single interaction than were their elders, and that the use
of multiple varieties was constitutive of a local Tucson identity.
Barker’s work suggested strong links between language and identity. For
example, Barker hypothesized, “for individuals both inside and outside the ethnic
group, the ethnic language comes to symbolize the group and its cultural
background, or in terms of its social function, to identify the group as a group”
(186). He therefore reasoned that the use of Mexican Spanish functioned as an
7
important marker of Mexican American identity for Tucson residents. At the same
time, residents used English to assert an American identity. Barker writes, “The
individual’s skill in using the language of a second or adopted culture comes to
symbolize his status in the new society” (187). Finally, Barker found that it was
common for younger people born in Tucson to use both English and Spanish
varieties within a single interaction.
In the field of informal relations among bilinguals, a field in which

most of the social life of the younger Mexican-Americans takes
place, we find that rapid shifting from one language to another is
common. Often two languages may be used in the same sentence
or phrase…. [I]t may be said that the mixing of the two languages
is indicative of the speaker’s participation in the urban social life
of the younger native-born group in the Tucson Mexican
community. [Barker 1947:195-96]
Barker’s study of the social functions of multiple linguistic varieties4 thus
anticipates later descriptions of code switching, social networks, and language and
identity.
An important base for code switching research in the field of linguistics is
Uriel Weinreich’s (1953) Languages in Contact, despite Weinreich’s caution
toward what he called bilingual speakers’ “switching faculty” (72). One of those
inspired by Weinreich’s book was Hans Vogt, whose “Language Contacts”
4
Barker identifies four varieties of Spanish (Southern Arizona, standard Mexican, Pachuco, and
Yaqui) and at least two varieties of English (labeled “standard” and “sub-standard”).
8
(1954) is cited as the first article to use the term “code-switching” in the field of
linguistics (Alvarez-Cáccamo 1998; Benson 2001).5
Weinreich was interested to describe the effect of language contact on
languages, in addition to describing the activities of bilingual speech
communities. He suggested that Barker’s (1947) description of Tucson was
insufficient, since it listed only four speech situations: intimate, informal, formal,
and inter-group discourse. Weinreich argued that Barker’s taxonomy was
“insufficiently articulated” (87) to describe all potential organizations of bilingual
speech events. Weinreich contended that anthropology should look to linguistics –
particularly to structuralism – in order to properly describe the practice of
bilingual speech, and the language acquisition/socialization process that takes
place in bilingual communities. At the same time, he suggested, “Linguists on
their part need the help of anthropology to describe and analyze those factors
governing linguistic interference which, though lying beyond the structure of the
languages in contact, do fall within the realm of culture” (6).
Weinreich’s description of switching codes6 suggests that bilingual
individuals possess two separate linguistic varieties, which (ideally) they employ
on separate occasions.
5
Benson actually cites Vogt’s review of Weinreich 1953 as the first piece to use the term “code-
switching.” Both the review and “Language Contacts” appeared in volume 10 of the journal Word
(1954).
9
The ideal bilingual switches from one language to the other

according to appropriate changes in the speech situation
(interlocutors, topics, etc.), but not in an unchanged situation, and
certainly not within a single sentence. If he does include
expressions from another language, he may mark them off
explicitly as ‘quotations’ by quotation marks in writing and by
special voice modifications (slight pause, change in tempo, and the
like) in speech. [Weinreich 1953: 73-74]
This description provides for situational switching (Blom and Gumperz 1972; see
below), the use of different language varieties in different situations. The
discussion of ‘quotations’ also prefigures the notion of footing (Goffman 1979,
1981; see below), both in the suggestion that a change in language may bring
other voices into a speech event7, and in the association between code switching
and other (mostly phonological) signals of changes of stance. However, it
specifically rules out conversational code switching (Gumperz 1982), and
especially intrasentential switching, for “ideal bilinguals.”
6
The notion of “switching codes” appears to have been borrowed from information theory.
Weinreich refers to Fano 1950, a paper also referenced by Jakobson (1971a [1953], 1971b [1961];
Jakobson and Halle 1956) in his discussions of code switching. Fuller exploration of these links is
unfortunately beyond the scope of the present paper. See Alvarez-Cáccamo (1998, 2000) for more
detail.
7
Several scholars (e.g. Hill 1985; Rampton 1995; Lo 1999; Pujolar 2001; Woolard 2004) have
discussed code switching in terms of voicing or heteroglossia (Bakhtin 1981). Woolard (2004)
suggests that Bakhtin’s socially oriented voicing may be more useful to understanding code
switching behavior than Goffman’s individually oriented footing (see below). She writes,
“Bakhtin’s approach better enables codeswitching analysts to articulate the linkages of linguistic
form, social context, macrosocial identity, and consciousness of all of these” (Woolard 2004:87).
Consideration of Bakhtin’s analyses of literary code switching is, however, beyond the scope of
this paper.
10
The “ideal bilingual” Weinreich refers to appears to be a rather speculative
figure, based on expectations or preconceived notions of bilingual behavior, rather
than empirical observations. Weinreich “suspects” that individuals may differ
from this ideal. He suggests that individuals who do alternate languages “in early
childhood, were addressed by the same familiar interlocutors indiscriminately in
both languages” (74).8 He provides no evidence, however, for the ubiquity, let
alone the propriety, of such “indiscriminate” language alternation. He also
predicts (though, again, with some disapproval) that the degree of switching may
differ among different societies. “If excessive switching should be demonstrated
to be the result of too early and unspecialized use of two languages,” a footnote
reads, “the possibility of social causation is all the more far-reaching” (83).
Vogt’s (1954) article, though very much inspired by Weinreich (1953), is
much less apprehensive about bilingual code switching.
Bilingualism is a universal phenomenon, since no languages we

know have been spoken over long periods of time in complete
isolation. It is even possible that bilingualism is one of the major
factors in linguistic changes—a point of view which could be
defended by good arguments. Code-switching in itself is perhaps
not a linguistic phenomenon, but rather a psychological one, and
its causes are obviously extra-linguistic. But bilingualism is of
great interest to the linguist because it is the condition of what has
been called interference between languages. [Vogt 1954:368]
8
For discussion of the one-person-one-language ideology in language acquisition see Romaine
(1989).
11
Vogt assumes that code switching is not only natural, but common. He suggests
that all languages – if not all language users – experience language contact, and
that contact phenomena are an important element of language change.
Vogt also argues that variation must be common within a language, since
languages change over time.
What therefore in the history of a linguistic system appears as a

change will in a synchronic description appear as more or less free
variation between different forms of expression, equally admissible
within the system. [Vogt 1954:367]
Vogt is not specifically interested in the distribution of variables among
individual language users, an aspect of variation and change that would later
fascinate sociolinguists (e.g. Labov 1963). He is, however, interested in the effect
that different degrees of code switching practiced by different speakers have on
language change. Thus Vogt foreshadows sociolinguistic studies of code
switching.
The phenomenon of diglossia, described by Charles Ferguson in the
journal Word (1959) is another precursor to linguistic analyses of code switching.
Ferguson’s notion of superposition influenced Gumperz’s work on dialect
variation and networks. (See Ferguson and Gumperz 1960; Gumperz 1961.
Gumperz 1982 draws a distinction between diglossia and conversational code
switching.) Ferguson’s definition describes two varieties of a language having
very different distributions within a community of language users.
Diglossia is a relatively stable language situation in which, in

addition to the primary dialects of the language (which may
include a standard or regional standards), there is a very divergent,
12
highly codified (often grammatically more complex) superposed

variety, the vehicle of a large and respected body of written
literature, either of an earlier period or in another speech
community, which is learned largely by formal education and is
used for most written and formal spoken purposes but is not used
by any section of the community for ordinary conversation.
[Ferguson 1959:336]
In the four “defining languages” Ferguson describes (Arabic, Swiss German,
Haitian Creole, and Greek), an H (high) variety is superposed above standard or
regional varieties, labeled L (low). While L is acquired naturally and used for
many daily interactions, H is learned via formal education and restricted to
formal, especially written usage. “In one set of situations, only H is appropriate
and in another only L, with the two sets overlapping only very slightly” (Ferguson
1959:328). Diglossia as defined by Ferguson thus does not allow for
conversational code switching – the alternating use of multiple varieties within a
single conversation. It is, however, reminiscent of situational switching (Blom &
Gumperz 1972; see below).
Although Ferguson limits diglossia to varieties of the same language,
subsequent scholars have discussed diglossia in situations where unrelated
languages comprise the H and L varieties, most notably Fishman (1967). Fishman
seeks to tease apart diglossia from bilingualism. He describes bilingualism as a
characteristic of individuals, “whereas diglossia is a characterization of linguistic
organization at the socio-cultural level” (1967:34). Fishman describes functional
divisions between unrelated languages, such as Guarani (L) and Spanish (H) in
13
Paraguay, which are very similar to the Haitian Creole (L)-French (H) division
described by Ferguson (1959).
Neither Ferguson (1959) nor Fishman (1967) cite examples of alternation
between varieties within a single interaction or discourse. However, as mentioned
above, their descriptions of diglossia bear on the notion of situational switching.
Further, Fishman, citing an unpublished paper by Blom and Gumperz, mentions
that varieties may be employed for humor or emphasis in a process of
metaphorical switching (Fishman 1967:36). Thus, Fishman’s account of diglossia
at least seems to have been inspired by the nascent theory of situational and
metaphorical switching (Blom and Gumperz 1972; see below).9
2.2. Goffman: Footing and code switching

Erving Goffman (1979, 1981) described footing as a process in interaction
similar to some functional descriptions of code switching. Indeed, Goffman cites
several of Gumperz’s descriptions of code switching as examples of footing. The
difference he draws between his own theory of footing and Gumperz’s and others’
descriptions of code switching is a formal one. Whereas code switching (at least
9
Fishman also credits Gumperz for expanding the notion of diglossia to include multilingual
societies. However, at least one of the studies Fishman cites as a description of diglossia
(Gumperz 1964a) was labeled by Gumperz as a study of code switching.
14
for Goffman) necessarily includes a shift from one language to another,10 footing
shifts may also be indicated by changes in prosody, pitch, body alignment, etc.
Even so, Goffman writes, “For speakers, code switching is usually involved” in
footing shifts, “and if not this then at least the sound markers that linguists study:
pitch, volume, rhythm, stress, tonal quality” (Goffman 1981:128).
For Goffman, footing is the stance or positioning that an individual takes
within an interaction. Goffman explicates these stances primarily by unpacking
the notions of speaker and hearer, and also by exploring the multifaceted
character of the speech event. Within a single interaction – even within a short
span of talk – an individual can highlight any number of different roles. Goffman
suggests that changes in purpose, context, and participant role are common in
interaction, and offers footing as a useful theory of the multiple positions taken by
parties to talk in interaction. During the course of an interaction, an individual is
likely to display a number of different stances; much of Goffman’s discussion of
footing is thus dedicated to switches in footing. Alternating languages, among
other linguistic markers, can serve to mark these shifts in footing.
Scholars of bilingualism and code switching have taken up Goffman’s
notions of footing as an explanation or extension of code switching in at least two
10
It is far from clear that early code switching research assumed such strict separation of
languages. Blom and Gumperz 1972, for example, focus on two dialects of spoken Norwegian
(dialects, moreover, which Maehlum (1996) suggests cannot be strictly separated). Similarly,
Fishman states explicitly, “A theory [of diglossia] which tends to minimize the distinction
between languages and varieties is desirable for several reasons” (1967:33).
15
ways. For Zentella (1997), footing provides a functional rationale for much of the
code switching she sees practiced by Puerto Rican children in New York.
Goffman’s concept of footing provides the principle that underlies

a broad variety of switches…. The children of el bloque used code
switching primarily to signal a change in footing via two
approaches; they switched languages to underscore or highlight the
re-alignment they intended (Realignment), or to control their
interlocutor’s behavior (Appeal/Control). [Zentella 1997:93]
Footing is thus a (functional) explanation for the (formal) description of language
alternation; this matching of form and function constitutes code switching. Most
instances of code switching can be seen as language alternation11 which signals a
change in footing. Zentella thus makes footing part of the definition of code
switching.
In a slightly different, but certainly related vein, Woolard (2004) sees
footing as superordinate to code switching. She suggests that footing is a bridge
between the specific study of code switching and broader understandings of
related linguistic phenomena. The description she provides of footing relates it
both to Gumperz’s (1982) discussion of contextualization cues and to Myers-
Scotton’s (1993b) rights-and-obligations sets (see below). Woolard sees footing
as similar to, but distinct from discussions of code switching, and suggests that
recognition of the nuanced roles that footing deals with will enrich discussions of
11
Zentella describes seven language varieties: Standard Puerto Rican Spanish, Nonstandard Puerto
Rican Spanish, English-dominant Spanish, Standard NYC English, African American Vernacular
English, Puerto Rican English, and Hispanized English.
16
code switching beyond Gumperz’s or Myers-Scotton’s models. I will suggest
below that code switching research in sociocultural linguistics has begun to step
beyond static models in just this way.
2.3. Gumperz: Code switching and contextualization

Perhaps no sociocultural linguist has been more influential in the study of
code switching than John J. Gumperz. His notions of code switching developed
alongside the emerging ethnography of communication model (Hymes and
Gumperz 1964; Gumperz and Hymes 1972), and his work has been influential in
the fields of sociolinguistics, linguistic anthropology, and the sociology of
language.
Much of Gumperz’s early work was carried out in northern India
(Gumperz 1958, 1961, 1964a, 1964b), 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). Thus the relationship between speakers
affects the choice of language variety. Although his primary interest in this study
is to map particular formal (especially phonological) elements of speech to
particular social categories (especially caste), Gumperz recognizes that different
aspects of identity become salient in different speech situations, and thus affect
17
the form of language used. Subsequent studies (Ferguson and Gumperz 1960;
Gumperz 1961) expanded on the relationship between linguistic form and social
positioning. Just as the relationship between speakers was found to affect
language variety, other aspects of the speech situation, such as topic and setting,
were reflected in formal aspects of language use.
Linguistic features characteristic of an argot are generally

recognized as such within the speech community. As a matter of
fact, they are the signals by which natives judge and receive
advance information about the nature of a communication
situation. [Gumperz 1961:985]
Gumperz was critical of static descriptions of a “linguistic and cultural whole”
(Kroeber 1939, qtd in Gumperz 1961) that ignored variation among individuals
and across social situations. This interest in formal variation and its effect on the
communicative event would serve as a basis for descriptions of code switching as
well as contextualization.
The idea that linguistic form is affected by setting and participants as well
as topic was emergent during this period. Ervin-Tripp’s (1964) 12 definitions of
setting, topic, and function provide an important base for the work of Gumperz
and others. Her study of bilingual Japanese-born women living in the United
12
Dil (1971) suggests that Gumperz, Hymes and Ervin-Tripp deserve joint credit for development
of the ethnography of communication. The three were co-organizers of the meeting that
occasioned the papers collected by Hymes and Gumperz 1964.
18
States observed considerable correlation between language choice and discourse
content, providing an example of “semantic” analysis of language choice that,
while influential (e.g. Myers-Scotton 1993b), would be criticized as only partial
and approximate (e.g. Auer 1984, 1995). Ervin-Tripp also noted that topic and
participants’ identities could affect form (including style, register, and
sociolinguistic variants). This work influenced scholars of language and social
interaction, including Gumperz (1964b; Blom and Gumperz 1972).
In 1963, while working with the Institute of Sociology at Oslo University,
Gumperz met Jan-Petter Blom (Dil 1971). Together, Blom and Gumperz
undertook a study of verbal behavior in Hemnesberget, a small settlement of
about 1,300 people in Northern Norway. Gumperz (1964b) compared the use of
two dialects, standard literary Bokmål and local Ranamål, in Hemnesberget 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” (148) –
barriers of caste, class, and village groupings in India, and of academic,
administrative, or religious setting in Norway. On the basis of these comparisons,
Gumperz argued that verbal repertoire is definable in social as well as linguistic
terms. Distinct repertoires are identified in terms of participants, setting, and
topic, and then described in terms of phonological and morphological
characteristics.
19
Blom and Gumperz (1972) expanded the analysis of the functions of
Bokmål and Ranamål in Hemnesberget in what has come to be a touchstone in
code switching research. They described Bokmål and Ranamål as distinct codes,
though not distinct languages. The two codes are mutually intelligible, and are
both regarded by the residents of Hemnesberget as acceptable forms of
Norwegian. The codes are distinguished by extensive though slight phonological,
morphological and lexical differences, as well as native speakers’ belief that the
two varieties are separate, and tendency to maintain that separation of form.
Choice among these [phonological, morphological, and lexical]

variables is always restricted by sociolinguistic selection
constraints such that if, for instance, a person selects a standard
morphological variant in one part of an utterance, this first choice
also implies selection of pronunciation variables tending toward
the standard end of the scale. A speaker wishing to ask for
another’s place of residence may, e.g., start his sentence either with
(R) [Ranamål] ke “where” or (B) [Bokmål] vor. In the first case,
the rest of the sentence will read hanj e ifrå “is he from?” In the
second case, it will be ær hann fra; vor and hanj do not co-occur.
[Blom and Gumperz 1972:416]
Blom and Gumperz asked why, despite their substantial similarities, and the fact
that most speakers commanded both varieties, the two varieties were largely
maintained as separate. “The most reasonable assumption,” they argued, “is that
the linguistic separateness between dialect and standard… is conditioned by social
factors” (417). Thus, each variety was seen as having low-level differences in
form, as well as somewhat distinct social functions.
Blom and Gumperz posited that social events, defined in terms of
participants, setting, and topic, “restrict the selection of linguistic variables” (421)
20
in a manner that is somewhat analogous to syntactic or semantic restrictions. That
is, in particular social situations, some linguistic forms may be more appropriate
than others. Among groups of men greeting each other in workshops along the
fjord, the variety of language used differed from that used by teachers presenting
text material in the public school, for example. It is important to recognize that
different social events may, for example, involve the same participants in the
same setting when the topic shifts. Thus, teachers reported that they treated the
shift from lecture to discussion within a class as different events. While lectures
were (according to teachers’ reports) delivered in the standard Bokmål, a shift to
the regional Ranamål was used to encourage open debate. Blom and Gumperz call
this type of shift, wherein a change in linguistic form represents a changed social
setting, situational switching (424).
It is worth noting that, although many scholars (e.g., Myers-Scotton
1993b; see below) have seized upon Blom and Gumperz’s seemingly
deterministic suggestion that language form indicates “participant’s definition of
each other’s rights and obligations” (1972:424), Blom and Gumperz’s description
of situational switching allows for considerable variation for different speakers,
settings, and types of events. The definition of metaphorical switching relies on
the use of two language varieties within a single social setting. Blom and
Gumperz describe interactions between clerks and residents in the community
administration office wherein greetings take place in the local dialect, but
business is transacted in the standard.
21
In neither of these cases is there any significant change in

definition of participants’ mutual rights and obligations. The
posture of speakers and channel cues of their speech remain the
same… The choice of either (R) or (B)… generates meanings
which are quite similar to those conveyed by the alternation
between ty and vy in the examples from Russian literature cited by
Friedrich [1972]. We will use the term metaphorical switching for
this phenomenon. [Blom and Gumperz 1972:425]
Blom and Gumperz suggest that the use of local (R) phrases in a standard (B)
conversation allude to other social events in which the participants may have been
involved. This allusion lends some connotative meaning, such as confidentiality,
to the current event, without changing the topic or goal.
The notions of situational and metaphorical switching were taken up by a
great many sociolinguists, linguistic anthropologists, etc. Whereas Blom and
Gumperz identified Ranamål and Bokmål as “codes in a repertoire” (414) and
went to some pains to describe the formal differences between the two, though,
many subsequent scholars have been content to equate code with language, and
focus their analyses on either functional distributions, or the definition of
situations. Consider the following from Heller (1988a).
They [Blom and Gumperz (1972)] proposed a basic type of code-

switching, situational codeswitching, which is rooted in a social
separation of activities (and associated role relationships), each of
which is conventionally linked to the use of one of the languages
or varieties in the community repertoire. [Heller 1988a:5]
In addition to shifting Blom and Gumperz’s situational switching to situational
codeswitching, this description equates code with languages or varieties. Heller is
far from unique in this respect. Compare Bailey’s (1999) definition of this key
concept in linguistic anthropology: “Code-switching is the use of two or more
22
languages in one speech exchange by bi- or multilingual speakers” (241). Rather
than defining or describing the codes of the speakers they study, many
contemporary scholars seem content to treat code switching as transparent, and to
focus on various functions.
The variety of understandings of metaphorical switching is no doubt
attributable to the brevity of Blom and Gumperz’s description of the phenomenon.
In three short paragraphs, they suggest that metaphorical switching “gives… some
of the flavor” of other situations, adds “special social meaning,” and functions
similarly to ty/vy (familiar versus formal second person reference) alternation. In
the one example where they contrast metaphorical and situational switching, an
alternative analysis seems possible. While they do not transcribe actual speech,
Blom and Gumperz describe a situation wherein a local resident and an
administrative clerk have conducted some business using “constant alternation
between the standard and the dialect during their business transaction”
(1972:425). This, according to Blom and Gumperz is clearly metaphorical
switching. Near the end of the interaction, the local asked the clerk – using the
local dialect – whether he had time to step aside to discuss another matter. Blom
and Gumperz classify this request as a situational shift. However, given the lack
of specifics, we have no way of knowing what particular effects the “constant
alternation” had in the “business transaction,” and thus no way to judge how what
is described as metaphorical versus situational switching actually functioned on
this occasion. If, for example, as Blom and Gumperz describe elsewhere, the two
23
speakers greeted one another in Ranamål, asked and answered questions in
Bokmål, and then concluded the conversation in Ranamål, we might identify
much more particular functions for each switch. Such a detailed description may
well call into question the general description of situational versus metaphorical
switching.
Critics have pointed out that Blom and Gumperz (1972) provide scant
detail of actual language use in their description of the verbal repertoire of
Hemnesberget. Maehlum (1996) is particularly critical of the suggestion that
Bokmål and Ranamål comprise separate codes. She argues that, in other rural
areas of Norway, local and standard dialects are not nearly as discrete as Blom
and Gumperz suggest. Thus, any suggestion that the verbal repertoire of
Norwegian speakers is comprised by two distinct codes is flawed.13 Maehlum
suggests that “local” and “standard” exist not as empirically identifiable, discrete
codes, but “as idealized entities: it is their existence as norms which is important”
(1996:753, original italics). Further, certain phonological or lexical/morphological
variables are particularly salient as indicators of particular dialects. This suggests
that sociolinguistic variants are available as indexes of various social meanings,
but that attempts to define particular codes, and thence the situations in which
they occur are problematic. It is perhaps preferable, then, to identify the formal
13
To be fair, however, Blom and Gumperz did not actually insist that the bipartite division they
suggest was a complete measure of the repertoire – merely an element of it.
24
signals of situation or identity available to a group of speakers, and the uses made
of these signals, rather than to assume a priori that dialects, varieties, or
languages will be equally salient across groups.
More than many subsequent scholars, Gumperz seems to have recognized
the imperfection of the description of switching as either situational or
metaphorical. By 1982, Gumperz’s preferred terminology was conversational
code switching. (The description and definition of conversational code switching
was, however, largely in terms of metaphorical switching.) Gumperz
acknowledged that it is generally difficult for analysts to identify particular
language choices as situational or metaphorical, and that native speakers generally
have few intuitions about or recognition of their own conversational code
switches. Except in cases of diglossia, the association between linguistic form and
settings, activities, or participants is highly variable, and rarely definable by static
models.
Since conversational code switching is not amenable to intuitive
methods14, and not strictly relatable to macro-sociological categories, Gumperz
(1982) argued that close analysis of brief spoken exchanges is necessary to
identify and describe the function of code switching. On the basis of his analyses
of Slovenian-German switching in Austria, English-Hindi switching in northern
25
India, and Spanish-English switching in the United States, Gumperz suggested a
list of six code switching functions which “holds across language situations” (75),
but is “by no means exhaustive” (81). Gumperz suggested quotation marking,
addressee specification, interjection, reiteration, message qualification, and
“personalization versus objectivization”15 (80) as common functions of
conversational code switching. It is noteworthy that the functions of code
switching that Gumperz identifies are quite similar to the contextualization cues
he describes elsewhere in the volume.16
Code switching signals contextual information equivalent to what

in monolingual settings is conveyed through prosody or other
syntactic or lexical processes. It generates the presuppositions in
terms of which the content of what is said is decoded. [Gumperz
1982:98]
Like other contextualization cues, language alternation may provide a means for
speakers to signal how utterances are to be interpreted, information beyond
referential content. Compare this description of code switching to Gumperz’s
definition of contextualization cues:
14
Gumperz (1982:62) points out that both subjects in Hemnesberget and Spanish-English
bilinguals in the United States denied any alternation of linguistic form, but even after listening to
recordings of themselves and “promising” to refrain from switching, persisted in code switching.
15
The category of “personalization versus objectivization” is somewhat fuzzy, but relates to
illocutionary force, evidentiality, and speaker positioning. According to Gumperz, “The code
contrast here seems to relate to things such as: the distinction between talk about action and talk as
action, the degree of speaker involvement in, or distance from, a message, whether a statement
reflects personal opinion or knowledge, whether it refers to specific instances or has the authority
of generally known fact” (1982:80).
16
Gumperz, it may be said, makes the comparison in reverse. His discussion of contextualization
conventions (Gumperz 1982, chapter 6) says that they are “meaningful in the same sense that…
the metaphorical code switching of chapter 4 [is] meaningful” (139).
26
Constellations of surface features of message form are the means

by which speakers signal and listeners interpret what the activity
is, how semantic content is to be understood and how each
sentence relates to what precedes or follows. These features are
referred to as contextualization cues. [Gumperz 1982:131, original
italics]
The meanings signaled and interpreted via code switching are, though, like all
contextualization cues, highly local and negotiated by speakers and listeners in
interaction.
Gumperz’s list of code switching functions inspired many subsequent
scholars to refine or propose their own lists of functions (e.g. McClure and
McClure 1988; Romaine 1989; Nishimura 1997; Zentella 1997). However, as
Auer (1995) suggests, the functions suggested by such lists are often ill defined.
The oft-cited category of reiteration, for example, fails to define exactly what is
repeated, or why. “A more in-depth sequential study… would make it clear that
this category [reiteration] subsumes a number of very different conversational
structures” (Auer 1995:120). Lists also tend to combine linguistic structures (such
as interjection) and pragmatic or conversational functions (message qualification,
addressee specification) without attempting to trace the relationship between
forms and functions. Although such lists may provide a useful step in the
understanding of conversational code switching, they are far from a satisfactory
answer to the questions of why switching occurs as it does and what functions it
serves in conversation. Noting a number of studies that have, following Gumperz
(1982) suggested similar taxonomies of functions, Bailey notes, “The ease with
which such categories can be created – and discrepancies between the code
27
switching taxonomies at which researchers have arrived – hint at the
epistemological problems of such taxonomies” (2002:77). Code switching may
serve any of a number of functions in a particular interaction, and a single turn at
talk will likely have multiple effects. Therefore, any finite list of functions will be
more or less arbitrary. Again, the suggestion is that it will be preferable to observe
actual interaction, rather than starting from assumptions about the general effects
of code switching.
3. Sociocultural studies of code switching
This section of the paper will describe sociocultural work produced in the
past decade or so describing the effects, functions, and norms of code switching.
Code switching scholarship within sociocultural linguistics may be divided into
several (sometimes overlapping) streams. For the purposes of this paper, three
broad areas will be discussed. First is the social psychological approach of Myers-
Scotton’s markedness model (1983, 1993b, 1998) and related work. This work
attempts to define, in general terms, the psycho-social factors that motivate a
bilingual speaker to choose one language over another for a particular utterance or
stretch of discourse. This work assumes a more or less consistent relationship
between linguistic form and social meaning. It further assumes that individual
speakers are rational actors, whose linguistic choices are based on an analysis of
the benefits and costs of particular language behavior.
Another area of inquiry analyzes identity and code choice in linguistic
interaction. Like the markedness model approach, this work analyzes the
28
relationship between the use of particular linguistic forms and the social
positioning they index. Unlike that work, however, scholars make no assumptions
about overarching cognitive or social structures governing language behavior.
Analyses of code switching within language and identity studies approach
linguistic interaction in a way that is perhaps opposite to the social psychology
approach of the markedness model. That is, rather than beginning from a
hypothetical model of linguistic interaction, and supporting it with data from
discourse, these scholars start with speech data and attempt to describe the
behavior of individual actors and its consequences for interactants and the
communities they form.
Finally, a number of scholars discuss the effect of code switching on talk
in interaction. This work, based on close analysis of interaction, tends to view
code switching behavior both as a method of organizing conversational exchange
and as a way to make knowledge of the wider context in which conversation takes
place relevant to an ongoing interaction. Since this wider knowledge is usually
analyzable at least partially in terms of identity, the separation between what I
here call “interaction and code switching” versus “identity and code switching” is
neither absolute nor unambiguous. Indeed, the three-part division suggested here
should be seen as one of analytic convenience, rather than significant theoretical
import.
29
3.1. Myers-Scotton’s Markedness

Carol Myers-Scotton described her markedness model in the book Social
Motivations for Codeswitching: Evidence from Africa (1993b).17 Myers-Scotton
had worked primarily in Kenya and also in Zimbabwe, Tanzania, Uganda,
Nigeria, and Malawi since the late 1960s, describing both the structural and
sociological aspects of language use in multilingual communities. The year 1993
saw the publication of two books: Dueling Languages: Grammatical Structures in
Codeswitching (Myers-Scotton 1993a) 18 was focused on formal issues, such as
constraints on code switching. Social Motivations (1993b) sought to “explain the
socio-psychological motivations behind CS [code switching]” (3).
According to Myers-Scotton, each language in a multilingual community
is associated with particular social roles, which she calls “rights-and-obligations”
(RO) sets (84). By speaking a particular language, a participant signals her
understanding of the current situation, and particularly her relevant role within the
context. By using more than one language, speakers may initiate negotiation over
relevant social roles.
17
Myers-Scotton discussed similar issues and developed the markedness model in code choice
prior to the publication of this book (e.g. Myers-Scotton 1972, 1976, 1983). Myers-Scotton 1983
actually laid out the Negotiation Principle and six maxims, including the unmarked choice and
exploratory choice maxims that figure in the refined model. However, as the fullest expression of
the model, it is Myers-Scotton 1993b that has influenced much subsequent work.
18
Unlike most of scholars described below, Myers-Scotton is known not primarily for
sociocultural theory, but equally as well for her formal model of code switching. The matrix
language frame model (MLF; Myers-Scotton 1993a) has been very influential in formal studies of
language contact, borrowing, and code switching. Such formal/structural studies are beyond the
scope of this paper.
30
CS in general is a type of skilled performance with communicative

intent. From the socio-psychological point of view, CS can be
characterized as symptomatic either (a) of an unwillingness or an
uncertainty on the speaker's part regarding the commitment to
indexing any single rights-and-obligations set between participants
in a conversation, or (b) of a negotiation to change the rights-and-
obligations set. This is so because each linguistic variety used in
CS has socio-psychological associations, making it indexical of a
rights-and-obligations set. [Myers-Scotton 1993b:6-7, original
italics]
Myers-Scotton assumes that speakers must share, at least to some extent, an
understanding of the social meanings of each available code. If no such norms
existed, interlocutors would have no basis for understanding the significance of
particular code choices.
Myers-Scotton seems unconcerned with the definition of the term code. In
her preface she states, “The use of the term ‘code’ in ‘codeswitching’ is
traditional, and nothing more” (1993b:vii). While she seems interested in the
preface to contrast the intentional meaning she sees in code switching with
“decipherable” meanings implied by the word code, she appears not to be
interested in definitions of code versus language. Languages such as Swahili,
Shona, Ndebele, etc. are at times called codes.
According to the markedness model, language users understand the
indexical value of each ‘code’ in their repertoire. For each situation, there is an
expected or unmarked language, indexing the appropriate values for participant
relations. A speaker will generally choose the unmarked language, since it
expresses the expected relationship between speaker, hearer, and setting.
However, speakers analyze the potential risks and benefits of all other potential
31
choices, and might make a marked choice on the basis of this calculation. Myers-
Scotton cites Accommodation theory (Thakerar, Giles and Cheshire 1982; Giles
et al. 1987) and Brown and Levinson’s (1978; 1987) work on politeness in her
argument that speakers act rationally to minimize costs and maximize benefits
when they engage in conversation.
The markedness model is stated in the form of a principle and three
maxims. The negotiation principle, modeled on Grice’s (1975) cooperative
principle, presents the theory’s central claim.
Choose the form of your conversational contribution such that it

indexes the set of rights and obligations which you wish to be in
force between the speaker and addressee for the current exchange.
[Myers-Scotton 1993b:113, original italics]
Three maxims follow from this principle. The unmarked choice maxim directs,
“Make your code choice the unmarked index of the unmarked RO set in talk
exchanges when you wish to establish or affirm that RO set” (114). The marked
choice maxim directs, “Make a marked code choice…when you wish to establish
a new RO set as unmarked for the current exchange” (131). The exploratory
choice maxim states, “When an unmarked choice is not clear, use CS [code
switching] to make alternate exploratory choices as candidates for an unmarked
choice and thereby as an index of an RO set which you favor” (142). Thus, the
social meanings of language (code) choice, as well as the causes of alternation,
are defined entirely in terms of participant rights and obligations.
While this description suggests a relatively inflexible arrangement of
language and identity mappings, Myers-Scotton argues that the model allows for
32
“dynamic variability” (1993b:84). Although the mapping between a code and the
social categories it indexes is fairly static, the salience of particular social
categories may vary from one community to another. Similarly, a category may
be salient for some types of interactions, but less salient for others. Thus,
determining which code is unmarked for a particular interaction requires attention
to the salience of all potential elements of identity, situation, and purpose.
Myers-Scotton asserts that in the following example, which takes place in
rural western Kenya, the local language, Lwidakho, is the unmarked choice. One
speaker’s use of Swahili and English are thus marked choices. (In this example,
Lwidakho is in ordinary font face, Swahili in italics, and English in bold.)
1 FARMER: Khu inzi khuli menyi As I live here, I have

2 hanu inzala- hunger
3 WORKER: Njaa gani? What kind of hunger?
4 F: Yenya khunzirila hanu- It wants to kill me here
5 W: Njaa gani? What kind of hunger?
6 F: Vana veru- our children
7 W: Nakuuliza, njaa gani? I ask you, what kind of hunger?
8 F: Inzala ya mapesa, kambuli. Hunger for money; I don’t have any.
9 W: You have got a land. Una You have got land. You have land.
10 shamba. Uli nu mulimi. You have land.
11 F: Mwan mweru- My brother,
12 W: Mbula tsisendi. Can’t you see I don’t have money. Can’t you see
13 how I am heavily loaded? how I am heavily loaded?
[Adapted from Myers-Scotton 1993b:82-83]
According to Myers-Scotton, the use of Swahili and English in this example has
shock value because they are marked choices for the occasion. Swahili is
associated with out-group encounters; since these two men live in the same region
and (she assumes) speak the same native language, Lwidakho would be the
33
unmarked choice. By choosing marked codes, the worker seeks to establish a new
rights-and-obligations set relative to his interlocutor. By speaking out-group
languages, the worker denies his in-group obligation to his neighbor, including
any obligation to give him money.
Some critics of the markedness model argue that it relies too heavily on
external knowledge, including assumptions about what speakers understand and
believe. The preceding example relies on the analyst’s understanding of the
meaning of Lwidakho, Swahili, and English for these speakers. Auer (1998)
reanalyzes this excerpt from a conversation analysis perspective (though, since he
is working from Myers-Scotton’s transcripts and has no access to more detailed
recordings, one might argue that he is not doing conversation analysis per se). He
analyzes the worker’s use of Swahili at lines 3, 5, and 7 as clarification requests,
separated from the indirect request/decline that is carried out in Lwidakho. Auer
thus argues that it is possible to account for code switching behavior without
appeal to “conversation-external knowledge about language use” (Auer 1998:10)
required by the markedness model. Of course, it is possible for the analyst to learn
which languages are typically used in particular situations via, for example,
ethnographic observation. Furthermore, one can argue that speakers learn these
norms as part of the language socialization process – just as they learn how to
make clarification requests and other insert sequences. A stronger criticism
remains, however: The markedness model requires the analyst to make
assumptions about each individual speaker’s knowledge and understanding of the
34
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.
It is possible to counter this internal-states critique by arguing that the
markedness model is not intended as a description of individual speakers in
particular conversations (despite the repeated references to individual speakers in
the formulation of the model), but as a model of social or structural norms. It is
certainly not uncommon for such normative models to take the form of
admonitions to individual speakers, while actually intended to represent general
social or cognitive standards (cf. Grice 1975, “Make your conversational
contribution such as is required, at the stage at which it occurs, by the accepted
purpose or direction of the talk exchange in which you are engaged” (45)).
However, as Auer (1995) points out, empirical studies have failed to reveal the
strong correlations between particular languages and speech activities that the
markedness model predicts. “Many speech activities are not tied to one particular
language, and even among those which have a tendency to be realized more often
in one language than in another, the correlation is never strong enough to predict
language choice in more than a probabilistic way” (Auer 1995:118).
Another critique of the markedness model, expressed by Woolard (2004),
is its poor definition of markedness. Myers-Scotton mentions the Prague School,
and particularly Jakobson and Trubetzkoy as influences on her notion of
35
markedness (though neither scholar is quoted, nor listed as a reference), but her
use of markedness seems to differ from those scholars.
Trubetzkoy first defined markedness in phonology: within a phonological
system, there are often pairs of phonemes differentiated by some feature, or
marking (e.g. voiced, round, nasal, etc.). This marked feature often appears on
several pairs of phonemes. In situations where the phonemic contrast is
neutralized, it is the unmarked variant which occurs. Thus, the marked form is
meaningful (in that it shows phonemic contrast), while the unmarked form is
neutral. Jakobson and others in the Prague school expanded this notion of
markedness to morphology and semantics. According to Jakobson,
In grammatical oppositions, the distinction between marked and

unmarked terms lies in the area of the general meaning of each of
the juxtaposed forms. The general meaning of the marked term is
characterized by the conveyance of more precise, specific, and
additional information than the unmarked term provides. [Jakobson
1990: 138]
Thus, we should expect the marked form in any opposition to have a specific
meaning not specified in the unmarked form. Within Myers-Scotton’s markedness
model, however, the unmarked code choice conveys information just as precise
and specific as a marked choice, since code choice indexes the speaker’s rights
and obligations, and thereby the situation of the interaction.19
19
One might go so far as to say that the unmarked choice is actually more specific, since it indexes
a specific situation, while a marked choice may be exploratory, not indexing a situation.
36
Myers-Scotton describes unmarked elements as “generally simpler” and
more frequent (1993b:80). However, the markedness model describes all possible
language choices as equally meaningful, though mapped to different sets of
indexes. Woolard charges,
Myers-Scotton defines the unmarked code as the “expected

medium” in a particular type of conventionalized exchange
(1993[b]:89-90). It is expected because it has been used most
frequently in such contexts. The markedness model thus reduces
markedness directly to frequency. [Woolard 2004:80]
Myers-Scotton disputes Woolard’s claim that markedness is defined by frequency
in the markedness model. She claims that frequency is useful for identifying
unmarked forms, but does not define the unmarked choice. Rather, the unmarked
choice “[conforms] to prevailing community views for an appropriate RO set”
(Myers-Scotton, quoted in Woolard 2004:91). The basis for such community
views is not specified, however, and may indeed be reducible to frequency.
Despite these criticisms, the markedness model is probably the most
influential and most fully developed model of code switching motivations. Myers-
Scotton continues to refine the model in ways that are consistent with current
research on contact linguistics (Myers-Scotton 1998; Myers-Scotton and Bolonyai
2001) and the so-called standard theory (Chomsky 1965) of linguistics (e.g.
Myers-Scotton and Jake 2001; Jake, Myers-Scotton and Gross 2002).
3.2. Identity and code switching

Whereas the markedness model and subsequent work seeks to provide a
systematic and generalizable account of the process of code switching, much
37
work in linguistic anthropology, sociolinguistics, and other areas of sociocultural
linguistics provide interpretive and interactional understandings of code switching
in particular contexts. Although this school of sociocultural linguistics has
produced its share of broad theoretical work (e.g. Milroy and Muysken 1995;
Alvarez-Cáccamo 1998, 2000; Woolard 2004), it is generally more closely tied to
the observation of behavior in particular settings than to generally applicable
explanations of linguistic capability. Such studies stand as illustrations of the
place of code switching in particular social and historical settings, rather than as
models for a universal practice or potential (Heller 1992).
Monica Heller’s ethnographic observations and sociolinguistic study in
Quebec and Ontario have led her to consider the economics of bilingualism, and
to view code switching as a political strategy (Heller 1988b, 1992, 1995, 1999).
Her earlier work on the politics of conversational code switching (Heller 1988b)
was influenced by Gumperz (1982) and Myers-Scotton (1983), particularly in the
discussion of rights and obligations. Since languages tend to become associated
with idealized situations and groups of speakers, the use of multiple languages
“permits people to say and do, indeed to be two or more things where normally a
choice is expected” (Heller 1988b:93). This strategic ambiguity allows
anglophones in Quebec, for example, to achieve a position in francophone
controlled corporate culture, while still laying claim to an anglophone identity,
with its associated value on the international market. Uniting Bourdieu’s (1977)
concept of symbolic capital with Gumperz’s (1982) discussion of verbal
38
repertoires20, Heller (1992, 1995) 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. Thus, Heller is not primarily interested in the motivation
or ‘cause’ of switching, but its potential transformative effects. Although some
critics charge that Heller’s focus on two internationally dominant languages,
French and English, make her work difficult to generalize to other settings (Baker
2001), in many ways the focus on locally situated practices and refusal to over-
generalize lend strength to her theoretical arguments.
While Heller and others describe the relationship between language and
identity in economic or class terms, many scholars have focused on social
categories such as ethnicity. Rampton’s (1995) work on crossing, a type of code
switching practiced by speakers across boundaries of ethnicity, race, or language
‘community,’21 examines the language behavior of Asian, Afro-Caribbean, and
Anglo adolescents in ‘Ashmead,’ UK. Language varieties, Creole, Panjabi, and
stylized Asian English, typically associated with an ethnic group, are used by non-
members to accomplish complex functions. While Rampton does find some of the
language-crossing-as-mockery discussed in earlier accounts, crossing in various
20
Nor is Heller unique in bringing such an economic perspective to discourse strategies. Compare
Gal (1979, 1988), Woolard (1985), Hill (1985), et alia.
21
Code switching or crossing as a means to negotiate or comment on ethnic or racial identities is
also seen in the work of Nishimura (1992), Bucholtz (1999), Lo (1999), Jaffee (2000), Torras and
Gafaranga (2002), et alia.
39
directions also serves to forge a common adolescent group, to dissociate from
parents or elders, and to resist endemic stereotypes.
Rampton defines crossing in terms of metaphorical switching (Blom &
Gumperz 1972), but in so doing he complicates the notions of situational and
metaphorical switching, and of contextualization, considerably. He defines
situational switching as language alternation (Auer 1984) which accomplishes
contextualization (Gumperz 1982). Rampton reminds us that the boundaries of
metaphor are not clear cut (cf. Lakoff & Johnson 1980); similarly, metaphorical
and situational switching cannot be easily delimited. His primary interest, though,
is in “figurative” code alternation, a category which, for Rampton, is identical to
double voicing (Bakhtin 1981).
Figurative/metaphorical code alternation is the same thing as

double-voicing/double-languaging, and the theoretical apparatus
accompanying the latter provides a very useful way of thinking
through the different empirical forms that ‘metaphorical code
switching’ can take. [Rampton 1995:279-80].
Unlike situational switching, which Rampton argues simply replaces the current
situational frame with a new one, crossing adds additional contexts through which
an interaction must be interpreted. Rampton further divides figurative code
alternation into metaphorical switching (which Rampton says is identical to
Bakhtin’s uni-directional double voicing) and ironic code switching (which equals
vari-directional double voicing).
Rampton’s treatment of crossing acknowledges that the functions of any
particular instance of language alternation are likely to be complex and
40
ambiguous. In different social and linguistic contexts and among different
interlocutors, meanings are highly variable. Additionally, since crossing occurs
across ethnic or linguistic boundaries, its functions may differ from the in-group
switching most commonly studied by code switching researchers. Rampton’s
work thus serves as a call for more locally situated understandings of the
functions of discourse within particular settings and groups of speakers.
Issues of race, ethnicity, and crossing, as well as economic issues of class
and domination are prominent in Bailey’s (2001, 2002) work on language and
identity among Dominican Americans. His analysis of code switching is also
based in conversation analysis, much like the interactional work discussed below.
Bailey’s work focuses on Dominican American youth – young people born in the
United States to parents from the Dominican Republic – living in Providence,
Rhode Island. Dominican Americans, according to Bailey (2001, 2002) define
their ethnic affiliation as at once non-White and non-Black. That is to say, while,
like their African-American peers, Bailey’s subjects view themselves as outside
the dominant racial category “White,” they also reject identification with African
Americans based on phenotype or ancestry. Bailey’s adolescent subjects most
often identify themselves as Spanish, a category based on language, or Dominican
– a category that references national origin. Bailey’s subjects also disaffiliate at
times from first-generation immigrants from the Dominican Republic, especially
contrasting their own urban upbringing with a perceived rurality of recent
immigrants. In discourse, this complex identity is indexed by shifting uses of
41
nonstandard Dominican Spanish, Caribbean Spanish, African American
Vernacular English, and other nonstandard English varieties.
Bailey (2002) provides examples of code switching such as the following,
which he argues serve two functions, one interactional, and one identity-relevant.
(A Dominican American male, Woody, is walking toward the exit

of a classroom when Isabella addresses him.)
Isabella: Where you going?
Isabella: Huh?
→ Isabella: Eh /ẽ/? (Huh?)
Isabella: Woody!
[Bailey 2002:78]
According to Bailey, a switch from English to Spanish in Isabella’s third turn
serves to specify Woody as the intended recipient (a commonly discussed
function of code switching) when he does not respond to her. At the same time,
the use of nonstandard English and Spanish indexes Isabella’s Dominican
American identity, and her connection with Woody.
An obvious objection presents itself in discussing this extract. Namely,
how valid is it to call Isabella’s usage here “code switching”? After all, the turn
Bailey identifies as containing the code switch consists of a single, brief discourse
marker. Not only is it grammatically simple, it differs from the English discourse
marker, eh, only by its nasalization. However, if, as Bailey suggests, the
Dominican American cohort he has observed uses this particular form, it may
indeed serve as an index of the group. Further, following Maehlum’s (1996)
suggestion that certain phonological markers can serve as salient markers of a
42
language variety, a feature such as the nasalization of Isabella’s /ẽ/ may function
as both an index of category membership and a cue for contextualization.
These and many other studies of identity and code switching show that
close observation of discourse within narrowly defined communities of practice
can yield both empirically and theoretically rich understandings of the functions
of language variation in social interaction. By tying observations to particular
speakers and social actors, rather than moving too readily to discussions of
cultural or linguistic norms, scholars can come to detailed, reliable understandings
of the place of language in the construction and transmission of social traditions.
3.3. Interaction and code switching

Close observation of discourse is also a hallmark of interactional
linguistics, which seeks to understand “the way in which language figures in
everyday interaction and cognition” (Ochs, Schegloff and Thompson 1996:2).
These studies tend to be greatly inspired by conversation analysis, as well as
functional linguistics and linguistic anthropology. A number of studies under this
broad umbrella describe both the place of code switching in the language of turn
and sequence and the ways that language alternations, like other contextualization
cues, make broader contextual knowledge relevant to an ongoing discourse.
Auer’s 1984 Bilingual Conversation presented a pioneering study of
interaction and code switching. Auer argued that Gumperz’s conception of
situation is problematic, in that it is defined externally, and from the perspective
of the analyst. While Auer acknowledged that Gumperz’s own uses of situational
43
and metaphorical are less clear-cut that some scholars have taken them to be, he
nonetheless disapproved of the distinction.
[Based on Blom & Gumperz 1972] one would either have to

conclude that (in the situational case) code-switching is without
social meaning because it is a necessary consequence of certain
situational parameters, or that (in the metaphorical case) it is
dependent on an (almost) one-to-one-relationship between
language choice and situational parameters which can be
purposefully violated. [Auer 1984:4]
Far from pre-existing and determining language choice, Auer argues that situation
is created by talk in interaction. The form of each speaker’s utterances helps to
define the unfolding situation. Further, this negotiation itself has social meaning.
Auer (1984) is interested in language alternation generally. Language
alternation is further divided into transfer, language alternation for a structurally
bounded unit (e.g., a word, phrase, or sentence) followed by return to the original
language; and code-switching, or language alternation without a predictable point
of return to the first language. Speakers may use language alternation to
accomplish any number of conversational functions. What separates transfer from
code-switching is not any predetermined function(s), but their effect on
subsequent turns. “Whereas code-switching invites other parties to switch
languages as well ‘until further notice’” (Auer 1984:29), transfer has no such
consequences for other speakers. Instances of transfer tend to highlight their form,
including their language choice (e.g. German versus Italian) and their
grammatical form (e.g. word versus sentence). They are therefore useful for such
activities as quotation, translation, word search, etc. Code-switching, on the other
44
hand, does not highlight form in this way. The occurrence of ‘new’ language thus
is available to mark changes in footing, participant structure, or other elements of
context.22 The difference between code-switching and transfer, then, is not
determined primarily by the analyst, but by participants in talk in interaction.
Auer’s analyses of Italian migrant children in Germany did not find
significant correlation between topic and language form. He suggests that code
switching is not essentially ‘semantic’ in nature, not derived from the ‘meanings’
of the available languages, but rather is “embedded in the sequential development
of the conversation” (1984:93). Auer found a great preference for subsequent
speakers to maintain the language of the previous turn. Language alternation was
then available to mark contrast, either to bracket a sequence from the preceding
discourse (what Auer calls discourse related code-switching) or to negotiate (or
re-negotiate) a common language (participant related code-switching). Auer
recommended this procedural analysis of language alternation over individualistic
analyses based on introspection, or macro-sociological approaches that define the
meaning of potential language choices outside of actual language use.
Two main arguments are apparent in critiques of conversation analysis-
based discussions of code switching, like that pioneered by Auer (1984). First, in
22
Auer (1984) cites both Goffman’s (1979) work on footing and Gumperz’s (1982)
contextualization cues. However, Auer cautions against analyses that regard changes in footing or
context as causing code switching. Language alternation is available as a contextualization
strategy, but context or footing is created by participants’ joint efforts, which may or may not
make use of language alternation.
45
situations where there is no moment-to-moment unfolding of talk in interaction,
such as written discourse (Jonsson 2005), CA methods are not applicable. While
this is true, it seems trivial. It may limit the potential body of code switching data
somewhat, but does not seriously call into question any of the analyses or
conclusions of Auer’s or other interactional scholars’ work. A more serious
criticism is Stroud’s (1998) suggestion that CA approaches, by proscribing
argument from ethnographic or macro-sociological evidence, cannot provide
satisfactory analysis of language behavior in non-Western settings. Stroud
observes, “[L]anguage use and patterns of code-switching both structure and are
structured by indigenous cultural practices” (1998:322), a suggestion that most
sociocultural linguists would presumably more or less accept. If analysts then
ignore cultural information not visible (to them) within discourse data, their
analyses risk missing important elements of function and meaning. Stroud
maintains, “My argument is that conversational code-switching is so heavily
implicated in social life that it cannot really be understood apart from an
understanding of social phenomena” (1998:322). This vital understanding is often
provided by analysts’ focus on populations that they are themselves a part of;
however, it may also be desirable to undertake some broader examination of the
social context within which discourse takes place.
Several subsequent studies have examined sequential or interactional
functions of language alternation. Conversation analysts have suggested that code
switching may serve to enhance turn selection (Li Wei 1998; Cromdal 2001) or
46
soften refusals (Bani-Shoraka 2005; Li Wei 2005), and is a possible resource to
accomplish repair (Auer 1995; Sebba and Wooten 1998) or mark dispreferred23
responses (Li Wei 1998; Bani-Shoraka 2005). In addition to these interactional
functions, empirical studies have examined how switches in language variety
make particular elements of situation, speaker identities, or background relevant
to ongoing talk.
Sebba and Wooten (1998) describe alternation between London English
and London Jamaican (Sebba 1993). They suggest that interactional function is
insufficient to account for the code switching behavior they observe, and that
additional consideration of identity is necessary to enrich the overall analysis.
Sebba and Wooten (1998) suggest that the division of language varieties
into an out-group, hegemonic they-code and an in-group, solidarity building we-
code, as suggested by Gumperz (1982)24, is complicated for Caribbean Londoners,
since both London English and London Jamaican serve we-code functions.
Indeed, the strict identification of a single we-code is highly problematic since, in
23
In conversation analysis terms, responses which serve to accomplish the projected action of a
previous turn are generally considered preferred, while those that work against such
accomplishment are dispreferred. For further explanation, see Sacks, Schegloff and Jefferson 1974
and Hutchby and Woofit 1998.
24
As is the case for situational versus metaphorical switching and the enumerated functions of
conversational code switching, the we-code and they-code functions suggested by Gumperz as
rough preliminary categories have been reified as essential categories by later scholars, then
subsequently critiqued as rough and preliminary (cf. Gafaranga 2005).
47
most multilingual communities, most if not all available language varieties
function in some in-group interactions (cf. Rampton 1995; Bentahila and Davies
1995; Barker 1947). Sebba and Wooten describe specific interactional functions
of switching from London Jamaican to London English, and other, only partially
overlapping functions for switches from London English to London Jamaican.
They go on to argue that particular switches highlight participant identities, but
these ‘identities’ are locally constructed and highly variable. Their analysis shows
that macro-sociological categories such as race, gender, or ethnicity may be less
salient than particular positions within a current interaction.
Evidence from London Jamaican data suggests that in addition to

local, sequential explanations of code-switching, it is also
necessary to look at the interaction as a whole, as well as the wider
context in which it is located. … Interpretations must take into
account the shifting and negotiated nature of social identities
within talk as well as the values attached to the different codes by
their speakers. [Sebba and Wooten 1998:281]
Sebba and Wooten’s arguments suggest that the interactional functions and the
identity construction functions of code switching are linked. The marking or
creation of identity is a locally situated function of talk in interaction.
In a similar fashion, Gafaranga (2001, 2005; Torras and Gafaranga 2002)
argues that macro-sociological ideologies of identity provide an insufficient
account of either the orderliness of language alternation25 or the relevance of
25
Gafaranga prefers the term “language alternation” to “code switching,” since no consistent
definition of the latter term is shared across subdisciplines.
48
particular identities in bilingual talk in interaction. Further, Gafaranga (2005)
argues, contra Sebba and Wooten 199826, that language choice should not be seen
as an index of social structures. Rather, he argues, language is itself a social
structure.
Therefore, relating conversational structure and social structure in

bilingual conversation becomes a nonissue, as the conversational
structure is ‘occasioned’ by the language-defined social structure,
which is locally established in and through the conversational
structure. [Gafaranga 2005:298]
Language choice or code switching are occasioned by the demands of
conversation itself; attempts to relate this linguistic structure to an overarching
view of ‘culture’ or ‘society’ are misguided.
In place of a language-external theory of culture or society, Gafaranga
(2001) suggests that language alternation can best be explained via membership
category analysis (Sacks 1992).27 According to MCA, speakers may evoke certain
membership category devices – ordered collections of social relationships – when
such relationships become relevant for ongoing discourse. Membership category
devices are called into being via a set of application rules. When a hearer
understands an interlocutor’s talk to place speaker and hearer in a bound
relationship (e.g. student/teacher, friend/friend), she will generally respond by
placing herself in an appropriate position in relation to the implied categorization
26
In this respect, Gafaranga (2005) also opposes Gumperz (1982), Myers-Scotton (1993b), et alia.
27
Sacks first described MCA in his (1966) sociology dissertation at the University of California,
Berkeley. I have been unable to locate a copy of this dissertation.
49
device. Failure to do so is accountable (cf. dispreferred responses in conversation
analysis). Thus, according to Gafaranga, the use of more than one language
variety within a discourse by a single speaker is a means to negotiate a locally
relevant linguistic identity in relation to one’s interlocutor. Furthermore, the
meaning and function of any particular linguistic identity must be negotiated by
parties in interaction.
It is not clear, however, that Gafaranga’s use of MCA adds anything
significant to understandings of language behavior. If, as Gafaranga (2001)
argues, assuming a linguistic identity means something (i.e., has some
implications for ongoing talk), speakers must share some understanding of what
the identity entails. If speakers share extensive knowledge about such
implications prior to the talk at hand, MCA is open to the same types of criticisms
leveled at the markedness model (Myers-Scotton 1993b) and other theories of
broadly shared social norms. If, on the other hand, all meanings of linguistic
identity are negotiated in the talk at hand, language preference as a categorization
device adds little beyond what is explicable by conversation analysis; ‘doing
linguistic identity’ is isomorphic to positioning oneself in conversation.
Another scholar whose analyses are based in conversation analysis, Li
Wei reconciles sequential interaction not only with ‘linguistic identity’ in the
sense of locally constructed positioning (e.g. Li Wei 2002), but also with
sociological context (1994, 1998) and ‘identity’ (2005) in the sense of rights and
obligations (Myers-Scotton 1993b).
50
Li Wei has been critical of “analyst-oriented, theory-driven, top-down
approaches” (1998:157) to the analysis of code switching, including Blom and
Gumperz’s (1972) descriptions of situational and metaphorical switching, and
especially Myers-Scotton’s (1983, 1993b) markedness model. Li Wei says of the
markedness model:
This theory 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. [Li Wei
1998:157]
Li Wei, like Auer, Gafaranga, and other analysts inspired by ethno-
methodological approaches, rejects the notion that languages as codes28
transparently index prior social meaning. Nonetheless, he does not deny the
existence of social meanings. Rather, Li Wei argues that these meanings are not
merely ‘brought along’ to conversation as part of some social competence, but
must be ‘brought about’ through interaction (cf. Auer 1992). Li Wei, like other
CA analysts, insists that observations about social roles and language norms be
tied to close observation of discourse, rather than simply being posited on the
basis of analyst intuitions. Thus, where Milroy, Li Wei, and Moffat 1991 (see also
28
Li Wei does not, however, attempt to define code or language. What is objected to is the
suggestion that the ‘meaning’ of forms pre-exists interaction.
51
Milroy and Li Wei 1995) suggested that Chinese adults in Tyneside prefer to be
addressed in Cantonese, Li Wei (1998) grounds this observation in particular
discourse episodes. He notes a number of cases where Chinese children address a
parent in English, to be met with delays and requests for repair. Through these
responses, the parents not only show, buy also create the language preference
noted in the earlier studies.
Li Wei (2005) contrasts the strength of observations based on sequential
analysis with those based on rational-choice models, such as the markedness
model (Myers-Scotton 1983, 1993b) using examples such as the following.
1 B: mama women keneng qu MetroCentre kan dianying

2 A: xing xing xing
3 (1.2)
4 A: kan sheme [nimen zhedao ma youmeiyou hao kan de
5 B: [Jenny shuo ta xian da dianhuo wenwen
6 (0.6)
7 → B: can I have some money pleeease
8 → A: what for
9 → B: wo he Jenny qu MetroCentre kan dianying
1 B: We’ll probably go to the cinema at the MetroCentre.
2 A: Fine, fine, fine.
3
4 A: Do you know what to see? Is there anything nice to see?
5 B: Jenny says she’ll phone and ask the cinema first.
6
7 B: Can I have some money please?
8 A: What for?
9 B: Jenny and I are going to the cinema at the MetroCentre.
[Li Wei 2005:383]
The participants here, a Cantonese-English bilingual mother and daughter
in Tyneside, have been discussing (in Cantonese) the daughter’s plans for the
weekend. At line seven, the daughter switches to English to ask for money, and
then switches back to Cantonese at line nine to offer her reason for the request. A
52
rational choice model would suggest that the daughter assumes a marked code as
a means to ameliorate the fact that she is making a demand on her mother.
(Compare Myers-Scotton’s actual analysis of a Kenyan worker assuming a
marked choice to refuse a farmer’s request for money, presented above.)
However, Li Wei points out that there is much more going on here. The entire
episode up to the request can be seen as a pre-sequence to the request. The full
transcript includes a number of turns hearable as indirect requests, including line
one presented here. When the indirect requests fail, the daughter makes her
request directly, marking it off from the pre-sequence with a change in medium.
When the direct request meets with an indirect refusal (line eight), the daughter
again switches to Cantonese to provide a reason for the request. Li Wei points out
that the rational-choice analysis may tell us something interesting about the
exchange (that is, that the daughter ameliorates her request for money), but it is
the sequential analysis that gives fuller evidence for this activity, and thus a
warrant for the claim. Thus, rather than throwing out notions of rights and
obligations, Li Wei recommends complementing such analyses with close
observation of discourse. “To focus on the interaction-external factors alone
means ignoring the richness of the interactive work speakers do in conversation
and risks imposing of the analysts’ interpretation without evidence” (Li Wei
2005:387). On the other hand, close observation both strengthens analyses, and
leads to greater reliability.
53
3.4. Précis
This section has divided sociocultural studies of code switching into three
major areas. First, the markedness model (Myers-Scotton 1983, 1993b) and
similar work attempts to define social-psychological motivations for code
switching by treating speakers as rational actors who choose language varieties
based on their more or less fixed social meanings and functions. According to this
approach, a speaker chooses the language variety that best indexes the social role
(or rights-and-obligations set) that she wishes to make relevant at a given
moment. Code switching is defined as the use of more than language, with
“language” defined prior to interaction. Such alternation is generally used to
negotiate social role, or may be simply an unmarked norm for a group of language
users. This model is probably the most widely used and fully developed theory of
code switching in sociocultural linguistics. However, it is criticized for its loose
definition of “markedness,” its reliance on knowledge external to language
interaction and, perhaps most significantly, its lack of specific empirical warrant
for the analyst’s conclusions.
Second, work in linguistic anthropology, sociolinguistics, and allied fields
here labeled “identity and code switching” examines the correlation between
social position, group membership, or other forms of social identity and the use of
particular language varieties. Like the markedness model, this work tends to see
language use as indexical of social role. Unlike that model, however, language
and social role are generally seen as less fixed. While the work described under
54
this heading does not necessarily form a defined school within sociocultural
linguistics, it is possible to make some general comparisons between this strain of
scholarship and the markedness model. Principally, identity and code switching
work as defined here tends to be more empirical and less tied to linguistic or
psychological theory. Further, rather than attempting to describe motivations or
other individual causes of language behavior, this work tends to describe effects
of language choice on such areas as economic, ethnic, national and other social
arrangements. Thus, work of this type can yield rich description and sound
theoretical understanding of language behavior in particular social, geographic,
and historical settings. On the other hand, this work may be criticized for its lack
of generalizable conclusions.
Third, work influenced by conversation analysis, interactional linguistics,
and close analysis of discourse, here labeled as studies of “interaction and code
switching,” relate language behavior and code choice to both emergent discourse
structure and larger social context. This work calls into question studies which
assign a single, predetermined “meaning” to the choice of language variety
(particularly when “variety” is defined by the analyst prior to a specific
interaction). Scholars engaged in such work suggest that “codes” (in the sense of
socially meaningful linguistic varieties) as well as social arrangements emerge
from human interact and may be best understood by grounding analyses as closely
as possible in the particulars of linguistic interaction. Such work can both reveal
new insights about linguistic interaction, and yield empirical strength and greater
55
reliability to conclusions about social psychology or identity and related social
structures. While this work is solidly grounded in specific discourse events, it is
criticized for its failure to include (or in some cases, principled rejection of)
ethnographic information outside of this discourse interaction.
Certainly, all three of the major traditions discussed here have something
to add to the study of code switching. Conclusions such as those offered by the
markedness model offer insight into language practices generally; however, they
would unquestionably be strengthened by greater empirical evidence. It seems
that, in order for observations about language use, including variation or
alternation, to have validity and reliability, they should be based on close
observation of discourse. At the same time, it should not be assumed that all
elements relevant to discourse and social interaction are visible to the analyst,
particularly when the analyst is not embedded in the particular social structures he
or she is studying. We should remember Stroud’s (1998) suggestion that discourse
analysis be grounded in an understanding of the society within which
communication takes place. The optimal approach to understanding these
phenomena would thus seem to include ethnographic observation with close
analysis of discourse, providing an empirical warrant for any theory of discourse
interaction.
While the approaches described above all treat code switching or language
alternation, there is a general lack of precision in the way the key terms “code,”
56
“language,” and “code switching” are defined. The next section will suggest a
definition for the last of these, “code switching.”
4. Integrated definitions
A great many scholars in sociocultural linguistics use a definition of code
switching similar to Heller’s: “the use of more than one language in the course of
a single communicative episode” (1988a:1). Auer and Myers-Scotton, who
largely disagree on how or why code switching occurs, nonetheless sound quite
similar in their definitions of the phenomenon. Auer (1984:1) refers to “the
alternating use of more than one language,” while Myers-Scotton (1993b:vii)
mentions “the use of two or more languages in the same conversation.” Romaine
(1989) cites Gumperz as the source of this definition. However, these definitions
introduce an element not strictly present in Gumperz’s (1982) definition.
Conversational code switching can be defined as the juxtaposition

within the same speech exchange of passages of speech belonging
to two different grammatical systems or subsystems. Most
frequently the alternation takes the form of two subsequent
sentences, as when a speaker uses a second language either to
reiterate his message or to reply to some else’s statement.
[Gumperz 1982:59]
After a few examples of this “subsequent sentence” switching, Gumperz adds,
“Often code switching also takes place within a single sentence” (60), of which he
provides additional examples.
Note that Gumperz’s original definition refers to “grammatical systems or
subsystems,” while the subsequent restatements refer to languages. While the
former is scarcely more concrete or less ambiguous than the latter, it need not be
57
assumed that the two terms are identical. The plural languages seems to suggest
discrete varieties (as English, Spanish, Kiswahili, etc.), while the more equivocal
“systems or subsystems” might equally imply languages or elements of a
language, such as lexical items, syntactic constructions, and prosodic
phenomena.29 This list of grammatical subsystems may seem familiar, as it is very
similar to Goffman’s (1979) list of footing cues and virtually identical to
Gumperz’s (1982) preliminary list of contextualization cues.
The attempt to define language and languages is a perennial controversy
in linguistics. By defining code simply as a language (or variety of language)
without first defining these basic terms, scholars have essentially put off what
might be a foundational question. Alvarez-Cáccamo (1990, 1998, 2000) provides
exceptional attempts to define code and code switching. His discussion relies in
turn on work by Jakobson (1971b; Jakobson, Fant and Halle 1952, inter alia) and
by Gumperz (1982, 1992, inter alia). Alvarez-Cáccamo (1998) points out that for
Jakobson, an early adopter of the term code switching who was influenced by
information theory, languages have codes; they do not comprise codes. A
language user thus makes use of a code or codes when speaking, listening, etc.
The precise nature of any language user’s codes cannot be ascertained by an
analyst nor by fellow speakers.
29
Of course, in sidestepping the perennial problem of defining language, grammar, or other key
terms, I am no better than Gumperz, Auer, Heller, Myers-Scotton et alia.
58
Internal individual codes (senders’ and receivers’) must necessarily

differ, as they belong to different minds. But all human minds are
also uniquely alike: they produce language and communication,
which are formidably universal. Therefore, the question whether
each person possesses “different”... codes is parallel to the question
whether speakers of the “same” language share a grammar, or
whether culture, ideology, etc., is also shared. There are no
absolute answers to this, only a pragmatic one: does
communication between two persons sufficiently work? [Alvarez-
Cáccamo, personal communication]
Speakers use communicative codes in their attempts (linguistic or paralinguistic)
to communicate with other language users. Listeners use their own codes to make
sense of the communicative contributions of those they interact with. Listeners
may need to shift their expectations to come to a useful understanding of
speakers’ intentions. Similarly, speakers may switch the form of their
contributions in order to signal a change in situation, shifting relevance of social
roles, or alternate ways of understanding a conversational contribution. In other
words, switching codes is a means by which speakers may contextualize
communication.
A useful definition of code switching for sociocultural linguistic analysis
should recognize it as an alternation in the form of communication that signals a
context in which the linguistic contribution should be understood. The ‘context’
so signaled may be very local (such as the beginning of a turn at talk), very
general (such as positioning relative to some macro-sociological category), or
anywhere in between. Furthermore, it is important to recognize that this signaling
is accomplished by the action of a participant in a particular interaction. That is to
59
say, it is not necessary or desirable to spell out the meaning of particular code
switching behavior a priori. Rather, code switching is accomplished by parties in
interaction, and the meaning of their behavior emerges from the interaction. This
is not to say that the use of particular linguistic forms has no meaning, and that
speakers “make it up as they go.” Gumperz’s (1992) discussion of contextual-
ization is relevant here.
If instead of attempting to discover direct and stable linguistic

reflections of social categories in clause level phonology,
morphology or syntax we begin by looking more closely at the
clustering of co-occurring variables in situated everyday discourse,
in terms of what sorts of linguistic signs are involved and how they
are distributed, we soon discover regularities that are demonstrably
socially conditioned. [Gumperz 1992:40]
Individuals remember and can call on past experiences of discourse. These
memories form part of a language user’s understanding of discourse functions.
Therefore, within a community of discourse practice (cf. Lave & Wenger 1991;
Eckert & McConnell-Ginet 1992) certain forms may come to recur frequently in
certain contexts. Nonetheless, it is less interesting (for the current author at least,
and probably for the ends of sociocultural linguistic analysis) to track the
frequency or regularity of particular recurrences than to understand the effect of
linguistic form on discourse practice and emergent social meanings.
To recapitulate, then, code switching is a practice of parties in discourse to
signal changes in context by using alternate grammatical systems or subsystems,
or codes. The mental representation of these codes cannot be directly observed,
either by analysts or by parties in interaction. Rather, the analyst must observe
60
discourse itself, and recover the salience of a linguistic form as code from its
effect on discourse interaction.
A definition of code switching which collapses all functions into
contextualization appears to diverge from a robust and long-term trend among
sociocultural linguists. An enduring theoretical goal of code switching researchers
has been to describe different types of switching. Bailey (1999) recognizes three
types of code switching: situational switching, metaphorical switching, and
“unmarked discourse contextualization switching” (242). The first two categories
are, of course, drawn from Blom & Gumperz 1972. The latter term, while clearly
influenced by Myers-Scotton’s notion of markedness (see especially Myers-
Scotton 1988 for unmarked code switching), includes functions such as quotation
and reiteration, enumerated by Gumperz (1982) and many subsequent scholars.
Much current work on code switching lists similar typologies, or attempts to
catalog other functional categories (cf. Rampton 1995; Zentella 1997). However,
as Bailey (1999) points out, assigning a particular utterance to one of these
functional categories can be problematic. For example, when linguistic form
changes at a transition from small-talk to a formal speech, Bailey suggests, the
alternation may be a situational switch – since the change in form effects a change
in situation; a metaphorical shift – since the speaker’s relevant role may be
changed; or discourse contextualization – since it may signal the context in which
listeners should understand the speech. A single act may serve all of these
functions at the same time.
61
Attempts to define categories of function, whatever their possible
advantages to linguistic or social theory, remain problematic for concrete analysis.
Rather than artificially strengthening the boundaries between such functional
categories or suggesting ever widening catalogs, it seems preferable to recognize
that particular functions arise from actual interactions.
Two brief examples may serve to illustrate the suggested approach. In the
first, the occurrence of an English word in Japanese talk is not oriented to as
affecting context.
1 M: [ni-juu kyu doru kyu juu It’s twenty dollars and

2 kyu sento desu ninety-nine cents.
3 Y: ee? (2.0) huh?
4 Y: ni doru ha? Two dollars?
5 M: ha? ni-hyaku go juu huh? For two hundred
6 minutes de and fifty minutes
7 Y: nn yeah
8 M: ni-juu kyu doru kyu-juu twenty dollars and
9 kyu sento (°desu°) ninety-nine cents.
[Nilep 2004:4]
The speakers in this excerpt, both bilingual Japanese-English speakers (see Nilep
2004 for details), are conversing in Japanese. At least three forms, doru (dollar),
sento (cent), and minutes (minutes) seem to come from English. Attempts to
separate borrowing from code switching in formal terms tend to refer to levels of
linguistic signaling, such as phonology, morphology, syntax, etc. Gumperz
(1982), for example, claims, “New [borrowed] lexical and grammatical items
assimilate phonetically and rhythmically so that the total conversational effect is
that of an utterance as spoken in a single variety” (67-8). The phonetically and
rhythmically assimilated doru and sento should not, according to Gumperz’s
62
theory, be treated as code switching, and indeed they are not. However, neither is
the unassimilated minutes. According to Gumperz:
There are some marginal cases where phonologically

unassimilated items from a high prestige foreign language are
inserted as marked expressions into an otherwise monolingual
passage… The semantic effect here is similar to that of code
switching. For the most part, however, these are isolated cases.
They occur most frequently in formulaic expressions and this is
quite different from the constant alternation which marks code
switching in bilingual communities. [Gumperz 1982:68]
While English may be a prestigious foreign language, ni-hyaku go-juu minutes is
not a formulaic expression in Japanese. The attempt to spell out general formal
characteristics of code and the meaning or effect of code switching prior to actual
interaction thus fails. If, instead, we confine our observations to this interaction
and notice that the speakers do not treat minutes as different from doru in terms of
contextualization, we need not analyze it as code switching. (cf. Gafaranga 2001
on “language blind” analysis)
On the other hand, Alvarez-Cáccamo (1990) describes a case of “speaking
Galician in Spanish,” that is, making elements of Galician regional identity
relevant while using linguistic forms that are not obviously Galician.
1 P: .h.h y es uh es muy importante ↑

2 >que por ejemplo el tiro neumático< a nivel europeo ↑
3 esté incluído como deporte escolar,
4 en todos uh los: centros de enseñanza/
5 e sem embargo eiqui ↑
6 .h.h.h como somos assi tam es:peciales ↑
7 pos resulta que em galícia ↑
8 eh que hasta el tiro neumático siempre estaba dentro d- del
9 deporte escolar ↑
10 ahora pos lo han - lo han quitao ↑
63
1 And it is, uh, it is very important

2 that for example airgun shooting at the European level
3 is included as a school sport
4 in all, uh, the educational centers.
5 And yet here
6 since we are like peculiar
7 well, it turns out that in Galicia
8 uh, where airgun shooting had always been included as a school sport
9
10 well, now they’ve, they’ve done away with it.
[Alvarez-Cáccamo 1990:8]
The speaker, the president of the Galician Federation of Sport Shooting, is
being interviewed on Galician television. While the interviewer (not transcribed
here) speaks Galician, this speaker, in an official guise, speaks Spanish30 at the
beginning of this fragment. The speaker’s intended language is unclear, though, in
subsequent lines. With the exception of the nonstandard southern Galician eiqui
“here” in line five, lines 5-7 could, from a grammatical perspective, be seen as
either Galician or Spanish, and lines 8-10 seem Spanish. There is, however,
reason to regard lines 5-10 as belonging to the same context: all are delivered with
extra high pitch on the last lexical item, and all refer in some way to Galicia (eiqui
“here,” somos especiales “we are peculiar,” galicia “Galicia,” que hasta el tiro
neumático siempre estaba “where airgun shooting has always …”). An approach
to code switching which equates code with Galician and Spanish could not easily
account for this contextualization as code switching, given the unclear
demarcation of the varieties used here. However, the approach described above
30
The relationships among language varieties of Spain are beyond the scope of this brief example.
See, for example, Siguan 1993.
64
relies not on exhaustively defining these varieties, but on determining whether the
parties in this interaction treat the speaker as having accomplished speaking
Galician. (Alvarez-Cáccamo 1990 suggests they do.)
After reviewing some of the literature on code switching in sociology,
linguistic anthropology, and sociolinguistics, this paper has suggested a definition
of the term that, while partial, seems useful for sociocultural analysis. Code
switching is defined as the practice of selecting or altering linguistic elements so
as to contextualize talk in interaction. This contextualization may relate to local
discourse practices, such as turn selection or various forms of bracketing, or it
may make relevant information beyond the current exchange, including
knowledge of society and diverse identities. This approach understands code
switching as the practice of individuals in particular discourse settings. Therefore,
it cannot specify broad functions of language alternation, nor define the exact
nature of any code prior to interaction. Codes emerge from interaction, and
become relevant when parties to discourse treat them as such.
65
References
Alvarez-Cáccamo, Celso. 1990. “Rethinking Conversational Code-Switching: Codes, Speech
Varieties, and Contextualization.” In Kira Hall, Jean-Pierre Koenig, Michael Meacham,
Sondra Reinman and Laurel Sutton (eds.) Proceedings of the Sixteenth Annual Meeting of the
Berkeley Linguistics Society, 3-16. Berkeley: Berkeley Linguistics Society.
--. 1998. “From ‘Switching Code’ to ‘Code-switching’: Towards a Reconceptualization of
Communicative Codes.” In Peter Auer (ed.) Code-Switching in Conversation: Language,
Interaction, and Identity, 29-48. London: Routledge.
--. 2000. “Para um Modelo do ‘Code-switching’ e a Alternancia de Variedades como Fenomenos
Distintos: Dados do Discurso Galego-Portuges/Espanhol na Galiza” (Toward a Model of
‘Code-switching’ and the Alternation of Varieties as Distinct Phenomena: Data from Galician
Portuguese/Spanish Discourse in Galicia). Estudios de Sociolinguistica 1(1), 111-128.
Auer, Peter. 1984. Bilingual Conversation. Amsterdam: John Benjamins.
--. 1992. “Introduction: John Gumperz’s Approach to Contextualization.” In Peter Auer and Aldo
di Luzio (eds.), The Contextualization of Language, 1-37. Amsterdam: John Benjamins.
--. 1995. “The Pragmatics of Code-switching: A Sequential Approach.” In Lesley Milroy and
Pieter Muysken (eds.), One Speaker, Two Languages: Cross-disciplinary Perspectives on
Code-switching, 115-135. Cambridge: Cambridge University Press.
--. 1998. Code-switching in Conversation: Language, Interaction, and Identity. London:
Routledge.
Auer, Peter and Aldo di Luzio. 1992. The Contextualization of Langauge. Amsterdam: John
Benjamins.
Azuma, Shoji. 1991. “Two Level Processing Hypothesis in Speech Production: Evidence from
Intrasentential Code-switching.” Papers from the Regional Meetings, Chicago Linguistic
Society, 27(1), 16-30.
--. 1996. “Speech Production Units among Bilinguals.” Journal of Psycholinguistic Research,
25(3), 397-416.
Bentahila, Abdelali and Eirlys E. Davies. 1995. “Patterns of Code-switching and Patterns of
Language Contact.” Lingua 96, 65-93.
Bailey, Benjamin. 1999. “Switching.” Journal of Linguistic Anthropology 9(1-2), 241-243.
--. 2001. “The Language of Multiple Identities among Dominican Americans.” Journal of
Linguistic Anthropology 10(2), 190-22
--. 2002. Language, Race, and Negotiation of Identity: A Study of Dominican Americans. New
York: LFB Scholarly Publishing.
Baker, Colin. 2001. Review. Linguistic Minorities and Modernity: A Sociolinguistic Ethnography
by Monica Heller. Journal of Sociolinguistics 4/5, 556-573.
Bakhtin, Mikhail. 1981. The Dialogic Imagination. Caryl Emerson and Michael Holquist (trans.).
Austin: University of Texas Press.
Bani-Shoraka, Helena. 2005. Language Choice and Code-switching in the Azerbaijani Community
in Tehran: A Conversation Analytic Approach to Bilingual Practices. Uppsala, Sweden: Acta
Universitatis Upsaliensis.
Barker, George. 1947. “Social Functions of Language in a Mexican-American Community.” Acta
Americana 5: 185-202.
Belazi, Heidi, Edward Rubin, and Almeida Jacqueline Toribio. “Code Switching and X-Bar
Theory: The Functional Head Constraint.” Linguistic Inquiry 25(2), 221-237.
Benson, Erica. 2001. “The Neglected Early History of Codeswitching Research in the United
States.” Language & Communication 21: 23-36.
Blom, Jan-Petter, and John Gumperz. 1972. “Social Meaning in Linguistic Structures: Code
Switching in Northern Norway.” In: John Gumperz and Del Hymes (eds.), Directions in
66
Sociolinguistics: The Ethnography of Communication, 407-434. New York: Holt, Rinehart,

and Winston.
Bourdieu, Pierre. 1977. “The Economics of Linguistic Exchanges.” Social Science Information
16, 645-668.
Brown, Penelope, and Stephen Levinson. 1978. “Universals in Language Usage: Politeness
Phenomena.” In Esther Goody (ed.) Questions and Politeness: Strategies in Social
Interaction, 56-289. Cambridge: Cambridge University Press.
--. 1987. Politeness: Some Universals in Language Usage. Cambridge: Cambridge University
Press.
Bucholtz, Mary. 1999. “You da Man: Narrating the Racial Other in the Production of White
Masculinity.” Journal of Sociolinguistics 3(4), 443-460.
Bucholtz, Mary, and Kira Hall. 2005. “Identity and Interaction: A Sociocultural Linguistic
Approach.” Discourse Studies 7(4-5).
Cenoz, Jasone, and Fred Genesee. 2001. Trends in Bilingual Acquisition. Amsterdam: John
Benjamins.
Chomsky, Noam. 1965. Aspects of the Theory of Syntax. Cambridge, MA: MIT Press.
Cromdal, Jakob. 2001. “Overlap in Bilingual Play: Some Implications of Code-Switching for
Overlap Resolution.” Research on Language and Social Interaction 34(4), 421-451.
Dil, Anwar S. 1971. “Introduction.” In John J. Gumperz, Language in Social Groups: Essays by
John J. Gumperz. Stanford: Stanford University Press.
DiSciullo, Anna-Maria, and Edwin Williams. 1987. On the Definition of a Word. Cambridge, MA:
MIT Press.
Eckert, Penelope, and Sally McConnell-Ginet. 1992. “Think Practically and Look Locally:
Language and Gender as Community-based Practice.” Annual Review of Anthropology
21:461–90.
Ervin-Tripp, Susan. 1964. “An Analysis of the Interaction of Language, Topic and Listener.”
American Anthropologist 66(6), part 2, 86-102.
Espinosa, Aurelio. 1980 [1911]. “The Spanish Language in New Mexico and Southern Colorado.”
In Carlos Cortés (ed.). Spanish and Portuguese Languages in the United States. New York:
Arno Press.
Fano, Robert M. 1950. “The Information Theory Point of View in Speech Communication.”
Journal of the Acoustical Society of America 22, 691-696.
Ferguson, Charles. 1959. “Diglossia.” Word 15, 325-340.
Ferguson, Charles, and John Gumperz. 1960. “Linguistic Diversity in South Asia, Introduction.”
International Journal of American Linguistics 26(3), part 3, vii-18.
Fishman, Joshua. 1967. “Bilingualism with and without Diglossia; Diglossia with and without
Bilingualism.” Journal of Social Issues 23(2), 29-38.
Fotos, Sandra. 2001. “Codeswitching By Japan’s Unrecognized Bilinguals: Japanese University
Students’ Use Of Their Native Language As A Learning Strategy.” In Mary Goebel Noguchi
and Sandra Fotos (eds.) Studies in Japanese Bilingualism. Clevedon: Multilingual Matters.
Friedrich, Paul. 1972. “Social Context and Semantic Feature: The Russian Pronominal Usage.” In
John Gumperz and Dell Hymes (eds.), Directions in Sociolinguistics: The Ethnography of
Communication, 270-300. New York: Holt, Rinehart and Winston.
Gafaranga, Joseph. 2001. “Linguistic Identities in Talk-in-interaction: Order in Bilingual
Conversation.” Journal of Pragmatics 33(12), 1901-1925.
--. 2005. “Demythologising Language Alternation Studies: Conversational Structure vs. Social
Structure in Bilingual Interaction.” Journal of Pragmatics 37(3), 281-300.
Gal, Susan. 1979. Language Shift: Social Determinants of Linguistic Change in Bilingual Austria.
New York: Academic Press.
--. 1988. “The Political Economy of Code Choice.” In Monica Heller (ed.), Codeswitching:
Anthropological and Sociolinguistic Perspectives, 243-261. Berlin: Mouton de Gruyter.
67
Giles, Howard, Anthony Mulac, James J. Bradac and Patricia Johnson. 1987. “Speech
Accommodation Theory: The First Decade and Beyond.” In Margaret L. McLauglin (ed.),
Communication Yearbook No. 10, 13-48. Beverly Hills: Sage.
Grice, H. Paul. 1975. “Logic and Conversation.” In Peter Cole and Jerry L. Morgan (eds.), Speech
Acts, 41-55. New York: Academic Press.
Goffman, Erving. 1979. “Footing.” Semiotica 25, 1-29.
--. 1981. Forms of Talk. Philadelphia: University of Pennsylvania Press.
Gumperz, John. 1958. “Dialect Differences and Social Stratification in a North Indian Village.”
American Anthropologist 60, 668-681.
--. 1961. “Speech Variation and the Study of Indian Civilization.” American Anthropologist 63,
976-988.
--. 1964a. “Hindi-Punjabi Code-switching in Delhi.” In H. Hunt (ed.), Proceedings of the Ninth
International Congress of Linguistics, 1115-1124. The Hague: Mouton.
--. 1964b. “Linguistic and Social Interaction in Two Communities.” American Anthropologist
66(6), part 2, 137-153.
--. 1982. Discourse Strategies. Cambridge: Cambridge University Press.
--. 1992. “Contextualization Revisited.” In Peter Auer and Aldo di Luzio (eds.), The
Contextualization of Language. Amsterdam: John Benjamins, 39-53.
Gumperz, John and Dell Hymes. 1972. Directions in Sociolinguistics: The Ethnography of
Communication. New York: Holt, Rinehart and Winston.
Halmari, Helena. 1997. Government and Codeswitching: Explaining American Finnish.
Amsterdam: John Benjamins.
Heller, Monica. 1988a. Codeswitching: Anthropological and Sociolinguistic Perspectives. Berlin:
Mouton de Gruyter.
--. 1988b. “Strategic Ambiguity: Code-switching in the Mangagement of Conflict.” In Monica
Heller (ed.), Codeswitching: Anthropological and Sociolinguistic Perspectives, 77-96. Berlin:
Mouton de Gruyter.
--. 1992. “The Politics of Codeswitching and Language Choice.” In Carol Eastman (ed.),
Codeswitching, 123-142. Clevedon: Multilingual Matters.
--. 1995. “Code-switching and the Politics of Language.” In Lesley Milroy and Pieter Muysken
(eds.), One Speaker, Two Languages: Cross-disciplinary Perspectives on Code-switching.
Cambridge: Cambridge University Press.
--. 1999. Linguistic Minorities and Modernity: A Sociolinguistic Ethnography. London: Longman.
Hill, Jane. 1985. “The Grammar of Consciousness and the Consciousness of Grammar.” American
Ethnologist 12(4), 725-737.
Hutchby, Ian, and Robin Woofit. 1998. Conversation Analysis: Principles, Practices, and
Applications. Malden, MA: Polity Press.
Hymes, Dell. 1964. “Introduction: Toward Ethnographies of Communication.” American
Anthropologist 66(6), part 2, 1-34.
Hymes, Dell, and John Gumperz (eds.). 1964. The Ethnography of Communication (special issue).
American Anthropologist 66(6), part 2.
Jaffe, Alexandra. 2000. “Comic Performance and the Articulation of Hybrid Identity.” Pragmatics
10(1), 39-59.
Jake, Janice, Carol Myers-Scotton and Steven Gross. 2002. “Making a Minimalist Approach to
Codeswitching Work: Adding the Matrix Language.” Bilingualism: Language and Cognition
5(1), 69-91.
Jakobson, Roman. 1971a. “Results of a Joint Conference of Anthropologists and Linguists.” In
Selected Writings, volume II, 554-567. The Hague: Mouton.
--. 1971b. “Linguistics and Communication Theory.” In Selected Writings, volume II, 570-579.
The Hague: Mouton.
--. 1990. On Language. Linda Waugh and Monique Monville-Burston (eds.), Cambridge, MA:
Harvard University Press.
Jakobson, Roman, and Morris Halle. 1956. Fundamentals of Language. The Hague: Mouton.
68
Jonsson, Carla. 2005. Code-switching in Chicano Theatre: Power, Identity and Style in Three
Plays by Cherríe Moraga. Umeå, Sweden: Institutionen för moderna språk.
Joshi, Aravind. 1985. “How much Context-sensitivity is Necessary for Assigning Structural
Descriptions: Tree Adjoining Grammars.” In D. Dowty, L. Karttunen, and A. Zwicky (eds.)
Natural Language Parsing. Cambridge: Cambridge University Press.
Kroeber, Alfred. 1939. “Cultural and Natural Areas of Native North America.” University of
California Publications in American Archeology and Ethnology 38, 1-242.
Labov, William. 1963. “The Social Motivation of a Sound Change.” Word 19, 273-309.
Labov, William and Joshua Waletzky. 1967. “Narrative Analysis.” In June Helm (ed.) Essays on
the Verbal and Visual Arts. Seattle: University of Washington Press.
Lakoff, George and Mark Johnson. 1980. Metaphors we Live By. Chicago: University of Chicago
Press.
Lave, Jean, & Etienne Wenger. 1991. Situated Learning: Legitimate Peripheral Participation.
Li Wei. 1994. Three Generations, Two Languages, One Family. Clevedon: Multilingual Matters.
--. 1998. “The ‘Why’ and ‘How’ Questions in the Analysis of Conversational Code-Switching.”
In Peter Auer (ed.), Code-Switching in Conversation: Language, Interaction and Identity,
156-176. London: Routledge.
--. 2002. “‘What do you Want me to Say?’ On the Conversation Analysis Approach to Bilingual
Interaction.” Language in Society 31(2), 159-180.
--. 2005. “‘How can you Tell?’ Toward a Common Sense Explanation of Conversational Code-
Switching.” Journal of Pragmatics 37(3), 375-389.
Lo, Adrienne. 1999. “Codeswitching, Speech Community Membership, and the Construction of
Ethnic Identity.” Journal of Sociolinguistics 3-4, 461-479.
MacSwann, Jeff. 2000. “The Architecture of the Bilingual Language Faculty: Evidence from
Intrasentential Code Switching.” Bilingualism: Language and Cognition 3(1), 37-54.
Maehlum, Brit. 1996. “Codeswitching in Hemnesberget – Myth or Reality?” Journal of
Pragmatics 25, 749-761.
Milroy, James and Lesley Milroy. 1985. “Linguistic Change, Social Network and Speaker
Innovation.” Journal of Linguistics 21(2), 339-384.
Milroy, Lesley, and Li Wei. 1995. “A Social Network Approach to Code-switching.” In Lesley
Milroy and Pieter Muysken (eds.) One Speaker, Two Languages: Cross-disciplinary
Perspectives on Code-switching, 136-157. Cambridge: Cambridge University Press.
Milroy, Lesley, Li Wei, and Suzanne Moffat. 1991. “Discourse Patterns and Fieldwork Strategies
in Urban Settings.” Journal of Multilingual and Multicultural Development 12(4), 287-300.
McClure, Erica, and Malcolm McClure. 1988. “Macro- and Micro-sociolinguistic Dimensions of
Code-switching in Vingard (Romania).” In Monica Heller (ed.) Codeswitching:
Anthropological and Sociolinguistic Perspectives, 25-51. Berlin: Walter de Gruyter.
Myers-Scotton, Carol. 1972. Choosing a Lingua Franca in an African Capital. Edmonton:
Linguistic Research.
--. 1976. “Strategies of Neutrality: Language Choice in Uncertain Situations.” Language 52(4),
919-941.
--. 1983. “The Negotiation of Identities in Conversation: A Theory of Markedness and Code
Choice.” International Journal of the Sociology of Language 44, 115-136.
--. 1988. “Codeswitching as Indexical of Social Negotiations.” In Monica Heller (ed.),
Codeswitching: Anthropological and Sociolinguistic Perspectives. Berlin: Mouton de Gruyter.
--. 1993a. Dueling Languages: Grammatical Structure in Codeswitching. Oxford: Oxford
University Press.
--. 1993b. Social Motivations for Codeswitching: Evidence from Africa. Oxford: Clarendon Press.
--. 1998. Codes and Consequences: Choosing Linguistic Varieties. New York: Oxford University
Press.
Myers-Scotton, Carol, and Agnes Bolonyai. 2001. “Calculating Speakers: Codeswitching in a
Rational Choice Model.” Language in Society 30, 1-28.
69
Myers-Scotton, Carol, and Janice Jake. 2000. “Four Types of Morpheme: Evidence from Aphasia,
Code Switching, and Second-Language Acquisition.” Linguistics 38(6), 1053-1100.
Nilep, Chad. 2004. “Practice and Domination: Toward a Theory of Political Micro-Economy.”
Colorado Research in Linguistics 17. www.colorado.edu/ling/CRIL/Volume17_Issue1/
index.htm
Nishimura, Miwa. 1992. “Language Choice and In-Group Identity among Canadian Niseis.”
Journal of Asian Pacific Communication 3(1), 97-113.
--. 1997. Japanese-English Code-switching: Syntax and Pragmatics. New York: P. Lang.
Ochs, Elinor, Emanuel Schegloff and Sandra Thompson. 1996. Interaction and Grammar.
Poplack, Shana. 1980. “Sometimes I'll Start a Sentence in Spanish y Termino en Espanol: Toward
a Typology of Code-switching.” Linguistics 18(233-234), 581-618.
Pujolar, Joan. 2001. Gender, Heteroglossia and Power: A Sociolinguistic Study of Youth Culture.
Berlin: Mouton de Gruyter.
Rampton, Ben. 1995. Crossing: Language and Ethnicity among Adolscents. London: Longman.
Romaine, Suzanne. 1989. Bilingualism. Oxford: Basil Blackwell.
Sacks, Harvey. 1992. Lectures on Conversation. Gail Jefferson (ed.) Oxford: Blackwell.
Sacks, Harvey, Emanuel Schegloff and Gail Jefferson. 1974. “A Simplest Systematics for the
Organization of Turn Taking for Conversation.” Language 50, 696-735.
Sankoff, David, and Shana Poplack. 1981. “A Formal Grammar for Code-Switching.” Papers in
Linguistics 14(1-4), 3-45.
Sapir, Edward. 1929. “The Status of Linguistics as a Science.” Language 5(4), 207-214.
Sebba, Mark. 1993. London Jamaican: A Case Study in Language Contact. London: Longman.
Sebba, Mark and Tony Wooten. 1998. “We, They and Identity: Sequential Versus Identity-Related
Explanation in Code-Switching.” In Peter Auer (ed.), Code-Switching in Conversation:
Language, Interaction and Identity, 262-286. London: Routledge.
Siguan, Miguel. 1993. Multilingual Spain. Amsterdam: Swets & Zweitlinger.
Stroud, Christopher. 1998. “Perspectives on Cultural Variability of Discourse and some
Implications for Code-switching.” In Peter Auer (ed.), Code-Switching in Conversation:
Language, Interaction and Identity, 321-348. London: Routledge.
Thakerar, Jitendra, Howard Giles and Jenny Cheshire. 1982. “Psychological and Linguistic
Parameters of Speech Accommodation Theory.” In Colin Fraser and Klaus Sherer (eds.),
Advances in the Social Psychology of Language, 205-255. Cambridge: Cambridge University
Press.
Torras, Maria-Carme, and Joseph Gafaranga. 2002. “Social Identities and Language Alternation in
Non-Formal Institutional Bilingual Talk: Trilingual Service Encounters in Barcelona.”
Language in Society 31(4), 527-548.
Vogt, Hans. 1954. “Language Contacts.” Word 10(2-3), 365-374.
Weinreich, Uriel. 1953. Languages in Contact. The Hague: Mouton.
Woolard, Katherine. 1985. “Language Variation and Cultural Hegemony.” American Ethnologist
12(4), 738-748.
--. 2004. “Codeswitching.” In Alessandro Duranti (ed.) A Companion to Linguistic Anthropology,
73-94. Malden, MA: Blackwell.
Zentella, Ana Celia. 1997. Growing Up Bilingual. Malden, MA: Blackwell.
70

"Code Switching" in Sociocultural Linguistics

Uploaded by

Copyright:

Available Formats

"Code Switching" in Sociocultural Linguistics

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

"Code Switching" in Sociocultural Linguistics

Uploaded by

Copyright:

Available Formats

“Code Switching” in Sociocultural Linguistics

The term code switching (or, as it is sometimes written, code-switching or

codeswitching)1 is broadly discussed and used in linguistics and a variety of

Linguistics and Language Behavior Abstracts database published in virtually

because of it – scholars do not seem to share a definition of the term. This is

perhaps inevitable, given the different concerns of formal linguists,

psycholinguists, sociolinguists, philosophers, anthropologists, etc. This paper will

suggest useful definitions for sociocultural work.

labeled “code switching” is interested in syntactic or morphosyntactic constraints

Alternately, studies of language acquisition, second language acquisition, and

speakers’ or language learners’ cognitive linguistic abilities, or to describe

studies seem to use code as a synonym for language variety. Alvarez-Cáccamo

Praticamente todo o trabalho sobre o “code-switching”, ou

Practically all work on “code-switching,” or changing codes, has

decades. The identification of various constraints, though controversial, has

inspired a great deal of work in syntax, morphology, and phonology. A structural

Scotton and Gross 2002). By ignoring questions of function or meaning, though,

Auer (1984) warns, “Grammatical restrictions on code-switching are but

effect of a particular alternation. If linguists regard code switching simply as a

product of a grammatical system, and not as a practice of individual speakers,

language itself” (1929:213).

This paper is thus positioned within the discipline of sociocultural

linguistics, an emerging (or one might say, revitalized) interdisciplinary approach

functions and meanings of language use.

conduct in general” (1929:207). He suggested that anthropology, sociology,

psychology, philosophy and social science generally would be enriched by

drawing on the methodologies as well as the findings of linguistic research. He

It is peculiarly important that linguists, who are often accused, and

the development during the 1930s-1950s of what has come to be known as

sociolinguistics and psycholinguistics suggests that, at least for some linguists,

structures of language itself.

for a more socially engaged linguistics. In his introduction to a special issue of

American Anthropologist dedicated to the Ethnography of Communication

revitalization of socially and culturally oriented linguistic analysis. Bucholtz and

sociocultural linguistics, “the broad interdisciplinary field concerned with the

Sociocultural linguistics is thus suggested as a broader term, to include

sociolinguistics, linguistic anthropology, discourse analysis, conversation analysis

and sociology of language, as well as certain streams of social psychology,

What follows is a look at some foundational studies of code switching in

linguistic anthropology, sociolinguistics, and sociology, as well as a brief survey

serve as a basis and context for sociocultural discussions of the contextualizing

functions of language alternation and modulation.

The history of code switching research in sociocultural linguistics is often

and sociolinguistics prefigured later code switching research in sociocultural

linguistics. Section 2.1 briefly reviews some of the important “pre-history” of

systems in discourse. Section 2.2 considers sociologist Erving Goffman’s notion

of footing, a theoretical construct closely related to (and influenced by) code

2.1. Pre-history of code switching research

description of language use among Mexican Americans in Tucson, Arizona. In

Barker suggested a continuum between intimate family relationships,

extending through increasingly public situations, to formal interaction with

English speaking neighbors. He found that interactions among family members

of multiple varieties was constitutive of a local Tucson identity.

background, or in terms of its social function, to identify the group as a group”

(186). He therefore reasoned that the use of Mexican Spanish functioned as an

individual’s skill in using the language of a second or adopted culture comes to

varieties within a single interaction.

In the field of informal relations among bilinguals, a field in which

An important base for code switching research in the field of linguistics is