STUDY OF THE LINGUISTIC VARIATION OF RHOTICS IN CAMPINAS NATIVE ACCENT

Abstracts

This article presents the results of a research study that investigated a variation in a linguistic segment of the Campinas native accent found in the city of Campinas, SP, Brazil: the /R/ in the syllabic coda position. The hypothesis that guided this research is that the rhotic in the Campinas native accent is in a more advanced state in relation to its attenuation than in speech in other towns in the interior of São Paulo state, causing it to be either vocalized or erased. This attenuation would account for the auditory impression of the study’s informants, who believe they utter a variation of /R/ assessed as “intermediate” and which they view as proper to Campinas native accent. The selected corpus comprises data collected from twelve informants from the interior of the state of São Paulo. The data were recorded and analyzed acoustically and statistically. The theoretical foundation adopted was that of Fant’s (1970) Acoustic Theory of Speech Production in conjunction with a sociolinguistic framework. The results show that there is linguistic variation of the rhotic and that the most frequent variety is the caipira/R/. These findings contradict the alleged existence of an /R/ characteristic of Campinas native accent.

Linguistic variation; Rhotics; Acoustic phonetics

Este artigo apresenta resultados de uma pesquisa que buscou investigar a variação de um segmento linguístico do falar campineiro: o /R/ em posição de coda silábica. Como hipótese principal deste estudo, assumiu-se que o rótico produzido pelos informantes campineiros estaria em um estado mais avançado, se comparado a outras cidades do interior paulista, no que se refere ao enfraquecimento desse segmento, tendendo à vocalização ou ao apagamento. Esse enfraquecimento seria o responsável pela impressão, de oitiva, dos informantes que julgam pronunciar uma variante de /R/ avaliada como “intermediária” e indicada como característica do falar campineiro. O corpus selecionado consta de dados coletados junto a doze informantes naturais do interior paulista. Esses dados foram gravados e submetidos à análise acústica e estatística. Para análise dos dados, o referencial teórico adotado foi o da Teoria Acústica de Produção da Fala, conforme Fant (1970), somado aos pressupostos da Sociolinguística. Os resultados alcançados nesta pesquisa mostram que há variação linguística do rótico e que a variante mais frequente é o /R/ caipira. Esse resultado contraria a alegada existência de um /R/ característico do falar campineiro.

Variação linguística; Róticos; Fonética acústica

Initial Considerations

All variations of a language are well structured and form a complex linguistic system that is certainly suitable for the needs of those who use it. However, in spite of this appropriateness, linguistic studies report on the extent to which some linguistic variations are the target of evaluations, both valorative and depreciative. The impact caused by such assessments may contribute to the maintenance or disappearance of linguistic traces; thus, these issues are always current and meaningful to those who study linguistic variations and/or changes. In reference to the Paulista accent (the accent of the inhabitants of the city of São Paulo), it is known that the /R/¹ occurring in the medial and final syllabic coda positions², called the caipira /R/, is largely stereotypical, a fact that has been demonstrated by studies conducted by Amaral (1982), Head (1973, 1978), Leite (2004), and Castro (2006), among others. The pronunciation of the /R/ that characterizes the speech³ of the city of Campinas⁴ (one of the most prosperous cities in the interior of São Paulo state) is defined as “intermediary” by informants from the city of Campinas interviewed in this study.

During individual interviews with residents of Campinas, informants were asked to state their opinions on a possible characterization of the Campinas native accent. They reported that:

LF, M, 26 years old⁵: it’s a middle term between the caipira and the way a person from the city of Sao Paulo speaks / ((laughter)) /.../ it’s intermediate (++), not very long but not very right (+) it is not very:: I don’t know / I can’t explain it right a:: / the way the people from Campinasspeak (+) but it is / it is something you don’t say: a really strong /R/ (+) it is:: really STRONG like:: the speech of:: someone from the interior and also, they don’t pronounce a very SHORT /R/ like someone from the / from the capital (+) I notice that the:: / the: people from the interior of Sao Paulo state like to make the consonants longer and the people from the city of Sao Paulo like to cut them in half

Or:

SA, F, 37 years old: I think it is more from the interior (+) slightly more from the interior (+) I believe it is more influenced by the interior

The city of Campinas is known as the capital of the interior of the state of São Paulo. This nomenclature can be found in articles published in the Correio Popular ⁶ newspaper as, for example, in its special issue celebrating the city’s anniversary. The Metropolitan Region of Campinas is the only Metropolitan Region whose nucleus is not also the capital of the state. Economist Ulysses Semeghini (1991) stated that the city of Campinas functions as an “interface” (between the capital, Sao Paulo, and the interior), unlike other cities in the state of São Paulo. It has the function of anticipating, for the countryside, changes in social and urban patterns typical of industrialization and transformations that would normally appear first in the state capital.

When considering the statements of the informants concerning “intermediate” pronunciation in conjunction with the position that Campinas has in the economic, political, and social scene, it is relevant to ask (in relation to the linguistic aspect discussed here, i.e., the variation of /R/ in the syllabic coda position) if Campinas is at the forefront of other cities in the interior of São Paulo state; that is, is it in a “more advanced stage” with regard to the process of the rhotic’s linguistic variation believed to have occurred, much like what has occurred in other Brazilian capital cities according to a study by Callou, Moraes & Leite (2002). The hypothesis that guided this study, therefore, is that the rhotic occupying the syllabic coda position in the Campinas native accent is in a more advanced stage when compared to other cities in the interior of São Paulo state in relation to the attenuation of this segment, and with a trend toward vocalization or weakening. This attenuation would account for the auditory impression of the informants who believe they utter a variation of the /R/ assessed as “intermediate” and which they view as characteristic of the Campinas native accent.

The Campinas informants interviewed argued that they pronounce an “intermediate” /R/. Is there any basis for this statement? The informants were sensitive enough to capture the rhotic variation and, by stating the above, could they be referring to possible variations of the /R/ in the Campinas native accent? It is also possible to formulate the hypothesis that the natives of Campinas idealize this pronunciation, similarly to what was registered by Trudgill (1974) and by Labov (1982), where informants look up to those who use this form, which then becomes prestigious.

Based on these suppositions, an attempt was made to analyze the process of linguistic variation of the /R/ in the medial and final syllabic coda positions as they occur in Campinas, and to investigate the language attitudes presented by informants with a view toward describing the stereotypes and attitudes related to the linguistic variation in question. As a result of these main objectives: (i) auditory and acoustic analyses, with the purpose of verifying which variants compete in the syllabic coda position of the Campinas native accent, were conducted; and (ii) whether the analyses conducted provide any information that could support an understanding of the reports of the Campinas informants (who stated that they pronounce an “intermediate” /R/ which they distinguish from the accent of the interior of São Paulo state) was verified.

The section below first discusses issues that are relevant to the theoretical framework adopted. Next, a description and an analysis of the data are presented with the purpose of responding to questions concerning the issues that guided this study and to prove or refute the proposed hypotheses.

Rhotics: the caipira /R/ in focus

The study of rhotics, a class of sounds of the archiphoneme /R/, has given rise to many discussions in the field of linguistic investigation⁷. Sociolinguistic studies demonstrate that rhotics show a high degree of polymorphism and are excellent examples of social and regional stratification. On the other hand, phonetic descriptions highlight the great variability of this class of sounds, making it difficult to group them under the same set of characteristics.

Among the possible forms of rhotics in the various speeches present in the interior of São Paulo state, we highlight the occurrence of the caipira /R/ variant – one of the remaining traces of the caipira dialect, and which can be considered the most evident characteristic of this dialect, as well as one of its most stigmatized features. Therefore, it is a constant target of comments, ridicule and denigration – mostly as a result of negative evaluations by communities beyond the speech boundaries of the caipira /R/.

In relation to the possible forms of the rhotics, this study will focus on the variants that compete for the syllabic coda position in the Campinas native accent.

The caipira /R/: noticeable even to the unaware

Amaral (1982), in his work “The Caipira Dialect” (O Dialeto Caipira), describes the dialect of the old province of São Paulo⁸. According to the author, the caipira dialect presented a distinct and unmistakable system and was spoken by the majority of the population. The influence of this way of speaking extended to the educated minority of the capital city and, since then, it has been evaluated pejoratively, to the extent of being singled out as corrupting the language, as well as considered a vice of the language.

In addition to the great number of phonetic, syntactic, and morphological characteristics, as well as the extensive vocabulary listed by Amaral (1982) that specify the caipira dialect when compared to the Portuguese spoken by the educated population in Brazil in early 1900s, there are also behavioral features that distinguish the caipira way of life⁹. Thus, according to the author, “[…] caipirismo was not limited to language, but permeated all the manifestations of daily provincial life.” (AMARAL, 1982, p.41)

Like Amaral (1982), language researchers such as Rodrigues (1974) and Head (1987), among others, point to the pronunciation of the retroflex /R/ as a characteristic of the caipira dialect. The authors’ studies of the caipira dialect describe the articulation of the /R/ typical of the dialect differently. However, as Head (1987) warns, although not all the researchers call it a retroflex /R/, in the example provided by Amaral¹⁰ (1982), the word “retroflex” used to describe the typical rhotic of the caipira dialect appears alongside the “caipira /R/” as synonymous terms for characterizing the typical pronunciation of this dialect.

The retroflex feature of the typical “r” of the caipira dialect, as described by Amaral (1982), requires a more retracted position of the tongue for its production. It is possible to see that the posterior position of the tongue required to produce this retroflex pronunciation is not limited to the production of rhotics, but extends itself to the articulation of other segments, producing retroflex sounds¹¹ and giving a retroflex voice quality to the dialect in question. In this manner, in describing the variants of phonemes in the caipira dialect, Amaral (1982, p.47) points to the presence of the caipira /R/, both in the intervocalic and postvocalic position. Therefore, even in the intervocalic position, belonging to the alveolar tap (which is what occurs in the word arara[aRaRa]), the author registers the occurrence of the caipira /R/.

It is known that many of the features of the caipira dialect, especially its retroflex voice quality, have been lost over time. For Amaral (1982), the development of the population and increased trade and contact of the province of São Paulo with other places, among other causes, have contributed to great changes in the caipira dialect. Thus, if in 1920 Amadeu Amaral had already indicated that the caipira dialect could only be found in small towns, what we can see today is that one of the markers that has remained representative of this dialect is the pronunciation of the caipira /R/¹². This pronunciation is found in the interior of the state of São Paulo and is pointed to as being typical of caipira speech, “characteristic enough to be noticed by the most unaware”, in the words of Amaral (1982, p.41). In addition to the accent found in the interior of the state of São Paulo, the caipira /R/ may also be observed – although with some changes – in the speech of populations residing in other states¹³ such as Minas Gerais, Paraná, Mato Grosso, and Goiás, as stated by Rodrigues¹⁴ (1974).

Social assessment, linguistic insecurity, and stereotypes: triggers for linguistic variation and/or change

Among the many unanswered questions that remain as objects of investigation in linguistic studies, we highlight the issues concerning the mechanism by which languages change, as well as the processes that lead to the great diversity of languages in today’s world (LABOV, 1974). In relation to linguistic variation and change, sociolinguistic studies have shown that, far from being accidental, this process is strongly conditioned by social, stylistic, and evaluative factors, for example.

Regarding such evaluative factors, Labov (1974, p.50) points out that these are complex, as speakers react to speech as a whole and hardly ever show awareness of the speech patterns of others. Despite this difficulty, he asks: “Is there any connection between the performance patterns, the attitudes, and value judgment of speakers?”

This issue is relevant to the purposes of this study. Therefore, the conclusions reached by researchers who take these aspects into consideration will be discussed later in this paper.

When social evaluation and linguistic insecurity trigger variation

The sociolinguistic literature contains a great number of studies portraying the complexity of linguistic systems of languages, and the relevance of such systems to the convenience of its users. As a result of the purposes of this study, three of these studies will be discussed later in this paper. There are three representative studies: two conducted by William Labov (published in 1963¹⁵ and in 1966) and one conducted by Peter Trudgill from 1974.

On Martha’s Vineyard, an isolated island in the state of Massachusetts, Labov (1972) observed that the social changes occurring in that community triggered a linguistic change: the alteration of the position of the first elements of the diphthongs /aj/ e /aw/. Thus, for example, he observed that there were two different pronunciations for the words right and house. One of the non-prestigious pronunciations typical of the island’s natives – [rəɪt], [həʊs] – is a central pronunciation, while another more recent pronunciation – [raɪt], [haʊs] – is similar to the pronunciation found in prestigious variations of American and British English (also called Received Pronunciation: RP). To Labov’s surprise, analysis of the data indicated that there had been an increase in the number of occurrences of the non-prestigious pronunciation. The central pronunciation, therefore, occurred more frequently and even became exaggerated, especially in the speech data of younger informants, characterizing hypercorrection.

Labov’s proposed interpretation of the change in this sound is based on understanding the social forces that affect the life of natives of Martha’s Vineyard. The investigation revealed that there was a social distribution of diphthongs, in such a way that those who wished to stay on the island adopted a more central pronunciation, conservative and non-prestigious, while those who wished to leave and had no close relationship with the islands or its history, adopted the prestigious pronunciation, that is, non-central. Therefore, in order to interpret the centralization of the above-mentioned sounds, Labov proposes the consideration of, among other factors, those issues related mainly to the islanders’ social aspirations and subjective attitudes in relation to life, lack of work, and their reaction to the summer tourists. Thus, he concludes: “In summary, we can then say that the meaning of centralization, judging from the context in which it occurs, is positive orientation towards Martha’s Vineyard.” (LABOV, 1972, p.38).

This study shows how phonetic differences may mean much more than what is captured when only a strictly formal analysis is conducted: in this case, the examination of the structure of formants that characterize the centralization of diphthongs. Language may be a significant factor in identifying a community, in establishing bonds among the members of a group, as well as preserving and setting up boundaries for the spaces of these groups in a situation where they feel threatened by “others”.

Unlike what took place on Martha’s Vineyard, it is the prestigious pronunciation that is desired by New Yorkers, as shown by Labov (1982). However, the prestigious and desired form is not often the one that occurs more frequently in the speech of informants.

When Labov (1982) analyzed the stratification of English in New York City, he showed that linguistic variants are determined by patterns of social and stylistic norms. The post-vocalic /R/¹⁶ at the end of a word and in the pre-consonant position (as in beer, board) is one of the five phonological variations selected. The investigation conducted by Labov (1982) demonstrated that the non-realization of the post-vocalic /R/ indicates casual and low-social-status speech; therefore, it is not the prestigious pronunciation, nor that required for formal styles, because the manner in which rhotics are pronounced is the main manifestation of the pattern of prestige prevalent in New York City.

Patterns of prestige change as a result of socio-historic events. Labov’s study (1982) also provides a good example of an occurrence of a reversion in the patterns of prestige of a specific linguistic variant. According to the author, in the 18th century, the New York City English dialect was characterized by the presence of the /R/ and, in the 19th century, it became an accent without the /R/, probably due to the influence of London’s accent¹⁷. However, since the end of World War II¹⁸, the use of the post-vocalic /R/ has increased in the speech of the upper-middle-class population as a result of the large influx of people coming to New York City, where the standard pronunciation was marked by the realization of the /R/.

The positive evaluation given to the pronunciation of /R/ caused, therefore, an increase in the use of this variant among upper-middle-class speakers, according to Labov (1982). In this case, the change occurred in the direction of the norm – the prestigious form – which, as we know, is not always the direction of change.

The examination of the subjective reactions of speakers in relation to the pronunciation or attenuation of the final and pre-consonant /R/ (car, card, etc.) shows that these reactions are more uniform than performance. Data analysis demonstrated that all social groups agreed that the pronunciation of a constrictive /R/ in words such as car, for example, was proper in more formal contexts. However, the great majority of New Yorkers did not pronounce the /R/, so only the upper middle-class speakers showed any degree of pronunciation of rhotics in casual speech. Findings like these have led to the following interpretation:

New Yorkers also showed a systematic tendency to report their own speech inaccurately. Most of the respondents seemed to perceive their own speech in terms of the norms at which they were aiming rather than the sound actually produced. (LABOV, 1982, p.336).

Due to such results (which were repeated in Trudgill’s 1974 study), Labov (1982) proposed that the variations found in speech communities should be understood to result from variations in subjective evaluations and not as fluctuations in speakers’ performances.

Trudgill (1974) conducted a study in the city of Norwich, U.K., in which one of the phonological variants investigated was the presence or absence of the glide [j] in words such as tune, music, among others, focusing on two co-existing variants in the city: /ju:/ and /u:/, with /tju:n/ being seen as more prestigious than /tu:n/. During data recording, he was able to observe which variant was more pronounced among the interviewees and, using a self-evaluation test, he verified what the opinion of informants was with regard to their own pronunciation. When he compared the results of the tests with those obtained through the recordings, he noticed that there were discrepancies between the variant produced and the one the informant believed was produced. The results indicated that, among the informants that did not use the glide, 16% argued that they had pronounced this segment, while 84% stated they had not. On the other hand, among the informants who actually did pronounce the glide, 60% said they had, while 40% believed they had not. Based on these results, he concluded that 16% of those who pronounced the “discredited” form tended to over-evaluate their own pronunciation, while 40% of the interviewees who presented the more “prestigious” pronunciation, tended to under-evaluate their pronunciation. With this, Trudgill (2000, p.76) concluded that: “Speakers, that is, report themselves as using the form at which they are aiming and which has favourable connotations for them, rather than the form they actually use. (No conscious deceit is involved, it seems.).”

These evaluations (which correspond not to the linguistic aspects delivered but rather to the norms to which each speaker aspires) show great linguistic insecurity. According to Labov (1972), in addition to the inaccurate perception of one’s own pronunciation, indications of linguistic insecurity are identified when there are wide fluctuations in stylistic variation and when there are hyper-sensitivity and stigmatized traces on the part of speakers. The excerpt below illustrates a situation of profound linguistic insecurity observed in New York City:

In general, New Yorkers show a strong dislike for the sound of New York City speech. Most have tried to change their speech in one way or another, and would be sincerely complimented to be told that they do not sound like New Yorkers. Nevertheless, most of the respondents have been identified by their speech as New Yorkers whenever they set foot outside of the metropolitan area. They firmly believe that outsiders do not like New York City speech, for one reason or another. Most New Yorkers show a strong belief in correctness of speech, and they strive consciously to achieve such correctness in their careful conversation. (LABOV, 1972, p.132).

Changes in the evaluation of speakers may be the cause and not the effect of change in patterns of speech. Thus, linguistic aspects evaluated positively may be evidenced, imitated, and adopted as standard forms by members of the language community, for example. It is necessary to be attentive to these issues when attempting to understand and not merely diagnose a linguistic variation.

Linguistic stereotypes: another phase in the evaluation of linguistic variables

Not all linguistic variables involved in variations and/or linguistic change in a speech community are prestigious and not all of them are evaluated in the same manner. According to Labov (1974, 1972), information obtained from social evaluations of these linguistic variables may be employed to place them in one of three categories: indicators, markers, and stereotypes.

Indicators are linguistic traces that reflect a social variation (age, social group) but generally do not show stylistic variation and have little effect on the listener’s judgment of the social status of the speaker. Markers are traces that show both social and stylistic variation and have a consistent effect on the conscious or subconscious judgment of the listener of the speaker’s status. Stereotypes are external topics that have a social impact on the speech community. They may be socially labeled and may or may not correspond to actual linguistic behavior.

Although stereotyped linguistic traces are stigmatized, they may be very durable and persistent. Labov (1972) points out that dissemination of these traces may occur towards numerous meanings and requires a considerable period of time – and this allows for numerous social changes to occur during that period of time, which may boost or deter such dissemination. In the event there is a strong social reaction against these linguistic traces, this reaction may trigger a fast attenuation process and, consequently, cause them to disappear. On the other hand, if the group or speech community using such traces begins to be evaluated positively and, therefore, obtains social standing, a movement in the other direction may occur, and the linguistic trace that had once been stigmatized may become the target of positive evaluations and even be imitated. With regard to the propagation of these traces in the linguistic system, the author states that:

As the original change acquires greater complexity, scope, and range, it comes to acquire more systematic social value, and is restrained or corrected in formal speech (amarker). Eventually, it may be labeled as a stereotype, discussed and remarked by everyone. The future prospects of this stereotype depend upon the fortunes of the group it is associated with. If the group moves into the mainstream of society, and is given respect and prominence, then the new rule may not be corrected but incorporated into the dominant dialect at the expense of the older form. If the group is excluded from the mainstream of society, or its prestige declines, the linguistic form or rule will be stigmatized, corrected, and even extinguished. (LABOV, 1972, p.320).

Social evaluations of the caipira /R/ allow for the classification of the variable as belonging to the stereotype category. This is the last level in the evaluation scale proposed by Labov (1972).

The following examples present the broad stereotyping of the rhotic in the accent of speakers from the interior of the state of São Paulo.

Stigma and stereotypes around the caipira dialect, the caipira /R/, and the “intermediate” /R/

The caipira/R/ is widely stereotyped in the Paulista dialect, as demonstrated in the studies conducted by Amaral (1982), Head (1973, 1978), Leite (2004, 2010), and Castro (2006), among others.

By 1920, Amadeu Amaral had highlighted the pejorative manner in which the caipira dialect was viewed and, consequently, the linguistic aspects typical of this dialect, among which the /R/ stands out and which he called “caipira.” The caipira dialect, considered distinct and unmistakable, was stigmatized and viewed as non-standard language. Because it was considered non-standard¹⁹, people were concerned with the possible diffusion of this linguistic trace, something that had to be avoided. They felt that, under its influence, the speech of even the most educated might be corrupted. Evidence of this concern can be found in Amaral’s statement (1920) below, wherein he reports that there were concerns regarding the establishment of law schools in the province of São Paulo due to the influence of the caipira dialect. According to Amaral:

For some time now, this is what has caused, for the people of the interior of the province of São Paulo, the fame of tainting the vernacular speech with numerous and ugly corruptions of the language. When the Imperial Senate discussed the creation of law schools in Brazil and considered São Paulo as one of the locations, there were those were against it who argued that the language of the locals would inconveniently contaminate future graduates, who would come from various regions of the country [...] (AMARAL, 1982, p.41).

Studies conducted by Head (1973, 1978) discuss the stigmatization of the caipira /R/. In the first study, the author stated that the decrease in the relative frequency of occurrences of this segment was related to an increase in the degree of reflexion. Thus, in the style of monitored speech among informants of the same socio-economic class and age group, the number of realizations of rhotics was smaller, a fact that confirms the stigmatized nature of this pronunciation. In the second study, these findings were confirmed, and the following considerations were added: (i) the informants from urban areas presented a greater decrease in the frequency of the stigmatized variant when compared to rural informants; (ii) the variation in the frequency of the /R/ is greater in cities than in rural areas; and (iii) the informants from urban areas who are in higher economic classes are more sensitive in relation to the stigmatized nature of the caipira /R/ when compared to informants of lower classes and/or rural areas.

Leite (2004) also demonstrates the wide stigmatization and stereotyping of the /R/ of the caipira accent, both in data collected from inhabitants of the city of São José do Rio Preto, SP, and in television advertisements. In this study, the above-mentioned segment is always negatively labeled: ugly, marked, and dragged, as well as being called a strong accent.

In addition to these denominations given to rhotics, the data collected via surveys show a less marked pronunciation, called “intermediate.” Is this just another stereotyped belief or is there actually a variation in progress that has been noticed by informants more sensitive to linguistic issues?

The social psychology studies selected by Hewstone and Giles (1997) show that stereotypes are more easily noticed, stored in memory, and activated when compared with evidence that contradicts them. They also generate expectations, and those who perceive them want these expectations to be confirmed. This is why they state that:

People tend to see behaviour that confirms their expectancies, even when it is absent. When stereotypes set up expectations of behaviour, disconfirming evidence tends to be ignored, but confirming evidence remembered. (HEWSTONE & GILES, 1997, p.276).

These features of stereotypes compound one another: “stereotypes become self-fulfilling prophecies,” as Snyder, Tanke & Berscheid (1977)²⁰ state. According to Watzlawick’s definition (1994, p.97), “[…] a self-fulfilling prophecy is a supposition or prediction that, just because it was made, converts the supposed event into reality, expected or prophesied, confirming, therefore, its own ‘accuracy’.” Thus, this is about an initially false statement that evokes a new behavior capable of rendering the initial formulation true.

Stereotypes present self-justification and self-perpetuation dynamics that cause those who are the objects of stereotyping to behave in ways that correspond to the stereotyped image they project. These dynamics are attested in several studies conducted in the field of human sciences, especially social psychology, such as the study by Snyder, Tanke and Berscheid (1977). These studies also show that the traditional line of thought of cause and effects not obeyed²¹ when the issue is a self-fulfilling prophecy. In relation to this issue, Watzlawick (1994) states that:

[...] an act deriving from a self-fulfilling prophecy first creates the conditions for the expected event to occur and, in this sense, creates a reality that would not have been produced without it [the act]. This act, therefore, is neither true nor false: it merely creates a situation and, with it, its own “truth”. (WATZLAWICK, 1994, p.98–99).

The studies described above, as well as those cited by Watzlawick (1994) and Hewstone and Giles (1997), for example, indicate how individuals put stereotypes into action and how these stereotypes may become self-fulfilling prophecies. However, only a few prophecies self-fulfill, explains Watzlawick (1994). According to the author, only when the faith in a prophecy is such that the person believes it will occur in the “future”, will it be capable of influencing the present and, therefore, of confirming itself. In the words of the author: “Absent the factor of belief – this conviction factor – the effect is also absent.” (WATZLAWICK, 1994, p.101)

Considering that stereotypes have an important role in language maintenance, variation and change, and that the caipira /R/ is clearly stereotyped, I attempt to verify which variants of rhotics occur in the syllabic coda position in the accent of the people from Campinas who were interviewed, and investigate the value judgment they give to these variants. Is stigma attributed to the caipira /R/ a determinant capable of triggering the variation and, consequently, causing the weakening of this segment? In relation to the stereotyped image of the “intermediate” pronunciation of the Campinas native accent, will it be confirmed?

Before proceeding with data analysis, the next section provides a description of the methodology employed in this study.

Characterization and acoustic description of the data

Ladefoged and Maddieson (1996) state that the similarities among the rhotics may be more of an acoustic and auditory nature than articulatory. They add that it is not the point of articulation that defines the class of rhotics because segments of different articulations, such as dorsal and coronal, for example, are grouped as rhotics. Therefore, they clarify: “The most important evidence that they belong to a single class, at least from a phonological point of view, is the fact that rhotics of one type often alternate with other rhotics.” (LADEFOGED; MADDIESON, 1996, p.216).

Rhotics and vowels are similar, according to Ladefoged and Maddieson (1996), because they present syllabic variants or because they merge (co-articulate) with contiguous vowels. The authors also present other evidence of this similarity when they point out that in Germanic languages (Danish and Swedish) the vowels followed by rhotics tend to be long as well as acquire the “colored” quality of the following rhotic, presenting, therefore, acoustic modifications. In the end, they conclude:

Thus, the rhotics form a heterogeneous group from the phonetic point of view, exhibiting a wide variety of manners and places of articulation. We find rhotics that are fricatives, trills, taps, approximants, and even ‘r-colored’ vowels, as well as articulations that combine features of several of these categories. The most common places of articulations are in the dental-alveolar area, although post-alveolar (retroflex) /r/’s are not unusual, and in some languages /r/’s have a uvular articulation. (LADEFOGED; MADDIESON, 1996, p.217).

The r-colored vowels mentioned by Ladefoged and Maddieson (1996) are observed by Ladefoged (2001) in American English. This author states that these vowels invoke a traditional trace called rhotacization, since the traces – high/low, front/back, and rounded/unrounded – are not enough to describe them. Ladefoged (2001) defines a rhotacized vowel as:

Rhotacization is an auditory quality, which, like height and backness, is most appropriately defined in acoustic terms. In a rhotacized vowel (or portion of a vowel) there is a marked lowering of the frequency of the third formant. (LADEFOGED, 2001, p. 212).

Rhotics form a class of sounds with a great number of phonetic differences and present formant frequencies²² related to the points of constriction in the vocal tract. Lindau (1980a) discusses some of the phonetic differences among several Nigerian languages, addressing the classes of rhotics, implosives, and vowels. In her description of the r sounds, she states that the first and second formants appear to reflect the quality of the vowel that comes before or after this segment, while a third and fourth formant are important indicators of the point of constriction in this type of sound. In relation to this description, the author draws attention to the acoustic correlate that occurs when there is a constriction in the posterior-most or anterior-most regions of the vocal tract:

According to acoustic theory, a lowered third formant, close to the second formant, indicates a constriction fairly far back in the postalveolar-midpalatal region with strong retroflexion. This happens in Izon. As the constriction moves forward in the mouth the third formant increases. A relatively high third formant, close to the fourth formant, indicates a dental place, as happens in the Kalabari example. (LINDAU, 1980a, p.107).

Lindau’s study (1980a) indicates that the lowering of F₃ is characteristic only of rhotics produced in the postalveolar region presenting, therefore, strong retroflexion. On the other hand, rhotics produced in a more anterior region of the vocal tract, in relation to the mid palate area, present an elevation of F₃ and not a lowering.

According to Linday (1980b), in English and Spanish it is possible to verify, acoustically, a lowering of F₃ in the realization of rhotics²³. However, the lowering of F₃ is not a feature of rhotics in other languages. According to the author, in the realization of tap and approximant related to the pronunciation of speakers of Degema, a language originating in Nigeria, it was not the lowering of F₃ that was observed, but rather an elevation, so that the third formant presented itself closer to the fourth formant.

Through these studies, Lindau (1980b) points out that the lowering of the third formant may be obtained by retroflexion and by a constriction in the post-alveolar palatal regions. This lowering may also be produced with the tip of the tongue lowered and by a constriction of the pharynx. The studies also confirm the fact that the rounding of the lips contributes to the lowering of F₃.

In summary, Lindau’s research (1980a, 1980b) indicates that it is the family or similarity relationships – more than the acoustic characteristic – that allows the classification of the class of rhotics. Thus, trills and taps are similar in regard to closure duration; apical and uvular trills are similar in relation to the fast and consecutive pattern of the trills, apical taps; and approximants present similarities in relation to the articulatory scale of closure and possibly in relation to the position of the third format; while uvular and fricative trills present a similar pattern of formants. In her words:

Thus there is no physical property that constitutes the “essence” of all rhotics. Instead, each member of the class of r-sounds resembles some other member with respect to some property, but not with respect to the same property across all r-sounds. (LINDAU, 1980b, p.118).

Similarities among rhotics are due much more to acoustic and auditory reasons than to articulatory reasons because there is a broad range of modes and points of articulation that involve these segments.

Considering the acoustic parameters described by the theoreticians mentioned above, the methodological criteria adopted and the results are described, and part of the data that forms the body of this work is discussed.

Methodological procedure

The corpus of the study that originated this paper is composed of data collected from 12 informants that were inhabitants of the city of Campinas. Social variables considered for the selection of informants were: gender, age group, and level of education, as detailed below:

1. Gender: male, female;

2. Age group: (1) 20 to 30 years of age, (2) 37 to 47 years of age, (3) above 54 years of age. An interval of 7 years was defined between age groups;

3. Level of education: secondary education (S), or higher education (H) completed or in progress.

A summary of the characteristics of informants are presented in Table 1 below:

Table 1
– Description of informants by the social variables selected

Informant’s Initials	Gender	Age Group/ Age (years)	Level of Education
MC	F	1 / 23	S
GP	M	1 / 20	S
CL	F	2/ 47	S
JC	M	2 / 46	S
LH	F	3 / 54	S
JP	M	3 / 56	S
TG	F	1 / 20	H
OE	M	1 / 29	H
LB	F	2 / 37	H
EG	M	2 / 37	H
JB	F	3 / 56	H
AL	M	3 / 68	H

• Source: Made by the author.

In this paper, the data of 9 of the 12 informants²⁴ are discussed. Also, only data reflecting the general pattern of configuration of the variants found will be considered. Therefore, differences due to the social variables cited above will not be explored²⁵.

The experimental design, formed by real words and pseudo-words, was developed with the purpose of obtaining the same contexts for the CVR sequence (consonant, vowel, rhotic) in the medial /´tVR.CV/ and final /CV.´tVR/ coda positions. The words selected, shown in Table 2 below, were inserted in a vehicle-sentence (“say ______ to her”) and in sentences. The words and sentences were read by each informant and repeated 3 times.

Table 2
– Words and pseudo-words selected as part of the experimental design

Vowels	Medial Coda /´tVR.CV/	Final Coda /CV.´tVR/	Vowel in CV Syllable
/a/	Tarto	Qatar26	Pato
/e/	Terto	Bater	Beco
/i/	Tirto	Tatir	Tipo
/o/	Torto	Tutor	Toco
/u/	Turta	Tutur	Tuco

• Source: Made by the author.

Data was recorded in an acoustically treated room in the Recording Studio of the Institute of Language Studies at the University of Campinas (IEL, Unicamp). Data were captured by a microphone (AKG: model 420 PP headset) connected to a pre-amplification board (M-Audio: model MobilePre USB). This board was connected to a notebook computer (Toshiba: latest generation model SatelliteM45). For the recording, Audacity, a free, open-source, audio editor and recording program (22,050-Hz sample rate, 24-bit resolution; digitalized in pulse modulation code: PCM) was used.

The data were transcribed²⁷, manually segmented, and analyzed with the aid of Praat software. The parameters adopted for the acoustic analysis were based on the frequency of the first three formants of the VR sequence (in three positions: IP, MP and FP²⁸). The frequencies of the first three formants were verified based on Fast Fourier Transformation (FFT) algorithms and Linear Predictive Coding (LPC), overlapped. The values obtained via these analyses were submitted to acoustic analysis. Results of the acoustic analyses are presented below.

Results and discussion

The most frequent variant in the Campinas native accent is the caipira /R/. This variant is stigmatized and negatively evaluated by most of the informants interviewed. Acoustic analysis of the 540 repetitions²⁹ of the /R/ in the medial and final syllabic coda positions revealed a predominance of the caipira /R/ variant, both in the pronunciation of the informants from age group 1 and those from age group 3. The caipira /R/ presents a low F3 (whose average is 2,096 Hz in the medial coda position and 2,070 Hz in the final coda position) in the medial or the final position, even when it occurs after anterior vowels. As an example of the realization of this variation, see spectrograms 1 and 2 below³⁰. In the first spectrogram (where the realization of two words are concatenated: torto and tutor), it is possible to verify the occurrence of the variant after posterior vowels while the second spectrogram (where there are two concatenated words: terto and tutor) shows the realization of the caipira /R/ after an anterior and a posterior vowel.

Figure 1
– Informant TG’s spectrogram 1 (concatenated torto and tutor).

Figure 2
– Informant GP’s spectrogram 1 (concatenated terto and tutor).

In addition to the caipira /R/, the occurrence of a rhotacized vowel variant was observed. Among the features of rhotacized vowels listed by Ladefoged and Maddieson (1996) and Ladefoged (2001), only the first was present in the data analyzed: the low F3³¹. The second characteristic (corresponding to the quality of rhotacization present at the beginning of the vowel) was not observed in all of the occurrences classified in this study as rhotacized vowels because it is the quality of the vowel rather than the rhotic that prevails. Figure 3 below shows the occurrence of this variant.

Figure 3
– Informant LB’s spectrogram 1 (concatenated tirto and tatir)

In the statistical analysis of the means of the formant frequencies of the caipiraVRt (rhotacizedVRt) and the vowel in the CV syllable, the Kruskal–Wallis test indicated a statistically significant difference in the rank sum of F1 and F3. Because it is a trajectory, it is the final position that we are most interested in observing. The mean F1 frequency in the final positions found for VRt is located between the means found for the caipira /R/ and those corresponding to the vowel in the CV syllable. In relation to F3 (a formant that better characterizes the caipira /R/), the analysis showed statistically significant differences among the variations that were compared (which corresponded to the F3 of the caipira /R/, the vowel in CV syllable, and the VRt, or rhotacized vowel). The lowest mean values of the third formant frequency were those corresponding to the final position of the caipira /R/ variant, followed by the final position of VRt and, finally, the vowel in the CV syllable. If, auditorally, the quality of the rhotics was not noticed at the beginning of the trajectory, the results obtained with the F3 frequency means show that, starting at the initial position, the VRt frequency is lower than that found for the vowel in the CV syllable. It is also possible to verify that the VRtfrequency means, in all positions, are higher than the means for the caipira /R/ variant. Figure 4 illustrates these results.

Figure 4
– F3 frequency means for the initial position (IP), medial position (MP), and final position (FP) for the vowel in the CV syllable, VRt, and caipira /R/

The findings of this study show that there is a linguistic variation of rhotics and that the most frequent variant is the caipira /R/. These results contradict the alleged existence of a typical Campinas /R/.

Acoustic analysis of the data showed that it is the caipira /R/ that is produced by people from Campinas, even when they try to establish distinctions between the different forms of pronunciation of the rhotic. Thus, there is a contradiction between the statements made by most informants and their pronunciation of the rhotic. It may be said that this contradiction is only apparent because speakers usually refer to linguistic forms that they consider socially prestigious and not to those that are actually part of their linguistic repertoire. Informants from the city of Campinas view themselves as producing the pronunciation that they consider more prestigious. This finding confirms the second hypothesis of this study.

Despite the stereotyping of the caipira /R/ of the accent in the interior of the state of Sao Paulo, this pronunciation is kept active in the Campinas native accent. We must remember that stereotypes have great influence in relation to linguistic maintenance or change and that some of their main features are persistence, rigidity, and resistance to change. Therefore, stereotypes related to the caipira /R/ tend to continue to be active. However, we also need to remember that another stereotyped belief is found in the informants’ reports – one which attributes an intermediate pronunciation to the Campinas native accent.

Stereotyped opinions also have an important role in society because they protect people from seeing reality whenever this reality is disturbing or disconcerting, according to Lippmann (2008). Thus, it may be argued that this belief in the existence of a soft, intermediate, or less marked pronunciation functions as a shield, protecting those who find shelter behind it. It is not a coincidence that this stereotype is sustained because its maintenance functions as a defense for a position that one wishes to ensure in society; in this case, for the majority of the people from Campinas, it is a way to establish the boundaries between Campinas – the capital of the interior – and other cities in the interior of São Paulo state. The issues relevant to this topic are closely related to the linguistic aspect discussed here and will be the object of future research.

Final remarks

In their reports, informants stated that the Campinas /R/ is distinct from the rhotic characteristic of the accent in the interior of Sao Paulo state (i.e., the caipira /R/) because it is softer or “intermediate.” However, acoustic analysis of the data has shown that it is the caipira /R/ they produce, even when they attempt to establish a distinction between the forms of pronunciation of rhotics. Therefore, there is a contradiction between the statement found in most reports and the informants’ pronunciation of the rhotic. It may be said that this contradiction is only apparent because speakers usually refer to linguistic forms they consider socially prestigious and not to those that are actually part of their linguistic repertoire. The informants from the city of Campinas view themselves as producing the pronunciation they consider prestigious. This result confirms the second hypothesis of this study.

REFERÊNCIAS

• ABAURRE, M. B. M.; SANDALO, M. F. S. Os róticos revisitados. In: DA HORA, D.; COLLISHONN, G. (Org.). Teoria linguística: fonologia e outros temas. João Pessoa: Editora Universitária: UFPB, 2003. p.144-180.
• AMARAL, A. O dialeto caipira: gramática, vocabulário. 4.ed. São Paulo: HUCITEC, 1982.
• BONET, E.; MASCARÓ, J. On the representation of contrasting rhotics. Universidade Autônoma de Barcelona, 1996.
• BRANDÃO, S. F. Nas trilhas do –r retroflexo. Signum: Estudos da Linguagem, Londrina, v.10, n.2, p.265-283, 2007.
• BRANDÃO, S. F. A geografia linguística no Brasil. São Paulo: Ática, 1991.
• CAGLIARI, L. C. Elementos de fonética do português brasileiro. 1981. 185f. Tese (Livre Docência) – Instituto de Estudos da Linguagem, Universidade Estadual de Campinas, Campinas, 1981.
• CALLOU, D.; MORAES, J. A.; LEITE, Y. Processo(s) de enfraquecimento consonantal no português do Brasil. In: ABAURRE, M. B. M.; RODRIGUES, A. C. S. (Org.). Gramática do português falado. v.VIII. Campinas: Ed. da UNICAMP: FAPESP, 2002. p.537-555.
• CANO, W.; BRANDÃO, C. A. (Coord.). A região metropolitana de Campinas: urbanização, economia, finanças e meio ambiente. v.2. Campinas: Ed. da Unicamp, 2002.
• CASTRO, V. S. A resistência de traços do dialeto caipira: estudo com base em atlas linguísticos regionais brasileiros. 2006. 285f. Tese (Doutorado em Linguística) – Instituto de Estudos da Linguagem, Universidade Estadual de Campinas, Campinas, 2006.
• CUNHA, C. Língua portuguesa e realidade brasileira. 3.ed. Rio de Janeiro: Tempo Brasileiro, 1972.
• DUBOIS, J. et. al. Dicionário de linguística. São Paulo: Cultrix, 1998.
• FANT, G. Acoustic theory of speech production. 2.ed. The Hague; Paris: Mouton, 1970.
• FERREIRA, C.; CARDOSO, S. A. A dialetologia no Brasil. São Paulo: Contexto, 1994.
• HARRIS, J. Flaps, trills, and syllable structure in Spanish. MIT Working Papers in Linguistics, Cambridge, MA, v.42, p.81-108, 2002.
• HARRIS, J. Spanish phonology. Cambridge, MA: The MIT Press, 1969.
• HEAD, B. F. Propriedades fonéticas e generalidades de processos fonológicos: o caso do “r caipira”. Cadernos de Estudos Linguísticos, Campinas, v.13, p.5-39, 1987.
• HEAD, B. F. Subsídios do atlas prévio dos falares baianos para o estudo de uma variante dialetal controvertida. Cadernos de Estudos Linguísticos, Campinas, v.1, p.21-34, 1978.
• HEAD, B. F. O estudo do “r-caipira” no contexto social. Revista de Cultura Vozes, [s.l.], v.67, n.8, p.43-49, 1973.
• HEWSTONE, G.; GILES, H. Social groups and social stereotypes. In: COUPLAND, N.; JAWORSKI, A. (Org.). Sociolinguistics: a reader. New York: St. Martin’s Press, 1997. p.271-283.
• LABOV, W. The stratification of English in the New York City. Washington, D.C.: Center for Applied Linguistics, 1982.
• LABOV, W. Estágios na aquisição do inglês standard. In: FONSECA, M. S.; NEVES, M. F. (Org.). Sociolinguística, Rio de Janeiro: Eldorado, 1974. p.49-85.
• LABOV, W. Sociolinguistic patterns. Philadelphia: Pennsylvania University Press, 1972.
• LADEFOGED, P. A course in phonetics. 4.ed. New York: Harcourt College Publishers, 2001.
• LADEFOGED, P.; MADDIESON, I. Rhotics. In: LADEFOGED, P.; MADDIESON, I. The sounds of the world’s languages. Oxford: Blackwell Publishers, 1996. p.215-245.
• LEITE, C. M. B. O /R/ em coda silábica no falar campineiro. 2010. 225f. Tese (Doutorado em Linguística) – Instituto de Estudos da Linguagem, Universidade Estadual de Campinas, Campinas, 2010.
• LEITE, C. M. B. Atitudes linguísticas: a variante retroflexa em foco. 2004. 149f. Dissertação (Mestrado em Linguística) – Instituto de Estudos da Linguagem, Universidade Estadual de Campinas, Campinas, 2004.
• LINDAU, M. Phonetic differences in Nigerian languages. UCLA Working Papers in Phonetics, Los Angeles, n.51, p.105-113, 1980a.
• LINDAU, M. The Story of /r/. UCLA Working Papers in Phonetics, Los Angeles, n.51, p.114-119, 1980b.
• LIPPMANN, W. Opinião pública. Trad. J. A. Wainberg. Petrópolis: Vozes, 2008.
• MARCUSCHI, L. A. Análise da conversação. São Paulo: Ática, 1986. (Série Princípios).
• MELO, G. C. A língua do Brasil. Rio de Janeiro: Livraria Agir Editora, 1946.
• RODRIGUES, A. N. O dialeto caipira na região de Piracicaba. São Paulo: Ática, 1974.
• ROSSI, N. et al. Atlas prévio dos falares baianos. Rio de Janeiro: Instituto Nacional do Livro, 1963.
• SELKIRK, E. The Syllable. In: HULST, H.; SMITH, N. (Ed.). The structure phonological of representations: part II. Dordrecht: Foris, 1982. p.337-383.
• SEMEGHINI, U. C. Do café à indústria: uma cidade e seu tempo. Campinas: Ed. da UNICAMP, 1991.
• SNYDER, M.; TANKE, E. D.; BERSCHEID, E. Social perception and interpersonal behavior: on the self-fulfilling nature of social stereotypes. Journal of Personality and Social Psychology, Washington, v.35, n.9, p.656-666, 1977.
• TRUDGILL, P. Sociolinguistics: an introduction to language and society. 4.ed. London: Penguin Books, 2000.
• TRUDGILL, P. The social differentiation of English in Norwich. London: Cambridge University Press, 1974.
• WATZLAWICK, P. Profecias que se autocumprem. In: WATZLAWICK, P. (Org.). A realidade inventada: como sabemos o que cremos saber? Trad. J. P. Santos. Campinas: Editorial Psy II, 1994. p.95-116.

• 1
  As a function of the variability of the rhotic sounds, it is difficult to quantify exactly what the rhotic phonemes in intervocalic positions are and what the loss of contrast is between these sounds in syllabic coda positions, as pointed out by phonological studies; thus,the archiphoneme /R/ will be adopted in this text to represent this neutralization.
• 2
  According to Selkirk (1982), a syllable consists of an onset (O) and a rhyme (R); the rhyme, in turn, consists of a nucleus (Nu) and a coda (Co). Any category, except Nu, may be empty.
• 3
  For Dubois et al. (1998, p.266): “Speech is a form of language used in a specific social group or as a sign of belonging or wishing to belong to this social group. Each one of these speeches hasits own specific syntactic rules and vocabulary that are unique and many others that are common with other dialects of that language or even of all of them.” In this paper, the term “Campinas native accent” is used to refer to the variation of the Brazilian Portuguese language spoken in the city of Campinas.
• 4
  The Metropolitan Region of Campinas (MRC) comprises 19 countries (the nucleus of which is the municipality of Campinas, located in the Brazilian state of São Paulo). The MRC has an area of 3,673 km²and has 2.3 million inhabitants. According to Cano and Brandão (2002, p.403), among the 12 Metropolitan Regions (MRs) listed in the study “Characterizations and Trends of Urban Networks in Brazil”(Caracterização e tendências da rede urbana do Brasil, IPEA, IBGE, NESUR), Campinas is the only MR whose nucleus is not also the state’s capital.
• 5
  Informants’ identities were kept confidential and, consequently, they are identified only by their initials, followed by their gender (M, male; F, female) and age at the time of the interview.
• 6
  Correio Popular, July 14, 2003, p. 19.
• 7
  As discussed in Leite (2004), phonological studies also give rise to discussions of the class of sounds of rhotics because there is no consensus on the quantity of rhotic phonemes. In relation to this issue, read or consult Harris (1969, 2002), Bonet and Mascaró (1996), and Abaurre and Sandalo (2003), among others.
• 8
  In the preface of “The Caipira Dialect” (O Dialeto Caipira), Paulo Duarte indicates that the informants in Amaral’s (1982) study came from the regions of Capivari, Piracicaba, Tietê, Itu, Sorocaba, and São Carlos.
• 9
  The genuine caipiras are described by Amaral (1982) as ignorant and backward small farmers. Therefore, with this denomination, Amaral (1982) attempts to characterize the speech of those inhabitants in the interior of São Paulo state who have simple, rural habits.
• 10
  This is how Amaral (1982, p.47) describes the /R/, one of the variants typical of the caipira dialect: “the inter- and postvocalic r (arara, carta) has a peculiar value: it is lingual-palatal and gutteralized. During its prolation, instead of projecting the tip of the tongue against the upper teeth, a movement that produces the Portuguese modality, the sides of the tongue are raised in the direction of the small molars of the maxillary arch and its tip is turned upwards, without touching the hard palate. There is almost no trilling vibration. For the ear, this caipira r sounds very similar to the postvocalic English r.” Although it is not called retroflex, the description presented, according to Head (1987), is clearly dealing with the retroflex aspect of articulation. Thus, the production of the caipira /r/ may occur via contact between the tip of the tongue (turned upward) and the palate or through a posterior articulation made by raising the tongue blade (dorsum), that is, gutteralized pronunciation. In relation to this description by Amaral (1982), Head (1987) adds that, despite the difference between the realization of a retroflex consonant (lingual-palatal) and a guttural consonant, it is possible to point out similarities between these articulations because “[…] both represent production processes in a posterior direction, with realizations more retracted than they would be without the turning up of the tip of the tongue or without raising the dorsum.” (Head, 1987, p.10)
• 11
  Cagliari (1981, p.43) states that we can find retroflex sounds in the accent found in the interior of the state of Sao Paulo and especially in the caipira dialect. In his words: “In the caipira dialect, in addition to constrictive sounds, it is not unusual to also find occlusive, nasal and retroflex lateral sounds.”
• 12
  Perhaps examples of the caipira /R/ may still be found – particularly in the lingual-palatal articulation, as described by Amaral (1982) – as well as the retroflex position that characterized the accent of the original speakers of this dialect, in the interior of São Paulo state, as part of the idiolect of older speakers. Therefore, it is not our intention to state that the retroflex variation (which today can be pointed to as representative of the accent found in the interior of the state of Sao Paulo) has the same features as the variation described by Amaral (1982).
• 13
  The fact that speakers in different regions of the country produce the caipira r does not mean that these speakers are representatives of the dialect in question. What can be observed is that only one of the caipira dialect’s features (the realization of the retroflex /r/) can still be found as a variant in some dialects.
• 14
  In relation to the origin, the geographical distribution, and the vitality of the pronunciation of the caipira/R/, it can be said that there is no consensus among authors. For an in-depth investigation of this topic, it is worth reviewing the works of Amaral (1982), Melo (1946), Rossi (1963), Cunha (1972), Rodrigues (1974), Head (1987), Brandão (1991, 2007), and Ferreira and Cardoso (1994), among others.
• 15
  The study conducted on Martha’s Vineyard was initially published in 1963. In the present paper, references to the 1963 study will be made based on the version of the research published in Labov (1972).
• 16
  The following were excluded: cases in which the /R/ at the end of a word is followed by a word that begins with a vowel, as well as the occurrences in which the rhotic is preceded by a mid-central vowel, such as in her and bird.
• 17
  In England, the accents in which the /R/ is not pronounced are more prestigious and are seen as more “correct” than those where the rhotic is pronounced. The prestigious pronunciation is adopted by the media and is considered Standard English pronunciation, also called RP. On the other hand, the pronunciation of the post-vocalic /R/ is adopted on television and in theater for denoting rural characters with low levels of education and also for the purpose of denoting humorous characters. In some American English accents (as well as in some from Central America, England, New Zealand, Australia, and South Africa), the post-vocalic /R/ is not pronounced, either, according to Trudgill (2000).
• 18
  According to Labov (1972, p.64), before World War II, the anglophile tradition predominated in New York’s schools. As a result, children were taught that pronunciation of the /R/ was provincial and, therefore, the proper pronunciation of car, for example, was without the /R/.
• 19
  Amaral (1982, p.41) cites author Joaquim M. de Macedo where Macedo states that Judge José Antonio Pimenta Bueno, the Marquis of São Vicente (Santos; December 4, 1803 – February 19, 1878) had “unpleasant tainted pronunciation” due to “being sloppy and poorly mannered” and not through any defect of his speech organs. This judge, a native of the interior of the province of São Paulo, was part of the first group of graduates from the Law School of São Paulopontuar Macedo also observed that, despite the peculiarities of his pronunciation, the judge when “speaking in Court, imposes silence and demands respect…”.
• 20
  In this study, the authors demonstrated that stereotypes can become self-fulfilling prophecies. The experiment consisted of observing the behavior of a group of men who initiated a telephone conversation with various women after seeing a supposed photo of them: some photos portrayed very attractive women and others who were less attractive. The men who thought they were speaking with attractive women were more sociable, warm, and outgoing. The women, in turn, also became more sociable, lively and confident. According to the authors, stereotypes may create their own social reality, in such a way that in social interactions, the stereotype of the men who took part in the experiment was confirmed by the behavior of the stereotyped target (the woman). Consequently, they state that a physically attractive person may behave in a friendlier and more pleasant manner, not because they have those attributes but because the behavior of others triggers and maintains behaviors considered manifestations of these features.
• 21
  Examples of the failure of the linearity of the cause-and-effect relationship are presented by Watzlawick (1994). The author says that the inversion of the temporal flow of this relationship may occur when we are dealing with self-fulfilling prophecies. The following excerpt may clarify this statement: “In March 1979, newspapers in California started to broadcast blatant news about the imminent reduction in the supply of gasoline. Californian drivers rushed to the gas station to fill up their tanks. The refuelling of 12 million tanks (approximately 75% of which were empty at that time) exhausted the enormous reserves of the product and caused the predicted shortage overnight; on the other hand, eager to fill up as many vehicles as possible (instead of keeping them nearly empty, as was common then), drivers formed long lines of cars, and the population spent hours waiting at gasoline stations: and this increased the panic. Soon afterwards, when the situation calmed down, it was verified that the supply and distribution of gasoline in the state of California had not suffered any reduction.” (WATZLAWICK, 1994, p.98) The author argues that examples like this show how much a future event can determine effects in the present.
• 22
  Formants are peaks of acoustic energy that represent the resonance frequencies of the air in the vocal tract and are classed F₁, F₂, F₃, etc., starting with the lowest resonance frequency, according to Ladefoged (2001). The formantic structure is the main trace of speech sounds. All vowels and consonants have formants, and it is the pattern of formants (especially the disposition of the first two formants) that allows us to differentiate vowels or recognize repetition of a vowel and to classify it, even when it is produced by different speakers.
• 23
  Lindau (1980b) points to a lowering of F₃ in the realization of trills, taps and approximants.
• 24
  The variations found in the data of the 12 informants were classified into two samples: sample 1 and sample 2. Sample 1 comprises the data of 9 informants, collected through the reading of the words and sentences; and sample 2 is composed of data collected through interviews with all 12 informants. In this paper, only data from sample 1 is presented. For further information about the totality of the data, see Leite (2010).
• 25
  For further information about these differences, see Leite (2010).
• 26
  Refers to the Arabian country, officially recognized as an emirate, located in the Middle East.
• 27
  According to Marcuschi (1986).
• 28
  The formantic frequencies of rhotics are difficult to measure. Approximant rhotics and rotacized vowels present even greater challenges because they are co-articulated with the vowels they precede. Because of this interaction, a methodological strategy, aimed at delimiting the vowel and the rhotic, was established to measure the values of the formant frequencies with the purpose of observing the major points in the VR trajectory. Thus, the trajectory was subdivided into three positions – initial position (IP), medial position (MP), and final position (FP) – from which the values of the formant frequencies were obtained. For further details, see Leite (2010).
• 29
  This number corresponds to the total occurrences of /R/ in the data of nine informants. Data from three informants were discarded due to the variants found. The 60 repetitions of the /R/ for each informant (30 in the medial coda position and 30 in the final coda position) are distributed according to the five vowels selected, as detailed in Table 2.
• 30
  The spectrograms show data from three informants: TG, GP and LB, described in Table 1.
• 31
  It would be appropriate to say that the value of the F3 frequency is relatively low because it is not as low as the one observed for the caipira /R/ variant.

Publication Dates

• Publication in this collection
  Jan-Apr 2015

History

• Received
  Mar 2012
• Accepted
  Feb 2014