| 361
‘International Journal of Bilingualism’ • Volume 9 • Number 3 & 4 Running
• 2005, 361–
375|
Head
Linguistic change and social meaning:
Codeswitching in the media
Shahrzad Mahootian
Northeastern Illinois University
Acknowledgements
I’d like to thank the editors of this special volume, in particular Ad Backus, for their many helpful
comments. I’d also like to thank Lewis Gebhardt for his useful comments and suggestions.
Abstract:
Key words
This article investigates the relationship between code choice, bilingual
codeswitching
identity and language change. Code choice and codeswitches in a bilingual
Spanish-English publication are examined with two questions in mind.
language change
The first asks the extent to which stylistic and social variables, including
identity, govern code choice. The second looks at the relationship of code
media
choice to language change. Based on the pattern of code choice found in
a popular women’s magazine, I conclude that codemixed discourse is one
of three varieties of code available to the bilingual, and where this variety is used intentionally,
it is meant to emphasize the speaker’s bilingual identity. I also consider structural aspects of
codemixed discourse and determine that although the L1-L2 structure of this discourse is distinct
from the monolingual L1 and L2 structures, it nevertheless follows from universal principles of
grammar and does not require the positing of a third grammar. I further argue that the use of
mixed-code discourse, especially in written media, as a pragmatically and structurally distinct
variety available to bilingual speakers, falls within the Weinreich, Labov, and Herzog (1968)
and Labov (1972 a,b) conceptions of language change. Thus, I suggest that in some contexts,
codeswitching, is itself an instantiation of language change.
1Introduction
That language changes over time is apparent to linguists and nonlinguists alike. Accounts
of how language changes and descriptions of change-in-progress have been successfully
developed by Bynon (1977); Hock (1996); Weinreich, Labov, and Herzog (1968); Labov
(1972a,b) among others. Weinreich et al. sum up the nature of language change in its
social context as follows (pp.100 – 101): “The key to a rational conception of language
change —indeed, of language itself—is the possibility of describing orderly differentiation in a language serving a community.” Weinreich et al. (1968) and later Labov (1972a,b)
showed us that change-in-progress can be observed. Moreover, they effectively argued
that the process of language change is part of a social dynamic that offers a glimpse of
Address for correspondence
Linguistics Department, Northeastern Illinois University, 5500 N. St. Louis Ave., Chicago, Illinois, 60625, U.S.A.;
e-mail: < [email protected] >.
‘International Journal of Bilingualism’ is ©Kingston Press Ltd. 1997 – 2005
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
362
S. Mahootian
the relationship between language and social identity on the one hand, while providing
evidence for language theories, on the other.
Lexical borrowing is often cited as the most common type of language change
resulting from language contact, where borrowings typically enrich the language but
do not affect the overall structure of the host language (Hock, 1996). By and large, the
status of codeswitching, another consequence of language contact, has been unclear
vis-à-vis language change. With the exception of its role in the language shift/language
death cycle, switching between languages has mostly been discussed in terms of social/
stylistic strategies used in response to setting, audience and topic, and as a means of
group identification (Bell, 1984, 2001; Grosjean, 1982; Gumperz, 1982; Kachru, 1982;
Mahootian, 2000; Romaine, 1995, among others).
I propose that in some contexts codeswitching is itself an instantiation of language
change. I analyze code choice as it appears in a conventionalized format, printed media,
where language choice is made consciously: copy is written, proofread and approved
by a number of people before it is set to print. I claim that mixed code discourse, used
by a subset of Spanish-English bilinguals in the U.S.A. to mark identity with other
Spanish-English bilinguals, is a variety with a distinct communicative function which
has achieved “official” status. Particularly, I suggest mixed-code discourse is used to
underscore a bilingual identity that is connected to, yet distinct from, the identity of
speakers in their monolingual contexts. I examine the relationship between code choice
and bilingual identity and the degree to which this relationship has brought about
language change. Finally, I address whether the spreading of the mixed-code variation
from its original scope (spoken discourse) into a standardized, conventional discourse
context (print media) is evidence of language change in the tradition of Weinreich et. al
and Labov. Simply, the questions to be answered are (1) to what extent are code choice
and codeswitching governed by social factors? and (2) to what extent does codemixed
discourse constitute language change?
Towards answering these questions, in Section 2, I discuss the relationship between
code choice, codeswitching and the conscious and unconscious use of each. In Section 3,
I argue that intentional codeswitching is a means to signal bilingual identity. In Section 4,
I turn to codeswitching data in the media to further establish the link between intentional
codeswitching and bilingual identity. The structure of codeswitching is analyzed in
Section 5 where I maintain that although structurally it is governed by universal grammatical constraints that apply to all languages, L1-L2 mixed-code structure is distinct
from both L1-only and L2-only structures because it results from the interaction between
two separate grammatical systems. In Sections 6 and 7, I return to the two questions
that I posed above. I argue codemixed discourse is functionally and structurally in
variation with the L1 and L2 varieties available to the bilingual speaker and conclude
that codeswitching is a form of language change.
2 Code choice, codeswitching and intentionality
Backus and Eversteijn (2002) point to the difficulties of distinguishing language choice
from codeswitching. Among other things, they conclude that, “it may be justified to
define language choice as what speakers do when deciding in which language to conduct
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Linguistic change and social meaning
363
a conversation and codeswitching as alternating between languages within a conversation” (p. 14). In agreement with this hypothesis, in my analysis I distinguish between
mixed-code discourse, which I claim is itself a variety available to the bilingual speaker,
and codeswitching, the alternation between languages found within the mixed-code
discourse. “Discourse” refers to written text or spoken conversation.
To understand the role of code choice as an actuation point for language change
and the conventionalized use of mixed-code discourse as the manifestation of that
change, we need to expand the discussion to include “intentionality” in code choice.
The question that arises is the degree of intentionality or conscious choice one should or
should not ascribe to the use of mixed-code and to individual instances of codeswitching.
A related question concerns the degree to which these choices are sociolinguistically
significant or in relatively free and random variation. Here, again, we must be careful
to distinguish between mixed-code as a discourse mode, and codeswitching, as the
alternation between two languages at the utterance level, within the discourse mode.
Intentional code choice as a harbinger of structural change and, eventually, as
change itself has been largely overlooked. In fact, codeswitching is often cited as a
sign of language shift on the way to L1 death. I suggest that at the discourse level, code
choice is at its most intentional or conscious and, consequently, socially meaningful.
The bilingual speaker has three choices: monolingual L1 code, monolingual L2 code
or mixed L1-L2 code, and each code is associated with its own set of sociopragmatic
functions (Backus & Eversteijn, 2002; Blom & Gumperz, 1972; Kachru, 1982; Mahootian,
2000, 2002). Bilingual speakers choose one of the three options, in part guided by what
Bell (1984) calls “audience variables.” In Bell’s theory of audience design, “speakers
accommodate primarily to their addressee” (1984, p.145). Topic and setting, two other
variables in language or style choice, are secondary influences. Applying Bell’s audience
design model, we expect that the felicitous bilingual will not choose mixed code when
speaking to a monolingual English or Spanish speaker, but may choose to use mixed
code with another bilingual Spanish-English speaker.
When the speaker chooses either the L1 or the L2 monolingual codes, the discourse
simply proceeds according to the syntactic and pragmatic rules of that code. But what
happens when the speaker chooses the mixed L1-L2 code? Choosing the mixed-code
means that the speaker allows himself/ herself to codeswitch. What rules does the
speaker apply to produce codeswitched utterances? To what extent are the codewitches
motivated by sociopragmatic considerations? The first question is really a syntax question. Structural accounts of codeswitching abound, with all of them concluding that
switching is a rule-governed phenomenon and almost all agreeing that codeswitching
does not entail postulating a third grammar. The structural aspects of codemixed
discourse will be addressed more fully in Section 5.
The second question is more complex. At utterance level, there are a variety of
reasons for codeswitching, only some of which can be termed socially meaningful,
conscious choices. Some switches may very well be devoid of any social motivation,
even when “chosen” consciously. The following examples provide clarification. The
data presented is from the women’s lifestyle magazine Latina where, presumably, all
the codeswitching is conscious. The writers choose where to switch and where not to
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
364
S. Mahootian
switch. And yet along with many examples such as the one in (1), we also get some
switches like the one shown in (2).
In (1), the Spanish word mujer is used in place of the English ‘woman’. Mujer easily
conjures up ethnic identity, the image of a woman of Latina heritage, rather than an
American or Americanized woman.
(1) even if you are the kind of MUJER who thinks…
even if you are the kind of woman who thinks
‘even if you are the kind of woman who thinks…’
In the utterance in (2) however, there is no extralinguistic message in the use of
the Spanish words y cuales meaning ‘and which’. No convincing argument can be made
in favor of these words carrying any cultural weight when rendered in one language
versus the other. It’s simply a switch that fits into a pattern of switching as a whole and
can be accounted for by various syntactic models of codeswitching.
(2) It helps to know which items are worth the sticker shock, Y CUALES not.
It helps to know which items are worth the sticker shock, and which not.
‘It helps to know which items are worth the sticker shock, and which not.’
Finally, we know (and bilingual speakers acknowledge) that codeswitching may
also be an unconscious performance level phenomenon, where speakers are unaware
they are switching. Of interest to this study is the conscious choice of mixed code and the
conscious codeswitches within that mode. The intentional use of mixed-code discourse
in a conventionalized medium, juxtaposed with L1-only and L2-only discourse, signals
acceptance of the L1-L2 mix as a variety. In the following section, I show the link between
this new variety and bilingual identity.
3 Codeswitching and bilingual identity
The relationship between language and identity has long been established and documented (Antaki & Widdicombe, 1998; Carbaugh, 1996; Gumperz, 1982; Gumperz &
Cook-Gumperz, 1982; Tabouret-Keller, 1997). Language is both coconstructor and
a reflection of social identity. This relationship becomes more complex in bilingual
communities where more than one language is involved, and the languages and cultural/
ethnic values and identities bear unequal social prestige.
A case in point is the plight of Hispanics’ language and ethnic heritage and identity
in the United States. Bourhis and Marshall (1999, p.251), write that “Spanish speakers
are aware that knowledge of English is crucial for their economic survival in the United
States while also being concerned with the maintenance of their Hispanic linguistic and
cultural identity.” They also comment that although Spanish speakers constitute the
largest linguistic minority in the U.S.A., “Spanish speakers are caught in a dilemma:
To appear to be ‘good Americans’, they feel pressured to shift to English as quickly
as possible; but they also recognize that in the long run this shift may be achieved at
the cost of losing their mother tongue and their Hispanic cultural and ethnic identity.”
(p.250).
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Linguistic change and social meaning
365
It has also been noted (Gumperz, 1982; Romaine, 1995) that when political ideology
changes and a group becomes more conscious of their ethnicity, attitudes toward code
mixing change. For example, Romaine points to the reversal of the previously pejorative
use of pocho and caló used in California and southwestern United States to refer to
the variety of Spanish-English spoken by Chicanos. She notes, “With the rise in ethnic
consciousness, these speech styles have become symbolic of Chicano ethnicity. Pocho
and caló are now increasingly used in modern Chicano literature.” (pp.291 – 292).
These two realities, that on the one hand Hispanics in America are aware that
their cultural and linguistic identity is threatened, while on the other hand the rise of
ethnic consciousness is linked to an increased use of code mixing, serve as a backdrop
to the data I present here.
Examination of code choice and codewitches in a Spanish-English bilingual
women’s magazine Latina reveals an interesting pattern of premeditated switching.
As I discussed earlier, we know that switching is not always a matter of conscious
choice. However, we will see that switching can be used consciously to evoke a sense
of cultural identity, unity and camaraderie. Here it is used as a direct and undeniable
assertion of the bilingual identity. It is a way for speakers to underscore their ethnicity,
their connection to their heritage and to others who share that heritage and the values
associated with it, within the majority culture and language.
Bell (1984) observes that the language of mass media draws from the norms of
the population with which it wishes to identify.1 It is therefore, not surprising that the
format of Latina is such that all three codes —L1, L2, and mixed L1-L2—are used,
consciously chosen to represent and to identify with the readership.
On the other hand, Sebba (2002), points out that, “in written text monolingualism
is the norm,” and when two languages are used, they are kept apart, each rendered in
monolingual code. This phenomenon he attributes to an “orthographic standardization [which] is imposed to varying degrees.” He also suggests an “orthographic regime”
where publishing, as an institution, adheres most to orthographic standards such that
texts for publication are the most regulated with respect to language separation, and
noninstitutional types of writing, such as personal letters, diaries, and so forth fall
into less and less regulated domains. One may conclude then, that if language-mixed
texts are found in institutionalized publications such as journals and magazines, it is a
reflection of a community norm which has found acceptance.
Every article in Latina is presented in English or in English-Spanish, and in Spanish.
The Spanish versions are abbreviated versions of the English or English-Spanish versions.
The L1-L2 mixed versions are predominantly in English, peppered with Spanish nouns,
determiner phrases, conjunctions, prepositional phrases. Only a small percentage of
the switched utterances are words where there is no single equivalent lexical item. The
majority of the switches have common, simple English counterparts. It is interesting
to note that in the L1-L2 mixed-code versions, the predominant or host language is
always English. There are no mixed-code versions where the host language is Spanish
1
In a recent front-page article in the Tempe, Arizona, daily newspaper The Arizona Republic, entitled “Englishonly Latinos on rise,” the writer suggests that to stay connected with their Latin culture, young Latinos should
“subscribe to magazines that cater to Latinos, such as Hispanic Magazine, Latina and People en Español.”
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
366
S. Mahootian
and occasional switches are made into English. There are potentially a number of
interrelated explanations for this one-way mixing.
One explanation may be that Spanish-only version reflects the “language purity”
attitude prevalent in many bilingual communities (Grosjean, 1982; Hock, 1996; Kachru,
1982; Mahootian, 1993; Romaine, 1995). Another explanation may be that it reflects the
fear of the loss of mother tongue and cultural identity that Bourhis and Marshall mention.
English is not the endangered language, hence there is no reason to “protect” it; rather it
is the Spanish language and the Latino culture and identity that need reinforcement.
So then what is meant by a bilingual identity among Latino-Americans? Someone
who is at least partly of Latino origin, someone who is fluent or dominant in English,
may be fluent in Spanish or speak some Spanish. This identity is less dependent on
full knowledge of English and Spanish than it is on a sense of belonging to the Latino
culture. Bilingual identity, then, is the assertion of Latino heritage into the mainstream
majority culture. It appears that one way this identity is asserted is through the use of
codemixed discourse.
4 The data: Mixed code in the media
The data was culled from a popular life styles publication called Latina. The first
issue of this magazine hit the stands in 1996. The primary target audience of this
publication is American Hispanic women between the ages of 15 – 40 who can function
in both English and Spanish.2 Latina is a national publication with a circulation of
approximately 175,000. Many of the features are written by a pool of staff writers. Other
pieces are submitted by freelance writers. All articles and features are translated into
Spanish. However, since the Spanish version are often shorter, abbreviated versions of
the English texts, there seems to be an assumption that readers of Latina are, at least
with respect to literacy, English dominant. Writers are not directed to use or not use
codeswitching in the articles/features. As mentioned earlier, switching only occurs
in the English texts. All switched items are rendered in italic type-face. On the cover
of the magazine, references to the feature articles are typically codeswitched phrases:
“Seduce him en la cocina,” “Dinner in 29 minutos,” “Tienes Lupus? You may not know.”
In the table of contents, most titles are in a single language, usually English, although
some titles may include codeswitching into Spanish. Subtitles describing the articles
appear in English or English-Spanish mix, followed by the Spanish version: “The sky’s
the limit? Despite barriers, workplace advancement is up to you. Hay barreras para el
avance profesional de las latinas.”
However, in the body of the articles the titles sometimes occur in both English and
Spanish. For example, “Maxwell bares all” is also provided in Spanish as “Maxwell al
desnudo.” In other instances the translation is less literal. Sometimes, an English title
serves for both the English and Spanish texts. And often a single codeswitched title is
used: “Novelas with heart.” Advertisements may appear in Spanish, English or both.
There is not a magazine policy directing the language used in the ads.
2
The publication information about Latina is from a phone interview with the advertising sales manager for the
Midwest.
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Linguistic change and social meaning
367
Two Spanish-English bilingual graduate students, one male (non-Hispanic), one
female (Mexican-American), created a database containing all 435 codeswitches found
in two issues of Latina: February, 1999 and May, 1999. Page numbers, translations,
syntactic category and grammatical function were provided for all utterances. The
students were then asked to provide an explanation for the codeswitches. Based on
their explanations, I designated the examples as either, (1) idiomatic, which included
examples such as la musica tejana as well as food names, titles of Spanish songs, movies,
and so forth, (2) attention-getting, such as y cuales in example (1), repeated below as
example (14), or (3) emotionally/culturally evocative/bonding. A subset of the data (56
tokens) that included all three types of switches and randomly ordered, was emailed to
three Spanish–English bilinguals of Hispanic descent (2 females one of Puerto Rican
and the other of Cuban heritage, 1 male of Puerto Rican heritage). They were told the
sentences were from Latina and asked to indicate why they thought the Spanish words
or phrases in each example were in Spanish instead of in English. They were also asked
to say what they thought the difference between the term Latino/a and Hispanic was.
As it turned out, all three were familiar with the magazine.
With respect to use of the term Latina/Latino versus the term Hispanic, all three
stated that the term Latino/a is favored by younger, more socially/politically active
individuals, who are as one respondent put it, “seeking equality and opportunities.”
Another respondent noted that, “Hispanic is a more old-fashioned term associated with
the Hispanic heritage. My mother is more comfortable with that term. Young people
see themselves more as Latinos. More closely related to Latin Americans who live in
the U.S.A. ‘Latino’ is closer to the use of ‘African-Americans’ instead of ‘Blacks’.”
As for the 56 examples, all three agreed on the switches that were idiomatic usages.
There was also a general consensus that those that I had drawn from the list of evocative switches (as in 3 – 11) were “more meaningful” in Spanish, that they carried “more
emotional power,” (“emotional statements will be said in Spanish. English is not sufficient”) that they were used to “create solidarity,” that they were “much stronger” when
said in Spanish (as e.g., la revolucion). Responses to the switches that had been drawn
from the “attention-getting” list (examples 14 –15) were mixed between, “Draws your
attention because it’s not English” to “I’m not sure why.”
Following, are some examples of all three types of codeswitches from Latina.
(3) Lately, you’d be hard-pressed to find a mainstream show without HERMANOS in it.
lately, you’d be hard-pressed to find a mainstream show without brothers in it.
‘Lately, you’d be hard-pressed to find a mainstream show without Hispanic men
in it.’
(4) even if you are the kind of MUJER who thinks…
even if you are the kind of woman who thinks
‘even if you are the kind of woman who thinks…’
(5) First time you meet his FAMILIA
first time you meet his family
‘First time you meet his family’
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
368
S. Mahootian
(6) A tale of two AMANTES splashing among the colors and textures of spring
A tale of two lovers
splashing among the colors and textures of spring
‘A tale of two lovers splashing among the colors and textures of spring’
(7) ROPA INTIMA for every body
clothes intimate for every body
‘Intimate clothes for every body’
(8) Forget those FLACA cover girls! What Latinos love most about Latinas SON
forget those skinny cover girls! What Latinos love most about Latinas are
LAS CURVAS
‘the curves’
‘Forget those skinny cover girls! What Latinos love about Latinas are the curves.’
(9) LA REVOLUCION has begun, and Spanish-language TV will never be the same.
the revolution
has begun, and Spanish-language TV will never be the same
‘The revolution has begun, and Spanish-language TV will never be the same’
(10) Other tricks for keeping your NIÑAS noticeable.
other tricks for keeping your breasts noticeable
‘Other tricks for keeping your breasts noticeable.’
(11) Say your words softly and slowly in ESPAÑOL.
Say your words softly and slowly in Spanish
‘Say your words softly and slowly in Spanish.’
(12) I can remember one Latino actually questioning our identity when we first moved
into a Denver suburb: “You don’t want to live with the RAZA, eh?” he said.
into a Denver suburb: “You don’t want to live with the race, eh?” he said.
‘I can remember one Latino actually questioning our identity when we first
moved into a Denver suburb: “You don’t want to live with Latinos, eh?” he said.’
(13) Send him UNA CARTA DE AMOR.
Send him a
letter
of love
‘Send him a love letter.’
(14) It helps to know which items are worth the sticker shock, Y CUALES not.
It helps to know which items are worth the sticker shock, and which not.
‘It helps to know which items are worth the sticker shock, and which not.’
(15) Start your new business before quitting your full-time TRABAJO.
Start your new business before you quit your full-time job
‘Start your new business before you quit your full-time job.’
(16) The ultimate TOSTONES
the ultimate mashed plantains
‘The ultimate mashed plantains’
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Linguistic change and social meaning
369
(17) 1 leaf CULANTRO or 2 sprigs cilantro
1 leaf (type of cilantro) or 2 sprigs cilantro
‘1 leaf culantro or 2 sprigs cilantro.’
With the exception of the switches in (16) and (17), linguistic necessity is not the
driving force behind the codeswitches in the English version, since clearly there are
simple equivalents to the Spanish terms used in all the other examples given above.
Nor can we use the psycholinguistic argument of the most common word phenomenon,
which is certainly relevant when speakers are in conversation and need to keep the flow
of conversation going. Here, in written discourse, where the writers have time to choose
their words, the psycholinguistic argument does not apply.
Instead we need to look to social factors to understand the motivation behind the
switches. Gumperz (1982) identifies a number of social motivations for codeswitching,
with the idea that switching can be a conscious choice meant to create a certain environment between speaker and listener. Of these, switching to emphasize group identity
(Blom & Gumperz, 1972) and to create unity are most consistent with the three respondents’ replies and provide the most intuitively satisfying account of the codeswitches.
What needs to be added to these motivations is the emotional/evocative dimension of
language that the respondents mention. As one of the contributors to Latina observes,
“Our language gives us powerful and enticing sounds: A breathless ‘te deseo’ is far
more provocative than a matter-of-fact ‘I want you’.” In this case, the writer seems to
be saying that emotions are better expressed in Spanish, because words sound more
powerful in Spanish, that the sounds of Spanish are somehow more emotive/sensuous
/provocative than matter-of-fact sounding English. Obviously, there’s also the cultural,
ethnic emotionally-charged connotative aspects of meaning that cannot be easily teased
out from the “sounds of the language.” Take, for example, the use of hermanos in (3).
In the preceding sentences, the article was talking about the appearance of Latinos in
recent Broadway productions. Although the writer could have used “Latino actors” or
“Hispanic male actors,” the use of the codeswitched hermanos has a greater psychological
impact and more emotional appeal to ethnic identity. The word hermanos captures the
same sense of solidarity as the word ‘brother’ does in English when used by AfricanAmericans to refer to other African-Americans. The Spanish words in examples (4) – (13)
also fall within the category of codeswitching to evoke ethnic bonding. They serve as
reminders of the Latino culture. The words, mujer, familia, raza, and una carta de amor,
and so forth are not rare or difficult words and could have been expressed easily in
English. Yet their appearance in Spanish creates an intimate or personal domain that
non-Spanish speakers or monolingual speakers of either English or Spanish language
could not share. Of the word familia, one respondent commented that it has “more
emotional power when you use the Spanish word. La familia is very different than an
American family. Familia means the close knit Latino family.” The switches in example
(8) are from a mixed-code article called, “Gentlemen prefer gorditas.” Gorditas refers to
“pleasantly plump” or “plus-size” Latinas. The use of flaca instead of “skinny” and son
curvas instead of “their curves” continues to feed the image of the admired full-figured
Latina that gordita implies.
The switches in examples (14) and (15) are different than those in (3) through (13).
They are neither idiomatic or ethnically significant. As mentioned earlier, they fit in the
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
370
S. Mahootian
category of switches that are syntactically allowable, and hence fit a general pattern of
codeswitching, but are not otherwise “meaningful” as switches. The largest proportion
of switches, 77%, were of the emotive, ethnically bonding type. Of the remaining 23%,
8% were idiomatic and 15% were either from functional categories or from nonemotive
lexical items. This pattern of switching is consistent with my claim of a relationship
between code choice and Latino bilingual identity.
5 The structure of codeswitching
Thus far I have shown that intentional use of mixed-code discourse is functionally distinct
from the L1 and L2 varieties available to the bilingual speaker. In this section, I look
at the structure of codeswitching and determine that at this level, too, codeswitching
is distinct from its L1 and L2 counterparts.
Over the last two decades, we have seen numerous syntactic accounts of codeswitching. Most have introduced constraints specific to codeswitching (Belazi , Rubin,
& Toribio, 1994; Halmari, 1997; Joshi, 1985; Myers-Scotton, 1993; Pfaff, 1979; Sankoff
& Poplack, 1981). A few have appealed to universal principles of syntax to account for
switches within utterances (Bentahila & Davies,1986; Mahootian, 1993, 1996, 1999;
Woolford, 1983), and within words (Mahootian, 1993, 1996, 1999). In all cases, researchers
agree that mixed utterances are a product of a rule-governed system that recognizes
two linguistic systems as its input. The appeal of the nonconstraint-based models is in
their application of language universals coupled with the implication that the “ideal
speaker” is not necessarily monolingual.
Here, I use the model developed in Mahootian (1993) to discuss intrasentential
codeswitching. The basic assumptions of my model are (p.139):
1. syntactic structure is projected from the lexicon and bilingual speakers have
access to the lexicons and therefore syntactic structures of both of languages,
2. some grammatical categories are universal (i.e., there are equivalent node labels
across categories).
Implicit in the first assumption is that codeswitching does not require a third
grammar. Rather, the two monolingual grammars of the bilingual are kept separate and
intact. These assumptions in conjunction with the fundamental relationship between
heads and complements in standard X-bar theory and the minimalist program (Chomsky,
1995) yield the Head-Complement Principle (Mahootian, 1993, 1996, 1999):
The language of a head determines the syntactic properties of its complements in
codeswitching and monolingual contexts alike.
Specifically, I mean that where there is a head-complement relationship, the
language of the head imposes its syntactic requirements on its complements, determining
their phrase structure position, category and feature content. For example in an SOV
language such as Japanese or Farsi, the verb which is the head of the VP requires its
object complement to be on the left. In codeswitching, where two languages are involved,
the language of the head dictates the order of complements, the complements may be
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Linguistic change and social meaning
371
in either language. In the Farsi example in (18), the object DP xormalu, in typical Farsi
word order, precedes the verb.
(18) mæn xormalu
dust daræm
I
persimmon love
‘I love persimmons.’
Examples (19) and (20) illustrate the head-complement principle’s application to
codeswitching data. The two languages involved, Farsi and English, are structurally
incongruent. In sentence (19), the head of the VP, “love XORMALU,” is the English
verb ‘love’. In English verbs take their complements to their right. As predicted by the
HCP, despite the fact that XORMALU is a Farsi DP, it takes its place to the right of
the English verb.
(19) I love XORMALU
I love persimmon
‘I love persimmons.’
The HPC also accounts for switches between bound and free morphemes, as
shown in example (20):
(20) nemidune MOTIVATION-eš
či-ye.
doesn’t know motivation–3rd sing poss pron what-is
‘He doesn’t know what his motivation is.’
In Farsi, bound possessive pronominal inflection heads subcategorize for leftbranching DP complements. The complement itself may be drawn from the English
lexicon or the Farsi lexicon, as long as the subcategorization requirements of the inflectional head are met.
As a syntactic model reflective of linguistic competence, the HCP model does not
distinguish between “codeswitching” and “borrowing.” In fact, as I have claimed in the
past, the distinction made between borrowing and codeswitching is neither necessary
nor well motivated for syntactic analyses of mixed utterances.3
In the same way, the HCP model accounts for all of the head-complement
codeswitches in the mixed-code discourse in Latina. In example (13), rewritten here as
(21) the English verb send takes two complements. As predicted by the HCP, the two
complements may be in either English or in Spanish. In this case, one is in English and
the other is in Spanish:
(21) Send him UNA CARTA DE AMOR.
Send him a
letter
of love
‘Send him a love letter.’
But what about switches involving pre- and post-nominal adjectives? In Mahootian
and Santorini (1994, 1996) and Santorini and Mahootian (1995), I account for such
switches by showing that adnominal adjectives are phrasal adjuncts rather than heads.
3
For a more detailed discussion of switches between free and bound morphemes and the head-complement principle,
see Mahootian (1993, 1996).
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
372
S. Mahootian
In languages such as Spanish speakers have access to structures for both pre- and
postnominal adjectives. Since neither the AP or NP is a head, both are able to be in
either language. Thus, we can get switches such as in example (22), where a Spanish
postnominal adjective is used in pronominal position with an English noun.4
(22) Forget those FLACA cover girls
Forget those skinny cover girls
‘Forget those skinny cover girls!’
In summary, the structure of codeswitching, while governed by universal constraints
that apply to every language, is nevertheless unique in that the output is neither L1or
L2 at discourse, utterance and word level. Must we then hypothesize a third grammar
specific to codeswitching? I maintain that such a grammar is unnecessary. Rather, I am
suggesting that, at the syntactic level (as with the sociopragmatic level), the mixed-code
discourse and its instantiation, codeswitching, is a variety like the L1-only and L2-only
varieties available to bilingual speakers. At the abstract point of input, codeswitching
means that the rules, lexicon, and so forth of two grammatical systems, rather than
one, must be drawn from and negotiated to produce the L1-L2 variety. The outcome of
this negotiation is mixed-code utterances which are considered grammatical by the
speakers of the codemixed variety.
6 Codeswitching as language change
Traditionally, language change in-progress was considered an unobservable phenomenon
(de Saussure, Bloomfield among others). Only through the comparison of two synchronic
grammars of a language at different historical periods could one observe change. Lay
commentary on the “corruption” or “decay” of a community’s language brought about
by the introduction of new slang or borrowed words or deviation from prescriptive forms
was not considered evidence of observable language change.
Labov’s seminal work on language variation introduced variability as a window
through which one could observe change-in-progress. He states, “the origin of a change
is its ‘propagation’ or acceptance by others” and that “we can say that language has
changed only when a group of speakers use a different pattern to communicate with
each other.” (1972a, p.277). Although Labov’s studies were based on observations of
monolingual communities, similar arguments for change-in-progress can be made for
some bilingual communities.
Here, I have shown that codemixed discourse and codeswitching are (1) accepted
and propagated by others, and (2) that a group of speakers uses a pattern different
from other preexisting patterns (i.e., L1-only and L2-only) to communicate with each
other. I’ve shown that codemixed discourse is itself a speech variety and that codemixed
discourse and codeswitching may be used intentionally to index the “bilingual identity.”
4
One consequence of the phrasal adjunct analysis is that it predicts that codeswitching between languages with
pre- and postnominal adjectives we may yield switches where the noun and adjective of L1 appear in L2 order.
This prediction is, in fact, borne out in a number of examples in the literature. One such example appears in Narty
(1982): e hé HOUSE RED ò (‘S/he bought the red house’. Adangme-English). Other examples can be found in
Mahootian and Santorini (1996).
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Linguistic change and social meaning
373
For Spanish-English bilinguals, codeswitching reinforces their affiliation with two
cultures, asserts their identity as bicultural bilinguals while allowing them to maintain
strong ties with the home culture.
The appearance of codeswitching in written text is solid evidence that this variety
has been accepted: where there were two options, Spanish and English, there is now a
third, Spanish-English. Discourse which includes intentional codeswitching is not only
pragmatically but also structurally different than discourse which is conducted in allEnglish or all-Spanish. Producing mixed utterances entails two grammars negotiating
their head-complement relationships consistent with each grammar’s subcategorization
rules. The resulting mixed forms are fully grammatical, adhering to the principles of
grammar in general while maintaining the integrity of the separate rules of each of the
two grammars. The structural change referred to involves mixed language output at
word, utterance and discourse levels.
7Conclusion
I’ve examined the relationship between intentional codeswitching and language change
with two questions in mind. The first asked the extent to which stylistic and social
variables govern code choice and codeswitching. The second looked at code choice and
codeswitching as an instantiation of language change.
Use of codeswitching in a mainstream national publication is evidence of its
acceptance and propagation. Accordingly, codemixed discourse, especially in written
media, signals language change as it is described by Weinreich et al. and by Labov: a
variety (mixed-code discourse) with a distinct structure and sociopragmatic function
is used in variation with other varieties (Spanish-only and English-only) in the Latino
community in the United States.
Based on use patterns found in a popular women’s magazine, Latina, I conclude
that intentional code choice serves to emphasize and to promote an identity which falls
somewhere between ethnic group and what Fishman (1972, p.3) describes as a nationality:
“sociocultural units that have developed beyond primarily local self-concepts, concerns,
and integrative bonds.” Mixed Spanish-English code is another means for Latinos in the
U.S.A. (a) to show unity among themselves as a subset of all Hispanic-Americans, (b)
to identify themselves as a group separate from their predecessors’ generation, and (c)
to continue to maintain strong emotional ties with their heritage. I’ve further suggested
that the use of mixed-code as a variety pragmatically and structurally separate and
distinct from coexisting monolingual Spanish and English texts and codeswitching is
an indication that language change has been brought about.
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
374
S. Mahootian
References
ANTAKI, C., & WIDDICOMBE, S., (Eds.). (1998). Identities in talk. Sage Publishers, London.
BACKUS, A., & EVERSTEIJN, N. (2002). Pragmatic functions and their outcomes: Language
choice, codeswitching, and nonswitching. Forthcoming in the Second University of Vigo
International Symposium on Bilingualism Proceedings.
BELAZI, H. M., RUBIN, E. J., & TORIBIO, J. A. (1994). Codeswitching and X-bar theory: The
functional head constraint. Linguistic Inquiry, 25, 221 – 237.
BELL, A. (1984). Language style as audience design. Language in Society, 13, 145 – 204.
BELL, A. (2001). Back in style. In P. Eckert & J. Rickford (Eds.), Style and sociolinguistic variation. Cambridge: Cambridge University Press.
BENTAHILA, A., & DAVIES, E. E. (1983). The syntax of Arabic-French codeswitching. Lingua,
59, 301 – 330.
BLOM, J. P., & GUMPERZ, J. J. (1972). Social meaning in linguistic structure: Codeswitching
in Norway. In Li Wei (Ed.), The bilingualism reader. London: Routledge.
BOURHIS, R. Y., & MARSHALL, D. E. (1999). The United States and Canada. In J. A. Fishman
(Ed.), Handbook of language and ethnic identity. New York: Oxford University Press.
BYNON, T. (1977). Historical linguistics. Cambridge: Cambridge University Press.
CARBAUGH, D. (1996). Situating selves: The communication of social identities in American
scenes. New York: SUNY Press.
CHOMSKY, N. (1995). The Minimalist Program. Cambridge, MA: The MIT Press.
CROWLEY, T. (1992). An introduction to historical linguistics. Auckland: Oxford University
Press.
FISHMAN, J. (1964). Language maintenance and language shift as fields of inquiry. Linguistics,
9, 32 – 70.
FISHMAN, J. (1972). Language and nationalism: Two integrative studies. Rowley, MA: Newbury
House.
GROSJEAN, F. (1982). Life with two languages: An introduction to bilingualism. Cambridge, MA:
Harvard University Press.
GUMPERZ, J. (1982). Discourse strategies. Cambridge: Cambridge University Press.
GUMPERZ, J. J., & COOK-GUMPERZ, J. (1982). Introduction: Language and the communication of social identity. In J. J. Gumperz (Ed.), Language and social identity. Cambridge:
Cambridge University Press.
HALMARI, H. (1997). Government and codeswitching: Explaining American Finnish. Studies in
Bilingualism 12. Amsterdam: John Benjamins Publishing Company.
HOCK, J. (1996). An introduction to historical and comparative linguistics. New York: Mouton
de Gruyter.
JOSHI, A. K.(1985). Processing of sentences with intrasentential codeswitching. In D. Dowty, L.
Kartunen & A. Zwicky (Eds.), Natural language processing: Psycholinguistic, computational
and theoretical perspectives. Cambridge: Cambridge University Press
KACHRU, B. B. (1982). The bilinguals linguistic repertoire. In B. Hartford, A. Valdman, & C.
Foster (Eds.), Issues in international bilingual education. New York: Plenum Press.
LABOV, W. (1972a). Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press.
LABOV, W. (1972b). The recent history of some dialect markers on the island of Martha’s
Vineyard. In L. Davis, (Ed.), Studies presented to Raven McDavid. University of Alabama:
University of Alabama Press.
MAHOOTIAN, S. (1993). A null theory of codeswitching. Doctoral dissertation, Northwestern
University. Evanston, IL.
MAHOOTIAN, S. (1996). A competence model of codeswitching. In Sociolinguistic variation:
Data, theory, and analysis. Selected papers from NWAVE 23 at Stanford University. Stanford,
CA: CSLI Publications.
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016
Linguistic change and social meaning
375
MAHOOTIAN, S. (1999). Codeswitching and universal grammar: Insights from bilingual
language acquisition. 2nd International Symposium on Bilingualism. Newcastle upon Tyne,
England.
MAHOOTIAN, S. (2000). The medium is the message. Seventh International Pragmatics
Conference. Budapest, Hungary.
MAHOOTIAN, S. (2002). Sending a message: Codeswitching and the bilingual identity. Paper
presented at the Second University of Virgo International Symposium on Bilingualism, Virgo,
Spain, October.
MAHOOTIAN, S., & SANTORINI, B., (1994). Adnominal adjectives, codeswitching and
lexicalized TAG. In 3e Colloque International sur les Grammaire d’Arbre Adjoint(TAG + 3)
rapport Technique TALANA-RT-94-01.
MAHOOTIAN, S., & SANTORINI, B. (1996). Codeswitching and the complement/adjunct
distinction. Linguistics Inquiry, 27(3), 464 – 479.
MYERS-SCOTTON, C. (1993). Duelling languages: Grammatical structure in codeswitching.
Oxford: Clarendon Press.
NARTY, J. (1982). Codeswitching, interference or faddism. Language use among educated
Ghanaians. Anthropological Linguistics, 24, 183 – 193.
PFAFF, C. (1979). Constraints on language mixing. Language, 55, 291–319.
ROMAINE, S. (1995). Bilingualism. Blackwell Publishers
SANKOFF, D., & POPLACK, S. (1981). A formal grammar for codeswitching. Papers in linguistics 14, 3–46.
SANTORINI, B., & MAHOOTIAN, S. (1995). Codeswitching and the syntactic status of adnominal adjectives. Lingua, 96, 1 – 27.
SEBBA, M. (2002). Regulated Spaces: language alternation in writing. Forthcoming in the Second
University of Vigo International Symposium on Bilingualism Proceedings.
TABOURET-KELLER, A. (1997). Language and identity. In F. Coulmas (Ed.), The handbook
of sociolinguistics. London: Blackwell Publishers.
WOLFRAM, W., & SCHILLING-ESTES, N. (1998). American English: Dialects and variation.
Malden, MA: Blackwell.
WEINREICH, U., LABOV, W., & HERZOG, N. (1968). Empirical foundations for a theory of
language change. In W. H. Lehman & Y. Malkiel (Eds.), Directions for historical linguistics.
Austin: University of Texas Press.
WOOLFORD, E. (1983). Bilingual codeswitching and syntactic theory. Linguistic Inquiry, 14,
520 – 536.
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at PENNSYLVANIA STATE UNIV on May 12, 2016