A Configuração Plural Da Frase Nominal Em Uma Comunidade De Candomblé E Suas Implicações Para Práticas Educativas

2. Theoretical background and literature review

This section is divided into two parts. We first situate the field within which this study sits. Then, in the second part, we review previous work that has examined the plural configuration of the NP in the attempt to show broadly at which stage sociolinguistics is in the understanding of such configuration.

2.1 Theoretical background

This study sits within the field of Variationist Sociolinguistics (Weinreich, Labov & Herzog 2006[1968]; Labov 2008[1972], 1994, 2001a, 2001b). Variationist Sociolinguistics, as a field, arose as a reaction against traditional structuralist views of language, particularly those of Saussure and Chomsky, towards an observation of language variation and change, and the social meaning that this variation carries with it.

Labov’s Masters’ dissertation, which was partially published in Labov (2008[1972]), aimed at studying linguistic change by analysing speech. He noticed that language variation preceded language change, and that although previous authors had attributed a non-systematic nature to speech, he found that it was actually organised, a finding that came to be named “orderly heterogeneity” (Weinreich, Labov & Herzog 2006:101). There were patterns that could be made explicit and explained considering internal (linguistic) and external (social) factors, and these factors should be considered in the study of both, language variation and change.

Labov created a methodological model which studies the existence of linguistic variants and their probability of use (Etto & Carlos 2017). The model deals with statistical analyses of collected data, and for this reason it is also known as Quantitative Sociolinguistics. Variationist Sociolinguistics studies the systematic variation of spoken languages and their variants, i.e., the different ways of saying the same thing in the same context and with the same meaning. And “a set of variants is called linguistic variable” (Tarallo 1994:08). For example, in the present study, the marking and the non-marking of the plural in the NP are the variants, while the plural configuration of the NP is the variable. There are patterns of variation, and it is the sociolinguist’s goal to explain these patterns where before it sounded like ‘linguistic chaos’ (Tarallo 1994).

The aim of sociolinguistics is not limited to investigating variation and change. After all, there must be the social component in it, otherwise it would not be (socio)linguistics. One of the main goals of sociolinguistics is to examine language variation and the social meaning such variation carries with it (Levon 2010). In this sense, the use of a given variant may correlate with the speaker’s social class for instance, or schooling, sexuality, gender, race, ethnicity, professional life and so on. On taking sociolinguistics like this, it is also possible to understand that language variation is a means of social mobility, as one is judged as being part of a given social layer by the way they speak, and judged differently if they change the way they speak. Language variation is also a means of maintaining a status quo within society, since there will always be a variant which enjoys prestige - generally used by the upper classes, white or highly educated people - over more stigmatised variants. In light of this, sociolinguistic studies also analyse linguistic prejudice.

2.2 Literature review

This subsection is a review of the linguistic variable, i.e., the plural configuration of the noun phrase, as well as of the social factors we considered. We start by describing the variable and exemplifying it with noun phrases from the dataset. Then we review previous work in the field, focusing on social factors that constrain the marking or non-marking of the plural in the noun phrase.

For the purposes of this study, the concept of a NP is a set of words without a verb and with a noun as its head (Payne 1997). The head of the noun phrase “is the one word in the phrase which refers to the same entity that the whole phrase refers to” (Wáu 2020:10). Other words may accompany the head, before or after it, clarifying and/or modifying it, such as determiners, numbers, adjectives, and indefinite quantifiers, for example. A noun phrase can be singular or plural, and the focus in this article is the noun phrase in its plural form.

In standard Brazilian Portuguese (STA), all plural words in a plural NP are marked, i.e., the marking of their plural is redundant (Salomão 2010). This, however, is optional in informal language (Mendes 2016): not all items need to be plural marked and may be nonredundant (-Ø). Phrase (1) below is an example of STA marking, where all items receive the plural marking (-S). As for (2) and (3), one or more items are not marked (-Ø), respectively. Nearly always, once a word in the noun phrase is unmarked, all others to its right (the rest of the noun phrase) are also not plural marked. In either case, the semantic component is kept, i.e., the NP is understood as fully plural. The plural configuration of the noun phrase in the speech of our informants, whether it is STA or -Ø, is the linguistic variable in this study. For the sake of terminology in this article, standard, redundant, marked, and STA will be used interchangeably, and so will nonstandard, nonredundant, unmarked, and -Ø.

NPagr is one of the most widely studied phenomena in Brazilian sociolinguistics (Antonino 2007; Brandão & Vieira 2012; Oushiro 2015), so much that it is difficult to write a review and not risk leaving important publications out. Therefore, this review is, by no means, exhaustive. One of the reasons for so much interest in the phenomenon may be the fact that there is a real “combat” between the variants in the speech of Brazilians across the country. This phenomenon has taken place throughout Brazilian history through the contact of people who spoke different languages and has been termed irregular linguistic transmission (Lucchesi 2008; 2012; 2015), which is today in a process of sociolinguistic polarisation (Lucchesi 2019).

In this process of sociolinguistic polarisation, at least in part of the country, a change in progress is taking place, with STA forms slowly taking over again. When reviewing research done in three isolated Afro-descendant communities in the interior of the State of Bahia, Lucchesi (2019) included “time out of the community” as a social factor, considering those who had spent at least 6 months outside their communities and then returned. He found that there was a process of change among those who had been outside the community, especially among the younger members of the community, as they had spent time in more urban areas, had studied more and had had access to other means of communication. He termed this change in progress as “top down” and “from outside to inside” the communities, as the participants were assimilating other forms of the language, from -Ø to STA, and because they were returning home after some time in urban areas. Nominal number agreement was one of the variables analysed by Lucchesi. Findings suggest that the factor “time out of the community” is one of the most important aspects for the change in progress in the researched communities, weakening the context of sociolinguistic polarisation present in Brazil, the context of “combat” between variants, one of them prestigious and the other stigmatised.

This seems to be true only for some parts of the country though, especially in small communities and rural areas. Oushiro (2015), based on 118 interviews with informants grouped by gender, schooling and social class from the city of São Paulo, concluded that the nonstandard variant is an index both of masculinity and of belonging to some neighbourhoods of the city. The analysis of that study also pointed out that the correlation between its use and the general feeling of being paulistano (from São Paulo) is gradually disappearing. In her analyses, however, the presence of the redundant and the nonredundant forms of pluralisation of the noun phrase remains stable (i.e., no change in progress).  These results contrast with what Lucchesi (2019) found, showing the apparent change in progress he found is not homogeneous across the country.

When considering social factors, the majority of studies suggest that the nonstandard form is associated with low schooling and masculinity (Mendes 2016). Brandão (2013), for instance, conducted a study involving participants from Brazil and São Tomé, in Africa. She interviewed 18 informants from Brazil, grouped by gender, age and schooling; 22 informants from São Tomé, grouped by gender, age and schooling; and 9 students from São Tomé from different levels of education. She found that “low percentages of the presence of the plural marker in some communities [are] formed of individuals who are illiterate or have a low education level (in particular, those who live inland or are Afro-descendants)” (p. 52). Potentially for these reasons, the nonstandard form associated with low schooling and race is stigmatised, as the standard is not followed by some of her informants.

External (social) factors have been extensively examined in various Brazilian sociolinguistic studies. Findings of variation and change vary, and are sometimes conflicting. Guy (1981) found no systematic result that showed an effect of age over plural configuration of the noun phrase. He found that women tend to use the standard form significantly more than men. He concluded that these findings are constrained by two internal factors: the next stressed context, and the position of the word in the noun phrase.

Variation in age groups, however, is one of the most common factors found to have an effect on the NPagr variation. The findings sometimes point to a stable variation, whereas others suggest a change in progress in apparent time. In any case, Guy’s finding that age had no overall effect on NPagr is much more of an exception to the rule.

Scherre (1988), for instance, who grouped her interviewed informants in Rio de Janeiro by age, schooling, and gender, reached different results. Her analyses showed that informants with more schooling tended to use STA more often than those with lower schooling, a finding which is not surprising in sociolinguistic studies (Oliveira 1982) and which was confirmed in Sherre and Naro’s (2006) study. The age group presented a curvilinear distribution not accentuated, with younger and older informants favouring the use of -Ø and the intermediate age favouring STA, a context proper of stable variation.

Salomão (2010) analysed interviews and self-recordings done previously to her research and which were available in a database in São José do Rio Preto, Southeastern Brazil. She grouped the informants by age, schooling, and gender. Surprisingly, she did not find gender as a constraint to the pluralisation of the noun phrase in her study. Schooling and age, however, showed an effect. Lower schooling favoured the use of -Ø, whereas higher education favoured STA. As for age, younger participants showed more use of STA than older participants, tending to decrease again after 55 years of age. She argues that informants between 26 and 35 years old tend to use STA more often exactly because it is within this age range that informants attend higher education and start working, a fact she also considers relevant for the prevalence of the standard form.

Dália & Lucchesi (2021) assume that the variation in the pluralisation of the noun phrase has its origins in the acquisition of Portuguese as a second language by Brazilian indigenous people, African slaves and other people who colonised the region they researched, the process mentioned earlier in this text, termed irregular linguistic transmission (Lucchesi 2008; 2012; 2015). They conducted research in the 3º Distrito de Nova Friburgo, interior of the State of Rio de Janeiro, and claimed that linguistic transmission has perpetuated over time due to strict access to formal education by the biggest portion of the population. However, they reported that from the mid-20^th century there has been an increased use of the STA, which they attributed to increased schooling, influence of mass communication, as well as to the fact that many inhabitants moved from that rural area to urban centres and then returned to the rural community.

The specific social factors chosen in this study are age and schooling, and these factors were chosen in line with previous work (e.g. Scherre 1988; Oushiro 2015; Lucchesi 2019) and observations during fieldwork. Age is a factor which has traditionally been considered in sociolinguistic studies, both in terms of stable variation (e.g. Guy 1981; Scherre 1988; Antonino 2007) and others with regards to processes of change (e.g. Naro & Scherre 2011; Dália & Lucchesi 2021), as correlations between language variation and age, for example, may be a sign of change in progress (Labov 2001).  The informants in this study belong to two age groups: “older” were participants who were 30 years of age (or above) at the time of Lima’s (2023) data collection. Informants who were 18-29 years old were considered “younger”. By dividing participants into two groups, we could investigate whether what was reported in previous work (e.g. Guy 1981; Scherre 1988) held true among our informants and whether any change in progress could be identified.

The second social factor we investigated was schooling, which previous sociolinguistic research has reported to interact, at least to a certain extent, with linguistic choices (Salomão 2010). We coded schooling as “higher” if the informant was attending or had completed a university course at the time of Lima’s (2023) fieldwork, and “lower” if the informant had only attended compulsory education. Attending university is not as common in Brazil as in Europe, for example, and, indeed, only a minor part of the population does attend. In addition, taking schooling into account was also important because it relates to issues of linguistic stigma and prestige, as people with higher education tend to use more standard forms or, at least, the forms they use tend to be considered better than those used by people not educated at university level. It is worth noting that, whether this factor by itself may explain the variation attributed to it, it may also be questionable (Oushiro 2015).  Although schooling may play a role in linguistic variation, other aspects not strictly related to it may be at stake as well, such as the contact speakers have with different social networks and social media.

3. Methodology

In this section, we describe fieldwork conducted in Lima (2023) in Salvador (Bahia, Brazil), as the data we re-analyse here resulted out of such fielwork. The section begins with a brief review of the concept of ethnography and sociolinguistic ethnography in particular, which was the methodology followed in Lima’s (2023). Thereafter, we will provide a description of the terreiro, the informants and how they were grouped. Finally, we will present an overview of the methods used to conduct the interview in the data collection this study is based on.

3.1 Ethnography and Sociolinguistics

Ethnography can be understood as a social research methodology whereby the researcher immerses themselves in the social setting of their informants in order to capture social phenomena (Bernard 2006).  It requires the extensive presence of the researcher in the field for a long period of time (Blommaert 2007), which enables the researcher to make ‘thick descriptions’ of the mundane practices of the community (Geertz 1973). It also allows for the analysis and interpretation of their observations to show the reader what it means to be part of the studied community (Bernard 2006).

One of the main goals of ethnographic studies is to describe the cultural scene from the perspective of the researched participants (Agar 1980, Spradley 1980). From an ethnographic point of view, participants’ accounts on their own practices are important for the ethnographer to better understand their perspective of life and how they interpret the cultural and social scene they are embedded in.

Blommaert (2007) maintains that ethnography is built on two interdependent positions: the first position is ontological (i.e., social events are contextualised spatially, temporally and historically) and the second is epistemological (i.e., knowledge of social events are situated within the individual, the group or the community where the events take place, being therefore subjective) (Levon 2010). This is to say that “ethnography rejects the notion of an objective understanding of social action, and instead insists that knowledge is to a certain extent always “slippery” and contingent” (Levon 2010: 66).

Branislaw Malinowski is known for his contribution to ethnography, especially in English-speaking settings very much influenced by the British legacy. He was the first to do participant-observation in the way it is done nowadays (Ugwu, 2016). Rampton (2007) maintains that ethnography only had its origins in the work by Malinowski at the beginning of the twentieth century. From Malinowski, we have learned that ethnography involves a set of ethnographic practices, like participant-observation and field notes, so that a range of sociocultural phenomena can be recorded.

In the recent past and up to the present days, a great body of linguistic research interested in examining the social meaning of variation has increasingly combined sociolinguistics and ethnography (e.g. Eckert 2000, Buchotz and Hall 2005, Mendoza-Denton 2008). In these studies, “scholars have sought to understand the social meaning of variation emically, rather than identifying broad correlations with etic categories that are assumed to affect populations in similar ways” (Ilbury 2019: 68). All in all, it appears that, although ethnography was the main feature of the enterprise of second wave variationist sociolinguistics, it is being fruitfully used in third wave studies as well.

Having located as ethnographic the data collection based on which we proceeded to the present study, in the next subsection we describe the site where such collection took place, including the participants who provided the data we analyse here.

3.2 Looking within the larger setting

Inside the city of Salvador, land of parties, good food, poverty and extreme social inequality, there is a big ranch which is distinctive from the urban architecture surrounding it. A great deal of people attends this place every day, and several of them live there. It is run by Babá (i.e. a male leader of a terreiro), a 50-ish year old man who has been the leader for decades. The social structure is very strict; those who have been longer in the house are respected by the ones who have arrived more recently. The story of this house mingles with the origins of Candomblé, and this was the terreiro chosen for data collection.

In the data collection that resulted in data we here analysed, 18 informants were interviewed, all men, considered in this article and grouped by age and schooling (see below). All informants were filhos to Babá (sons to Babá in the religion), and they had all been initiated in Candomblé at the time of data collection. Half of them were ogans (those believed not to be able to get into a trance-like state) and half of them were rodantes (those who are believed to be able to ge into a trance and be inhabited by orixás, the gods in the religion). Informants who were at the terreiro on a daily basis were privileged in data collection, regardless of whether they lived there or not. We agree with Lima (2023) and believe that investigating the speech of initiated filhos of the terreiro who were there on a regular basis would allow one to understand how the terreiro is sociolinguistically structured.

3.2.1 Recordings

The audio-recordings done with participants and re-considered here are sociolinguistic semi-structured interviews. Interviews lasted 60 minutes on average, and they were recorded with a Zoom H4nsp with two lavalier BY-M1 microphones. All recordings were stored in WAV format. Each interview was divided into 4 parts: (i) personal history of the interviewee’s life, in which they would talk about topics such as infancy, teenage, adulthood, whether they were married or not, whether they had children or not; (ii) personal history of the interviewee’s life in Candomblé, in which they would talk about topics related to their involvement in the religion, what their role was, how being in Candomblé had changed their lives,; (iii) topics related to Candomblé , where the informant would talk about their knowledge of the religion, its structure, its origins, or the history of the terreiro; and (iv) more personal details, and self-identifications of the interviewee such as their income, age, and schooling. As explained in Lima (2023), these four parts were just a guide for the interview, not a straitjacket, and not all participants were asked all the questions.  Additionally, for example, often items in (i) and (ii) were talked about together, as many interviewees saw the story of their lives as inherently linked to their story in Candomblé. In any case, the interviews were conducted in such a way that information about the social factors of interest could be collected with informants.

3.3 The model building process

With data in hands, we isolated each plural NP in each of the 18 interviews, both STA and -Ø. Working with the R software (Kuznetsova et al, 2017), a full factorial model was built, that is a model including the social factors and their potential interactions or as fixed effects, and participant as a random effect. Participant was included as random effect to take into account individual-level variation in speech production, therefore obtaining a better picture of the role of each fixed factors and their potential interactions on the dependent variable of interest. The significance of the factors in the, at that point, best fit model obtained was then analysed with Anova(). More factors were shown not to be significant, and they were left out of the model. This last version of the model gave us less, yet solid and non-misleading results, and was therefore considered the best model.

4. Analyses

In this section, we report the quantitative analyses of the interaction age-schooling on the dependent variable, in addition to also providing our interpretations on important aspects considered in the field notes taken when data collection was carried out for Lima’s (2023) study.

4.1 Analyses of social factors at phrase level and at word level

This subsection reports the quantitative analyses of the effect of the social factors described above on the dependent variable ‘plural configuration of the NP’ (which could have been either -Ø or STA) considering the NP as a whole, and then considering each word separately. The NP was marked nonstandard if there were one or more unmarked plural forms, and standard if all plurals were marked according to standard Brazilian Portuguese. Table 1 below shows the results of the best fit model for this analysis. Note that, although the best fit model also included role as a social factor, showing role-schooling and role-age interactions, our analyses here lie only on the interaction age-schooling.

Table 1: Regression model for external factors at phrase level (Ref. STA)

In the literature, schooling has been widely shown to have an effect on plural configuration (Fernandes 1996; Biancardi 2011; Brandão 2013; Lucchesi 2019). The unmarked form of plural NPs in Brazilian Portuguese is perceived, among other aspects, as indexical of lower education (Mendes 2011, 2016), and this has also been shown in production studies (Narro and Scherre 2011, Oushiro 2015). It is not, therefore, surprising that schooling has shown an effect on the variable in this study in interaction with age.

Pairwise comparison show that among those with higher education, the older participants use the unmarked plural form more often than the younger (p = 0.0107). This finding suggests that standard language was overall less valued in the past than it is now, as older participants who have long been in the religion and with higher education use -Ø more often than highly educated younger informants. This can be supported in the theorisation of Lucchesi (2008; 2013; 2015) on irregular linguistic transmission, according to which the emergence of variation in the pluralisation of the noun phrase resulted out of the contact of people who spoke different languages in colonial times. It is important to notice that Candomblé was, at first, a religion of black people only, originating among African slaves and, therefore, the contact of African languages with European Portuguese gave rise to the variation. It appears that for a long period of time, as people in Candomblé lived in the terreiro, had little education and little contact with wider society, the stigmatised variant (-Ø) was favoured by most of them and standard language was less valued than it is today due to a stronger effect of the irregular linguistic transmission. With the new trend of younger Candomblé followers not living in the house, taking courses and having contact with wider society, the STA plural form seems to be taking over within this age group. This is happening in a similar fashion as the one described in Lucchesi (2019), who considered “time out of the community” as a factor which showed an effect over the plural configuration of the NP (-Ø ~ STA). For these reasons, we can talk about a change in progress in apparent time.

As for the analysis at word level, considering each item in the NP separately, Table 2 below shows the result of the best fit model. Note that, although the best fit model included role as a social factor, also showing an interaction between role and schooling, in this study we examine only the interaction between age and schooling. The interaction between age and schooling found by the model stood statistically significant with Anova (p = 5.006e-05).

Table 2: Regression model for external factors at word level (Ref. STA)

Results revealed that younger participants use the unmarked form less frequently than the older ones. As discussed previously, this may be indicative of a change in time due to the fact that younger participants do not spend as much time at the terreiro as older participants did and still do.

Further analyses of the interaction between age and schooling with pairwise comparison showed that, among those with higher education, the older participants use the unmarked form more frequently than the younger ones (p = 0.0133). This is consistent with the analyses at NP level and, as argued earlier, this finding seems to suggest that the use of standard Portuguese was less valued in the past. Interestingly, among those with lower education, the use of the -Ø form is more common among younger than older participants (p = 0.0008). To explain this, it is worth noting that, for example, during the interview, Diego stressed that primary education in the past was stricter and more demanding than it is nowadays, and this is something the ethnographer who collected data on which this study relies heard from other informants whilst at the terreiro. Indeed, results obtained by an assessment conducted by the Brazilian Ministry of Education has demonstrated that the quality of education in the country has lowered in the past 10 years (Brasil 2020). This may indeed explain the current result; as the model shows that, overall, among older participants with less education, the use of standard language was more valued in the past than it is now.

The analyses reported so far have shown that younger participants use STA plurals more often than older ones. In addition, younger participants with higher education also use STA more often than older informants with higher education. All in all, STA/-Ø variation is spread across participants.

As demonstrated in the quantitative analyses along with field notes taken, younger participants have more contact with formal language. As per field notes, this takes place not only because of formal schooling (secondary and higher education, for example), but also because they take extra courses to prepare themselves for the job market, a trend that is followed by younger participants. The same pattern is not true for older participants, who have their professional lives limited to the terreiro. This helps explain the findings brought up in the analyses, as contact with formal language facilitates the use of formal language.

5. Discussion

In this section, besides providing further interpretations for the findings reached in the quantitative analyses along with the considerations of field notes, we aim to raise a brief discussion concerning the potential positive implications of conducting the present study for educational practices. Overall, the main finding the analyses allowed us to reach is the fact that the -Ø variant is favoured among older participant more than among the younger ones at the terreiro where the data collection based on which we draw our analyses took place. The interaction age-schooling is clearly demonstrated by the quantitative model, but what this brings us in terms of social meaning is made more explicit when we consider important field notes that were taken at the time of the data collection detailed in Lima (2023).

As we presented in the introduction of this article, our hypothesis was that the STA/-Ø variation among informants could be related to the irregular linguistic transmission. Recall that, as presented in the literature review, this was a process of contact between people who spoke different native languages in Brazil’s colonial times that gave rise to the variation here under study. However, the data – and the best fit model that helped us analyse them – imposes the need to consider not only age as a factor, but its interaction with schooling. This interaction, when looked at with the aid of field notes, helps us clarify how it explains the linguistic variation under study and show that what we hypothesised holds true, i.e., the interaction that shows an effect over our dependent variable relates to the irregular linguistic transmission, involving participants’ age and schooling at the same time.

As per field notes taken when the data on which we drew our analyses, having a life outside the terreiro, with more contact with other layers of society, using of of social media, having more schooling, overall a life not nearly completely limited to the walls of the place is a new trend at the terreiro. Consequently, older participants in the research hold what they (socio)linguistically inherited from their ancestors more so than do the younger ones. Having this in mind, even considering older participants with higher education, it holds true that their contact with wider society is more limited than what happens among the younger ones. This helps in the process of keeping their sociolinguistic practices, to a given level, away from the influence of the modern world. Consequently, it becomes clear that it is access to what the terreiro itself does not offer, e.g. contact with formal language, access to more education in the form of extra-courses, for example, that makes younger participants more prone to have an influence of STA on their linguistic choices, which includes NPagr. Living more freely and having access to other forms of communication more so than it used to be for most older people at the place, the younger participants receive less of the inheritance coming from the times of the irregular linguistic transmission process, which clearly influences the older participants in a more direct fashion.

By looking at the data and analysing the phenomenon from this perspective, it becomes clear that, as explained, age and schooling work in interaction to produce the STA/-Ø variation, being STA a result of the influence of contact over the limits of the terreiro on younger participants, especially education, and less of this same influence older participants have. In this way, the effect of the interaction age-schooling on our dependent variable is directly linked to how much influence or linguistic inheritance the irregular linguistic transmission has brought to the place, more to the older, less to the younger participants. Furthermore, it explains the variation under study from an impartial perspective, i.e., not over-valuing one of the variants to the detriment of the other. On analysing the data in this way, we see that both STA and -Ø are equally communicative, both say the same thing, in the same context, just in a different way. This perspective shows that language varies, why it varies and what social meaning this variation carries and constructs, instead of creating a hierarchy between STA and -Ø that can only promote linguistic prejudice and stigma.

Schooling practices have traditionally favoured the STA variants, starting from literacy. We do not consider that there is a problem in teaching STA at schools. After all, learners should be prepared to use STA when this variant is considered more appropriate than -Ø. The issue, however, lies not in the fact that schools teach STA, but how -Ø is positioned whilst STA is taught. Students who favour -Ø in their daily lives and have just started school, for example, have throughout the history of Brazilian education system faced teaching practices that undervalue the variant they master, incurring in the potential risk of promoting low self-esteem, reinforcing the existing linguistic prejudice based on social beliefs that there is a way of speaking which is better than others, and such way which is considered better is clearly STA. Instead, school practices should aim at teaching STA and, at the same time, validating all the rich linguistic repertoire that students already have in their vernacular.

It is in this sense that we argue that studies such as the one here reported have the potential to bring positive implications for education. In our analyses, there is search to explain why variation takes place and what it means, what it tells us about people who use the different variants considering the variable, the social meaning of variation after all. What there is not is any attempt to judge one variant as superior or inferior to the other and, consequently, there is no judgment of the individual who realises NPagr as STA or as -Ø. By historically and sociolinguistically showing the variation just as it is, with no judgmental perspective, the study may shed light on how school-teachers see language variation as it relates to their teaching practices. By having access to analyses such as the ones we present here, education professionals may raise productive and positive reflections on how they conceptualise language and how they understand the importance of the process of teaching and learning to the life of an individual.

Consequently, the study may raise more awareness on the part of teachers, helping them in the construction of a more critical perspective to language teaching, open to language variation, one of the most well-known phenomena in linguistics. Such awareness has the potential to promote changes in schooling, not abandoning the practice of teaching STA, but including in such practice the validation and valuing of other forms of speaking, including the nonstandard variant we consider here, that it, -Ø. All this process works for the construction of a more inclusive and less stigmatising education and, as a result, helps fighting the existing linguistic prejudice that is (still) at play in society.