University of Massachusetts Amherst

From the SelectedWorks of Benjamin Bailey

2007

Heteroglossia and boundaries

Benjamin Bailey, University of Massachusetts-Amherst

Available at: https://works.bepress.com/benjamin_bailey/55/

 I! I
12
heteroglossia and boundaries
___________ _
b�
n
e l��Mk¥
processes of linguistic and social distinction

Language is the primary semiotic tool for representing and negotiating social
reality, and it is thus at the centre of social and political life . Among its myriad
social and political functions is to position speakers relative to a wide variety
I
of phenomena including co-present interlocutors, the activities in which
speakers are engaged, and various dimensions of the wider world, including
II I
social identity categories and their relative value. To speak is thus to position
oneself in the social world, i.e. to engage in identity practices (d. Le Page

IIII'I
and Tabouret-Keller 1985 ) .
All language provides l inguistic a n d discursive forms rich in social
connotations for the negotiation of identity. Monolingual individuals exploit

I1I,I
various registers, accents, sociolects, word choices, etc. for the omnipresent
tasks of positioning themselves and others within social categories and the
larger social world. Members of bilingual communities typically have an
expanded set of linguistic resources for these ongoing social negotiations
and often a broader range of relevant social categories to enact or contest.
On the linguistic level, they can draw forms from two languages as well as
hybrid forms resulting from language contact. On the social and cultural level,
many straddle social and cultural boundaries and are familiar with relatively
diverse cultural frameworks for interpreting and evaluating the world and
positioning themselves and others within it. I
jl
In this chapter, I use the notion of heteroglossia (Bakhtin 1981) as a
conceptual entree to social meanings of bilingual speechl and related identity
negotiations. Heteroglossia addresses (a) the simultaneous use of different kinds

I
of forms or signs, and (b) the tensions and conflicts among those signs, based
on the sociohistorical associations they carry with them (d. Ivanov 2001).
The first part of this definition of heteroglossia subsumes formal definitions of
bilingualism (as the coexistence of two linguistic systems) or code-switching
(as the alternation of codes within a single speech exchange) (Gumperz 1982;

257
 )'>/1
259
identity practices

11('11('1 I (IKK), Wllik IlclL

heteroglossia and boundaries

'roglossia denotes the use of different kind

(II \li�ll\ 111l' 1\'1'111 dol's
s of forms The same semiotics of distinction involved in social boundary work can be
Ilot refer, particularly, to the 'dist inct

I
lang uage s' that applied to ways of speaking (ct. Irvine 2001 on 'style'). Ways of speaking are
( ()IiI\'d 111\' l{lIssi;1I1 lerm
,11(' ( )I1I1I1(1If1y seell as cons
titut ing bilin gual ism. To the con
trary, Bakh tin constituted as distinct through contrast, rather than through any inherent
raznorechie to refer to intra-lang

11I
uage varieties with in characteristics, just as identities are constituted through boundary marking
1('1111 i.\ .\(l/llL'liIIiCS tran
/{I1,\\!.III, v.lrivlil's with
com peti ng social and poli tical imp
licat ions , and the processes (or lack thereof). As is the case with identity categories, what counts
slate d as 'the soci al diversity of
spee ch type s' rather as a socially meaningful opposition among linguistic forms is subj ective,
111.111 'IIvlcroglossia'. The
fact that heteroglossia encompa
sses both mon o. shifting, and ideologically infused. To a linguist whose perspective privileges
01 1;lllguage that is not poss
,)fld IIll1ll ilingual form s
allow s a leve l of theo rizin g abou
t the soci al natu re formal categories, for example, any bilingual speech may be highly salient
ible with in the con fine s of a focu
\wllcilillg .
s on code ­ because of the alternation of two codes, thus constituting a distinctive style.
Whilc code -swi tchi ng research com mon To the monolingual majorities in the US and most of Western Europe, mono­
ly treats the disti ncti vene ss of
codc s as a given, from a phen ome nologica lingualism is an emblem of citizenship and belonging, and any language
l perspective, languages or codes
Gill only be und ersto od alternation is an exercise in distinctiveness. To a bilingual child of international
as disti nct objects to the extent
to which they are migrants, however, code-switching in intragroup peer interaction may not
I rca ted as such by soci al
acto rs. From the soci ally- infu sed
pers pect ive of commonly be perceived by members as very distinct from speaking to such
hctcroglo ssia, judg eme nts about wha
t counts as 'different kinds of form
s ig n s' are based on the way s or peers without alternation.

I
social actors appear to disti ngui sh
among forms, These subjective processes of differentiation are linked to power, and in
rather than anal ysts' a priori claim
s.
The seco nd part of the defin ition relatively stable social and linguistic situations, the social and linguistic
of hete rogl ossia capt ures the inhe
rent
1 1' 1
poli tical and soci ohis toric al asso categories favoured by dominant groups come to be seen as natural through
ciati ons of any ling uisti c form
inde xica l mea ning s (Peirce 195 5), , i . e . its processes of hegemony or symbolic domination (Bourdieu 1991; Heller
or socia l conn otat ions . Thes e inde
meanings, or histo rical voic es, are xica l 1995) . In the United States, for example, White, English monolinguals are
not expl icit or stati c, but rather mus
interpreted on the basis of cons tella t be the dominant group, both economically and politically. This economic and
tion s of forms in part icula r interactio

I1I 1
and socio histo rical cont exts. Such nal political power is extended, via symbolic domination, to broader standards for
meanings are thus shift ing, subj
and negotiated. ectiv e social evaluation. Even though the majority of historical immigrants to the US
I appr oach iden tity in simi larly have not been English-speaking, and even though the country is increasingly

I
processual term s. Follo wing Barth's
sem inal work on ethnic groups, (1969) Latino and non-White, being a monolingual Anglophone, speaking a variety of
I approach identity as constituted
the boundaries that groups cons truc through Standard English, and being White constitute one as an unmarked American.
t between them selves, rath er than
the
I
characteristics of group members. Highly naturalized categories of race, language and national identity thus
The term 'identity' comes from Latin

III
mea ning 'the same ', and iden titie , idem, merge in the popular mind into an essentialized unity.
s are cons titut ed by SOCially coun
'the sam e' as others or coun ting as ting as Social change, through m i gration, can serve to denaturalize and
'different' from others. This form
foregrounds the subjective, social ulati on problematize boundaries and essentialized unities. Bilingual children of
reality of individual actors, in that
judgements and activities, rather than it is their international labour migrants, for example, problematize boundaries through
static characteristics of individuals,
that straddling linguistic and social worlds in their language and identity practices

I
serve to cons titut e categories. Soci
al identity is a function of two subj
proc esses : 'self -ascr iptio n' - how ectiv e (Auer 1984; Poplack 1988; Gal 1 988; Zentella 1997; Bailey 2002) . Similarly,
one defin es ones elf - and 'ascr iptio
othe rs' - how others define one (Bart n by in urban, postcolonial contexts in Africa, migration has blurred traditional
h 1969: 13). Thes e processes of ascri

II I
are not based on the objective sum ption social boundaries and associations of particular identities with particular ways
s of differences or simi larit ies amo
groups: 'som e cultu ral features are ng of speaking (Myers-Scotton 1993). Code-switching in such contexts is often
used by the actors as signals and emb
lems

II
of differences, others are ignored, frequent, intra-sentential, and unmarked in intragroup peer interactions,
and in some relationships radical diffe
are played down and deni ed' (ibid rences serving as a form of unmarked, discourse contextualization or serving no
: 14). All indi vidu als have mult iple
acteristics and allegiances, so it is char­ identifiable function at all. By failing to treat two or more languages as a
the situa tion al and selective high
of com mon altie s and differences light ing meaningful opposition in certain contexts, such social actors effectively erase
that is characteristic of identity grou
( Moerman 1965). Analysis of iden ping s the boundary that constitutes the two languages as distinct. This calls into
tity thus revolves around the ques
how, when and why indi vidu als tion s of question the very foundation of bilingualism as a rubric or perspective for
count as members of particular grou
ps. SOCially-oriented research on language.
 260 identity practices 261
heteroglossia and boundaries

Ch i ldren or i m m igran t s a l so undermine naturalized language-race-nation Language is popularly understood as denotation, in which words stand
un i t i es hy (a) assert i ng or e n acting identities
that cross-cut received categories for, or represent, things or ideas. The function of language is seen as the
i n ways t h a t expose and controvert the assumptions on which the categories
a re ha sed, a n d (h) b r i n g i n g with them social classification systems that
communication of propositional information, as in a conduit metaphor

w e re d o m i n a n t in the countries of origin but that contradict host country

(Reddy 1979), and the maximally efficient transfer of such information is
seen as 'good communication'. This folk theory of language, communication
ca t egories (Ba i l e y 2001 ) . In the United States, for example, the categories and efficiency is intertwined with living in capitalist, industrialized, bureau­
Bl acK <Jn d W h it e have been a central organizing principle of society for cratized, widely literate societies, which privilege certain types of productivity

Ii I
cel1t uri es, b u t many post-1965 immigrants from Latin America and Asia do and efficiency. Social functioning of language is seen as epiphenomenal in
110t lit n e,lt l y into these categories. Latinos from the Caribbean, in particular, this folk model, except when social variation is perceived as impairing the
hridge l h e categories Black and White in ways that undermine the popular efficient transmission of propositional information.

!I I
Ameri can notion of Black and White as representing unbridgeable distance.
�uch i mmigrants not only bridge such categories in terms of phenotype, they
This folk understanding is layered with hegemoniC ideologies that privilege
language varieties that are associated with powerful and privileged groups in
a l so maintain understandings of themselves as essentially Latino or Spanish,
thereby countering the primacy of phenotype-symbolized race in US social
society. These varieties are seen as being 'accentless' and 'correct', and ideal
for 'good', i.e. propositionally efficient, communication. D ivergences from
claSSi fication. This problematizing of essentialized boundaries can occur in this ideal standard - whether associated with class, ethnicity/race, or region
a n y situation in which the macro-social categories in countries of origin, _ are considered marked and less desirable. In the United States, for example,
the result of speCific histories of social relationships, do not match those in varieties of English that are associated with written language and the speech
destination countries ( Mittelberg and Waters 1992) . of educated, middle- and upper-class White Americans in the Midwest (Lippi­
I n this chapter, I first contrast popular and formal linguistic approaches to Green 1997; Silverstein 1996) are identified as normal and desirable, even by
language with more heteroglossic, social ones to contextualize the historical speakers whose speech does not approach these varieties.
privileging and constitution of code-switching as a discrete object of study. I Formal linguistics also treats the social and political functioning of language
then briefly review research on code-switching in order to introduce the type as marginal, approaching language as a semiotic system in and of itself. The
of code-switching - code-switching as a discourse mode (Poplack 1980) - that primary interest is in relationships among elements of this system, abstracted
particularly problematizes assumptions of difference in bilingual talk. A short from any actual uses or instances of language. The boundaries of the system
segment of bilingual speech from Dominican American peer group interaction are implicitly taken to be the boundaries of the language, an idealization
is presented to highlight the utility of the notion of heteroglossia relative to a that is not necessarily compared to actual speech, which may or may not be
narrower focus on code-switching, in analysing identity negotiations in talk. bilingual (see Auer, this volume) . Linguists generally focus on meanings that
Finally, I argue that code-switching and bilingualism might best be approached remain stable across time, speakers and contexts - i.e. denotational meanings

I' I
as social constructions, as is common with identity categories. The social _ and pay relatively little attention to actual use of language, i.e. social action.

I
constructionist perspective affords analytical insights while recognizing the In taking a formal, synchronic approach to language, formal linguists thus
power of the on-the-ground hegemonic social reality that bilingual speakers neglect relationships between linguistic forms and the social and political
face in societies with monolingual language ideologies. worlds that are described and negotiated through those forms.
Both folk and formal linguistic models of language may reflect more general
popular and linguistic approaches to language cognitive predispositions. Silverstein (2001), for example, has shown that
awareness of pragmatic function of language is closely tied to the degree
Among the intertwined reasons for the distinctive salience of bilingualism to which forms are referential, segmentable and context-reflecting. Thus,
in Western societies are a) the naturalization of monolingualism in the
ongOing nation-building projects begun in Europe during the last several
people can articulate form-meaning relationships relatively well when
the relationship is referential; when the forms are segmentable (as are I
centuries, and b) the referentialist nature of modern, Western language morphemes) rather than scalar (as is pitch); and when the forms refer to some
i deologies, both popular and academic. As the historical development of pre-existing phenomenon or situation rather than establishing it through
language-nation ideologies is explicitly discussed in Part One of this volume,
I briefly review contemporary popular language ideologies and epistemology
the act of speaking. The popular notion of language as a system of discrete
(segmentable) symbols for describing (referentially and context-reflecting) a

1'1
in formal linguistics. pre-existing world, as well as the linguistic emphasis on form, may be partly a
 /(,
263
identity practices

1 \'\1 ill ,)1 I It <I I

heteroglossia and boundaries
I Ill' I.I( I

I f l an
dimensions of language are simp
�Oille
10 ("11.\\11111' :IW:lrCIlCSS
ly more accessible form. From these perspectives, language is never a neutral instrument of pure
"'" IIII II n l l l iI popu Iii r perspective
others.

I H '1','1 II '\ II v(' 01 la nguage-as-synchroni

reference, as actual speech always occurs in a social context, which is never
of language-as-denotation and
a linguistic
,c1 11'1 I" I II II l. Frolll I he linguisti
neutral or ahistorical. Talk and texts thus need to be understood in terms of
c-system, cod e-switching repr
esents an past and ongoing social and political negotiations of which they are a part,
'y'II'III.\ II I:I[ arc 1/ priori conceptu
c point of view, it involves the over
lapping of two not as forms in isolation:
vivw, II :qllK'ars redu nda nt
alized as distinct and discrete . From
a popular
and irrationa l. I f lang uag e is
vvllick (II propositi ona
an inst rum enta l The living utterances . . . cannot fail to brush up against thousands of living
l information, then why use mor
e than one lang uage ,
1111 il',\S :J linguistic defic
iency prevents one from making dialogic threats, woven by socio-ideological consciousness around the given
III Olll' language (Woolard 200 utte ranc es enti rely object of an utterance; it cannot fail to become an active participant in
(; ()t i l'-s w i tched speech also dive
4: 75; Meeuwis and Blommae
rt 1998: 78)?2 social dialogue. After all, the utterances arises out of this dialogue as a
rges radi cally from the ideo logi
of spcaking, which is mon cal idea l way continuation of it and as a rejoinder to it - it does not approach the object
such as pho nological effects of
olin gual and shows no effects
of lang uage contact from the sidelines. (Bakhtin 1981 : 276-7)
other languages (l.e. 'foreign acce
The pop ular and linguisti
nt').
c perspectives that monolingual
and natu ral subtly inform code speech is norm al Non-referential indexical meanings are linked to history just as Bakhtinian
-switching research even in mor
politically oriented disciplines such e socially and voices are. Such indexical form-meaning relationships arise from historical
as anthropology. Researchers in
anthropology and related disciplin linguistic usages by speakers in particular social positions. It is through recurrent
es regularly attend to bilingual
ways that they do not necessari spee ch in connections between a social phenomenon or context and a linguistic form
ly attend to mon olin gual speech.
question of much anthropologi The driv ing that non-referential indexical meanings are constituted to begin with (Peirce
cal cod e-switching research, 'Wh
do that?' is not put with equ al y do they 1 955 ) . Particular phonetic patterns index a speaker's regional origins, for
force to the myriad other things
in talk that are not bilingual. By speakers do example, only through associations with actual speakers and their talk.
casting bilin gual speech as a mar
that calls for exp lana tion of a ked form The emphasis on the inherent political dimension of all talk separates the
type that mon olin gual speech
research imp licitly reproduces does not, such Bakhtinian notion of heteroglossia from many sociolinguistic approaches
the folk and linguistic beli efs of
speech as the natural form (Wo mon olin gual to language variation, in which correlations between forms and categories
olar d 200 4: 75) .
of speakers is the focus, rather than the sociohistorical connotations or
voices of the forms. In US correlational sociolinguistics (d., e.g. Rickford
heteroglossic and anthropological approaches to language and Eckert 2001), social categories have been treated as given, i.e. pre-existing
any interaction, and the agency of individuals and the role of ideology in
In contrast to formal and folk-models of language, heteroglossia takes as its
language use has been downplayed, or even denied (d., e.g. Labov 1979: 328) .
starting point the social and pragmatic functioning of language: ' ... verbal
Emphasizing the political and historical dimensions of language also stands
discourse is a social phenomenon - social through its entire range and in
in stark contrast to formal linguistics, of course, which takes a synchronic
each and every of its factors, from the sound image to the furthest reaches
approach to language and meaning.
of abstract meaning' (Bakhtin 1981: 259) . The notion of heteroglossia is
While indexicality and Bakhtinian heteroglossia encourage social and
thus congruous with traditions in anthropology of approaching language as
political readings of language and interaction, they are very difficult to opera­
essentially social, e.g. 'the main function of language is not to express thought,
tionalize. Indexicality, for example, has been described in relatively concise
not to duplicate mental processes, but rather to play an active pragmatic part
theoretical terms, but it encompasses a very large range of phenomena. Non­
in human behaviour' (Malinowski 1965: v2: 7). Thi s orientation does not
referential indexical forms are highly varied, ranging from phonetic features,
deny the referential, or denotational, function of language, but conceptual­
to word choice, to visual features, to other stylistiC dimensions of talk. The
izes reference as merely one pragmatic function among many in a system
nature of indexical objects and the spatial and temporal distance between
that is intrinsically pragmatic (Silverstein 1976: 20).
the indexical form and its object can also vary greatly (Silverstein 1992).
The Bakhtinian notion of heteroglossia overlaps in Significant ways with the
While the scope of this semiotic dimension of talk makes it an attractive area
semiotic and linguistic anthropological notion of non-referential indexicality
for theorizing, it is difficult to make such theory operational because of the
as developed by Peirce (1955) and Silverstein (1976 ) . Both are types of inter­
social, contextual and individual specificity of form-meaning relationships
textuality, in which meanings of forms depend on past usages and associations
in the indexical mode (Hanks 2001 : 120).
of those forms rather than the arbitrary referential meaning carried by the

IIIII
 I�
264
identity practices
heteroglossia and boundaries 265

code-switching research: from degeneracy, did not generate metaphorical meanings. Auer (1984, 1 988, 1995, 1998)
to metaphorical strategy, to local discourse pioneered an influential and ongoing tradition of work that linked such
contextualization, to no identifiable function socially unmarked code-switching to the omnipresent local exigencies of

Researc h Oil soci a l

coordinating interaction. In this conversational management, or discourse
functions and meanings of code-switching since the
contextualizing, mode, switches do not necessarily co-occur with external

I !I
IlJSOs h,IS repea t edly shifted in emphasis. Many language analysts through
changes in the context (situational switches) or effect significant shifts in
t i l l' I%Os, i ncluding those focusing on bilingualism (e.g. Weinreich 1953:
sociocultural framework (metaphorical switches). Individual switches instead
73) t rea t ed ( i nt ra-sentential) code-switching as linguistic interference, with
COllllota lions of linguistic, social or cognitive deficit. This orientation simply
serve local contextualization functions. In such unmarked discourse contex­
tualization switching, conventionalized associations between particular codes
m i r rors t he dominant Western language ideology of monolingual speech
as I l or m a l and code-switching as a mixed-up jumble that reflects speakers'
i nability to speak properly.
Sta rting in the early 1970s, J ohn Gumperz (Blom and Gumperz 1972;
and social worlds are at least partially suspended by participants (although
not necessarily by co-present overhearers). Shifts in codes function as signals
that there is a concurrent shift in speech activity - e.g. to a repair sequence or
to a rhetorical question - but not necessarily in sociocultural framework. The
,[
Gumperz 1982) led research that countered this orientation, presenting code­
act of alternation itself, rather than the direction of the shift, is the important
switching as a form of skilled performance through which individual social
dimension for discourse contextualization. This use of code-switching as a
actors could communicate various social and pragmatic meanings. Rather
tool for local discourse management has been documented by numerous
than representing a linguistic deficit or degeneration, code-switching was
researchers, often with disparate labels. Zentella (1997) refers to such switching
presented as complex, systematic and socially strategic behaviour.
in terms of 'conversational strategies', 'in the head communicational factors',
Gumperz's (Blom and Gumperz 1972) early formulation focused on two
and, following Goffman (1979), 'footing'. Myers-Scotton (1993) calls this
functions of switches. Situational switches were made in response to changing
Situations, e.g. the need to accommodate to co-present speakers in one
'code switching as the unmarked choice', and Gumperz (1982) subsumes it
I'
Ii ,
under a larger category of 'conversational code switching'.
language or another. Metaphorical switches were more SOCially indexical. By In these contexts of unmarked code-switching among peers, it is not always
partially violating an expected situation-code correlation, some aspect of possible to ascribe any function to a particular switch. In a corpus of 1 ,685
meaning (parallel to a Bakhtinian voice) associated with the switched-into switches among young New York Puerto Rican girls, for example, Zentella
language is brought into the conversation. In such metaphorical switching, (1997: 101) assigns fewer than half of the switches to specific conversational
then, changes in language effect changes in context and social roles, without strategies, or functions, because most of the individual switches do not have
apparent prior changes in the physical or outward context. a clear, analytically defensible function and do not co-occur with particular
At one level this formulation tends toward an essentialized linking of interactional patterns. Similarly, Meeuwis and Blommaert (1998: 76) find
language, identity and sociocultural worlds. It is only when there are relatively that multilingual talk among Zairians in Belgium can represent 'one code in
conventionalized aSSOCiations between language and social meanings that its own right', and that the insistence on two distinct languages as the frame
switches generate metaphorical meanings. Empirical research has shown that of reference for this form of speech is not helpful in terms of interpreting it.
such metaphorical meanings are most often generated in situations where Such work fundamentally challenges the taken-for-granted distinctiveness of
codes are relatively compartmentalized (e.g. Kroskrity 1993) or politically the languages used in code-switching and the taken-for-granted distinctive­
charged (e.g. Heller 1 992) and have strong 'we' vs . 'they' associations (see ness of code-switching as a communicative phenomenon.
also Maehlum 1996 for a critique of Blom and Gumperz's original work) . Key to such research have been methods and epistemologies associated
More generally, empirical work shows that analysts' ideologies of languages with conversation analysis and with anthropology. Conversation analysiS,

I
as discrete systems are only sometimes consonant with the i deologies to which shaped Auer's seminal work, insists on empirical, electronically
which speakers appear to orient when talking. The empirical variation in recorded data. This insistence, in conjunction with the availability of portable
ideologies to which speakers demonstrably orient illustrates the shortcomings electronic recording devices by the late 1 970s, is one reason for the increased
of privileging the formal category of code-switching. A more socially oriented documentation of unmarked, frequent and intrasentential code-switching.
perspective, such as heteroglossia, can better account for this variation across This type of code-switching is most common in informal, intragroup peer
social and historical contexts. interaction, and portable recording devices allowed taping in the natural
Beginning i n the early 1980s, many researchers began documenting settings where this type of talk spontaneously occurs. Recordings - and
code-switching that was not done in response to changes in situation and resultant transcripts - made such switching patterns impossible to ignore.
 )(I() Identity practices
heteroglossia and boundaries 267

( '0111'1 '1\,111. III ,\ I I II y\1\ 1I11' l ltods also Ii mit the effect
I 'Y I lrl',\ni hi ng very narro w bases
.
s of individual subjectivity heteroglossia can encompass socially meaningful forms in both bilingual and
,IIIIi .1\\111111,11011\
for expla natio n and monolingual talk; (b) it can account for the multiple meanings and readings
,1I1;1I1I1l'liI.lII(III: 111l' .lIlalys l

II ill'll, 1\ 11101\ III\,II Iselves demo

can invoke only those socia l const
ellati ons that of forms that are possible, depending on one's subject position; and (c) it
W Iii it' IIIi\ \ .II I kad l o a ustere analy
nstrably treat as mean ingful in the
interaction . can connect historical power hierarchies to the meanings and valences of
\1111\ IIIr\,\ (ll ('ollVersation, the
ses that are limite d in scope to form al
particular forms in the here-and-now.
rigou r of the method has encouraged more
d('I('II�lilk l'iaillls aiJout socia

IlIll'rluel/tors p u b licly display, and

l meanings withi n code -swit ching resea
rch. [OS #2 12:26:40) Isabella (came to the US around age 7) and Janelle (US­
continuously update for each other, their
(lilgoillg understandings of born) are sitting on steps outside of the main school building at the end
talk. Implied in each turn is an understan
ding, of their lunc h period. Isabella has returned from eating lunch at a diner
or uptake, of the prior
turn. Thus, the turn 'Fine' in respo nse to
the turn, near the school, and she has been describing the turkey club sandwich and
'ilow are yo u ?' displa ys a partic ular
unde rstan ding of that prior turn. Because
interlocutors must make these negot cheeseburger she had just eaten.]
iations visible to each other to achieve
iI degre e of inters ubjectivity
, analysts can 'look over their shoulders' to I
il
gain Janelle: Only with that turkey thingee Ilya yo estoy /lena. ['I'm already
a wind ow onto the unde rstan dings
that interlocutors them selve s displa y
full']
or these proce sses (ct. Herit age
not treat code switches in their
and Atkin son 1984: 11) . If indiv idual s
do
Isabella: IITwo dollars and fifty cent. II
interlocutors' prior turns as social ly loaded,
metaphorical actions, the analyst has no empir Janelle: That's good. That's like a meal at II Burger King.
ical basis for arguing that such
switches encode significant metaphorical Isabella IIThat's better than going to
meanings, even though the analyst
maya priori assume that they do. Burger King, you know what I'm saying? And you got a Whopper,
A paral lel methodological orientation in anthr french fries, and a drink. And the french fries cost a dollar over
opology also contributes to
effort s to see past popular and forma l lingu there.
istic ideolo gies. The fieldw orker
consc iously works to bracket cultural assum Janelle: For real?
ption s and understand subjects'
worlds throu gh members' categories, i.e., Sf, sf c6mo no? ['Yes, really.']
Mirale el ombligo. Miralo. Se Ie ve, ya se 10 tap6. «looking at a passerby))
from an emic or native's pOint Isabella:
of view. The active suspe nsion of comm onsen Janelle:
se assum ption s helps one to
partially overcome the omnipresent social ['Look at her belly button. Look. You can see it, she already covered
science problem of subjectivity,
e.g., to see mean ings of code- SWitc hing as it .']
a code- switc hing subject might
see them .
Seguro porque se 10 en sefi6. [' She must h ave s h owed it . ']
(.5)
Isabella:
heteroglossia and identity in a bilingual segment of talk «la ughing))
(1.5)
To illustrate heteroglossia and its application to issues of identity, I present Isabella: But it's slamming, though, oh my God, mad ['a lot of'] turkey she
a segment of transcript below that documents a few moments of bilingual puts in there.
talk between two Dominican American high school students in an everyday Janelle: That's one thing 11-, I love the way como 1- ['how th-'] the American
interaction during a break at school. The two students, Janelle and Isabella, be doing sandwich, they be rocking ['are excellent'] them things,
'
define themselves, and are defined by others, in multiple ways. Janelle was yo, they put everything up in there, yo.
US-born and raised, and therefore a citizen, and had been to the Dominican Isabella: De pla:ne, de pla:ne. «an airplane passes overhead))
Republic only once, as a baby. She identified herself as Spanish or Dominican,
but outsiders frequently took her to be Black American until they heard her The notion of code-switching is less useful in analysing this interaction
speaking Spanish. Isabella, who came to the United States at age 7, identified than the broader and socially-infused concept of heteroglossia. Janelle and
herself as Dominican or Spanish, and was occaSionally seen by outsiders as Isabella code switch into Spanish several times in this segment, but none of
Black American or Cape Verdean American. the switches generates any obvious metaphorical meanings. One switch into
The socially based construct of heteroglossia has several advantages over Spanish M[rale eL omb/igo . . - functions to hide the meanings of their talk
_ .

the narrower and formally defined construct of code-switching as a means from a non-Span ish-speaker passerby about whom they are talking, but it is
to understanding social identity negotiations in this talk. Specifically, (a) not obvious what social functions, if any, are served by the other switches.

I I
 268
identity practices heteroglossia and boundaries 269

While s u c h sw i t che� a r c locally unmarked, the concept of heteroglossia elision of syllable final /s/ (e.g. in e(s)toy), for example, are characteristic of
a ffords a t t e n lion t o m ca n ing s in a larger sociopolitical field. Janelle and
Isa be l l a 's sw i t ches a rc loudly unmarked in terms of identity negotiations, but
Caribbean Spanish, particularly Dominican and lower class varieties (Lipski
1994). Word order used by Isabella - mad turkey she puts in there - suggests
frequen t s wi t c h i ng as a d i scourse mode is always socially marked in a wider
US soc i e t y in which being a monolingual English speaker is an ideological
the influence of Spanish discourse patterns on her English. She preposes the

II I
direct object of the verb in this segment in what has been called fronting, focal
defil u l t agaillst which d ifference or distinctiveness is constructed (Urciuoli
1 (96). Th e p e r spe cti v e of heteroglossia allows one to distinguish between local
object construction (Silva-Corvalan 1983: 135), or focus-movement (Prince
1981). These heteroglossic forms in both English and Spanish - whether
fUl l c t i o n s of parti c ular code switches and the functions in the larger socio­ at the phonological, lexical, grammatical or discourse level - index social

!III
pol i t ica l flcld of identity formation in ways that a more formal perspective of histories, circumstances and identities in ways that a binary perspective of
co d c -s w i t c h i n g does not (cf. Myers-Scotton 1993: 149; Zentella 1997: 101).
code-switching commonly neglects.
Mea n i ngs of code-switching are contested in this larger SOCiopolitical field. A further strength of heteroglossia as a perspective is that it directs the
Va rious nativist English-only groups, for example, have sponsored legislation

I1II1
analyst to historical social relations, rather than just details of surface form, to
t o cr i m i nalize the use of languages other than English in many contexts, interpret language meanings. Each of the locally unmarked linguistic features
inclucli ng school, government and workplace. They portray such language enumerated above, for example, draws its meaning from contrast with an
alternation as undermining American unity, citizenship and decency. Many implicit field of alternate forms, the relative valences of which are a function

II
academics of the last 25 years, in contrast, have treated such code-switching of historical power relations. Code-switching and other language contact
as a discourse mode which can be seen as a form of resistance to dominant features are meaningful only in contrast to monolingual speech. Caribbean

I
discourses of unquestioning assimilation (Gal 1988: 259), or as a means of Spanish phonology is meaningful in contrast to Mexican, South American
constructing a positive self in a political and economic context that disparages and peninsular varieties, particularly Castilian. African American English is
immigrant phenotypes, language, class status and ethnic origins (Zentella meaningful in contrast to varieties associated with White Americans, and the
1997: 13). The ways in which these meanings vary depending on one's subject youth vocabulary contrasts with the vocabulary of adult speakers.
position - speaker, nativist American, or socially oriented language analyst In each case, the meaningful opposition is between an unmarked form

I
- illustrate the subjective and social construction of heteroglossic meanings
that is characteristic of talk more generally.
associated with groups historically or currently in power (monolingual I
Americans; speakers of Castilian varieties of Spanish; White Americans; and
The focus in code-switching research on constellations of linguistic features adults) and marked forms that index lower pOSitions in social hierarchies.
that are officially authorized as codes or languages, e.g. 'English' or 'Spanish', Meaningful oppositions arise in this instance not on the basis of formal
can contribute to neglect of the diversity of socially indexical linguistic distance among forms, but on the basis of historical power differentials
resources within cocl e s. The English that Isabella and Janelle use in the with which particular forms are associated. The perspective of heteroglossia
exchange above, for example, includes prescriptivist standard American explicitly bridges the linguistic and the sociohistorical, enriching analysis
English forms, non-standard vernacular forms, lexical forms associated of human interaction.
with African American English, and grammatical forms that Occur only in The variety and juxtaposition of linguistic resources by Janelle and Isabella
African American English. Their talk also includes an explicitly intertextual in the above exchange reflect their negotiations of social boundaries and
reference to popular culture: Isabella's De pla:ne, de pla:ne as she observes a meanings. Their alternation of English and Spanish and contact forms diminish
plane passing overhead . This utterance represents an example of Bakhtinian a linguistic boundary that others have created. Their ongoing use of forms
double-voicing, in which words are spoken as if they are to be understood (e.g. code-SWitching, African American English, and Dominican phonology)
as being in quotation marks. Her utterance is an intertextual reference to that are disparaged by dominant groups suggests resistance to hegemoniC
the words spoken by actor Herve Villechaize at the opening of each episode belief systems regarding language. Maintenance of non-prestige forms can
of the television show Fantasy Is/and, which premiered on US television serve as a vehicle of resistance to disparaging discourses on language, race
in 1977. Villechaize spoke these word s with strong second-language and identity from dominant groups in society and reproduce local solidarity.
phonology as de plane, de plane, and the phrase entered the popular culture Their use of forms associated with urban African American youth, particularly
vocabulary of many Americans, who still use it to verbally mark the passing Janelle's use of African American English syntax, suggests both longer-term
of a irplanes overhead. contact with African Americans and identification with African American
Th eir Spanish similarly i ndexes particular linguistic histories. Their experiences. Janelle's use of the term 'American' to refer to a group that is
pronunciation of word-initial y as an affricate /d3 / (e.g. in ya yo), and their implicitly distinct from her suggests (although she is US-born and a US citizen)
 270 identity practices
heteroglossia and boundaries 271

that s h e i d e n t i fi es h e rse l f with reference to another nation-state or in terms code-switching a s a social construction
o f rac i a l /e t h n i c ca t eg o r i es i n which she doesn't count as 'American'.3 The
j u x t a po�i t i o n o f d i verse linguistic elements in single utterances, e.g. I love the
While the notion that identity categories such as race or ethnicity are socially
way COII/O 1- tile AllleriCiln be doing sandwich reflects social negotiations and a
constructed i s now an academic commonplace, bilingualism, as b o t h a
popular and analytical category, is not generally seen as a social constructi o n .
soc i a l rea l i t y i n wh i c h neither linguistic practices nor social identities fit into
There are fundamental parallels, however, in the social and political processes
s t a t i c, u n i t a r y ca t eg o r i es of language and identity.
I t i s t c m p t i n g to l a bel Isabella and Janelle's identities and ways of speaking
through which difference is constructed among social identity categories a n d
as ' h y b r i d ', ' m i xed' or 'syncretic', but I argue against the analytical use of
among the linguistic forms that count a s bilingual talk. Both, for example, are
popularly seen as having self-evident, empirical bases, and both form parts
s u c h t e r m s here because they so easily express and reproduce dominant
of the highly naturalized assumption of a language-race-nation unity. I n
i d c o l og i es of essentialism. The term 'hybridity' is only meaningful against
both cases, however, the conceptualizations, salience and social significance
a b a c k d ro p of essentialism that analysts generally claim to have rejected.
of the categories are a function of social and pol itical processes rather than
W h i l e many postcolonial theorists use the term to refer to novel cultural
inherent, or essential characteristics of members of the categories. The fact
forms with roots in seemingly disparate experiences, others struggle with the
terms. Critics argue that 'hybridity', like popular folk-terms such as 'diversity', that bilingual speech draws both popular and academic attention may tell us
'multicultural', 'heterogeneous' and 'pluralism', can all pay lip service to relatively little about the nature of code-switching, and relatively more about
certain types of social difference, while implicitly reinforcing the political and popular and academic language ideologies of Western nation-states.
economic boundaries that constitute those groups as different and unequal Conceptualizing bilingual speech as a social construction does n o t
(Hutnyk 2005; Young 1995; Chow 1998). m inimize i t s on-the-ground social implications. An example from Soc ial
In labelling a system as syncretic, an analyst highlights the discrete heritage identity categories can help make this clear: the fact that Black-White race
of individual components of a system and suggests some incommensurability in the United States is a social construction, for example, does not make race
of those parts. Since all social systems are a function of multiple influences and an illusion or socially insignificant (Omi and Winant 1994). Race has been,
histories, what counts as a relevant opposition within a system is a contested, and remains, a central organizing principle in the United States and a way
subj ective and shifting question. As with the term 'hybridity', 'syncretism' of representing, rationalizing and reproducing tremendous social inequa lity.
carries with it connotations of pure and coherent anterior systems. Beliefs or Approaching race as a social construction, however, allows one to see t hat
practices are most often termed syncretic when they violate Western analysts' race is not about essential biological difference (which is how race is popularly
implicit assumptions of purity and inherent discreteness. construed) but about social history. What is socially Significant about rac e is
In anthropology, for example, the term 'syncretic' has often been used a distinctively violent history of coercion and inequality, not details of hair

,
to describe the religious practices of the Afro-Caribbean, as if European texture, skin shade or other morphological features. The social constructionist
Catholicism and African religious beliefs and practices were each pure and perspective directs attention to the political and historical processes through
fundamentally different. When Christianity includes pagan practices such which race has been constituted and given such significance in the US.

�
as celebration of evergreen trees around the time of the Winter solstice, Similarly, approaching monolingualism and b ilingualism as socia l ly
however, it is not seen as syncretic. Linguistically, English is not commonly constructed does not change their social force at the level of lived experience,
I
seen as a hybrid, or creole language, despite the readily apparent effects of but it does show that this social force is not a function of formal, or inherent
Norman French on the language(s) of Britain from the period following the linguistic differences among what count as languages. If bilingual talk is an
Norman Conquest in 1066. Languages, or language change, that develop from especially meaningful mode of speaking, it is not the nature of the forms that
European colonization of other parts of the world, in contrast, are typically make it so but rather particular social and political histories.
seen as hybrid or creole, e.g. Jamaican creole or Haitian creole. Whether Studying bilingual talk can be a route to understanding social boundary Work
one counts two aspects of a system as discrete and not-entirely-compatible, not because of the formal nature of bilingual talk, but because all talk is SOCial
and therefore syncretic, depends on one's subject position and historical and political. In contexts such as Western societies where code-SWitching has
power relations rather than the nature of the forms or systems in and of been made to count as particularly socially meaningful, insights into identity
themselves. While analysts can use these terms as parts o f projects that are not negotiations can come from attention to the social and political processes that
essentializing (e.g. Hill and Hill 1986), these words are intertextual, carrying have made monolingual-versus-bilingual speech a meaningful opposit ion.
with them connotations of usage in essentializing projects. Analytical constructs that are based on form, such as code-switching, or that
imply anterior, pure essences, such as hybridity, divert attention from the social
 272 identity practices 273
heteroglossia and boundaries

a n d pol i [ i c a l n a t u re of l a n g uage,behaviour and meaning. More processual Bourdieu, P. ( 1 9 9 1 ) . Language a n d Symbolic Power (G. Raymond a n d M . Adamson, trans.).
a n d soc i a l l y- i n fused co n s t ruct s such as heteroglossia and indexicality better Cambridge, MA: Harvard U niversity Press.
d i rec t a t t e n t i o n to t h e i rr e ducibly sociohistorical and ideological bases of Ch ow, R. ( 1 998). Ethics After Idealism: TheOly-Culture-Ethnicity-Reading. Bloomington,

l a ngu age I l l c a n i ngs a n d i dentity construction. Heteroglossia and indexicality IN: University of Indiana Press.
Gal, S. ( 1 988). The political economy of code choice. [n M. Heller (ed.), Codeswitching:
a rc fu n d a m e n t a l l y a b o ut intertextuality, the ways that talk in the here-and­
A nthropological and Sociolinguistic Perspectives, p p . 245-64. Berlin, New York and
now d raws m ea n i ng s from past instances of talk. Such terms encourage us to
Amsterdam: Mouton de G ruyter.
i n [ e rpre[ t h e mea nings of talk in terms of the social worlds, past and present,
Goffman, E. ( 1 9 79 ) . Footing. Semiotica 25: 1 -29.
of w h i c h words are part-and-parcel, rather than in terms of formal systems, Gumperz, j .j . ( 1 982). Discourse Strategies. Cambridge and New York: Cambridge
s u c h as codes, that can veil actual speakers, uses and contexts. U niversity Press.
Hanks, W. (200 1 ) . Indexicality. I n A. Duranti (ed.), Key Terms in Language and Culture,
pp. 1 1 9-2 1 . Malden, MA and Oxford: Blackwell Publishers.
notes
Heller, M. ( 1 988). Introduction. I n M . Heller (ed.), Codeswitching: A nthropological and

1 . In t h i s chapter I use the terms 'bilingualism', 'bilingual speech' a n d 'code-switch ing' Sociolinguistic Perspectives, pp. 1-24. New York: Mouton de Gruyter.

i n terchangeably. My interest in bilingualism here is only in its instantiation in -- ( 1 992) . The politics of code-switching and language choice. In e. Eastman (ed.),
speech, not as a state or capability. Codeswitching, pp. 1 23-42. Cleveland and Avon: Multilingual Matters.

I
2. Thi s language ideology is also reflected in commonly expressed folk explanations -- ( 1 9 9 5 ) . Language choice, social institutions, and symbolic domination. Language

other language'; 'they/we switch so that person X can understand better'; and 'they
for code-switching: 'they/we switch because they/we don't know the word in the in Society 24( 3 ) : 3 7 3-406.
Heritage, L and ]. M . Atkinson ( 1 984) . Introduction. In j .M. Atkinson and ] . Heritage
switched to hide what they were saying from person Y'. (eds), Structures ofSocial A ction: Studies in Conversation Analysis, pp. 1 - 1 5 . Cambridge:
3 . In everyday Dominican American usage, the term 'American' (or americano) refers U niversity of Cambridge Press.
to 'White American' (cf. UrciuoIi 1 996 among New York Puerto Ricans) . United
States-born Dominican Americans such as Janelle identify themselves as 'American'
Hill, ] . H . , and K. Hill ( 1986). Speaking Mexicano: Dynamics ofSyncretic Language in Central
Mexico. Tucson: University of Arizona Press.
Il :
i n some contexts, e.g. i n referring to citizenship or the passport they have, but Hutnyk, j. (2005) . Hybridity. Ethnic and Racial Studies 28( 1 ) : 79-102.
they identify 'what they are' as Dominican/Spanish/Hispanic. These categories are Irvine, j .T. (200 1 ) . 'Style' a s distinctiveness: The culture and ideology of linguistic dif­
mutually exclusive from the category W hite/American in local terms. ferentiation. In P. Eckert and J . R . Rickford (eds), Style and Sociolinguistic Variation,

Ivanov, V. (200 1 ) . Heteroglossia. In A. Duranti (ed.), Key Terms in Language and Culture,
pp. 21-43. Cambridge and New York: Cambridge University Press.

references
Kroskrity, P.V. ( 1 993). Language, History, and Identity: Ethnolinguistic Studies ofthe Arizona
pp. 95-7. Oxford and Malden, MA: Blackwell Publishers.

Auer, P. ( 1 984) . Bilingual Conversation. Philadelphia and Amsterdam: John Benjamins.

-- ( 1 988) . A conversation analytic approach to codeswitching and transfer. I n M. Tewa. Tucson: UniverSity of Arizona Press.

Heller (ed.), Codeswitching: A nthropological and Sociolinguistic Perspectives, pp. 1 8 7-2 1 3 . Labov, W. ( 1 9 79 ) . Locating the frontier between social and psychological factors in

New York: Mouton d e G ruyter. linguistic variation. I n Wong (ed.), lndividllal Differences in Language A bility and

-- (1 995) . The pragmatics of code-switching: A sequential approach. In L. M il roy Language Behavior, pp. 327-40. New York: Academic Press.
and P. Muysken (ed.), One Speaket; Two Languages: CroSS-disciplinary Perspectives on Le Page, R.B., and A. Tabouret-Keller ( 1 985). Acts ofIdentity: Creole-based Approaches to
Code-Switching, pp. 1 1 5-35. Cambridge: Cambridge University Press. Language and Ethnicity. Cambridge and New York: Cambridge U niversity Press.
-- (ed.) ( 1 9 9 8 ) . Code-Switching in Conversation: Langllage, Interaction and Identity. Lippi-Green, R. ( 1 99 7 ) . English with an Accent: Language, Ideology, and Discrimination in
London and New York: Routledge. the United States. London and New York: Routledge.
Bailey, B. (200 1 ) . Dominican-American ethnic!racial identities and United States social Lipski, j . M . ( 1 994). Latin A merican Spanish. London and New York: Longman.
categories. International Migration Review 3 5 ( 3 ) : 6 7 7-708. Maeh lum, B. ( 1 996) . Codeswitching i n Hemnesberget - Myth or reality? Tournai of
-- (2002). Language, Race, and Negotiation ofIdentity: A Study ofDominicnn Americans. Pragmatics 25(6): 749-6 1 .
New York: LFB Scholarly Pub. Malinowski, B . ( 1 965). Coral Gardens and their Magic. Bloomington: Indiana University

Meeuwis, M . , and J . Blommaert ( 1 998). A monolectal view of code-switching: Layered

Bakhtin, M.M. ( 1 9 8 1 ) . The Dialogic Imagination: Four Essays (M. Holquist, trans. ) . Austin: Press.

Barth, F. ( 1 969). Introduction. In F. Barth (ed.), Ethnic Groups and BOllndaries: The Social
UniverSity of Texas Press.
code-switching among Zairians in Belgium. In P. Auer (ed . ) , Code-Switching in
Organization of Culture Difference, pp. 9-38. Boston: Little Brown and Co. Conversation: Language, Interaction and Identity, pp. 76-98. London and New York:
Blom, J . P. , and j.J. Gumperz ( 1 9 72). Code-switchin g i n Norway. I n j. Gumperz and
Routledge.
D. Hymes (ed . ) , Directions in Sociolinguistics, pp. 407-34. New York: Holt, Rinehart
Mittelberg, D., and M.e. Waters ( 1 992). The process of ethnogenesis among Haitian and
and Winston.
Israeli immigrants in the United States. Ethnic and Racial Studies 1 5 ( 3 ) : 4 1 2-35 .
E

274 identity practices

Moer m a n , M . ( 1 965 ) . Et h n ic identification in a complex civilization. American A nthro­

C. ( 1 99 3 ) . Social Motivations for Codeswitching: Evidence from Africa.

/lolos is t 6 7 : 1 2 1 5-30.
M ye rs-Scot t o n ,
O x ford: C l a re n d o n Press.

to til(' 1 990s ( 2 nd
O l l l i , M . , and I I . W i n a n t ( 1 994). Racial Formation in the United States: From the 1 960s
eel n ) . New York: Routledge.
I 'c i rce, C.S. ( 1 95 5 ) . I ,ogi c as semiotic: The theory of signs. I n ]. Buchler (ed.), Philosophical
Writills.� o(l'circc, pp. 98- 1 1 9 . New York: Dover Publications.
I 'oplack, S. ( 1 980) . Sometimes I ' l l start a sentence i n Spanish y termino en Espanol
t owa rd a typology of code-switching. Linguistics 1 8( 7-8): 58 1-6 1 8 .
( 1 988). Contrasting patterns of codeswitching i n two communities. I n M . Heller
Ii
--

( cd . ), Cudeswitching: Anthropological and Sociolinguistic Perspectives, vol. 48, pp. 2 1 5-44.

E. ( 1 9 8 1 ) . Topicalization, focus-movement and Yiddish-movement: A pragmatic

Ilerl i n , New York and Amsterdam: Mouton de G ruyter.
Pri n ce,
differentiation. Proceedings of the Seventh Annual Meeting of the Berkeley Linguistics
SOciety, 249-64.
Reddy, M. ( 1 9 7 9 ) . The conduit metaphor: A case of frame conflict in our language
about language. [n A. Ortony (ed.), Metaphor and Thought, pp. 284-384. Cambridge:
Cambridge University Press.
Rickford, l.R., and P. Eckert (200 1 ) . I n troduction. In P. Eckert and ].R. Rickford (eds),
Style and Sociolinguistic Variation, pp. 1 - 1 8 . New York and Cambridge: Cambridge
University Press.

constructions in Spanish. In F. Klein-Andreu (ed.), Discourse Perspectives on Syntax,

Silva-Corvalan, C. ( 1 983). On the interaction of word order and intonation: Some OV

pp. 1 1 7-40. New York: Academic Press, Inc.

Silverstein, M. ( 1 9 76 ) . S h i fters, li nguistic categories, a nd cultural descripti o n . I n
K . H . Basso and H . A . Selby (ed s ), Meaning in A nthropology, pp. 1 1-56. Albuquerque:
University of New Mexico Press.
-- (1 992). The indeterminancy of contextualization: When is enough enough? In P.
Auer and A . D i Luzio (eds), The Contexualization ofLanguage, pp. 55-76. Amsterdam :
John Ben j am i n s .
-- ( 1 996). Monoglot 'standard' in America. I n D . L . Brenneis a n d R. K.S. Macaulay
(eds), The Matrix of Language: Contempormy L inguistic Anthropology, pp. 284-306.
Boulder, CO: Westview Press.
-- (200 1 ) . The limits of awareness. [n A . Duranti (ed.), Lingllistic A nthropology: A
Reader, pp. 382-40 1 . Malden, MA and Oxford: Blackwell Publishers.
Urciuoli, B. ( 1 996). Exposing Prejudice: Puerto Rican Experiences of Language, Race, and
Class. Boulder, CO: Westview Press.
Weinreich, U . ( 1 9 5 3 ) . L angllage in Contact: Findings and Problems. New York: The
linguistics Circle of New York.
Woolard, K . A . (2004) . Codeswitching. In A. Duranti (ed . ) , A Companion to Linguistic
Anthropology, pp. 73-94. M a l den, MA: Blackwell.
Young, R. ( 1 9 9 5 ) . Colonial Desire: Hybridity in Theory, Cultllre and Race. London:
Routledge.
Zentella, A.C. ( 1 99 7 ) . Growing up Bilingual: Puerto Rican Children in New York. Malden,
MA: Blackwell.