FILE 10.
Language and Identity
10.5.1Using Language to Mark Identity
Language exists only because people have created it and use it on a daily
basis to communicate, and it is therefore a social phenomenon. As was
discussed in Chapter 1, and as is further elaborated in Chapter 14 on animal
communication, language is a strictly human phenomenon: no other creatures
use language as humans do—yet all humans, in all societies, use language.
Language therefore sets us apart from the rest of the world and unifies us: it
is an indicator of the unique place humans have in the world, a marker of our
identity as humans.
At the same time, given the diversity of human languages, language has
often also been used as a marker of identity within the greater human society,
and this use of language is the focus of this file. Because every typically
developing human acquires language, and acquiring language depends on the
ambient language varieties surrounding the learner (see Chapter 8), language
is readily available as a way to show which social groups a person identifies
with or dissociates from.
In Files 10.3 and 10.4, we described some of the lines along which
languages vary: nationality, ethnicity, age, gender, socioeconomic class, and
so on. Throughout those files, we pointed out that there is not usually
anything inherent about any particular group that might condition the
members of that group to speak one way as opposed to some other way. For
example, there’s nothing about being younger in itself that makes some
younger English speakers use more rising intonation than older English
speakers—one can easily imagine a world in which the older speakers use
more rising intonation and the younger speakers use less, or a world in which
older and younger speakers’ language is differentiated in some other way
entirely. What matters here is that age is one factor that differentiates groups
of speakers who identify with each other in some manner, and once these
groups are formed, the members of the group may develop particular
language characteristics that distinguish their group from other groups. These
characteristics are often not consciously developed; they simply happen
naturally in the course of language use and are then picked up and propagated
by new acquirers of the particular language variety.
10.5.2Defining “Identity”
The notion of “identity” is obviously a complex one. Most people have more
than one group whom they identify with. For example, you may identify
yourself in the classroom as a student but at the camp you work at in the
summer as an instructor. Or you may identify yourself broadly as an
American, but more specifically as someone from northwest Ohio rather than
some other part of the country or the state.
To a large degree, your “identity” at any given moment actually depends
on the circumstances of the situation—where you are, whom you are talking
with, what you are talking about, how you want to be perceived, what you
hope to accomplish, who else is present, and so on. Most scholars agree that
identity is not a static thing that anyone “possesses” but rather is a dynamic
construct that emerges from social interactions. It revolves around
establishing the relationship between one individual and the rest of society.
One of the focuses of sociolinguistic research has been to determine
both what elements define a person’s identity and how identity is established
in linguistic interactions. Clearly, identity is not expressed solely through
language use—other actions, practices, and characteristics also create a
person’s identity. It is common, however, to use language to establish or
indicate the identity of yourself or of others.
It is important to remember that this use of language is not always
conscious or intentional, or even within your control as the speaker. You may
use particular lexical items that mark you as coming from a particular region,
without even realizing that people from other regions might use different
terms. Or the person you are talking to may have ideas about what your
pronunciation of a particular word signifies that you do not intend—for
example, not pronouncing the /r/s at the end of syllables, in words such as in
four, card, paper, and so on, might be taken as a sign of your low prestige by
a listener from New York but as a sign of your high prestige by a listener
from London. The perception of your identity by those listening to you is
arguably just as important as your perception of it, as identity is rather
meaningless in isolation. While you can certainly use your own knowledge of
society’s associations between particular linguistic phenomena and elements
of identity to bolster, downplay, or separate yourself from certain types of
identities, you do not have complete control over your linguistic identity.
10.5.3Signaling Identity
How exactly can particular aspects of identity be signaled through language?
You will probably not be surprised to learn that identity can be indicated at
every level of linguistic structure, from phonetics to pragmatics.
The most obvious use of language to signal identity is when people
overtly state affiliation with or dissociation from a particular group (e.g., I’m
from Ireland, or Oh, no, you shouldn’t think I’m British).
Another overt use of language to establish identity that is slightly more
subtle than simply stating it is the use of forms that mark a particular identity.
For example, when there are different words for male and female versions of
the same profession, the use of one or the other signals the speaker’s
association with a particular gender (I’m an actor/actress). In a language such
as French where adjectives must agree in gender with the noun they modify,
this association is even stronger (je suis une actrice sportive ‘I am an (fem.)
athletic (fem.) actress’). Other speakers also signal their views of someone’s
identity this way—for example, using the pronoun he or she indicates that a
speaker assumes that the person they are referring to is of a particular gender.
More obliquely, identity can be signaled by making use of linguistic
characteristics that society associates with particular social groups. For
example, using monophthongal [ɑː] instead of diphthongal [ɑɪ] in words such
as tide or I is associated with the southeast United States. Using like as an
interjection (I was, like, going to the store) is associated with younger
speakers. Using be to mark habitual action, as in I always be late to school, is
associated with African-American English. So, using such forms can be taken
as a sign of belonging to the associated group. 1
On a broader level, in societies where multiple languages are used, the
choice of one language over another can be used to indicate a certain
affiliation or identity. For example, in Canada, both French and English are
official languages at the federal level (see File 11.3 for more on official
languages). However, most provinces are almost exclusively anglophone
(English-speaking), while Quebec is francophone (French-speaking)—New
Brunswick is the only officially bilingual province in Canada. Because of
various political tensions between Quebec and anglophone Canada (e.g.,
there has long been an independence movement to separate Quebec from the
rest of Canada), the choice of language is sometimes seen as making a
political statement. In 1978, Soma Day told of an encounter she had: “I
stopped in a garage and struggled to explain that my windshield wipers were
congellé [frozen over] and I wanted to make them fonctionner [function]. The
man listened in mild amusement and then said, ‘You don’t have to speak
French to me, Madame. I’m not a separatist’” (reported in Fraser 2006: 144–
45). While it has become more socially acceptable for anglophones to speak
French in Quebec, the choice of which language to speak when two
bilinguals encounter each other is still influenced by a number of factors: “the
language the relationship was established in, where the conversation happens,
the presence of other people, the nature of the relationship, and a whole series
of other factors that can involve shadings of power (‘I’ll pick the language
here’), one-upmanship (‘I speak your language better than you speak mine’),
exclusion (‘I speak my language only with my people—and you’re not one of
them’), complicity (‘We speak this language and they don’t’), solidarity (‘I’m
one of you’ or ‘You’re one of us’)—or simply convenience” (Fraser 2006:
144).
It is interesting to note in the above discussion that identity can be
marked either by showing the inclusion of the speaker in a particular group or
by showing the dissociation of the speaker from a different group. That is,
sometimes the usage of linguistic forms directly marks a particular identity
(e.g., using habitual be might mark the speaker as being African American).
But sometimes the usage indicates identity by showing that the speaker is
choosing not to speak in some other way that might be expected. For
example, Elaine Chun’s 2001 study showed that some Korean-American men
use characteristics of African-American English, not to show that they are
African American, but instead to show that they are not Euro-American.
Signaling identity through language can have various consequences.
There are plenty of social stereotypes that accompany ideas of identity, so
your language use may cause listeners to form ideas about you, your
personality, your abilities, and so on. For example, in the United States,
British dialects are often considered to be a mark of a more educated person,
while speakers of southern US dialects are considered uneducated. Of course,
these associations are usually unfounded—someone from Alabama may be
more highly educated than someone from Oxford—but they do have
consequences for communication and social perception. Thomas Purnell,
William Idsardi, and John Baugh did a study in 1999 that showed
discrimination by landlords based on the perceived ethnicity of a potential
renter, as determined through a telephone conversation. John Baugh, who has
fluent command of Standard American English, African-American English,
and Chicano English (see Section 10.4.5), called various landlords who had
advertised housing for rent. In housing districts that were largely Euro-
American in population, when he spoke using either the African-American or
the Chicano English dialects, he was often told that the housing was
unavailable—and then, using the Standard American English dialect within
30 minutes of the other calls, he was told that it was. This kind of dialect
discrimination (which is illegal in the United States) is an unfortunate
consequence of the use of language to mark identity; fortunately, John Baugh
and his colleagues are working with the US government to fight against it. 2
10.5.4Studying Identity
In the early days of sociolinguistic research, scholars often tried to isolate
“essential” elements of speakers’ identities. They would determine certain
sociological factors that could be defined relatively easily (such as the factors
described in Files 10.3 and 10.4: region, age, gender, etc.) and then pick a
particular linguistic variable (such as rhoticity, use of pop versus soda, use of
the needs washed construction, etc.). Next they would see how the two were
correlated (e.g., whether younger speakers use more /ɹ/s at the end of
syllables than older speakers do). These studies were extremely valuable at
establishing broad characteristics of different types of speech, and the studies
formed the foundation of modern sociolinguistic methodology. They are
easily replicable and are well-suited to doing statistical analysis. Most of the
discussion of sociolinguistics that we have presented has been based on such
studies.
At the same time, however, these studies are somewhat limited in
determining a person’s sociolinguistic identity. First, they make use of
sociological factors that are predetermined by the researcher and that may or
may not have any actual relevance for a particular speaker’s identity. Second,
they are tied to the sociological variables that are salient in a society; it is
therefore difficult to do meaningful comparative studies across societies.
Third, some of these variables are difficult to isolate or establish for a
particular person—for example, what do you do about someone who has
moved around several times? Someone who is of mixed ethnicity? Someone
who grew up with wealthy parents but has since rejected that lifestyle?
While the factors used in such studies clearly do play a role in identity
formation (and establishing particular correlations can help determine a
person’s identity based on her speech patterns), they do not tell the whole
story. More recently, researchers have expanded their studies to try to include
factors that speakers themselves may identify as being more relevant. For
example, does a speaker feel that her involvement in the service organization
Habitat for Humanity has created for herself a particular community with
which she identifies? Do people involved in Habitat for Humanity have
unique speech patterns—for example, the use of certain home-building
jargon? Similar groups can be identified for all sorts of communities. One
obvious location of such groups is schools—high schoolers often form
cliques that distinguish themselves linguistically. Although two speakers
could both be 17-year-old, Euro-American, female, middle-class students at
the same high school, they may belong to radically different social groups
and have different speech patterns that mark those groups. Penelope Eckert’s
study of Belten High School showed the large linguistic differences between
“Jocks” and “Burnouts” (see File 11.5 for a discussion of this study). Jock
girls used fairly standard vowel pronunciations, while Burnout girls tended to
use more extreme versions of the Northern Cities Shift discussed in File 10.3.
The members of the two groups had distinctly different linguistic patterns
that reflect their different social groups within a fairly narrowly defined
environment.
One problem with many sociolinguistic studies of identity is known as
the “observer’s paradox.” Often, it is impossible to get an accurate picture of
what speakers do “naturally” among themselves precisely because observing
them makes them change their speech. Knowing that they are being watched,
recorded, or studied may make speakers self-conscious, and many will try to
speak the way they think the researcher wants them to, rather than how they
would normally speak. The perceived identity of the researcher may also play
a role in determining the speech of a person being studied: John Rickford and
Faye McNair-Knox found in a 1994 study that the same African American
participant talked markedly differently when being interviewed by a Euro-
American researcher than when being interviewed by an African American
researcher. This type of differing language use also points out the ever-
changing notion of identity: clearly, the participant was the same person in
both situations but projected a different identity each time.
One way of at least partially avoiding the observer’s paradox is to use
larger databases of speech that have been recorded for other purposes or in
such large quantities that speakers seem to forget that they are being
recorded. These data can be used for subsequent analysis using techniques
known as discourse analysis or conversational analysis to see how particular
identities are established in the course of a conversation. In discourse
analysis, the researcher breaks down a conversation between two or more
people into its various component parts (e.g., the types of turns that are taken
by the participants, the information conveyed, and the linguistic forms at all
levels in which it is conveyed). Of course, this kind of study often does not
give the researcher the same sort of flexibility as doing a direct interview
with someone, because the researcher is reliant on whatever the speakers
happened to be talking about.
No matter what technique a researcher uses, there will be advantages
and disadvantages. These days, a more complete picture of sociolinguistic
identity can be obtained by doing multiple types of studies that complement
each other and each bring out a different aspect: for example, starting with a
broad ethnographic study of a community and using surveys to begin to study
language patterns that naturally occur, and then narrowing down the research
with one-on-one interviews and using techniques of discourse analysis that
pinpoint particular phenomena.
10.5.5Martha’s Vineyard: A Case Study in Language and Identity
In 1961, William Labov conducted a sociolinguistic study on the island of
Martha’s Vineyard in Dukes County, Massachusetts, to investigate the impact
of social patterns on linguistic variation and change. The linguistic feature
chosen for analysis was centralization of the diphthongs /ɑɪ/ and /ɑʊ/, as in
why and wow, to [əɪ] and [əʊ], respectively. In a preliminary investigation,
Labov discovered that after all phonetic, prosodic, and stylistic motivation
had been accounted for, there was still variation in speakers’ use of
centralized diphthongs. His subsequent study was designed to investigate the
motivation underlying this residual variation. Toward this end, Labov set out
to test a number of different variables, many of which were introduced in
Files 10.3 and 10.4.
Was centralization related to geography? The island was, by universal
consensus, divided into up-island (strictly rural) and down-island (consisting
of the three small towns where 75% of the population lived). United States
Census reports were consulted for information on the population distribution
of the island.
Was ethnic group a factor in centralization? Native Vineyarders fell into
four ethnic groups: (1) descendants of old English families, (2) descendants
of Portuguese immigrants, (3) a small Native American population, and (4) a
miscellaneous group from a number of origins. Another group, not
considered in the study, was the summer population.
Were the economic background and the current economic situation of
the island in any way correlated with linguistic behavior? In comparison to
the rest of the state, the Vineyard had higher unemployment, lower average
income, and no industry, and thus was heavily dependent on the summer
tourist trade. This heavy reliance on tourism was viewed by some islanders as
a threat to independence. As a result, many islanders displayed resistance to
the summer visitors and took pride in being different from the tourists, the
greatest resistance being felt in the rural up-island areas.
The results of the study revealed that, first of all, centralization was a
linguistic feature of Martha’s Vineyard and thus regional in character. That
is, residents of the island pronounced /ɑɪ/ and /ɑʊ/ as [əɪ] and [əʊ], while
summer tourists and mainland residents did not centralize the diphthongs. But
within the island population, some residents centralized, while some did not.
Analysis of centralization by age indicated an increase of centralized
diphthong use with age, peaking between thirty-one and forty-five years and
then decreasing. It was also interesting to note the economic situation of this
particular group. Members of this age group seemed to suffer the greatest
degree of economic pressure, having chosen to remain on the island while
supporting their families, even though employment opportunities were not
abundant. Additionally, high school students planning to go to college and
then return to the island exhibited greater centralization than those going to
college but not planning to return to the island.
With respect to ethnic group, the Portuguese population, which for years
had been attempting to enter the mainstream of island life, showed a high
degree of centralization. And those of Native American descent, having
battled discrimination from the other groups for more than 150 years and also
desiring acceptance, also displayed a high incidence of centralization.
Although it is clear that each of these regional and social factors that we
might expect to influence linguistic variation does play a role, we can reach a
deeper understanding of the effects at work here by thinking about language
and identity. Specifically, we can summarize the effects of these different
factors on centralization in terms of group identification. How closely
speakers identified with the island—e.g., wanted to remain, wanted to enter
into the mainstream, saw themselves as Vineyarders and were proud of it—
was positively correlated with degree of centralization. The use of
centralization was not necessarily conscious on the part of these speakers;
centralization was, however, associated with being a Martha’s Vineyard
native, so those who felt especially close to the island tended to increase their
use of this linguistic variable. Thus, speakers use and manipulate language to
show their own identity. Remember, though, that there are many aspects of a
speaker’s identity and speakers often desire to signal different identities, or
different combinations of identities, at different times and in different
contexts.
______________
1
An interesting twist on this is the use of identity to affect language perception. Elizabeth Strand found
in a 1999 study that listeners will actually classify phonemes (see File 3.2) differently depending on
whether they think they are hearing a man or a woman. Strand synthesized a continuum of fricatives
from [s] to [ʃ] and had people categorize each sound on the continuum as one or the other. The listeners
heard the sounds while looking at either a male face or a female face—and the categorization of the
continuum was different depending on the face!
2
To find out more about this study, and to hear recordings of John Baugh, follow the link to his
website on the Links page for Chapter 10.
