Uruguayan Portuguese (UP) is a rural and stigmatized variety spoken in
several bilingual and diglossic communities along the
Uruguayan–Brazilian border since colonial times. It has coexisted with
Spanish, the national language1
The Uruguayan
constitution does not indicate Spanish as the official language, but rather
refers to it as the national one.
and the one preferred by groups
of higher socioeconomic status, despite educational policies and language
planning aimed at promoting Spanish monolingualism over the last two
centuries. Lack of schooling in Portuguese and little access to written
models led previous studies to conclude that UP is a monostylistic,
rural, and heavily mixed dialect with no standard model (
Behares, 1984b;
Elizaincín, 1992). However, the present
study suggests that recent urbanization of border communities has
allowed greater reception of and sensitivity to urban Brazilian
Portuguese (BP), the variety spoken in the neighboring country, which
has caused local UP to be pulled in the direction of the more
prestigious variety. This tendency can be seen through the
incorporation of new phonological variants in the speech of certain
groups, which, by borrowing from urban BP, lead a linguistic change
from extremely stigmatized varieties of UP to urban varieties, which
are close to an ideal standard. Urbanization of UP, thus, entails a
movement away from its rural and hybrid origin, toward an assimilation
of linguistic features that are stereotypically Brazilian, as the
result of a desire to emulate speakers of larger urban monolingual
communities in central Brazil, whose dialect is shown daily on
television.
This movement away from a focused UP (Le
Page & Tabouret-Keller, 1985) may be explained as dialect
diffusion (as defined by Bortoni-Ricardo,
1985), a process by which rural vernacular speakers' direct
or indirect contact with the standard language results in the formation
of a rurban dialect, in which the occurrence of typical rural
lexical items decreases and some nonstandard rules, which were almost
categorical, become variable rules. Dialect diffusion, thus, is a
result of dialect leveling once it involves the reduction of locally
marked variants as a consequence of dialect contact. Both processes
underlie the incorporation of features borrowed from urban BP into
rural UP, hereby classified as urbanization of UP.
This process is investigated through the distribution of
palatalization of dental stops (di, ti), a typically Brazilian
pronunciation that has entered UP. Based on the area's social
history and current sociolinguistic stratification, this article
discusses the factors that influence the acquisition and use of this
variable and its status as a change in progress. Finally, it argues
that in this particular context, in which speakers are linguistically
insecure and Portuguese is spoken as a minority language in domestic
domains, television plays a major role in the process of urbanization
of UP by providing a linguistic model for the community. Nevertheless,
the study concludes, along with Naro and Scherre (1996), that it is not television itself that
changes linguistic behavior, but it is the speakers' attitude
toward the surrounding cultures that enables television to become a
useful source of linguistic modeling.
DATA COLLECTION
Data were collected during five months of field work in Rivera, a
bilingual town on the Uruguayan–Brazilian border (Figure 1). The participants were all residents of
Rivera and were selected on the basis of availability. The conclusions
presented here are based on the output of 56 one-hour interviews in
Portuguese and intense participant observation. The interviewees are
grouped according to their socioeconomic status2
Socioeconomic status was measured based on a composite
index in which three factors were taken into consideration: average
household income, occupational status, and level of education. The
weight of each factor was, 0.5, 0.4, and 0.3, respectively. Scores from
each factor were multiplied by the factor's weight, and then added
together. Once the overall scores were computed, it was possible to
divide the sample population into discrete socioeconomic categories.
Conclusions regarding these categories were reviewed and confirmed by
several community members.
(working-class, lower-middle-class,
and mid-middle-class socioeconomic groups), age (15–29,
30–49, 50–70 years old), and gender. These categories were
defined after several weeks of observation and interviews, and after
being extensively discussed with members of the community. Both
socioeconomic status and age cohorts revealed themselves as important
social variables because of the community's social history.
Different socioeconomic groups have identified with the border culture
in different ways, and the three generations have experienced different
periods of urbanization.
The Uruguayan–Brazilian border (adapted from Peter Trudgill's
Dialects in contact (1986:84)).
The interviews were conducted in the homes of the participants.
Although the presence of a tape-recorder and the interview situation
itself prevented obtaining a reliable sample of casual speech, a
problem referred to by Labov (1972) as the
observer's parodox, the goal of the interviews was to record as
much spontaneous conversation as possible. To achieve this goal, the
interviewer (a middle-class Brazilian female in her 30's)
encouraged the participants to ask questions, introduce topics, produce
lengthy narratives, and maintain eye contact with her to minimize the
awareness of the recording. Except for the introductory questions and
the parts of the interview designed to obtain formal style, the
majority of the conversations focused on soccer, soap operas, families,
holidays, cooking, and individual histories. Everyone who was contacted
was willing to be interviewed, except the upper-middle-class women who
constantly denied their ability to speak Portuguese, which was found to
be untrue in most cases.
To access stylistic differences, a picture-naming task was used to
capture their most formal style.3
A
reading passage, a device used in several studies to elicit a formal
style, was not appropriate for this population for several reasons.
First, some of the participants were illiterate; second, UP is an oral
dialect with no written tradition; and finally, a text would be written
in Standard Brazilian Portuguese and would therefore impose a norm that
is unknown for several speakers.
The participants were shown
pictures of things with names that included the phonological variables
in question.
4 For variable (di,ti), the
pictures were of: tigre, ‘tiger’, tiara, ‘hair
band’, tinta, ‘paint’, elefante,
‘elephant’, sete, ‘seven’, dia,
‘day’, diamante, ‘diamond’, diabo,
‘devil’ dezoito, ‘eighteen’, verde,
‘green’, dente, ‘tooth’.
This proved
to be a reliable technique for eliciting an extremely formal style
(according to Labov's definition,
1972),
in that speakers paid close attention to the word they chose to name each
picture, as well as to its pronunciation. Several times, two forms were
given, one in the local form (referred to by locals as
brasilero,
bayano, or
fronterizo), and another in the
correct form (as put forth by the participants), Brazilian
Portuguese, in which case the second production was counted. This task
proved to be effective in eliciting a formal style and created
stylistic contrast in the production of several variables (Carvalho,
1998,
2003b).
PORTUGUESE IN URUGUAY
Throughout all Northern Uruguay, Portuguese was the only language
spoken by Portuguese and Brazilian settlers until the end of the 19th
century, when Spanish was imposed on the Portuguese-speaking
communities through state educational policies and language planning
(Behares, 1987a; Elizaincín, 1976, 1978, 1980, 1992; Elizaincín, Behares, & Barrios, 1987).
From that point on, several measures were taken to introduce the
Hispanic element to the border, as part of the Uruguayan
government's agenda of unifying the country and counteracting the
presence of Luso-Brazilian culture and language, which was seen as a
threat to the newly formed country's ideal of homogeneity (Elizaincín, 1979). In fact, early in the
20th century, Spanish had penetrated the Portuguese-speaking North,
because of the success of public education, migration of southern
Uruguayans, and the foundation of several border towns. Nevertheless,
despite language policies that attempted to inhibit the use of
Portuguese in northern Uruguayan territory, Portuguese has survived,
and bilingualism is widespread and diglossic. Spanish is the language
of school and public life, whereas local Portuguese is employed as a
vernacular in in-group interactions (Behares,
1984a; Carvalho, 1998; Elizaincín, 1992).
Rivera, a town of nearly 70,000 inhabitants, constitutes the largest
urban center on the border. Its closest neighbor is Sant'ana do
Livramento, a Brazilian town of the same size. There is only a street
dividing both towns, and movement of cars and people between the two
countries is uncontrolled. Rivera's downtown concentrates public
buildings, stores, and middle-class housing, where one may hear only
Spanish, except at bus stops and open-air markets. In the immediate
surrounding vicinities, one finds lower-middle-class housing and paved
streets. Here, one generally hears both Spanish and Portuguese,
depending on the nature of the interaction. On the city's
outskirts, in peripheral neighborhoods with low-income housing, unpaved
streets, and substandard sanitation, mostly Portuguese is used among
insiders.
Behares (1987b) claimed that one finds the
“real” border tradition in nearby rural areas. In the town
of Rivera, this border tradition is best preserved on the periphery
among the working-class. In downtown, however, there is a high
concentration of middle-class residents who usually idealize Spanish
monolingual behavior and rarely identify with their geographic origin
(Behares & Gabbiani, 1987).
Elizaincín (1978) noted that although
the use of Spanish intentionally differentiates middle-class members
from the working-class, maintenance of local Portuguese among family
and friends in the working-class functions as a unifying element among
its members. Carvalho (1998) detected an
ongoing shift from Portuguese to Spanish in the private domains of
society among the lower-middle-class members, whereas UP's
vitality is still strong among the lower social strata. Members of the
lower-middle-class assimilate the linguistic behavior of the
upper-middle-class by shifting from Portuguese to Spanish when
interacting with their extended family members. As such, this group has
also tended to raise its children in Spanish, thereby giving rise to a
language shift, at least in family domains. As Barrios (1996) pointed out, this shift is the result of
linguistic planning by the Uruguayan government, which affects not only
public domains, but also private ones. Language planning has entered
the homes of groups, which by eliminating Portuguese from their
households wish to remove themselves from the stereotype of a UP
speaker, and thereby assimilate the linguistic behavior of the upper
classes.
The varieties of Portuguese found in Uruguay are very similar to the
ones found in the countryside of the southern Brazilian state of Rio
Grande do Sul, except for some features resulting from Spanish
interference. Nevertheless, despite the sustained contact with Spanish,
UP has managed to maintain its phonological integrity, allowing for
relatively easy identification of languages in instances of code
switching, despite great similarities between their grammars.
Phonologically, the many features that UP shares with rural Southern
Brazilian dialects include denasalization of final unstressed nasal
vowels, vocalization of the palatal lateral /λ/, and
failure to apply the raising rule to final unstressed /e/. The
use of a trill /r/, instead of the uvular /R/, and the
pronunciation of a postvocalic lateral as an alveolar /l/ are
phonetic characteristics of UP also found in the interior of Rio Grande
do Sul, probably resulting from contact with Spanish.
Lexically, the most salient isogloss that differentiates UP from
urban BP is the use of words of rural origin, essential for the
categorization of UP as being very closely related to rural BP, and the
heavy presence of Spanish loans, the main reason for the popular
categorization of UP as a mixed language, or portuñol.
Morphosyntactically, lack of number agreement markers in the noun
phrase, simplification of the verbal paradigm, and nominative forms of
the pronouns in clitic positions are some among several features of UP
that follow the tendencies of vernacular Brazilian Portuguese, in
addition to sporadic cases of Spanish transfers into UP morphology,
such as an overuse of reflexive pronouns and some hybrid word
formations.
Urbanization of UP
UP is an oral variety with no institutional support, as all
schooling, media, and literature are in Spanish.5
There is, however, a scarce number of literary texts
written in UP, some of which appeared in Behares and Días (1997). Azevedo (2000) and
Coll (1997) examine literary works in Spanish
that attempt to portray Portuguese-influenced dialects of
Spanish.
The only standardizing forces are the ones
originating from Brazil through television or personal interactions
with urban monolinguals. The influence of urban BP on UP varieties has
increased in the last 30 years as a result of exposure to Brazilian
television and recent urbanization.
Although an updated history of the area's urbanization patterns
is lacking, the numbers found in Lombardi and Altezor (1986) and Klaczo and Rial (1981) show that between 1975 and 1985 the urban
population (concentrated in the town of Rivera) grew 15% while the
rural population decreased 9%, indicating migration to the city. In
addition to rapid population growth, frequent and fast transportation
from Rivera to Montevideo (an average of six hours to the South) as
well as to the Brazilian city of Porto Alegre (six hours to the North)
facilitated access to and communication with the nearby capitals.
Moreover, the opening of a large slaughterhouse in Sant'ana do
Livramento in the 1970's and of several duty-free shops in Rivera
in the 1980's created hundreds of jobs in the city. Finally, but
no less importantly, the arrival of television in the 1960's
brought images of idealized urban life to Rivera's homes. As
explained by a local social sciences teacher:
La realidad en Rivera ha cambiado mucho. La Rivera de hace un
cuarto de siglo era de vida calma, mucha ganadería y algunos
comercios. Hemos cambiado con el impacto de la televisión, del
turismo, y de la multitud de gente que ya no se conoce. (Ar.,
67-year-old mid-middle-class male)
‘The reality in Rivera has changed a lot. Rivera, a
quarter of a century ago, had a calm life, a lot of cattle, and a few
stores. We have changed a lot with the impact of television, of
tourism, and of so many people who do not know each other
anymore.’
Urbanization does not affect the population uniformly, as the
urban–rural polarity is associated with social class and age
dimensions. The younger generations are the ones most apt to identify
with more urbanized cultural patterns. The older population, in
contrast, has kept emotional ties to the rural life. By the same token,
members of the middle classes are more urban-oriented than the
working-class, whose members are usually former agriculture
workers.6
In my sample population, 73% of
the participants in the first age group (16–29) reported to have
no contact with the countryside (no family, no friends, no visits),
whereas 67% of the oldest generation (50–70) reported having
worked or lived in the countryside. Likewise, whereas 67% of the
mid-middle socioeconomic group reported having no contact with the
countryside, 62% of the lowest socioeconomic group reported having
worked or lived in the countryside.
Urbanization and exposure to television have resulted in an extension
of the community's linguistic repertoire. From a highly localized
UP variety, Riverans now have access to a dialect continuum that ranges
from rural UP to urban BP, and variation along this continuum responds
to social and stylistic characteristics that motivate speakers to focus
on rural UP or diffuse toward BP (Carvalho, 2003a, 2003b). The
assumption that urbanization of UP has been a result of greater
exposure to Brazilian language and culture is based on the high
prestige assigned to Brazilian Portuguese, a decisive factor in
determining dialect acquisition (Bailey et al.,
1993; Boberg, 2000; Surek-Clark, 2000). Uruguayan Portuguese, on the
other hand, is overtly stigmatized for several reasons. It is rural, it
is nonofficial, and it is heavily influenced by Spanish, which
contradicts the notion of linguistic purity. Consequently, similarly to
bilinguals in several other communities who feel that they speak
inadequately both languages, such as the ones discussed by Gal (1979) and Grosjean (1982), Riverans are extremely linguistically
insecure, and this insecurity is exacerbated by the constant presence
of the standard model. The following comment expresses the general
consensus among the border population:
Eu acho feio o brasileiro da fronteira. (Gl., 42-year-old
mid-middle class female)
‘I think border Brazilian is ugly.’
And maybe to compensate this ugliness, another said:
Nós tratemo de imitar o português do Brasil. (Ll.,
19-year-old lower-middle-class female)
‘We try to imitate Brazilian
Portuguese.’
The stylistic and sociolinguistic variation that results from this
contact between dialects is confined to a range of variation between
local and standard forms. The presence of lexical borrowings,
nonstandard forms, and UP phonology characterizes deep rural UP. Their
absence, on the other hand, is a result of careful substitution of
their equivalents in standard Portuguese.
The constant presence of urban BP through television exacerbates the
community's linguistic insecurity. The following quote belongs to
a 9-year-old girl, who when explaining her acquisition of Portuguese,
reveals, at this early age, her linguistic insecurity toward the
variety spoken by Brazilians on television:
Eu sempre falei assim, mas aprendi também na
televisão a hablar o português. Porque eu escuto mas
também aprendo. Eles falam mais correto que nós. Os
artista sabem muito mais porque eles nasceu no Brasil, nós
não.
‘I have always spoken like this, but I have also learned
from television to speak Portuguese. Because I hear but I also learn.
They speak more correctly than we do. The actors know better because
they were born in Brazil, but we weren't.’
This girl's comments are representative of a general feeling of
insecurity regarding the Portuguese spoken in this bilingual community.
All participants voluntarily judged their dialect inferior to the one
spoken in Brazil. When questioned if being bilingual was a good thing
for them, all participants said that it would be, if they
spoke real Portuguese. This 18-year-old lower-middle-class
woman explicitly expressed her willingness to speak a less local
dialect, and to acquire an urban version of it, or “real
Portuguese”:
Me molesta usar [s] cuando en portugués
verdadero es [z]. Me gusta usar
en vez de [di]. Nunca uso ‘você’, pero me
gustaría usarlo en vez de ‘tu.’
‘It bothers me to use [s] when in real
Portuguese it is [z]. I like to use
instead of [di]. I never use ‘você’, but I
would like to use it instead of ‘tu’.’
In this comment she explains how she would like to use a more
urbanized Portuguese, revealing salient variants that characterize both
dialects. Among the markers of BP cited by this woman, one finds the
palatalization of dental stops.
PALATALIZATION
In most Brazilian dialects, dental stops have palatal affricate
realizations before /i/, as in
instead of [dia], ‘day’, and
instead of [tia], ‘aunt’. This realization results
from the proximity between the area of articulation of the consonant and
that of the vowel.
Palatalization has been documented in several studies of BP.
Nascentes observed its occurrence in the dialect of Rio de Janeiro as
early as 1922. In 1970,
Mattoso Câmara found full palatalization in Rio de Janeiro, but argued
that in São Paulo the plain dental was more common. Although
palatalization is still variable in several areas of Brazil, there is
evidence that it has been spreading, thus representing a change in progress
(Azevedo, 1981). Rio de Janeiro dialect has been
responsible for disseminating palatalization to other dialectal areas. This
diffusion is facilitated by the fact that the Rio de Janeiro dialect is
considered prestigious (Elia, 1976),
especially among speakers of stigmatized dialects, and it reaches the
entire country through the Rio de Janeiro-based Rede Globo, the largest
national television network.
Geographical diffusion of palatalization is still underway in
Southern states, where dental realizations are common enough to
constitute a stereotype of Southern dialects. Bisol (1991) studied the palatalization phenomenon in four
dialect areas of Rio Grande do Sul, the state that borders Uruguay:
Porto Alegre, the state capital and largest urban center, Sant'ana
do Livramento, Rivera's twin city, a German–Portuguese
bilingual community, and an Italian–Portuguese bilingual
community. She concluded that palatalization is expanding in all these
areas, including the Brazilian border town, but that the contact with
Spanish (Italian and German) still hampers its application, as attested
by the fact that the speakers from Porto Alegre are the ones who
palatalize the most.
Palatalization of dental stops is extremely common in Brazilian
television programs and soap operas and in the speech of urban
monolinguals. Following a diffusion that started in Rio de Janeiro,
this variant derives its prestige from its social and geographic
origin, a typical case of language change caused by the spread of
prestige patterns of urban centers, which, in this context, has crossed
national borders. A comparison of my own data with previous studies of
UP shows that palatalization is slowly being incorporated into this
dialect.
Rona anecdotally observed in 1965 that the dental realization was far
more common in Rivera than the palatalized one. Hensey, in 1982, found
very little palatalization in his small sample of UP (6 Rivera
speakers), and concluded that the vast majority followed the
conservative tendency to produce the dentals as stops, which allowed
him to state that “[in] typical border Portuguese
pronunciation we can expect words like ‘tipo’,
‘digo’, to be pronounced [tipo],
[digo]” (1982:16). Nowadays, the
dental pronunciation is still a feature of UP, and its substitution by the
palatalized version constitutes a means for Riverans to sound more like
Brazilians. One teenager, who was trying to explain the difference between
her Portuguese and the Portuguese spoken by her parents, explained it
thus:
Eles dizem assim: ‘Tu sabe que outro [di]
a …’ Eu não, eu falo ‘outro
.
(Kr., 18-year-old lower-middle class female)
‘They speak like this: ‘ You know that another day
([di]a). But not me, I say ‘another day
.’
This comment hints that speakers are able to differentiate the local
dialect from the standard one according to the (non-) application of
the palatalization rule. We may infer that speakers are aware of this
difference, and choose one variant over the other based on the social
value carried by the variant, and whether or not they want to sound
like a local, a speaker of ‘portuñol’.
Quantitative results
Quantification was based on 2,248 tokens in the speech of 56
interviewees. Palatalization occurred in 32%, and dentals were realized
68% of the time. Raw data were applied to the statistical package
varbrul 2S for PC. Along with the factor probabilities,
varbrul binominal analysis gave raw numbers and percentages of
instances of palatalization used in cross-tabulations. Both linguistic
and extralinguistic factors were determined to be relevant to the
output of this variable.
Linguistic factors
The linguistic factors considered in the variable rule analysis were:
tonicity of the environment, preceding segment, following segment, and
juncture. In the final run, when varbrul provided the best
model, with a 0.043 significance, the group factors juncture and
preceding environment were eliminated. The only factors considered
relevant by varbrul were the following environment and the
tonicity of the syllable.
Table 1 shows varbrul analysis of
application of palatalization according to linguistic factors. The two
types of following segment shown to be relevant were nasals and
sibilants. Nasals were given a relatively high probability factor of
.63, which means that palatalization is more likely to occur when the
variable is followed by a nasal. This could be a result of the high
frequency of palatal nasals in the corpus, which because of regressive
assimilation, triggered anticipation of the point or articulation of
,
instigated by the environment [_i]. The
other factor detected to be relevant by varbrul in the
following segment factor group was the presence of a sibilant.
Sibilants are shown to have an extremely strong inhibiting strength,
with a probability weight as low as .13. The inhibitory tendency that a
sibilant exerts in the application of the palatalization rule has also
been detected in the study of Bisol (1991),
who rightly pointed out that the retention of the alveolar results from
the fact that the anticipation of /s/ articulation renders
difficult the raising and fronting of the body of the tongue, necessary
for the palatalization.
varbrul analysis of application of palatalization according
to linguistic factors
Tonicity of the environment was selected as the least important
factor in the distribution of (di, ti), with a contributing weight
value of .57 for stressed syllables, a barely significant value of .54
for pretonic positions, and an inhibiting value of .41 for posttonic
positions. Interestingly, the same tendency was detected by Bisol for
palatalization in Rio Grande do Sul. She pointed out that this order of
preference, that is, stressed, pretonic, and posttonic, implies that if
palatalization occurs in pretonic syllables, it will also occur in
stressed positions; if it occurs in posttonic syllables, it will also
occur in pretonic positions (1991:117). Bisol
explained that palatalization implies an increase in phonetic properties,
which applies preferably in stressed positions. Thus, it is possible to
infer that palatalization is entering UP mainly through stressed
syllables.
Extralinguistic factors
The extralinguistic variables considered were: style, socioeconomic
group, age, and gender. The style factor group was eliminated by the
analysis, because it presented no significant contribution to
palatalization. Table 2 summarizes the
results obtained from varbrul.
varbrul analysis of application of palatalization according
to extralinguistic factors
The numbers in Table 2 clearly show a
pattern of stratification across social groups. Palatalization is a
variable related mainly to age. In the youngest group, its application
is very common, with a probability weight as high as .91, whereas in
the oldest group of the sample population palatalization is nearly
absent, with an extremely low probability value of .05. The probability
value of .29 for the intermediate age group (30–49 years old)
shows that, at this age, application of the rule is strongly inhibited,
however to a lesser extent. The second most important factor is
socioeconomic status. Members of the mid-middle socioeconomic group
palatalize more than members of the other groups, showing a strong
contributing value of .81. The other two factors, lower-middle-class
and working-class, are shown to be strong inhibitors, with very low
factor values of .34 and .30, respectively. As expected, women tend to
palatalize more than men, showing a probability of .60 in the
application of the rule. Men, on the other hand, tend to produce the
more local stop variant, showing an inhibiting probability factor
weight of .39.
The tendency to palatalize among the middle-class, the young, and
women is representative of the fact that these groups are using the
symbolic value of this new pronunciation to show an urban orientation,
different from the traditions of the rural border community. The
working-class, the elderly, and men, who usually rely on local
connections and resources, on the other hand, favor a more conservative
form of Uruguayan Portuguese. The middle classes idealize the national
culture centered in the capital, travel to Montevideo as often as
possible, send their children to college there (the only option for
post-secondary education), and maintain personal and business relations
in Montevideo. These data show that this class, when using Portuguese,
may use an urbanized version of it. The underlying motivation for them
is their different mobility, lifestyle, and ideologies that lead them
to the more prestigious urban and national languages and cultures.
Likewise, young people in Rivera have closer contact with Brazilian
culture and urban values and are voracious consumers of Brazilian media
and music. By assimilating the norms of urban BP they attempt to align
themselves with a group that is different from that of their parents,
and to become active participants in a relatively new model of urban
life, different from the one that characterized Rivera in the past, and
with which the older generations had much more contact. The
identification of palatalization with a particular reference group, as
suggested by qualitative data, and the choice by some socially
identifiable groups to adopt this pronunciation, as suggested by
quantitative data, indicate that the adoption of palatalization
represents an act of identity (Le Page
& Tabouret-Keller, 1985) with the reference group.
Change in progress
The question that remains is whether the incorporation of
palatalization (and other BP features) represents an actual change in
progress or whether it is the result of an ephemeral accommodation,
which, according to Giles and Smith's (1979) definition, is a shift in one's dialect
in order to speak more similarly to the interlocutor. Age
differentiation in the distribution of a variable can be an indicator
of linguistic change in progress. Lack of stylistic differences may
hint that this change is still in its early stages (Labov, 1966, 1981, 2001a). As with many other linguistic changes
(Labov, 1972, 2001a), women are significantly ahead of men in
their use of palatalization. Real-time data corrobate the hypothesis of
change, as previous descriptions of UP mention little palatalization
(Hensey, 1972; Rona,
1965). However, in defining age differences as a potential
indicator of change in progress, it is important to verify that the
majority of rule application in the middle-class group is given by the
youngest participants, so that one can discard stable stratification. A
cross tabulation of socioeconomic and age groups with respect to
palatalization is shown in Table 3.
Cross tabulation of palatalization across age groups and
socioeconomic groups
The data in Table 3 and Figure 2 suggest that there may be a change in
progress, because the youngest generation in the mid-middle-class
socioeconomic group is responsible for most of the instances of
palatalization in this socioeconomic group, 89%. Rate of palatalization
decreases gradually among the next lower socioeconomic groups of the
same age cohorts. Palatalization rates continue to decrease among the
middle-aged participants down the social scale, a pattern that persists
in the next older group, the elderly. Finally, among the oldest members
of the working-class, there was not a single instance of
palatalization, which allows us to characterize this group as being the
most conservative in the use of local variants of (di, ti). Independent
of style shift and gender differences, the total absence of
palatalization in this group may suggest that this is a new variant in
UP, and that it is a prestigious change originating in the higher
socioeconomic groups. From these results, one can also infer that
socioeconomic stratification of palatalization among age groups becomes
evident only in the two younger groups. That is to say, that among the
oldest group, the tendency to pronounce (di, ti) as dentals is largely
generalized. Meanwhile, in the next two younger groups it becomes clear
that, within these age groups, it is the mid-middle-class members who
lead the innovation. Likewise, the cross tabulation of palatalization
among gender and age groups in Table 4 and
Figure 3 reveals that among the oldest group,
both men and women prefer the dental pronunciation. In the next two
younger age groups, however, women lead in bringing palatalization into
UP.
Cross tabulation of palatalization across age groups and
gender
Cross tabulation of palatalization across age groups and
socioeconomic groups.
Cross tabulation of palatalization across age groups and
gender.
Therefore, one could suggest, based on both apparent and real-time
data, that palatalization represents a change in progress in UP, which
has spread southward from Southeastern Brazil. The leaders of this
change are the young, the middle-class, and the women. However, the
investigation of a possible linguistic change in progress in a
bilingual context, where change is brought in through accommodation to
an outside dialect, poses the possibility of ephemeral accommodation,
as well.
Results of a survey on language choice show an ongoing shift to
Spanish in private domains of society among the groups that palatalize
the most when using Portuguese (the young and the mid-middle-class);
from which we conclude that the groups with the highest palatalization
rate are the ones that use Portuguese the least (Carvalho, 1998). They use Spanish among family and
friends and, therefore, solidarity is not constructed or evoked through
UP. As a consequence, their Portuguese is free to change and become
urbanized, as it may be used mainly for interactions with outsiders.
Several acts of ephemeral or short-term accommodation, defined by
Trudgill (1986) as transitory adjustments
made under particular circumstances, accumulate over time and may cause
semipermanent change in one's speech, or long-term accommodation,
if kept by the individual in all transactions in the contact area.7
Most studies investigate accommodation as the
result of sudden immersion in a different speech community where other
dialects are spoken (Chambers, 1992). My
study investigates a rather different situation. Acquisition of urban
BP by UP speakers takes place gradually through sporadic contact with
outsiders and passive exposure to television.
Palatalization,
then, can be the result of an ephemeral accommodation, once it is to be
used with urban Brazilians who palatalize, like the interviewer.
8 The fact that the interviewer was an
urban-educated Brazilian leads us to a crucial methodological problem.
Even though she made an effort not to palatalize and became a fluent UP
speaker during her stay in Rivera, as Bell (1997) pointed out, speakers tend to accommodate to
the speech of groups they think the interlocutor represents. The only
way to assess the extent to which the interviewer influenced the
speaker's palatalization rate is to interview the sample again
using other interviewers, as suggested by Rickford and McNair-Knox
(1994).
Nevertheless, the status of palatalization in UP as the result of an
ephemeral accommodation alone is unlikely. First, previous studies have
indicated the presence of palatalization among UP speakers in the past
(Hensey, 1982; Thun &
Elizaincín, 2000). Second, the data discussed here show a
systematic linguistic and social distribution of (di,ti) among the
sample population that is too robust to be dismissed as a result of
sporadic accommodation, and which better represents a community grammar
resulting from permanent acquisition of an urban BP feature. Thus, I
propose that palatalization has entered UP as a Southward diffusion
from the Southeast of Brazil that has crossed the border and has had a
permanent affect on UP. This case of dialect acquisition is fostered by
urbanization and greater contact with the standard variety and results
in an extension of the linguistic repertoire of UP speakers.
One intriguing point about the acquisition of palatalization by UP
speakers partially as a result of accommodation, is that borrowing from
the prestigious dialect can only occur through borrowing from an
outside dialect. This is because in Rivera, intergroup verbal
interaction is carried out exclusively in Spanish, as Portuguese is
only used during in-group interactions (Elizaincín, 1978). This counters the
situation discussed by Guy (1990), in which
“change from above may come from the existing linguistic
repertoire of the highest group, rather than contact with an external
community” (1990:51). In Rivera, it is not
possible that a linguistic change in UP brought about by the upper classes
could spread down the social scale through contact, simply because
interaction between groups is in Spanish, whereas Portuguese is reserved
for communication among family members and friends, especially among the
working-class. Thus, the source of this linguistic change is an outside
community, that is, urban Brazil. As a consequence, one wonders who is
bringing palatalization into the border, a question that leads us to
the problem of transmission, the central problem of the theory of
language change according to Labov (2001b).
Spatial diffusion is believed to be the result of social contact.
Face-to-face interactions are usually the channels through which
linguistic innovations spread. In Rivera, interaction among Uruguayans
and Brazilians across the border is not infrequent. Some Riverans have
family and friendship ties in Sant'ana do Livramento, and cross
the border sporadically to visit or go shopping, from which one can
hypothesize that palatalization is entering UP through interpersonal
communication. However, diffusion of palatalization into UP is a
phenomenon that would be difficult to explain solely by frequency of
contact with speakers of the standard dialect. First, it is important
to bear in mind that palatalization is still entering the Portuguese
spoken by border Brazilians as well (Bisol,
1991). Second, although 25% of the total sample population
reported going shopping or visiting friends at least once a month in
Brazil, 100% of the young people (the leaders in palatalization)
reported having no social life on the other side of the
border. They usually keep their friendships within their own
neighborhoods, a behavior very typical of this age group. It is the
adults who keep ties outside their communities, a practice known to be
responsible for transmission of new linguistic features (Milroy, 1987), and not the youth, who typically
maintain social ties in close-knit community territories.9
This is a rather impressionistic statement
based on answers to questions regarding social activities in Brazil, as
this study does not present a systematic social network analysis, from
which it would benefit.
Thus, it seems improbable that
interpersonal communication with Sant'ana do Livramento dwellers
is the sole factor responsible for the spread of palatalization and
other urban features into UP.
Lack of contact among speakers of UP and urban BP leads us to review
what Auer (1998) called the
“behavioral-frequency model” (in Kerswill, 2002), or acquisition due to frequent
accommodation acts, as the only explanation for diffusion of urban BP
into UP, and to propose, in addition, Auer's “identity
projection model.” In this model, Auer, based on Bell's
audience design and Le Page and Tabouret-Keller's acts of
identity, accounts for situations where accommodation may not be in
response to a particular interlocutor, but to stereotypes of the group
the interlocutor belongs to, or of a socially attractive group
not actually represented in the immediate context (Kerswill,
2002:680–681).
However, the “identity projection model,” as an
underlying motivation for urbanization of UP among certain groups,
still requires the presence of a linguistic model. Based on the
participant's impressions of the topic (emic) and the current
understanding of the community's profile (etic), I propose that
Brazilian television is used by certain groups as a corrective agent,
and thus indirectly contributes to the diffusion of palatalization
(among other features) among these groups.
TELEVISION
Among the UP-speaking sample population, every household had access
to a television, and regularly viewed Brazilian programs, especially
the soap operas. At the time of the interviews, Brazilian television
was preferred 100% of the time, as the Uruguayan channel's poor
transmission proved little competition for the more modern and diverse
Brazilian channels.10
It should be noted,
however, that the recent arrival of cable television in border
communities might alter this situation. Cable television is now
affordable to the middle-class and presents a wide variety of programs
from Argentina and Spain, in addition to offering improved reception of
the Montevideo channels. A recent informal survey on the topic among
school children in the border towns of Rivera and Artigas (Carvalho, 2002) revealed that although there seems
to be a greater exposure to Spanish-speaking television, Brazilian soap
operas remain extremely popular.
In fact, television has
erroneously been blamed for the presence of Portuguese in Uruguay by
educators and government agencies whose analysis disregards the roots
of Portuguese in colonial times. Consequently, among several measures
considered for the purpose of ridding Northern communities of
Portuguese, as mandated by the Uruguayan Department of Education in
1968, it was suggested that efforts had to be made to improve the
capital's media broadcasting in the North. In fact, several
interviewees recalled being advised by their teachers not to watch
television at all during the dictatorship years when Portuguese was
strongly repressed.
11 From 1973 to 1985,
a military-civilian government imposed rigorous political censorship
and control in Uruguay. See Carvalho (1998)
for a detailed review of the language debate during the
dictatorship.
Nevertheless, Brazilian television, particularly soap operas, is
still widely popular in Rivera. Behares and Gabbiani (1987:7) stressed that the very first contact that
UP-speaking children have with standard Portuguese is through
television:
Estos hablantes [UP native speakers] desconocen el
español hasta su ingreso en la escuela, pero desconocen
también el portugués estándar y toman contacto con
él a través de los medios masivos de comunicación,
el la medida que éstos logran penetrar los moldes comunicativos
de la infancia.
‘These speakers [UP native speakers] do not
know Spanish until they enter school, but do not know standard
Portuguese either and have contact with it through mass media as the
media manages to enter childhood's communicative
channels.’
Television continues to provide samples of “real
Portuguese” throughout the lives of UP speakers. This is not to
say that television viewing is able to implement standard features into
this community's speech repertoire. Rather, the role of television
is interpreted as a source of prestigious speech, which may or may not
be imitated. The determining factor for the use of television as a
linguistic model to be adopted is the speakers' negative attitude
toward their native dialect and their desire to change it. My argument
is that the impact of television in the urbanization of UP, although
indirect, is undeniable.
Several sociolinguists have doubted the role of television in an
audience's linguistic behavior. Trudgill (1986) argued against the hypothesis that television
can affect speech; according to him, face-to-face interaction is
necessary before diffusion takes place. His argument is supported by
Labov and Harris (1986) and Rickford (1988) who agreed that linguistic traits are not
transmitted across group boundaries simply by exposure to other
dialects in the mass media, but through interpersonal interaction.
Although there is evidence that the media may play a role in the spread
of vocabulary items, it is believed that it does not affect
pronunciation or syntax. Labov's results, which show that African
American vernacular speakers do not assimilate Standard English despite
several hours of daily exposure to it on television, add to this
hypothesis. Chambers (1998) admitted that
exposure to standard speech on television may add to its prestige, but
that change itself must be conveyed in face-to-face interactions among
peers, as “there is no evidence for television or other popular
media disseminating or influencing sound changes or grammatical
innovations” (1998:126).
Sociolinguistic studies on the influence of media on speech have
tended to investigate direct correlations between media consumption and
linguistic behavior, which, hardly surprisingly, are usually not found.
In a study that measures the effect of television on nonstandard
dialects, Saladino (1990) tested the direct
impact of television on the level of standarization of a Southern
dialect in Italy. She treated little exposure (0–6 hours per
week) versus more exposure (7–15 hours per week) as two variables
and correlated them to linguistic behavior. She found that television
has no effect on the level of Italianization of the dialect, and
suggested that, although television may improve comprehension, it does
not necessarily improve production. She reached the irrefutable
conclusion that the mere existence of a model does not guarantee that
it will be adopted, given that “covert factors such as feelings
of solidarity can have great influence on whether standard language
norms are adopted or not” (1990:67).
It is doubtful that exposure to television and adoption of standard
features may ever be directly correlated, because for this relationship
to happen there needs to be both psychological and social motives that
instigate dialect assimilation. All of the participants in the present
study were asked about their television-viewing habits and all of them
reported watching television for at least one hour every day (the
answers ranged from 1 to 5 hours daily). Interestingly, the group that
reported watching television the least was the young group, with an
average of 1–2 hours per day. Meanwhile, the elderly, mostly
retired and generally with more free time, reported viewing television
the most, with an average of 5 hours per day. Therefore, the number of
hours of television viewing does not correlate with palatalization
rate, but social variables such as age, social class, and gender do. It
is essentially the desire among the young, the middle-class, and women
to replace UP features with BP that motivates their use of television
as a model for dialect acquisition.
In what follows, several comments are transcribed to make explicit
the population's perceived direct relationship between “a
pure Portuguese” with the one spoken on television, stressing the
role of this mass medium as the major provider of the standard
dialect:
Gosto do jeito que o pessoal fala na televisão porque
é brasileiro em si. Aqui nós somo rompe-idioma. (RS,
21-year-old lower-middle-class female)
‘I like the way people speak on television because it is
real Portuguese. Here, we speak a broken language.’
Os brasilero fala bem brasilero, nas novela eles fala bonito.
Os brasilero fala perfeito, nós aqui fala entreverado. (ML,
23-year-old mid-middle-class male)
‘The Brazilians speak very Brazilian, in the soap operas
they speak beautifully. The Brazilians speak perfectly, we here speak
mixed.’
Eu acho feio o brasileiro da fronteira. O portugués bem
falado é lindo. Eu gosto do jeito da Globo falar
português. (Gl, 51-year-old mid-middle-class
female)
‘I think border Portuguese is ugly. The well-spoken
Portuguese is beautiful. I like the way Globo speaks
Portuguese.’
In the context of a semiborder, Boberg (2000) studied the adoption of American English
features by Canadians, and argued that the determining factor for
diffusion of American variants through mass media is the overt prestige
of certain variants associated with perceptions of correctness. In
Uruguay, Portuguese speakers value the dialect of Portuguese spoken on
television not only as more correct, but as the only acceptable form of
Portuguese. The speech patterns heard on television are reinforced by
sporadic contact with Brazilians, as can be seen in the following
comment by a teenager who works at an expensive bakery that attracts
many Brazilian tourists. She expresses her perception of the dialect
spoken on television as similar to the one spoken by these Brazilian
shoppers, the reference group:
Os brasilero que vem na padaria falam como os da
televisão. Em Rivera e na campanha falam mesclado. (Ll,
19-year-old lower-middle-class female)
‘The Brazilians that come in the bakery speak like the
ones on television. In Rivera and in the countryside they speak
mixed.’
Behares (personal communication, 2002) offered a quote from his data
in which a working-class 64-year-old woman changes from the nonstandard
pronunciation of a [j] in the word
trabalhar, ‘to work’, to the standard
[λ] explaining that, according to her grandson, she needs
to talk standard, like the television:
Os pobre trabaiava que nem escravo. Sabi que o destino de pobre
sempre foi trabaiá nunca pará … Trabalhar!!
(standard, very stressed and careful pronunciation), que meu neto diz
que temo que falá como na televisão! [my
emphasis]
‘The poor people used to work like slaves. You know that
the poor people's destiny has always been to work and never stop …
Work! (standard, very stressed and careful pronunciation), ‘cause my
grandson says that we have to talk like on television!’
[my emphasis]
‘[he] copies from the
actors?’
Sim, sim … aprendeu e copiou das novela.
‘yes, yes … he learned [it] and copied
[it] from the soap operas.’
Milroy and Milroy suggested that although mass media give rise to
awareness of a linguistic innovation, it is unable to promote
adoption (1999:25). My study reveals that
awareness is a precondition for adoption, and without exposure to Brazilian
television, both would be unlikely. The following quote, provided by a
7-year-old boy, shows that television promotes awareness of the
palatalization variable in particular:
D[d]iego. Em brasilero é
.
Eu escutei na novela.
‘D[d]iego. In Brazilian it is
.
I heard it in the soap opera.’
The following dialogue between a young lower-middle-class woman and
the interviewer makes it evident that awareness of the standard
provided by television is used as a means for adoption because of her
desire to assimilate it.
Meu pai e minha mãe falam diferente que eu. Eles falam
mais mesclado.
‘My dad and my mom speak differently from me. They speak
more mixed.’
I: Onde tu aprendeu a falar assim?
‘Where did you learn to speak like
this?’
De escuta, de escuta. Trato de imitar mas às vezes
é difícil.
‘From listening, from listening. I try to imitate but it
is hard sometimes.’
Like students who use television as a source of linguistic input in
their effort to learn a second language, young border bilinguals use
television as a corrective agent in their effort to learn a second
dialect. Both groups have an urge to “improve” their
language skills, are linguistically insecure, and willing to acquire
the new model.
Therefore, a substantial point to be made about the influence of
television on people's speech is that, besides exposure to
television, individual motivation to assimilate to a given model is
crucial. The indirectness of television's effect on linguistic
behavior was first suggested by Naro (1981),
in his seminal work on media's influence on speech. He found that
the cultural orientation variable, or the degree of penetration into
the culture of the surrounding higher socioeconomic levels, was the
determining factor allowing soap opera viewers to produce the standard
variant. In 1996, Naro and Scherre expanded the media variable by
taking into account contact with and attitude toward the media, showing
a correlation between the use of the standard form and increased
integration with the media. They do not argue for causality, as they
affirm that:
We do not feel that it can be concluded from our research that
involvement with the media causes changes in linguistic behavior.
Rather we feel that a third factor, such as a general orientation or
attitude toward the surrounding society, might be responsible for this
phenomenon here shown to be correlated (1996:228).
The differences in individual attitudes toward the standard model
heard on television explain why only part of the sample population uses
television as a linguistic model, although everybody watches television
daily. From the community's perspective, the lifestyle diffused by
television along with language input is primarily valued by certain
groups, who admire national and urban trends to which they have been
increasingly exposed because of recent urbanization, but from which
they have been kept apart because of their geographical origin. From
the individual perspective, personal histories and attitudes elucidate
the motivations underlying individual linguistic behavior.
‘Bts’, a mid-middle-class, 43-year-old woman, provides us
with an eloquent statement about her willingness to diverge her dialect
from the border variety, and assimilate urban BP:
Eu gostava da campanha e falava como falam na campanha. Agora
como eu estou trabalhando no comércio, eu tento assimilar os
vocábulos mais corretos. Eu já não digo
cojé, se não digo colher. Já não digo
calle, digo rua. É diferente, por mais que eu queira falar o meu
brasileiro de antes, eu já não falo porque eu estou
deixando de lado. Hoje eu acho feio o brasileiro da fronteira.
Muié, coié, é feio. O português bem falado
é lindo. Eu gosto do jeito da Globo falar português. Eu
sei que eu tinha que ter orgulho da minha língua materna mas
é uma lingua materna tão feia que dói de ouvir.
Não sei, não tenho orgulho de falar português, aqui
só usam essa língua para falar coisa feia, tipo
‘fio da puta!’. Cuando chica, yo tenía unos gustos
refinados. Me gustaría hablar correctamente el español
pero no podía porque no lo sabía. Yo siempre fue refinada
en mis gustos, siempre tuve un deseo grande de superación, y yo
nunca acepté el fronterizo por eso. Es una lengua de la
calle.
‘I used to like the countryside and to speak like they
speak there. But now since I am working at a store, I assimilate the
more correct words. I don't say ‘street’ (in Spanish),
I say ‘street’ (in Portuguese). It is different, even if I
want to speak the Brazilian I spoke in the past, I can't because I
am leaving it behind. Now I think that border Brazilian is ugly.
Muié, cuié, are ugly. Well-spoken
Portuguese is beautiful. I like the way Globo speaks
Portuguese. I know I should be proud of my mother language, but it is
such an ugly mother language! It is so ugly, it hurts to hear it. I
don't know … I am not proud of speaking Portuguese, here,
people use this language just to say bad things, like ‘son of a
bitch!’. [From now on she switches to Spanish]: When I
was little I had sophisticated tastes. I wanted to speak Spanish
correctly but I couldn't because I didn't know how. I always
had fine tastes, always had a great desire to succeed, and that is why
I never accepted fronterizo. It is a street language.’
This comment illustrates several points of my interpretation of
Rivera's speech community. By saying that she does not use rural
UP anymore, she indicates that she has replaced some of her native
rural UP with features of urban BP, because of a change in attitude
triggered by her experience working at the store. She confirms the
stigma of UP and the fact that her linguistic model is the television
network Globo. Even though she affirms having always been a
person of sophisticated tastes who never liked ‘fronterizo’
at the end of her narrative, in the beginning she states that she used
to like the countryside and to speak like they do there. Her exposure
to Brazilian television reinforces her linguistic insecurity and
provides a model for assimilation. Even though the presence of these
models may affect her speech, it is clear that the crucial condition
for dialect acquisition is her attitude.
By the same token, speakers who wish to emphasize their border and
rural cultural identity maintain the dental pronunciation in a process
of divergence, implicit in the following comment, where a working-class
57-year-old man compares his Portuguese with the one spoken on
television:
This participant acknowledges palatalization as a feature typically
used in television, but expresses no desire to adopt it. By calling
attention to his refusal to accommodate to the “Portuguese spoken
on television” he asserts his status as a fronterizo and
reacts to the pressures of the linguistic marketplace (Bordieu, 1991). His code-switching from Spanish to
Portuguese at the end of his statement also reaffirms his alliance to
his native UP. Therefore, even though dialect acquisition is a
consequence of contact, it is most importantly a consequence of
attitude and desire.
The corpus on which this study is based is replete with references to
television as a standardizing force, and to disregard this indirect
relation between television viewing and urbanization of UP is to fail
to consider emic social factors, which, as Lane claimed (2000:268), are necessary for the development of
explanatory models of change. This analysis accounts for both community
trends and individual choices taken as acts of identity. Quantitative and
qualitative data collected in the speech community of Rivera reveal ways in
which individuals negotiate their images against larger social constructs laid
by group trends.
CONCLUSION
In concluding, frameworks of speech accommodation and acts of
identity were integrated in the Labovian paradigm to explain the
current social distribution of palatalization among UP speakers,
contributing to a better understanding of this dialect, previously
thought of as monostylistic and free of standard models (Behares, 1984b; Elizaincín, 1992). It was suggested that
palatalization is one of the features acquired in a broader process of
urbanization of UP. If this process continues to take place among
groups that are shifting to Spanish in their everyday interactions, and
if this shift affects all social strata in Rivera in the long run, it
is possible that UP will be kept only in the rural areas of Northern
Uruguay, and in towns like Rivera, Portuguese use will be restricted to
interactions with Brazilians from the other side of the border. In that
case, dialect acquisition could be a first step toward dialect shift,
or if generalized, dialect death, resulting from the urbanization of a
rural variety. To confirm this hypothesis, it is essential that trend
or panel studies on both language choice and variation patterns be
carried out.
Moreover, these conclusions would benefit from a detailed network
analysis, which would reveal channels of diffusion masked by the
current grouping of social categories. More data on the way languages
and dialects are negotiated in everyday interactions would shed light
on how these innovations are spreading and what role they play in the
dynamic construction of identities and practices (Eckert, 2000), circumscribing choices along the
bilingual and multidialectal continuum, and contributing to language
change and shift.
This study constitutes a first step toward exploring the output of
this situation, which results from the juxtaposition of languages,
dialects, and sociocultural models. It identifies the uniqueness of the
situation by typifying a border context in which a low variety in a
diglossic community is subject to a change from above. Instead of
“sociolectal reduction” of minority languages encountered
in bilingual societies (Mougeon & Beniak, 1991, 1996; Williams, 1987), UP represents a case of
“sociolectal extension” resulting from constant exposure to
BP. In addition, this work shows that the source of the change is an
outside community, whose presence is mainly experienced through
exposure to television, countering the consensus among sociolinguists
about the inability of the media to affect linguistic variation. The
study of urbanization of UP through quantification of palatalization
offers insight into the analysis of language variation and change in a
borderland environment, revealing unique sociolinguistic aspects of
border communities.
