Applied Linguistics Review 2015; 6(3): 281307

Ricardo Otheguy*, Ofelia Garca and Wallis Reid

Clarifying translanguaging and

deconstructing named languages:
A perspective from linguistics
DOI 10.1515/applirev-2015-0014

Abstract: The concept of translanguaging is clarified, establishing it as a particular

conception of the mental grammars and linguistic practices of bilinguals.
Translanguaging is different from code switching. Under translanguaging, the
mental grammars of bilinguals are structured but unitary collections of features,
and the practices of bilinguals are acts of feature selection, not of grammar switch.
A proper understanding of translanguaging requires a return to the well known
but often forgotten idea that named languages are social, not linguistic, objects.
Whereas the idiolect of a particular individual is a linguistic object defined in
terms of lexical and structural features, the named language of a nation or social
group is not; its boundaries and membership cannot be established on the basis of
lexical and structural features. The two named languages of the bilingual exist
only in the outsiders view. From the insiders perspective of the speaker, there is
only his or her full idiolect or repertoire, which belongs only to the speaker, not to
any named language. Translanguaging is the deployment of a speakers full
linguistic repertoire without regard for watchful adherence to the socially and
politically defined boundaries of named (and usually national and state) languages.
In schools, the translanguaging of bilinguals tends to be severely restricted. In
addition, schools confuse the assessment of general linguistic proficiency, which
is best manifested in bilinguals while translanguaging, with the testing of proficiency in a named language, which insists on inhibiting translanguaging. The
concept of translanguaging is of special relevance to schools interested in the
linguistic and intellectual growth of bilingual students as well as to minoritized
communities involved in language maintenance and revitalization efforts.
Keywords: translanguaging, code switching, named languages, idiolects, language education and assessment

*Corresponding author: Ricardo Otheguy, Graduate Center, City University of New York, New
York, USA, E-mail: [email protected]
Ofelia Garca, Graduate Center, City University of New York, New York, USA
Wallis Reid, Graduate School of Education, Rutgers University, New Brunswick, USA

282

Ricardo Otheguy et al.

1 Introduction
Since first coined by researchers in Wales, the term translanguaging has been
increasingly adopted by scholars to describe the linguistic practices of speakers
labeled as bilingual or multilingual, and to describe as well the many ways that
those practices are leveraged for a variety of purposes, especially in education
(Blackledge and Creese 2010; Canagarajah 2011a, 2011b; Creese and Blackledge
2010; Garca 2009; Garca and Wei 2014; Garca and Otheguy 2015; Hornberger
and Link 2012; Lewis et al. 2012a, 2012b; Wei 2011; Lin 2014; Sayer 2013; Williams
1994). The new term aimed to overturn the conceptualization of the two languages of bilinguals (which for us includes multilinguals) as clearly distinct
systems normally deployed separately, but occasionally deployed in close,
alternating succession under a practice known as code switching. But the
attempt to overturn has not been successful. As the term translanguaging gained
adherents in sociolinguistics and education, it began to drift toward covering
essentially the same conceptual terrain as code switching. The notion of code
switching, understood by most informed scholars in a dynamic and creative
fashion as the expressive transgression by bilingual speakers of their two
separate languages, endows these speakers with agency and often finds in the
very act of switching elements of linguistic mastery and virtuosity (Auer 1999,
2005; Gumperz 1976, 1982; Myers-Scotton 2005). But no matter how broadly and
positively conceived, the notion of code switching still constitutes a theoretical
endorsement of the idea that what the bilingual manipulates, however masterfully, are two separate linguistic systems. It was the attempt to set aside this
dual conception that in part prompted the coinage of translanguaging, and that
in part motivates us to offer a new clarifying explication of the term here.
More broadly, the new explication is needed because the adoption of
translanguaging has not produced, in our view, a sufficiently strong challenge
to prevailing understandings of language and linguistic behavior in speakers
generally, and especially in bilinguals. Part of the problem is that the challenge
that we envision is not embodied in the concept of translanguaging alone; it
requires, additionally, that we, as others have before us, directly problematize
the notion of a language. For, as we have come to realize, it is the uncritical
acceptance of this foundational term that has kept us from fully grasping the
implications of translanguaging for our understanding of bilinguals and of the
schools charged with their education and assessment.
A second reason for the clarification of translanguaging proposed in this paper
is the need to dispel the belief, taking root in some quarters, that translanguaging
works against the efforts of minoritized communities to protect and revitalize their

Clarifying translanguaging

283

languages and linguistic practices. On the contrary, translanguaging, as we understand it, helps to disrupt the socially constructed language hierarchies that are
responsible for the suppression of the languages of many minoritized peoples. As
we will show here, translanguaging does this by insisting on the well-known, but
almost always forgotten, postulate that a named language is a social construct, not
a mental or psychological one, and on the radical implications of this position for
ones theoretical conceptualization of bilingual individuals and societies.
We will suggest that a full understanding of what is meant by translanguaging and an accurate take on what is meant by a language allows us to
graduate from the goal of language maintenance, with its constant risk of
turning minoritized languages into museum pieces, to that of sustainable practices by bilingual speakers that thrive in spatial and functional interrelation with
the sustaining linguistic practices of other speakers (Fishman 1964; Garca 2011).
Translanguaging, then, as we shall see, provides a smoother conceptual path
than previous approaches to the goal of protecting minoritized communities,
their languages, and their learners and schools.
As a presentational device, and because we have learned how difficult it is
to get our point across, we make use of a simile based on a culinary tale. We first
use the culinary tale to expose the fable of the named national cuisine that so
many hold dear, and proceed thereafter to debunk the myth of the lexically and
structurally defined named national language that so many hold dearer still. We
then use these tall tales of kitchen and tongue as the basis for restating the
definition of a named language as a socially constructed rather than a linguistic
object (Heller 2007; Jrgensen et al. 2011; Makoni and Pennycook, 2010). The
definition of a named language as a social construct, as we shall see, logically
precedes our new explication of translanguaging. With the culinary fable and
the restated definition of a named language in place, we will be able to define
translanguaging as the deployment of a speakers full linguistic repertoire without
regard for watchful adherence to the socially and politically defined boundaries of
named (and usually national and state) languages.
This more fully articulated definition of translanguaging will provide us,
then, with the framework that is necessary for a radical critique of the many
language assessment and teaching practices that work to the detriment of
bilingual children worldwide. The article ends by proposing translanguaging
practices for both assessment and teaching that would nurture all bilingual
students. These translanguaging educational practices, as we see then, will
further the growth of students in both content education and linguistic proficiency, including proficiency in the many situations where the watchful adherence that weve just mentioned ought to be relaxed, and in the relatively rare,
but still unarguably important situations where it is to be engaged.

284

Ricardo Otheguy et al.

The new explication of the concepts of translanguaging and a language

offered in this paper contributes most directly to the debate over the type of
cognitive representation of linguistic features that it makes sense to theorize for
bilinguals, and over the type of instruction that follows from assuming one or
another type of representation. But the discussion here, as we shall see presently, encompasses not only language in bilinguals and in the educational
institutions that serve them, but also language in general, in all speakers,
including monolinguals. It is in that sense that the proposal advanced here
about language and translanguaging is to be seen, for the limited range of
problems in its purview, as not only a theory of bilingual education and
bilingualism, but more generally also as a theory of language.

2 A culinary fable
A Japanese professor who had hosted Ofelia and Ricardo in Osaka came to visit
them in New York last year. While in Osaka she had served them wonderful meals
each evening. When she came to New York, Ofelia and Ricardo, who were born in
Cuba, served her a meal of ropa vieja (shredded beef), black beans, and rice,
ending with flan (custard) as dessert. The second evening they served a bowl of
New England clam chowder, then grits, fried chicken, and green peas, ending
with a dessert of pumpkin pie. The Japanese guest watched Ofelia in the kitchen
intently. Upon her return to Japan, she would have been able to describe the
meals to her friends, how they looked and tasted, and how they were prepared.
But she would not have been able to tell them what named national cuisines the
two meals represented. This is because named national cuisines such as Cuban
and American (as many New Yorkers, and perhaps many readers, may have
described the two meals) are not defined by observable similarities among dishes;
for instance, ropa vieja and black beans are not Cuban because they share
ingredients or taste alike. Instead, named national cuisines are culturally and
politically defined categories; each cuisine consists of a large collection of ways to
prepare dishes that are grouped together on the basis of the regional or national
affiliation of the people who initially developed them.
This fact can be seen more clearly in what Ofelia and Ricardo did on the
third evening. They told their Japanese guest that on the first evening she had
been served a Cuban meal and on the second evening an American meal. They
then served her a meal of blackened catfish, gumbo and french fries, followed
by a devils food cake. When they were finished eating, there was a pop quiz.
They asked her what national cuisine this third meal represented. Was it Cuban
like the first one, or American like the second one? Being told that the first meal

Clarifying translanguaging

285

was Cuban and the second was American did not help her to know that the third
one was also American. The reason is that the first American meal New
England clam chowder, grits, fried chicken, peas and pumpkin pie would
not allow her to extrapolate to general features of American cuisine that would
help her decide that the third meal blackened catfish, gumbo, french fries,
devils food cake was also an instance of American cuisine. What makes those
two meals American is that both consist of dishes developed and served in the
United States, not that they have features in common. Similarly, the only thing
that makes both bouillabaisse and frogs legs sauted in butter and garlic count
as French cuisine is that both dishes are prepared and eaten mostly in France.
As a cook, Ofelia almost never plans meals with an eye to making them
purely Cuban or purely American. For instance, on the fourth evening Ofelia and
Ricardo served a meal of turkey with gravy, congr with yucca, and mashed
potato. To foody New Yorkers in the know, it is clear that this meal does not
count as either purely Cuban or purely American; yet it did not strike the
Japanese guest as strange or poorly planned. She did not recognize this fourth
meal as being fundamentally different from the first three.
The first lesson here, then, is that the failure of the Japanese visitor to
identify the provenience of the meals was not because there was anything
wrong with her sense of taste; she simply lacked the cultural knowledge of
what counted as Cuban dishes and what counted as American dishes. We note a
second lesson. The ability to construct a meal of dishes that all represent a
single named national cuisine is rarely a consideration in deciding whether
Ofelia or anybody else is a good cook. This is because a good cook is someone
who can prepare attractive, exciting, delicious, intriguing meals in which each
dish complements and balances the others. After all, one could be very good at
adhering to cuisine boundaries and yet still be a lousy cook.
To recap: we have seen that particular dishes and named national cuisines
are quite different things.
National cuisines are large collections of recipes and ingredients for the preparation of particular dishes. The basis for grouping the recipes together as
Cuban or American (or French, Mexican, Russian or anything else) is the cultural
or national affiliation of the original cooks, not similarities among the recipes.
Whether or not a meal falls into any one single category of national cuisine
has little to do with how good the cook is or how enjoyable the meal.
So what does this culinary fable have to do with understanding the serious
limitations of the notion of a language that constitutes the necessary preamble
to our new explication of translanguaging? The next section makes the answer
to this question evident.

286

Ricardo Otheguy et al.

3 Deconstructing named languages

The point is simple: a named national language is the same kind of thing as a
named national cuisine. Like a named national cuisine, a named language is
defined by the social, political or ethnic affiliation of its speakers. Although the
idea of the social construction of named languages is old in the language fields,
it is often not understood. The point that needs repeating is that a named
language cannot be defined linguistically, cannot be defined, that is, in grammatical (lexical or structural) terms. And because a named language cannot be
defined linguistically, it is not, strictly speaking, a linguistic object; it is not
something that a person speaks.
The culinary analogy serves to press against named languages the charge of
lexical and structural essentialism. A particular named language, as we see it,
cannot be defined in terms of a set of essential lexical or structural features, any
more than a cuisine can be defined by a set of essential ingredients or recipes.
This charge of essentialism merits some elaboration, as it challenges the familiar
demarcation and enumeration of named languages, and the connection that
these have to the comfortable understanding of bilingualism that prevails in
many scholarly and educational circles, and that we seek to undermine.
We start by reaffirming the distinction between (a) languages as the names
of enumerable things that are socially or socio-politically constructed, maintained, and regulated (names such as Arabic, Basque, Bulgarian, English,
Mandarin, Navajo, Spanish, Swahili) and (b) languages as entities without
names, as sets of lexical and structural features that make up an individuals
repertoire and are deployed to enable communication. Due to its familiarity, the
distinction between senses (a) and (b) seems simple at first. Sense (a) of
language, the social sense, encompasses the countable and easily enumerated
entities that are tightly associated with established peoples or nations, and often
additionally with established or aspiring states (Arabic is the language of Egypt,
Syria, Tunisia; Spanish is the language of Cuba, Mexico, Uruguay, Euskara is
the language of the Basque nation, Hawaiian is the language of indigenous
Hawaiians). Sense (b) of language, the mental or psychological sense, encompasses the billions of individual linguistic competences of speakers the world
over, irrespective of whether we call them monolingual or multilingual.
Simple though it is, the distinction is seldom maintained by linguists or
educators with the vigilance and persistence that the seriousness of the matter
requires. Assenting to the distinction when it is raised, and doing so often with
impatience (and even annoyance!), linguists and educators tend nevertheless to
immediately set the distinction aside once the real work begins of analyzing

Clarifying translanguaging

287

linguistic structures, researching pedagogical approaches, or evaluating assessment policies. But we ignore the two senses of language at great intellectual
peril to ourselves, and at enormous practical peril to the populations whose
linguistic and educational practices we research and whose interests we aim to
protect.
As a way to solidify the idea that named languages are social constructs and
not lexical or structural ones, we now remind readers of one of the best known
consequences of this fact, namely the inability of grammarians, that is, of
students of lexicon and structure (which in our usage includes phonology and
morphosyntax), to be of much use in helping to adjudicate one-language-ortwo disputes. For example, grammarians interested in the Iberian world have
always known that they have no lexical or structural basis for deciding whether
the speech of people living in, say, Barcelona, and the speech of people living
in, say, Valencia, count as instances of the same language or different languages. Nor do linguists have a basis for answering the question about people
living in Madrid and Mlaga, Madrid and Havana, Madrid and Lisbon, or Madrid
and Ro de Janeiro. To be sure, laymen regularly give clear answers to these
questions, in some cases with unanimity and in others with disagreement. But
the grammarian can neither endorse the laymens unanimity nor mediate their
disagreement.1 The reason is that nothing purely lexical or structural went into
making Catalan Catalan, Spanish Spanish, or Portuguese Portuguese. And this
in turn follows from the simple fact that the student of the grammatical properties that underlie the speech of individuals, that is, the student of language in
sense (b) above, has no theoretical basis for adjudicating disputes about separability and namability, since these are social and political matters that pertain
only to language in sense (a).2
The grammarian cannot adjudicate these disputes (even though the layman
is often happy to) for the same reason that the Japanese guest couldnt decide
which of the meals counted as Cuban and which as American, even though most
laymen in New York can easily tell which is which. The question for the
1 The point has a long history, having been already clearly enunciated in the work of Ferdinand
de Saussure, who insisted that neither languages nor dialects have natural boundaries
(Saussure 1916 [1986], Chapter III, Sec. 34). So once again here, we are not advancing a new
idea, but rather insisting that our explication of translanguaging requires us to pay close
attention to an old one.
2 The point made here about the Iberian realm applies to situations all over the world,
including the United States, where a generation ago there were heated debates as to whether
what some took to call Ebonics was the the same language as English. Here too, no more
than in the Iberian examples, the student of lexicon and structure has no technical tools to
settle the debate.

288

Ricardo Otheguy et al.

Japanese guest was not a culinary one, was not about the meals themselves:
about what ingredients they consisted of, how they were prepared, how they
tasted. Rather, it was a question about the classification of meals according to
their place of origin. This was an item of cultural knowledge that the Japanese
guest lacked. Similarly, linguists cannot tell us whether or not to count the
speech of Barcelona and Valencia (or Madrid and Havana, or Madrid and
Lisbon) as versions of the same language because these are not really linguistic
questions. Rather, they are questions about the political and cultural identities
of speakers, an item of cultural knowledge that is widely shared: laymen usually
recognize Madrid and Havana as speaking the same language because of historical and cultural factors connecting Spain and Cuba; they usually recognize
Madrid and Lisbon as speaking different languages because historical and
cultural ties have become loosened for several centuries now; and they argue
about Barcelona and Valencia because the historical and cultural factors are too
complicated to allow for easy settlement. A linguist can join the laymen in
stating these bits of cultural knowledge, and may, as a citizen, take part in
these debates, but cannot, qua linguist, that is, qua student of lexicon and
structure, weigh in with a technical position on the matter.3

4 Idiolects
Given that grammarians cannot legitimately address matters relating to language
in the social sense of the term, it is useful for our explication to point out what it
is that grammarians can do. Grammarians can analyze speech and text and
describe linguistic features: such things as phonemes, morphemes, words,
nouns, verbs, grammatical constructions or rules, tense systems, pronoun systems, case distinctions, gender distinctions, syntactic structures, or what have
you. (And depending on their theoretical orientation, they can study whether
these things are universal, whether theyre innate, and so forth; moreover, and
depending on their interests, they can describe them synchronically or diachronically.) And they can study, at least under some theories, the ways people
deploy these linguistic resources in speech and writing, that is, they can study
the language users strategies of communication.

3 For an example in which some of us have played the role of citizen advisors on debates
relating to language names and boundaries, see Otheguy (2009), Otheguy and Stern (2011,
2013), and Urciuoli (2013).

Clarifying translanguaging

289

When properly understood, all these things belong to what linguists call
idiolects. An idiolect is for us a persons own unique, personal language, the
persons mental grammar that emerges in interaction with other speakers and
enables the persons use of language. An idiolect is language in sense (b) above,
language viewed from the internal perspective of the individual, language seen
separately from the external perspective of the society that categorizes and
classifies named national languages. Idiolects are structured lists of lexical
and grammatical features, that is, they are lists subdivided in components
(e.g. lexicon, phonology, morphosyntax) and subcomponents (words belonging
to one noun class or another, systems of tenses, systems of case endings or
pronouns, etc., and, in some theories, movement, feature checking, etc.). Given
the importance to the field of the writings of Noam Chomsky, the idiolect may be
regarded by some as parallel to the I-language (Chomsky 1986). But the parallel
is quite limited. For us the idiolect is a mental grammar that is acquired
primarily through, and deployed mostly in, social and personal interaction;
the idiolect is thus rather different from Chomskys I-language, which unfolds
primarily as an instinct and is mostly deployed to generate introspective intuitions of grammaticality.
An idiolect, then, is the system that underlies what a person actually speaks,
and it consists of ordered and categorized lexical and grammatical features.
Those are the things that linguists actually analyze and study, not a named
language even when, confusedly, they use named languages to report on their
research. For example, one of the authors of this paper wrote his dissertation on
two past tenses, the pass simple and the imparfait; Wallis later wrote a whole
book on the noun plural -s and the third person -s on verbs (e.g., the -s on walks
in he walks). Another one of the authors wrote a dissertation on the articles el, la
and lo; Ricardo later wrote a book on the use of the personal pronouns yo, t, l/
ella, etc. Now, linguists are quite aware that what we have just said is not how
works such as Walliss and Ricardos are normally described; one usually says
(as they themselves said) that their dissertations were respectively on the French
tenses and the Spanish articles, and that the books were respectively on English
nouns and verbs and on Spanish pronouns. But that is precisely our point, that
what these authors were doing, as linguists, and despite the titles of their own
works, did not really involve, properly speaking, the notions of French, English,
and Spanish. These categories are not linguistic but socio-cultural, and as such
are extraneous to the enterprise of analyzing the idiolectal features that occupied their analytical efforts.4
4 For some of Ricardos and Walliss works on these topics, see Otheguy (2015), Reid (1991,
2011), Otheguy and Zentella (2012), Otheguy et al. (2007), and Shin and Otheguy (2013).

290

Ricardo Otheguy et al.

As we understand them, idiolects are personal and unique; no two are

identical, even the idiolects of siblings. Speakers are mostly quite unaware of
this fact, for two reasons. First, the differences between idiolects rarely impede
communication and so escape notice; secondly, when they are noticed they are
dismissed as of no consequence because of the belief that we speak the same
language. Nevertheless, the ways idiolects differ are important linguistically
because they involve underlying structural differences in peoples mental grammars. Here are some examples of what we mean.
One speaker may have more phonemes than another. Speakers who pronounce the words which and witch differently do so because they have a
voiceless phoneme // in addition to a voiced phoneme /w/, whereas
speakers who pronounce which and witch exactly the same lack the voiceless
one.5
A speaker who says I couldve gone has different verb morphology from that
of a speaker who says I couldve went.
A speaker who says I might could go has a more finely calibrated modal
system than a speaker who only says I might go and I could go.
A speaker who says both he be talking to me and he is talking to me has a
more semantically complex aspectual system than does a speaker who only
says he is talking.
A speaker who distinguishes imply from infer, fewer from less, or blimp from
dirigible, has a lexicon differently structured semantically than that of a
speaker for whom the pairs are synonyms.
All of these differences in the speech of people who are usually presumed to be
speaking the same named language, English, are due to lexical or structural
differences in their idiolects.
While no two idiolects are exactly the same lexically or structurally, there
are, to be sure, large areas of overlap among the idiolects of people who
communicate with each other. That is, we all share thousands of linguistic
features with people with whom we interact a lot, such as family and friends,
and with people who live in the same place as we do, or who live in countries
that share a history with ours, or who more generally share some sort of
linguistically mediated cultural or historical identity with us. Correspondingly,
we do not share many lexical or structural features with people with whom we
5 Note this is not a purely local difference restricted to a single word, comparable to the fact
that some speakers use the vowel of seat for the first vowel of the word economics while others
use the vowel of set. Speakers who have both // and /w/ also distinguish where from wear,
whale from wail, why from Y, wheel from weal, whether from weather, in addition to which from
witch.

Clarifying translanguaging

291

interact less, or whose history is separate from ours. But this overlap gives us no
warrant to conclude that named languages constitute lexically or structurally
based categories. The reason is that the overlap is never coterminous with the
boundaries that these sociocultural categories aim to demarcate.
For instance, most of the words Ofelia and Ricardo know and use are also
known and used by their son and daughters: friend, mother, t, mueble, tree,
come, sala, subway, viejo, casa, take, hablar, happy, nio, high school, ta, amigo,
feliz, building, today, hoy, hola, hermano, uncle, primo, tenedor, tray. In their
conversations at home, these words are common, and they often appear in the
same utterance. When they do, some people might say that this is a case of
mixing English and Spanish words, or of code switching. True enough: that is
how it looks from the outsiders perspective. But seen from the point of view of
the speaker, that is, from the insiders perspective, all these words together
belong to Ofelia and Ricardo and their children, that is, to their idiolects, not
to the English or Spanish languages as nationally or culturally defined. From the
insiders view, the question of which words belong to English and which ones
belong to Spanish (and which ones to both) cannot be asked coherently. It is an
apples-and-oranges question. All words that Ofelia uses are part of her mental
grammar; they all belong, by definition, to the same idiolect. And an idiolect is a
linguistic object whose constitutive elements are lexical and structural units,
whereas a language is a cultural object defined by place, memory, identity,
history, and, of course, a socially given (though sometimes contested) name. A
question formulated about the former category cannot be answered in reference
to the latter. Idiolects are what exist before one introduces distinctions between
national languages that forcefully shoehorn peoples linguistically specified
idiolects into culturally specified language categories.
Our insistence here that Ofelia and Ricardo are not mixing their languages
when they use the words listed above in the same utterance is really the same
point we made earlier: that there is no basis for including the speech of Cubans
and Mexicans as instances of Spanish but excluding the speech of Brazilians.
Languages are not true linguistic entities because their boundaries are established on non-linguistic grounds. Rather, they are groupings of idiolects of
people with shared social, political or ethnic identities that, once so grouped,
are described using linguistic terms that tend to give the mistaken impression
that the grouping was based on linguistic grounds in the first place.6

6 Note that we are not denying that there is an important linguistic difference between, say, the
word friend and the word amigo: namely their history, their etymology. The word friend is a
continuation of frond, while amigo is a continuation of amicum. The word frond is said to
belong to the culturally defined category of Old English (culturally and anachronistically; the

292

Ricardo Otheguy et al.

5 Idiolects, bilingualism, and translanguaging

When Ricardo and Ofelia speak with their children, they use all the words they
know because their children know and use them as well. However when they go
to Cuba, Mexico, Puerto Rico or Spain to visit relatives, they monitor their
speech to be sure they only, or mostly, use the words of their idiolect that
their family tend to know and use. And when Wallis comes for dinner or to
work on a paper, they avoid all the words they use in these places because he
doesnt know any of them.
Notice we didnt describe Ofelia and Ricardo as speaking exclusively in
Spanish in Cuba, Mexico, Puerto Rico or Spain, or as speaking exclusively in
English to Wallis, or as mixing Spanish and English at home. We didnt say this
because we are describing their language practices from their own internal perspective in terms of their idiolects. To be sure, their linguistic skills include making
decisions as to which words can be usefully deployed in which settings: use this
word in Cuba, but not in Spain; use this word in Spain, but not in Cuba; use this
one when talking to Wallis, but not that one. (For one of thousands of examples:
use jugo in Cuba but not in Spain; use zumo in Spain but not in Cuba; do not use
juice in either Cuba or Spain; do not use either jugo or zumo with Wallis).
Now, and this is an important but often overlooked point, the relevance of
social and locational constraints on idiolectal feature deployment is also a
characteristic of the idiolects of the people we call monolingual. Use this word
at home, but not at work, use that word when talking to children, but this one
when talking to adults, and so on and so forth. (For one of hundreds, perhaps
thousands, of examples: say bugging me at home but bothering me at work, say a
booboo to a child but a cut to an adult). So in this respect the idiolects of
monolinguals and bilinguals are not qualitatively different, only quantitatively
different. The difference is that the idiolects of bilinguals contain more linguistic
features and a more complex socio-cultural marking of which features to use when
and where. Said another way, monolinguals and bilinguals are not that different
linguistically they both have a mental grammar that guides sociolinguistic

users of the word didnt call their language Old English). The word amicum is said to belong to
the culturally defined category of Latin. But if one groups words according to etymological
origins traceable to the culturally defined box where they originated, one still doesnt end up
with languages. For example, the words of the single named language Spanish are derived from
many different culturally named languages (Latin, Euskara, Arabic, French, English, etc.), just
as the words of the single named language English have been taken from many different
culturally (and again, often anachronistically) named languages (Anglo-Saxon, Old Norse,
Norman French, Italian, Spanish, Dutch, Latin and Greek).

Clarifying translanguaging

293

interaction and is susceptible from the start to external input, and whose items
are deployed selectively depending on interlocutor and context. Thus the difficulty in establishing the difference between a monolingual and a bilingual is not
due simply to the fuzziness of boundaries between named languages. There is
also the fact that in both types of speakers geographical and sociocultural
considerations constrain the deployment of idiolectal features. The difference
between a monolingual and a bilingual person cannot be established on purely
linguistic grounds, but is justified only in sociolinguistic terms.
To avoid misunderstanding, a brief parenthetical clarification is in order
regarding terminology. Because we recognize the importance of the scholarly,
political, and sociolinguistic distinction between monolinguals and bilinguals,
we are not simply abandoning the distinction or scuttling the concepts of
language and bilingualism, as Pennycook (2010), and Makoni and Pennycook
(2007) have urged. For the same reason, we continue to talk in some settings
about languages and even about a particular language. We recognize that these
notions have had, and continue to have, real and material consequences in the
lives of many people, especially in the lives of minoritized people and those who
engage with us in advocacy on their behalf and their ways of speaking. But we
stress that, in accepting terms like language, a language, monolingual, and
bilingual, we are using categories that have nothing to do with individuals
when seen from their own internal linguistic perspective, categories that have
nothing to do, that is, with the billions of the worlds idiolects, which exist in a
separate, linguistically unnamed and socially undifferentiated mental realm.
When it comes to the concepts of named languages, monolingualism (or monolingual), and bilingualism (or bilingual), we are not denying their existence;
rather, we are restricting them to their proper domain of discourse. These terms
are appropriate and legitimate in discussions of social identity and sociolinguistic behavior because they are defined socially. But these concepts are not
appropriate for discourse dealing with mental grammar. That is, they are not
appropriate for discourse concerning lexical and structural matters (concerning
linguistic or psychological matters narrowly defined) because they cannot be
discovered and defined using either purely lexico-structural or psychological
criteria.
When the three authors of this paper say that we speak English, we are
acknowledging the importance of an outsiders perspective and reflecting a
social norm having to do with how others perceive us. We are recognizing that
each of our different idiolects partially overlaps lexically and structurally with
those of other speakers who all think of themselves as English speakers, and
who count us as one of them. The same is true when some of us say that we
speak Spanish. And the same is true for anyone who says that he or she speaks

294

Ricardo Otheguy et al.

any language. For example, saying that Ofelia and Ricardo are English
Spanish bilinguals from the U.S. simply means that her idiolect and his share
many features with each other and with the idiolects of three groups of people
with different though partially overlapping cultural identities. One group thinks
of themselves as English-speaking Americans, Australians, British, etc.; the
other thinks of themselves as Spanish-speaking Latin Americans or Spaniards;
and the third thinks of themselves as English Spanish bilingual speaking U.S.
Latinos.
It follows from our discussion that from the standpoint of lexicon and
structure, the only thing anyone actually speaks is his or her own idiolect,
something that no one else speaks. In other words, given that a named language
is a collection of the only partially overlapping idiolects of people who share a
common cultural identity (and who manage to communicate with greater or
lesser success); and given that the idiolects that comprise a named language
are all ultimately different, it follows that no one really speaks a named language. For, as we have mentioned, the set of shared features of any two speakers
that the society dubs as belonging to the same named language is unlikely to
ever be the same set; each speaker-dyad, even among the closest of relatives or
friends, most likely shares a slightly different set.7
Moreover, and this is the point that is at the heart of the matter, the common
features that are seen to be shared by idiolects emerge only after the idiolects
have been classified on a cultural basis as belonging to the same named
language. Put another way, linguists couldnt discover (and they certainly
have not discovered) the features that constitute the Spanish language or the
English language, or any other language, by starting from scratch. They must
first be told whose idiolects count as Spanish, that is, they must first be told
that the idiolects of most people living in Cuba and Spain are to count but that
those of most people living in Brazil and Portugal are not. Only then can
linguists find and describe the lexical and structural features shared by those
idiolects of Spanish.
The idiolect, which as we have seen is similar if not the same in kind in
bilinguals and monolinguals, is the cornerstone sustaining the concept of translanguaging. The similarity between monolinguals and bilinguals, it will be
recalled, resides in that the idiolect is, in both types of speakers, comprised of
large numbers of organized lexical and structural features. The features have no

7 For example, speaker A and speaker B could share a common phonemic system and a
common tense system but operate on different pronominal systems, while speakers B and C
could share a common tense system and a common pronominal system but operate with a
different set of phonemes.

Clarifying translanguaging

295

inherent membership in any named language. And they are deployed selectively
depending on context and interlocutor. The difference between monolinguals and
bilinguals is that monolinguals are allowed to deploy all or most of their lexical
and structural repertoire mostly freely, whereas bilinguals can only do so in the
safety of environments that are sheltered from the prescriptive power of named
languages. This is simply another way of saying that monolinguals are almost
always and everywhere allowed to translanguage, whereas bilinguals are only
allowed to translanguage in a limited number of protected settings. We want to
clarify and elaborate this point because it will be crucial, when we turn our
attention to education, to our critique of current tests of linguistic proficiency.
The clarification and elaboration involves a brief historical excursus that will
be useful for both scholars and educators. Let us recall how what we call English
got its words. We have already mentioned what philologists anachronistically call
Old English, the now named language that covers the idiolects that the AngloSaxons brought to the British Isles around 450 AD. Then, as is well known, in
1066 William the Conqueror from Normandy defeated King Harold at Hastings. For
two hundred years England was ruled by kings who spoke what society at the
time referred to as French. As the language of the kings spread to the Courts and
the Church, the upper classes began to incorporate into their idiolects more and
more of the words belonging to culturally defined French. Consider the first four
lines of The Canterbury Tales, written in 1387.
Whan that Aprill with his shoures soote

When April with his showers sweet [with fruit]

The droghte of March hath perced to the

roote,

The drought of March has pierced unto the

root

And bathed every veyne in swich licour

And bathed each vein with liquor that has

power

Of which vertu engendred is the flour,

To generate [therein] and sire the flower;

The words in bold below all came from the idiolects of the kings that had come
from Normandy and that society referred to as French:
Whan that Aprill with his shoures soote
The droghte of March hath perced to the roote,
And bathed every veyne in swich licour
Of which vertu engendred is the flour,

Now what can we say about these four lines? Are they an example of mixing
English and French, of code switching? Or are they an example of what

296

Ricardo Otheguy et al.

philologists today call Middle English? As we have insisted, this question has no
linguistic answer (no answer based on lexicon and structure), for it assumes
societal judgments regarding the features that a named language should or
should not contain. At the level of the individual English poet, we can say
that Chaucers idiolect contained both words brought to England by the
Anglo-Saxons and words brought by the Normans; and he used them all with
no apparent consideration of their history, just as Ricardo and Ofelia freely use
all the words they know at home. Whether Chaucer is code switching or writing
in what is called by some philologists Middle English is a question about how he
and his poem are perceived by others; it is not really a question about either
Chaucer himself his idiolect or his poetry. This distinction between what
societies view as language and what individuals do with their lexical and
structural features leads us to the further elaboration of the notion of translanguaging to which we now turn.

6 Translanguaging
In his most famous speech, Martin Luther King proclaimed: I have a dream that
my four little children will one day live in a nation where they will not be judged
by the color of their skin but by the content of their character. The alliteration of
color, content and character emphasizes the parallelism of the two by phrases,
making it sound as if skin color and content of their character stand on a par.
King was using skin color as a metaphor for his childrens race, a socially
defined category. In contrast, the childrens character was something personal
and intrinsic to them. So King was saying that he wanted his children to be
judged for their unique internal identity who they really were not their
external, social identity. King knew well the point were making. Just as a named
cuisine is not a food based category and a named language is not a lexically or
structurally based category, and just as they are both socially based constructs,
so is race not a biologically based category but a social one that is external to
any person of any race.
This distinction between treating a person as an individual and treating a
person as a member of a socially defined category is central to what we are
trying to say about language. It is the distinction between describing a persons
speech in terms that uniquely pertain to that person, and describing it in terms
of external societal criteria. For the latter, society provides a plethora of terms,
some sensible ones and others of dubious validity: English, Spanish,
Portuguese, Spanglish, Portuol, code switching, language mixing, etc.
But what terms do we have for the former?

Clarifying translanguaging

297

This is where the notion of translanguaging comes in, where the analyst
transcends the named language, language in sense (a) above, and returns the
focus to the individuals language, language in sense (b) above, that is, to the
idiolect. Translanguaging, in the definition offered above that we now know is
applicable to bilinguals and monolinguals, refers to the act of deploying all of
the speakers lexical and structural resources freely. To repeat, translanguaging refers to using ones idiolect, that is, ones linguistic repertoire, without
regard for socially and politically defined language labels or boundaries. It is
true, as we have said, that there are no speakers, even monolingual ones, who
translanguage everywhere and at all times, that is, who regularly use their full
idiolects completely freely; all speakers, even monolinguals, monitor their
speech to some extent in order to adapt to the interlocutor and social situation
at hand. Our point has been that since bilinguals have idiolects with larger sets
of lexical and structural features than monolinguals, and since they are often
of necessity in situations where they must communicate with monolinguals,
more of the language features of their idiolects are often of necessity suppressed, that is, their monitoring is more intense than is usually found in
monolinguals. For most monolinguals, to deploy all, or nearly all, of their
linguistic resources, that is, to translanguage, is closer to normal everyday
behavior, because monolinguals are usually granted license to operate at full
or nearly full idiolect. But for bilinguals, the deployment of full linguistic
resources can run up against strong norms articulating the sharpness of
linguistic boundaries. Only in situations that overcome these norms, that is,
only in the translanguaging mode where their entire repertoire is appreciated
as a tool for rich and unfettered expression, are bilinguals granted license to
operate at full or nearly full idiolect.
We stress, then, that translanguaging and using a named language constitute two different relations. Translanguaging involves the relation of full deployment between a persons idiolect and his or her speech or writing production.
Speaking a named language is a relation of partial overlap between a persons
idiolect and the idiolects of others. This overlap enables, under the right social
and historical conditions, the establishment of externally named boundaries. We
have said from the beginning that the difference boils down to one of perspective. Translanguaging adopts the perspective of the individual, the view from
inside the speaker; it offers a description based strictly on the internal categories
of lexical and grammatical structure. The named language adopts the view from
outside the speaker, a perspective from which the speaker has to fit as a member
of a set group; it offers a description based on external categories that emanate
from, and in turn reaffirm, sociocultural or national (and often also political)
structures.

298

Ricardo Otheguy et al.

To repeat, there is nothing inherently wrong with the categories of named

languages; on many occasions and with many interlocutors, it may make sense
to talk in those terms. Such talk is perfectly acceptable so long as we remain
aware that the named language categories have been ultimately constructed
throughout history for social purposes that bear a well-known and well-documented connection to the imposition of political power. Similarly there may very
well be times when it may be legitimate to use terms like code switching and
language mixing, which are ultimately based on named language categories.
The use of these terms may be acceptable so long as we remain aware that they
express the outsiders perspective and cannot therefore be assumed to describe a
duality in the individuals personal linguistic competence.8
Translanguaging, then, is not for all discussions of all topics at all times. It
is to be used when the topic at hand and the participating interlocutors feel the
need for, and become invested in, theoretical coherence. It becomes most useful
when we want to loosen the grip of social categories and describe the speakers
idiolects and the actual linguistic behaviors, their language practices. There is
often less call to do this when studying monolinguals (though it can be very
helpful, especially when speakers idiolects have features that have been left out
of the named standard language). But when carrying out research on bilinguals,
and especially bilingual students, the big referential difference between translanguaging and named languages is of crucial importance, because failing to
uphold it leads to considerable confusion and redounds to the detriment of the
speakers whom we study.

8 The dual linguistic competence assumed in code switching is part of a long and productive
research tradition (for current work see references above). But we should bear in mind that this
dual competence is not usually arrived at by researchers as a result of linguistic analyses that
start out neutrally, that is, without any assumption of duality. Rather, the duality is assumed a
priori, and it is established in a manner that mirrors directly the bifurcation of the socially
named language categories. That is, from the inception of research in code switching, examples
of switches held to be grammatical like estamos typing the paper we are typing the paper or
Carlos y su esposa bought the house Carlos and his wife bought the house, or held to be
ungrammatical like estamos typeando el paper or ellos bought the house they bought the house
have operated under the assumption that the data come divided up into already-identified
stretches of one language or the other. In this case, the underlined segments are assumed to be
English and the others are assumed to be Spanish, an assumption that is clearly unwarranted
for bilingual speakers. It is an assumption, and this is the point, that makes the conclusion of
internal duality of competence to a large extent pre-ordained by the dual tagging of the data, a
tagging performed on the basis of the external named language categories. When there is so
little distance between assumptions and conclusions, the validity of the latter tends to be
diminished.

Clarifying translanguaging

299

The concepts of translanguaging and idiolect allow us to more clearly

connect our efforts on behalf of minoritized communities with the charge of
essentialism that we have leveled against the notion of the named languages; for
once we stop focusing on the task of preserving or strengthening an essentialist
set of lexical and grammatical features that has been given the name Euskara or
Mori or Hawaiian (or English, French, or Spanish, or whatever), we can more
clearly see the object of our advocacy. The struggle is not to preserve a pure,
well-bounded and essential collection of lexical and structural features, but
rather a cultural-linguistic complex of multiple idiolects and translanguaging
practices that the community finds valuable. It is toward the affirmation and
preservation of these complexes, and not of named essentialist objects, that
maintenance and revitalization efforts are properly directed.
The next two sections consider how a translanguaging perspective can
change the understanding that educators bring to language proficiency assessment and language education in general, and how it can contribute to the
educational growth of students and the intellectual growth of the field.

7 Translanguaging and language assessment

What are the implications for language proficiency assessment of the distinction between speaking a named national language and translanguaging? In a
nutshell, it is this: testing the proficiency of children in a language must be
kept separate from testing their proficiency in language. Assessing the size,
development, flexibility, richness, complexity, and agility of deployment of an
idiolect must be kept separate from testing the ability to recognize and adhere
to politically defined boundaries in the deployment of the idiolect. The former
is a true assessment of linguistic proficiency; the latter is an assessment of
cultural and political proficiency. When these two things are conflated so that
the student is tested for both, the results will be determined by, and be a
reflection of, the students cultural identity as much as his or her linguistic and
communicative proficiency. Accuracy of measurement is a bedrock value in
the context of educational testing. Yet forbidding bilinguals to translanguage,
or assessing it negatively, produces an inaccurate measure of their language
proficiency. If proficiency assessment is to be accurate and informative, it must
adopt the inside perspective that will reveal the linguistic condition of the
individual students idiolect, irrespective of the social rules that qualify or
disqualify some or all of the idiolect as belonging to a particular named
language.

300

Ricardo Otheguy et al.

To be sure, being able to communicate with audiences of different linguistic

backgrounds and being able to behave in linguistically acceptable ways in a
variety of social settings (including the settings that devalue translanguaging)
are valuable skills, especially for older students. But they shouldnt be confused
with general linguistic skill. This skill involves the ability to express complex
thoughts effectively, to explain things, to persuade, to argue, to give directions,
to recount events, to tell jokes. If schools want to test students linguistic
ability their ability to do these things with language it doesnt make sense
to ask them to perform using only some of their linguistic repertoire; it doesnt
make sense to ask them, that is, to deploy only a portion of their idiolect. And
it certainly doesnt make sense to compare them to monolingual children
and adolescents, who are permitted to use all or most of their linguistic
repertoire their whole idiolect in doing those things.
The negative valuation of translanguaging in assessment doesnt make
sense for the same reason that it doesnt make sense to judge the culinary skills
of cooks by asking them to create a meal while limiting their choice of dishes to
just those of a single named national cuisine. That skill, as we saw earlier, has
little to do with being a good cook and producing tasty, tasteful and satisfying
meals. A culinary restriction of that sort would give an unfair advantage to cooks
who, for whatever reason, normally cook under national cuisine restrictions,
while disadvantaging all those cooks, amateur and professional, who produce
their terrific meals heedless of national cuisine boundaries. The former, the ones
who are used to operating under the national culinary restriction, can choose
from what is in fact their full repertoire of recipes and ingredients, while the
latter must set aside what may be their best recipes. That is, they must keep in
the kitchen what could very well be the best dishes they could put on the table.
Similarly, when students are tested for their ability to do things with language, it
doesnt make sense to insist that some of them should, in addition, try to restrict
their repertoire (which amounts to insisting that they try to pass for people with
a different cultural identity and personal history). Yet that is exactly what is
being done, for example, when Spanish-English bilingual children are told that
while speaking or writing Spanish they are never to use their English words.
Whether one takes the internal individual perspective of the idiolect or the
external social one of the named language, and whether one assesses bilingual
students with reference to monolingual ones are matters that bear elaboration. In
schools in general, but especially during testing, bilingual students, to their great
disadvantage, are kept from using their entire language repertoire, are compelled
to suppress a big part of their idiolect; are not allowed to translanguage. In
contrast, monolingual students, to their great advantage, are forced to suppress
only a small fraction of their idiolect (the part that is interpersonally inappropriate);

Clarifying translanguaging

301

are regularly allowed to translanguage. Both types of students are asked to be part
of a teaching and testing game that each ends up playing under different rules. It is
small wonder that the monolingual side usually comes out on top.
To make sure that we are not misunderstood, we stress that we regard the
ability for the bilingual to perform as a fluent user of a named language as a
valuable skill, a worthwhile educational goal, and a legitimate thing to test for.
Our point is that this is a specialized ability that is independent of general
linguistic proficiency. Bilingual students who, irrespective of the labels that
society puts on the lexical and structural features of their idiolect, can express
themselves accurately, pleasantly, and successfully in the variety of communicative tasks we mentioned above, should be evaluated as possessing excellent
linguistic proficiency. This positive assessment would be separate and independent of whether or how much these students translanguage. That is, the positive
assessment would be issued regardless of how well these students, by suppressing a large part of their idiolect, are able to pass for people with a cultural
identity that, no matter how highly valued by others, is not their own.

8 Translanguaging and language education

Language education has focused on teaching a version of a named language
known as the standard. Such standards do not embrace all the features conventionally associated with the named language. Rather, they make room only
for those features that index social prestige, that is, only those idiolectal features
found in the speech of those who share a superior class membership, political
power, and, in many cases, an ethnic identity. The linguistic canon that rules the
teaching of what schools call native language arts, second languages, additional
languages, foreign languages, and bilingual education has been shaped by acts
of selective legitimation that license only linguistic features associated with
powerful speakers and states.
This means, then, that schools everywhere seek to limit translanguaging in all
students; even monolingual students are penalized for deploying idiolectal features
that fall outside the sanctioned set. But it is especially in the case of teaching
students who are learning to speak what are considered two languages that the
rejection of translanguaging has the most deleterious social and educational consequences. For in monolingual students, the school-sanctioned set of lexical and
structural features constitutes a very large proportion of the total monolingual
idiolect, whereas in bilingual students it is usually less than half. We have
expressed this point by saying that schools for the most part allow monolingual
students to operate at full (or nearly full) idiolect, but forbid bilinguals from doing

302

Ricardo Otheguy et al.

so. This deprivation of speaking at full idiolect keeps the bilinguals silent and
unengaged, in classrooms where they are placed at a clear disadvantage.
By insisting on an externally inspired partition of the bilinguals idiolect and
on the suppression of one of its parts, traditional teaching of what are considered new languages does not encourage learners to integrate the new linguistic
features and practices into their own repertoire of features and practices. The
result is that the new features fail to become integrated as part of the learners
idiolect. That is, they seldom reshape their mental grammar through these new
interactions, seldom shape their speech in ways that uniquely pertain to them
and come to enhance their repertoire.
Bilingual education programs have experienced considerable growth over
the past fifty years. But most of these programs have been built on standardized
versions of named national languages rather than on the idiolects that bilingual
children bring into classrooms. Bilingual education programs tend to separate
their two languages, in an effort not to contaminate the other named language.
In other words, these programs usually aim to discourage or eliminate translanguaging. This is true even in programs that are the product of struggles by
minoritized communities; here too the schools isolate minoritized speakers from
the full idiolects and translanguaging practices that already exist in the community. Minoritized idiolects and practices need protection in order to grow, but
growth cannot take place when we isolate them from the interactions of authentic speakers whose idiolectal repertoires hold much more than what the schools
are willing to license. The prevention of translanguaging in many language
classrooms has had the pernicious effect that one would unfortunately come
to expect. It has resulted in the schools inability to truly develop the multilingual capacities of students. And this in turn has prevented many bilinguals
from becoming successful creative and critical learners (Wei 2011).
Clearly, learning to deploy ones idiolect so as to be considered a speaker of
English or Spanish or Euskara or Hawaiian is an important sociolinguistic
accomplishment and a valuable social skill. But learners must first be allowed
to speak freely, so they can develop the lexical and structural features for the
different social contexts in which they are expected to interact. Only then, when
all semiotic practices are licensed as valuable, will there be new speakers of
minority languages (ORourke et al. 2015), as well as majority languages.

9 Summary and conclusion

This article has offered clarification of the term translanguaging as it applies to
both education and the use of language in general. We have defined

Clarifying translanguaging

303

translanguaging as the use of ones idiolect or linguistic repertoire without

regard for the socially and politically defined boundaries of named languages.
In order to clarify the meaning of translanguaging, we have distinguished
between two senses of language, one pertaining to the social construct of the
named language and the other to the collection of individual (ordered and
categorized) linguistic features of the mental grammar. These considerations
apply the concept of translanguaging not only to languaging in education, but
to languaging in any setting, and to monolinguals as well as bilinguals. Thus the
theoretical scope of the discussion is not only language in education but
encompasses language more generally.
Of the two senses of language mentioned above, we have compared the first,
the social construct of the named national language, to a named national
cuisine. We have shown that neither a national cuisine nor a national language
is defined on the basis of essential traits, be they cooking ingredients or
linguistic features, so that a named language is no more a linguistic entity
than a cuisine is a culinary one; they are both social entities. We have recalled
that the collection of individual lexical and structural features is traditionally
called an idiolect. And we have insisted that the only thing that anyone actually
speaks is his or her idiolect. Moreover, it is only the idiolect that the grammarian
studies, even when penning titles having to do with French tenses, English
number, or Spanish pronouns. The tenses, the number, the pronouns, and all
other linguistic features postulated by a linguistic theory are, in most cases, true
idiolectal properties that can be subjected to analysis by the linguist. The
French and the Spanish are familiar labels whose purported use is to communicate but whose more frequent actual outcome is to confuse, because they
reaffirm the false claim that the features being studied are part of a larger
linguistic object.
To help make the point, we have recalled the inability of the grammarian,
the student of lexicon and structure, to settle one-language-or-two disputes.
This inability stems from the simple fact that languages are not isolated, identified, or defined in lexico-structural terms. Rather, they are established on the
basis of the identity and shared history of the people who speak different but
partially overlapping idiolects that are, by convention, and only when certain
social and historical conditions obtain, gathered together under a single linguistic rubric. When the social and historical conditions inspire consensus, the
layman finds a language and the grammarian goes along; when they inspire
conflict, the ensuing dispute, being of a social and historical nature, cannot be
settled using the tools of the grammarian, which are lexical and structural, and
pegged to idiolects, and thus unsuitable for addressing matters of history and
identity pegged to social groups.

304

Ricardo Otheguy et al.

Since named languages are social constructs, and bilinguals are usually
defined as possessing two such constructs, it follows that the terms language,
monolingual and bilingual refer to social, not lexico-structural objects. Our
interest is not to deny their existence. Rather, we wish to preserve them for
the useful and important roles that they can play in the life of speakers and
scholars alike, while stressing that we need to understand the grounds that
justify their existence in the first place.
Once named languages have been properly allocated to the socio-historical
realm of boundary-making efforts by actual or emerging nations (or by existing
or aspiring states), and once idiolects are situated in the individual idiolectal
realm of lexico-structural repertoires, we are able to define translanguaging as
the full implementation of the latter without regard for the former; as the full use
of idiolectal repertoires without regard for named-language boundaries. That is,
once the proper allocation of named languages and idiolects is accomplished,
we are able to see that to translanguage is to set in motion all of ones idiolectal
features, heedless of the social walls erected by the named languages. We have
explained that, to be sure, neither monolinguals nor bilinguals ever engage in
fully unfettered translanguaging, for the trivial reason that monolinguals regularly suppress those features of their repertoire (usually lexical ones) that are
inappropriate for certain settings or interlocutors. But we have noted that the
suppression involved in refraining from translanguaging (in not acting at full, or
nearly full, idiolect) is necessarily much greater in bilinguals who, owning a
larger repertoire, are regularly called to set aside larger portions of it, including
structural parts, when not translanguaging. The larger the suppression, the
larger, and the more necessary, the gain and liberation represented by translanguaging, especially in educational settings.
We have also argued that translanguaging offers a way to fairly and accurately
sustain the named languages of bilingual minoritized groups. The traditional
named language approach to the protection of the language practices of minoritized
groups is constantly colliding with the power of the dominant named language (or
languages). This dominant language constitutes a powerful sieve that is interposed
between the student and the school to trap many idiolectal features and toss them
aside as inappropriate or illegitimate. In contrast, a translanguaging approach
invites speakers to deploy all their linguistic resources, both lexical and structural,
freely, incorporating new ones and using them without restraint as a part of their
idiolect. Because of this, translanguaging offers support for the sustainability of
what are considered threatened or endangered languages. In addition, in both
assessment and instruction, translanguaging provides a more accurate measure
of learners linguistic abilities, and a fairer and more expeditious way to teach both
content and the social construct of named languages.

Clarifying translanguaging

305

For scholars and advocates in language minoritized communities who may

still recall Max Weinreichs famous, and somewhat tongue-in-cheek, definition
of a language, we may say that we stress the connection between the language
and the army and the navy (we have stressed the social character and definition of the named language that is the object of advocacy). We then propose that
we need to, if not decommission the tanks and sink the ships, at least find the
proper place to quarter and mothball them, so as to be able to value the
individual soldiers and sailors instead. That is, we need to give the named
minority language a rest so that we can more urgently, and more effectively,
advocate on behalf of the minoritized speakers.
Clearly language education and language assessment are social matters. But
appealing to the social context has failed to convince educators that it is unfair
and inaccurate to teach and assess bilingual students, and especially language
minoritized students, with monolingual approaches that rely on the standardized version of the named language as the guiding category. Translanguaging
offers an epistemologically different alternative, one that is based on the individual learners idiolect, and it offers as well the potential to expand and free up
all the learners linguistic and semiotic resources. Translanguaging evens the
playing field, giving bilingual students the same opportunity that monolinguals
have always had, the opportunity to learn and grow while enjoying the intellectual and emotional benefits of all of ones linguistic resources.

References
Auer, Peter. 1999. Code-switching in conversation: Language interactions and identity. London:
Routledge.
Auer, Peter. 2005. A postscript: Code-switching and social identity. Journal of Pragmatics 37(3).
403410.
Blackledge, Adrian & Angela Creese. 2010. Multilingualism: A critical perspective. London:
Continuum Press.
Canagarajah, Sursh 2011a. Codemeshing in academic writing: Identifying teachable strategies
of translanguaging. Modern language Journal 95. 401417.
Canagarajah, S. 2011b. Translanguaging in the classroom: Emerging issues for research and
pedagogy. Applied Linguistics Review 2. 127.
Chomsky, Noam. 1986. Knowledge of language. London: Praeger Publishers.
Creese, Angela & Arian Blackledge. 2010. Translanguaging in the bilingual classroom: A
pedagogy for learning and teaching? Modern Language Journal 94. 103115.
Fishman, Joshua. 1964. Language maintenance and language shift as a field of inquiry.
Linguistics 9. 3270.
Garca, Ofelia. 2009. Bilingual education in the 21st century: A global perspective. Malden, MA
& Oxford: Wiley/Blackwell.

306

Ricardo Otheguy et al.

Garca, Ofelia. 2011. From language garden to sustainable languaging: Bilingual education in a
global world. Perspective: A publication of the National Association for Bilingual Education
Sept/Oct 2011, 510.
Garca, Ofelia & Li Wei. 2014. Translanguaging: Language, bilingualism, and education.
London: Palgrave Macmillan Pivot.
Garca, Ofelia & Ricardo Otheguy. 2015. Spanish and Hispanic bilingualism. In Manel Lacorte
(ed.), The Routledge handbook of Hispanic applied linguistics, 639658. New York:
Routledge.
Gumperz, John. 1976. The sociolinguistic significance of conversational code-switching. Working
papers of the language behavior research laboratory, 46. Berkeley: University of California.
Gumperz, John. 1982. Discourse strategies. Cambridge, UK: Cambridge University Press.
Heller, Monica (ed.). 2007. Bilingualism: A social approach. New York: Palgrave Macmillan.
Hornberger, Nancy & Holly Link. 2012. Translanguaging and transnational literacies in multilingual classrooms: A bilingual lens. International Journal of Bilingual Education and
Bilingualism 15. 261278.
Jrgensen, J. N., M. S. Karrebk, L. M. Madsen & J. S. Mller. 2011. Polylanguaging in superdiversity. Diversities 13. 2437.
Lewis, Glyn, Bryn Jones & Colin Baker. 2012a. Translanguaging: developing its conceptualisation and contextualisation. Educational Research and Evaluation 18. 655670.
Lewis, Glyn, Bryn Jones & Colin Baker. 2012b. Translanguaging: Origins and development from
school to street and beyond. Educational Research and Evaluation 18. 641654.
Lin, Angel. 2014. Hip Hop Heteroglossia as practice, pleasure, and public pedagogy:
Translanguaging in the lyrical poetics of 24 Herbs in Hong Kong. In Angela Creese &
Adrian Blackledge (eds.), Heteroglossia as practice and pedagogy, 119136. London:
Routledge.
Makoni, Sinfree & Alastair Pennycook. 2007. Disinventing and reconstituting languages. In
Sinfree Makoni and Alastair Pennycook (eds.), Disinventing and reconstituting languages,
141. Clevedon: Multilingual Matters.
Makoni, Sinfree & Alistair Pennycook. 2010. Disinventing and reconstituting languages.
Clevedon, UK: Multilingual Matters.
Myers-Scotton, Carol. 2005. Multiple voices: An introduction to bilingualism. Hoboken, NJ, USA:
Wiley.
ORourke, Bernadette, Joan Pujolar & Francisco Ramallo. 2015. New speakers of minority
languages: The challenging opportunity. International Journal of the Sociology of
Language Number 231 (entire issue).
Otheguy, Ricardo. 2009. El llamado espanglish. In Humberto Lpez-Morales (ed.), Enciclopedia del
espaol en los Estados Unidos, 222247. Madrid: Instituto Cervantes & Editorial Santillana.
Otheguy, Ricardo. 2015. Variationist sociolinguistics and linguistic theory in the context of
pronominal perseveration. In Rena Torres Cacoullos, Nathalie Dion & Andr Lapierre (eds.),
Linguistic variation: Confronting fact and theory, 319334. Routledge Publishers.
Otheguy, Ricardo & Nancy Stern. 2011. On so-called Spanglish. International Journal of
Bilingualism 15. 85100.
Otheguy, Ricardo & Nancy Stern. 2013. Scholars and citizens: Judging the unfortunate term
Spanglish. Anthropology News, Winter 2013.
Otheguy, Ricardo, Ana Celia Zentella & David Livert. 2007. Language and dialect contact
in Spanish in New York: Toward the formation of a speech community. Language 83.
770802.

Clarifying translanguaging

307

Otheguy, Ricardo & Ana Celia Zentella. 2012. Spanish in New York: Language contact, dialectal
leveling and structural continuity. New York, NY, USA: Oxford University Press.
Pennycook, Alistair 2010. Language as a local practice. London & New York: Routledge.
Reid, Wallis. 1991. Verb and noun number in English : A functional explanation. London &
New York: Longman Publishers.
Reid, Wallis. 2011. The communicative function of English verb number. Natural Language and
Linguistic Theory 29. 10871146.
Saussure, Ferdinand de. 1916 [1986]. Cours de linguistique gnrale. Translated as: Course in
general linguistics, by Roy Harris. La Salle, IL: Open Court Publishers.
Sayer, Peter. 2013. Translanguaging, TexMex, and bilingual pedagogy: Emergent bilinguals
learning through the vernacular. TESOL Quarterly 47. 6388.
Shin, Naomi Lapidus & Ricardo Otheguy. 2013. Social class and gender impacting change in
bilingual settings: Spanish subject pronoun use in New York. Language in Society
42. 429452.
Urciuoli, Bonnie. 2013. Is Spanglish a bad term? Anthropology News, Fall 2013.
Wei, L. 2011. Moment analysis and translanguaging space: Discursive construction of identities
by multilingual Chinese youth in Britain. Journal of Pragmatics 43. 12221235.
Williams, Cen 1994. Arfarniad o Ddulliau Dysgu ac Addysgu yng Nghyd-destun Addysg
Uwchradd Ddwyieithog [An evaluation of teaching and learning methods in the context of
bilingual secondary education]. Unpublished University of Wales (Bangor) Doctoral Thesis.

Bionotes
Ricardo Otheguy
Ricardo Otheguy is professor of linguistics at the Graduate Center of the City University of New
York in the Ph.D. Program in Linguistics and the Ph.D. Program in Hispanic and Luso-Brazilian
Literatures and Languages. His theoretical work is in sociolinguistics, language contact, and
sign-based linguistic analysis; his applied work is in the teaching of Spanish to U.S. Latinos.
Ofelia Garca
Ofelia Garca is professor at the Graduate Center of the City University of New York in the Ph.D.
programs in Urban Education and Hispanic and Luso-Brazilian Literatures and Languages. Her
most recent work has focused on the potential of translanguaging in language education
contexts, and particularly in bilingual education.
Wallis Reid
Wallis Reid holds a doctorate in linguistics from Columbia University. For thirty-three years he
taught at Rutgers University in the Graduate School of Education. His interests are in linguistic
theory and sign-based linguistic analysis.

Copyright of Applied Linguistics Review is the property of De Gruyter and its content may
not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.