THE DEVELOPMENT OF SOCIOLINGUISTIC AWARENESS AND SUCCESS IN READING PERFORMANCE

ABSTRACT

National and international assessments show that Brazil fails in the teaching of reading in its entire school system. This paper argues that the explanation for part of the students’ success in early learning of reading might be related to the development of sociolinguistic awareness. The rates of variation in eight phonological variables (-ow monophthongization, -aj, -ej monophthongization, diphthongization, coda -r deletion in nouns and in verbs, -d deletion in -ndo, cluster simplification, and vocalization of palatal lateral) in three different stylistic contexts (from less to more monitored speech: short interview - naming picture task - reading aloud task) collected in a 3^rdgrade classroom show a gradual relation between the increase in stylistic monitoring and the use of standard variant and a positive effect in the performance in a reading comprehension test. These results reveal that the development of sociolinguistic awareness can be measured by transposing the variant into spontaneous speech for reading aloud, suggesting automaticity in the decoding process by lexical route.

Sociolinguistic awareness; reading aloud; reading comprehension

RESUMO

Avaliações nacionais e internacionais mostram que o Brasil falha no ensino de leitura em todo o sistema escolar. Neste artigo defende-se que a explicação para o sucesso no aprendizado inicial da leitura de uma parte destes estudantes pode estar relacionada ao desenvolvimento da consciência sociolinguística. As taxas de variação de oito variáveis fonológicas (monotongação de -ow e de -aj, -ej, ditongação, apagamento de -R em coda em nomes e em verbos, apagamento de -d em -ndo, simplificação de cluster e vocalização da lateral palatal) em três diferentes contextos estilísticos (do menos para o mais monitorado: entrevista curta - tarefa de nomeação de figuras - tarefa de leitura em voz alta) coletados em uma sala de aula da 3ª série mostram uma relação gradual entre o aumento do monitoramento estilístico e o uso da variante padrão e efeito positivo no desempenho em um teste de compreensão em leitura. Esses resultados revelam que o desenvolvimento da consciência sociolinguística pode ser medido pela transposição da variante da fala para a leitura em voz alta, sugerindo a automatização no processo de decodificação por via lexical.

consciência sociolinguística; leitura em voz alta; compreensão em leitura

Introduction ¹

The results of national assessments such as the National Assessment of Alphabetization (ANA) and international assessments, such as the Program for International Student Assessment (PISA), show that Brazil fails in the teaching of reading through the entire school system. Just to mention an example, in Sergipe, the 2015 ANA results show that only 25% of children achieve adequate reading skills for the 2^ndgrade. This is not a new issue, nor is it unique to Brazilian society (LABOV, 1972a), but here it appears to be stagnant, without improvement.

The reading problem in Brazil may be partly due to differences among standard and dialectal varieties, which have not yet been fully explored in national reading education programs. Different methods and approaches have been presented and proposed as palliative measures to this problem; and Sociolinguistics can contribute to the reduction of asymmetries not as a palliative, but as an integral element, with the development of sociolinguistic awareness. This paper assumes that the explanation for a part of the students’ success in reading may be related to the development of sociolinguistic awareness: the awareness of existing linguistic variables in language and their social and/or stylistic meaning concerning contexts of use, social relationships, and personal identities ( VAN COMPERNOLLE; WILLIAMS, 2013 ). This article has two goals:

1. Compare the behavior of variable linguistic traits with performance in a reading comprehension test to identify the predictive power of certain sociolinguistic traces, to help reading literacy programs perceive the importance of developing sociolinguistic awareness.

2. Improve the scale of social evaluation from Sociolinguistics (indicators, markers, and stereotypes) by measuring the behavior of eight phonetic-phonological variables of Brazilian Portuguese in different stylistic contexts.

Considering the relationship between consciousness and stylistic values, observing the emergence of sociolinguistic awareness can provide feedback to the research of scales of social appreciation of linguistic variables, which in Brazilian Portuguese are commonly inferred from the distribution of frequencies in terms of social profiles ( FREITAG, 2016 ).

Linguistic and sociolinguistic awareness

The evaluation of language is crucial for the constitution of speakers’ identity in terms of social appreciation and consciousness. Different approaches try to explain this relationship, for example, Labov’s (1972b), with indicators, markers, and stereotypes; Bell (1984) , with audience design; Silverstein (2003) , with indexical orders; Eckert (2008) , with indexical fields. What these approaches hold in common is that speakers make choices during the speech, but these choices are not necessarily conscious. Sociolinguistic awareness is an explicit knowledge that emerges from experiences of recognizing linguistic differences at the time of interaction, leading to the realization that they are both linguistically and socially significant ( SQUIRES, 2016 ).

All speakers of a language develop some language awareness, a wide range of specific knowledge of language: at the first level, structure and grammar; at the second level, knowledge of the social aspects of a language, including linguistic variation, variability in the accommodation between speaker and audience, intentions of the speaker and code-switching (sociolinguistic awareness); at the third level, knowledge of how language can intentionally be manipulated for persuasive purposes; and the fourth level, critical awareness of language, including the recognition of how social and discursive patterns are mutually constitutive and how speakers are widely embedded and constrained by these community language patterns ( PIPER, 2003 ).

The school explores linguistic awareness in a particular way in the early reading learning, with the development of phonological and morphosyntactic awareness, which contribute to automaticity in decoding and reading comprehension. Sociolinguistic awareness is, however, not fully explored yet, not even by sociolinguistic approaches.

All speakers, to some extent, show sociolinguistic awareness when they try to explain why certain people have a different, peculiar way of speaking. What does the general public, not linguists, think about language? On the one hand, there is a “popular” sociolinguistic awareness: explanations, clarifications, and beliefs about the relationship between language and social context made by non-specialists. On the other hand, there is the force of prescriptivism, a set of normative metalinguistic practices, focusing on the value of correction, on the “correct” use of language, according to the norm established by grammarians. While prescriptivism tries to support itself in scientific models, the popular sociolinguistic awareness is ignored by science. McKenzie (2015) emphasizes the importance of the way the population commonly perceives, processes, and codifies linguistic variation in speech, including the identification of linguistic varieties as socially or regionally marked. By considering sociolinguistic awareness, perception studies can, for example, provide scientific information about labels given to specific social and regional variants by non-linguists, as well as help to identify the sociolinguistic features that determine their classification, and assist the researcher in his/her task of mapping and classifying varieties and linguistic features of some language.

As mentioned above, variable linguistic features can be categorized according to the level of social appreciation (LABOV, 1972b): stereotypes - linguistic features, socially and consciously marked by speakers; markers - social and stylistic linguistic features that allow consistent effects on the conscious or unconscious evaluation of the listener about the speaker; and indicators - socially stratified features, however, not subject to stylistic variation. Still, linguistic variation can be at the level of the consciousness or below it. Speakers are more aware of variables when in contact with phonetically different as well as phonologically contrasting variants ( TRUDGILL, 1986 ). The knowledge of variation is a consequence of a semiotic process during which speakers are exposed to linguistic differences, observe them, and systematize patterns of their use in connection with social facts ( SQUIRES, 2016 ). While many projects have investigated the state of knowledge of speakers about cases of variation in descriptive studies of linguistic production, the process through which speakers construct this knowledge (sociolinguistic awareness) is yet not fully explored. Differences between perception, recognition, and understanding of variation, as explained by Squires (2016) , have been widely discussed in the field of second language acquisition, but still need to be explored in the first language context.

Measuring sociolinguistic awareness can answer questions such as: what features (social, psychological) of individual speakers are activated by the features of their speech? How do speakers of different linguistic varieties classify and label the same sounds? How to account for the ability of speakers to perceive certain variable features of the language but not others? How to systematize the speakers’ ability to identify and categorize the varieties of a spoken language or the speakers’ social or geographical background?

Following Labov’s (1972b) proposal of social appreciation, the results of sociolinguistic studies of linguistic production are interpreted regarding the indexation to social groups of given variants; the variants are assumed as markers of social identity, whose use makes speakers apart or bring them closer.

The popular sociolinguistic awareness has a considerable clear construction of social stereotypes: the set resulting from the arrangement of certain variable linguistic features evokes attitudes towards their speakers. It is supposed that linguistic variants can denote social identities and elicit attitudes, but there is still no explanation for the links between social identity, social stereotypes, and language ( KRISTIANSEN, 2001 ).

Squires (2016) suggests that the level of awareness of stereotypes, markers, or indicators can be expressed as the level of knowledge that speakers have about variation. Squires’ (2016) proposal is that sociolinguistic awareness goes beyond implicit vs. explicit knowledge: as to stereotypes, speakers know that a linguistic variable is related to a type of speaker; as to indicators, speakers do not know which type of speaker the linguistic feature is related to.

The school allows reflection about the language usage; the sociolinguistic awareness is measured in tasks that lead the student to the recognition and identification of different varieties of a language and listen to examples of speech with a difference between formal and informal in many communicative situations, such as wedding ceremonies, meetings, cooking classes, etc. ( PIPER, 2003 ; ; GÖRSKI; FREITAG, 2013 ).

In the context of L2, Van Compernolle and Williams (2013) define sociolinguistic awareness as the awareness or knowledge of linguistic variables (e.g. lexical, grammatical, phonological variables) and their social and/or stylistic meaning in contexts of use, social relationships, and individual identities. The authors illustrate it with the example of the variation between the presence and absence of ne in French negation, a grammatical variable that carries social and stylistic meaning: the presence of ne is required in standard, formal French (particularly written), while the absence is characteristic of almost all European and North American French speakers in more informal, everyday (particularly spoken) French.

The development of sociolinguistic awareness presupposes first being aware of which linguistic features are variable (and which the variants are); and secondly, being aware of the social and stylistic values of the variants. Sociolinguistic awareness implies both linguistic and socio-cultural knowledge: conscious knowledge of socio-stylistic variation can entail the controlled use of linguistic constructions in different contexts.

Sociolinguistic awareness and reading performance

Reading in an alphabetic writing system can be performed through a direct visual recognition process (lexical route) or through a process involving phonological mediation (phonological route). The reading approach that recognizes two possible routes is known as the Dual Route reading model ( COLTHEART; RASTLE, 1994 ). Both reading routes start with the visual analysis system, which identifies the letters, and their position in the word and groups them. In the phonological route, first, the letters are decoded, then a possible pronunciation is searched, and then the meaning is conveyed; this is the only possible route to read new or unusual words in regular spelling and neologisms. In contrast, the lexical route uses orthographic maps and the representation of visual words activated in memory ( EHRI, 2014 ); this is the route used for frequent words and essential for irregular words. Orthographic maps form connections between read words and sound patterns, linking in memory the spelling, pronunciation, and meaning of words; visual words are the representations of all the words read in memory. When a word is read, its pronunciation and meaning are immediately activated in memory, without any sound or combination required. Visual words are read as whole units without pauses between sounds ( EHRI, 2005 ). It is in the visual representation of words that sociolinguistic awareness interacts with reading: the linguistic variety of a community is taken as reference for the graphophonemic conversion. In this sense, the presence of linguistic variation in reading cannot be considered an error, but rather a clue to the lexical route and, therefore, to automaticity in decoding, which is essential for understanding what is being read.

However, the relationship between linguistic variation and reading performance has been considered negative in the national scenario in the following way: linguistic variation in the student’s speech is inversely correlated to the performance in reading and writing tasks, as pointed by the studies of Barrera and Maluf (2004) , and Guimarães (2005) . By recognizing the effects of linguistic variation on early reading and writing processes, Barrera and Maluf’s (2004) study proposes to “identify and describe the different types of linguistic variation used, at the beginning of alphabetization, by a group of students coming from low-income classes; verify if the students with the largest linguistic variation at the beginning of the year are the same ones who face the greatest difficulty in acquiring written language at the end of the year” ( BARRERA; MALUF, 2004 , p.40).

Assuming that “linguistic variation used” means not adopting standard linguistic forms considered prestigious, it means that students who make use of stigmatized variants are the ones who will present the greatest difficulties in reading and writing tasks. To verify this hypothesis, the study followed a group of students from a municipal public school, “on the outskirts of the city of São Paulo (which) predominantly receives children from low socioeconomic families” ( BARRERA; MALUF, 2004 , p.40).

The data for linguistic variation analysis was provided by a narrative reported from a visual stimulus, documented in audio (a procedure that in the present study is called naming picture task, as described in the method section). Later, the “non-standard linguistic forms” were identified ( BARRERA; MALUF, 2004 , p.40). The sum of the “types of variation used” divided by the time of the task was computed as an index of linguistic variation for each child ( Figure 1 ).²

Image 1
– Distribution of linguistic features with predominance of the standard in speech

In the list of controlled linguistic variables, some considered as non-standard do not significantly differ from those considered as standard; this is the case of the suppression in -ndo, which in São Paulo (FERREIRA; TENANI; GONÇALVES, 2012), presents 76% of use of the variant with suppression; there is no linguistic argument either that justifies a distinction between the raising from “cumeu”, considered non-standard, to “cumida”, considered as standard ( Chart 1 ). Lambdacism, articulation of /l/ by /r/ in complex onset, of late acquisition, cannot be considered as a variation either (NICOLOSI; HARRYMAN; KRESHECK, 2004).

Chart 1
– Non-standard linguistic features

Description	Example	Process
The vowel [o] is pronounced as [u], at the beginning of the words	buné, tumati, cumeu.	Vowel raising
The final segment “am” [ãu] of the verbs in the 3rdpersonplural is pronounced as [o]	eles saíro, fizero, brigaro, ficaro, foro.	Denasalization/verbal agreement
The final nasal segment [em] is pronounced as [i].	homi, onti	Denasalization
Replacing [lh] with [i]	trabaia, paiaço, oio, oreia, muié, agüia, moiadu, vermeio, veia, véio	Vocalization of the palatal lateral approximant
Replacing [l] with [r]	armoço, pranta, bicicreta, compreto	Rhoticism
Replacing [r] with [l]	tlem, tliste, tlês	Lambdacism
Omission of [d] in the consonantal cluster [nd] of gerund forms	correno, falano, chorano, pulano, fazeno, brincano, conversano, pegano	Suppression of -d in -ndo
Omission of [r] in some consonant clusters	tabalha (trabalha) ôto/ôta (outro/outra), quato (quatro)	Complex onset simplification
Vowels [e] and [o] pronounced as [i] and [u] when occurring in the final non-stressed position of the words	alegri, sorveti trabalhu, cachorru, denti, passarinhu, árvuri	Final non-stressed syllable raising
Generalized omission of [r] in verbal infinitives	brincá, passeá, fazê, mordê, durmi, istudá, iscová	Coda -R deletion (verb)
Vowels [e] and [o] pronounced as [i] and [u] even not occurring in the final non-stressed position of the words	mintira, pudia, durmia, cumida, iscola, istuda	Vowel raising
Systematic reduction of diphthongs [ei], [ai] and [ou]	caxa, dexa, pipoquero, dinhero/u, poquinho, tornera, ôtro, fejão,	Monophthongization
Expansion of monothongs into diphthongs	mais (em lugar de mas), feiz, deiz, faiz, tambeim, ningueim	Diphthongization
Lack of nominal agreement (singular/plural)	as criança, minhas amiga, meus irmão, as menina, muitas letra, as pranta, às veiz, duas gota, ficamos duente, as coisa;	Nominal agreement
Lack of verbal agreement	elis volta, as pessoa gosta, as criança foi, é eu, elis pega, é essas, nóis brinca, elis si separô, elis vai, elis fica, os dois é.	Verbal agreement

• (In gray: observed variable features, but not added to the index of linguistic variation)
Barrera and Maluf (2004

Moreover, the index of linguistic variation only considers the frequency of non-standard features, and not their proportion in relation to the variable (type/token frequency), as it is the practice of sociolinguistics. Even so, the study found a negative association between the index of linguistic variation and the performance in reading and writing tasks (r = 0.27; p< 0.05).

Based on the same hypothesis, Guimarães’s (2005) study examines the influence of linguistic variation and morph-syntactic awareness on the differences in reading and writing performance of students in 2^nd, 3^rdand 4^thgrades, divided into two groups: Group 1, 18 students (4 in 3^rdgrade and 14 in 4^thgrade), with difficulties in reading and writing, who used resource rooms and attended regular classes; and Group 2, 17 students (3 in 2^ndgrade and 14 in 3^rdgrade), without difficulties in reading and writing, who attended regular classes only.

The measure of linguistic variation adopted in this study follows Barrera and Maluf’s (2004) study; and, among the measures of morph-syntactic awareness, what is really being measured is the use or not of variable linguistic features, as in the task of generative use of morphemes, which consists of the inflection of verbal forms according to the given context (stimulus). However, what is being measured is the variable phonetic realization of a morpheme:

• - pronunciation or omission of “r” in verbal infinitives (e.g. pronounce /passear/ or /passiá/, /cantar/ or /cantá/, /comprar/ or /comprá/) [ -R deletion in coda in verbs];

• - pronunciation or omission of the final “u” in past tense verbs, third person singular (e.g.: pronunciation /dançou/ or /dançô/, /falou/ or /falô/, /mostrou/ or /mostrô/) [monophthongization of -ow];

• - pronunciation or omission of “s” in past tense verbs, first person plural (e.g., pronounce /compramos/ or /compramo/, /perdemo/, /brincamos/ or /brincamo/);

• - pronunciation or omission of the nasalization of the non-stressed nasal diphthong in past tense verbs, third person plural (e.g., pronounce /cantaram/ or /cantaro/, /deram/ or /deru/). ( GUIMARÃES, 2005 , p.265, our translation).³

Or, still in the task of correcting grammatical violations, in which supposedly incorrect sentences should be corrected for verbal agreement, use of pronouns, and noun and adjective agreement, in which, according to the study, there are “morphemic anomalies” such as ‘ Nós vai ao circo no próximo domingo ’, ‘ Marta veste sua casaco ’, ‘ Juliana tem lindas olhos ’. While in the latter ones there is indeed a grammatical violation regarding gender agreement, in the former sentence there is linguistic variation. On the one hand, these inconsistencies weaken the validity of the test results in terms of morph-syntactic recognition, but, on the other hand, they may provide clues to sociolinguistic awareness.

Students in Group 1 had a higher rate of linguistic variation than those in Group 2 ( Chart 2 ); however, the hypothesis that students with lower rates of linguistic variation would conversely have higher scores in the tasks that evaluated the morph-syntactic awareness was not confirmed (r = - 0.29; p = 0.089). Guimarães’ (2005) study details the types of variation by frequency, and concludes that three quarters of the total occurrences of the non-standard variants are relative to non-stigmatized and frequent phenomena in all Brazilians’ speech, such as vowel raising, -R coda deletion (verbs), denasalization in verbs, deletion of -d in gerund, monophthongization, and diphthongization. There seems to be an effect on the lack of opacity of the orthographic system, but unlike Guimarães’s (2005) , it would not affect only speakers of linguistic varieties far from the standard, but by all speakers of the language.

Chart 2
– use of non-standard linguistic features and application rates

Types	Examples of tokens	Tokens identified in G1		Tokens identified in G2
		F	%	F	%
Pronunciation of vowels [e] and [o] as [i] and [u]	alegri, mininu etc.	476	40.64%	226	33.38%
Omission of [r] in verbs	brincá, passeá etc.	94	8.02%	65	9.60%
Omission of the [u] final (verbs in the past, 3rdsingular person	falô, mostrô etc.	198	16.91%	171	25.26%
Omission of the nasalization in non-stressed nasal diphthong (verbs in the past, 3rdperson plural)	saíro, fizero, brigaro, ficaro/u, pegaro etc.	5	0.43%	-	-
Reduction of [n] the consonantal group [nd] in the gerund	pensano, chorano, sorrinu, correnu etc.	62	5.29%	14	2.07%
Reduction of diphthongs [ei], [ai], and [ou]	caxa, dexa, ôtro etc.	20	1.71%	21	3.10%
Expansion of monophthongs into diphthongs	feiz, deiz, tambeim etc.	57	4.87%	59	8.71%
Contractions, agglutinations, and other variations related to the prosodic aspects of the general rhythm of speech	ca/cas (com a/as), cê (você), ni (em) etc.	152	12.98%	85	12.56%
Insertion of a [a] in the beginning of verbs	atropeço, avoar etc.	5	0.43%	7	1.03%
Reduction of a [a] in a few words	dimirado, sustô etc.	3	0.26%	3	0.44%
Replacing [lh] with [i]	óia, trabaia, muié etc.	4	0.34%	2	0.30%
Replacing [l] with [r]	craru, bicicreta etc.	-	-	1	0.15%
Nasal termination [in] pronounced as [i]	homi, onti etc.	24	2.05%	-	-
Omission of [r] in consonantal groups	pá (pra), ôto (outro) etc.	24	2.05%	9	1.33%
Lack of nominal agreement	dois gol, os aluno etc.	8	0.68%	2	0.30%
Lack of verbal agreement	as pessoa gosta etc.	3	0.26%	3	0.44%
Lexical variations (vocabulary)	tacá (jogar), coisá etc.	11	0.94%	3	0.44%
Variation involving adjectives	mais melhor etc.	1	0.09%	-	-
Other variations	ingual, tamém etc.	24	2.05%	6	0.89%
TOTAL		1171	100.00%	677	100.00%

Guimarães (2005

The review of studies that consider the relationship between linguistic variation and performance in reading and writing tasks in Brazilian Portuguese highlights the importance of refining the control of the index of linguistic variation regarding: i) recurrence of variants (what is standard and what is not standard), ii) social evaluation of variants (stigmatized vs. non-stigmatized, perceived vs. not perceived, etc.), and iii) sensibility to the frequency of variants (the unique occurrence of a variable feature vs. systematic use).

Considering these aspects, the relationship between the sociolinguistic awareness measured by the sensibility of the frequency rate of a linguistic variant regarding the stylistic context and the performance in reading tasks should be positive: the greater the sociolinguistic awareness, the better the performance in reading tasks is. The change in the direction of the hypothesis in relation to the previous studies is due to the change in focus: while Barrera and Maluf’s (2004) and Guimarães’s (2005) studies considered the rate of linguistic variation in a more controlled activity (narratives based on figures) and compared it with the performance on reading tasks; this study considers the rate of linguistic variation on a stylistic continuum (speech, in a controlled oral activity and in a reading task), and the performance on reading comprehension tests. The presence of the productive linguistic variant in the community in the reading aloud is a clue that signals the lexical route, and for the assessment of the reading proficiency, it signals the skilled reader profile, who has visual memory of the word ( EHRI, 2005 ). The transposition of linguistic variable features from speech to reading does not negatively affect reading; in contrast, it can give clues about reading skills ( MACHADO, 2018 ; MACHADO; FREITAG, 2019 ; FREITAG, 2020 ).

Clues of sociolinguistic awareness

The effects of linguistic variation awareness can be more directly observed in reading, particularly concerning the contrast between non-transparent graphophonemic relationships, i.e. , those that have no bi-univocal correspondence. Although Brazilian Portuguese is considered a semi-transparent language regarding the graphophonemic conversion aspect, in contexts of linguistic variation, is asymmetry accentuated and does it raise issues related to the correction of pronunciation, such as: “ cenoura ” or “ cenora ”?

The prescriptivism of normative instruments (such as grammars) encompasses linguistic variation in orthoepy, but only for specific items not predictable by rules or principles of the alphabetic system of the language. Although it is assumed that the graphophonemic conversion selects forms from the variety of the community, prescriptivist orientations direct the choice of variants, at some moments selecting a dialectal variety and converting it into a standard linguistic feature, and at other moments, evidencing a pattern anchored in the written tradition. This is what Bechara presents for the process of diphthongization:

As for diphthongs, it should be noted: in pronunciation, ai, ei , and ou should keep their integrity, not exaggerating the value of i or u , nor eliminating them, as the distended modality does: caixa, queijo, ouro . [...] We usually diphthongized, by the addition of i , the final stressed vowels followed by a - z or - s . Thus, we do not make the difference between pás, pais e paz ; mas e mais. ( BECHARA, 2009 , p.60, our translation).⁴

For crescent diphthongs, the process of graphophonemic conversion endorsed by the normative prescription is transparent to the orthographic system (and “cenoura”: should be pronounced with a diphthong, although in the vernacular speech of Brazil monophthongization is the standard realization); whereas for -s/-z at the stressed end of words, the graphophonemic conversation is not transparent, a diphthong is inserted which is productive and categorical in the speech of Rio de Janeiro, but not in the speech of Santa Catarina or Rio Grande do Sul, for example ( LEIRIA, 2000 ; CALLOU; LEITE; MORAES, 2012).

Considering the variability, the clues of sociolinguistic awareness controlled in this study were selected by taking into account linguistic features of Brazilian Portuguese that are already widely documented from a sociolinguistic approach, enabling the estimation of its standard pattern, its dialectal and social recurrence, and contextual constraints⁵ , such as the trends identified in Rio de Janeiro, with data from PEUL ( PAIVA; SCHERRE, 1999 ; PAIVA; PAREDES-SILVA, 2012): monophthongization of decreasing diphthongs -ej and -ow; the stigmatization of the switch between liquids or -R deletion in consonantal clusters; the stability of the deletion of /d/ in -ndo segment; the implementation of the -R final drop, especially in verbs.

The monophthongization process consists in the suppression of the palatal glide [j], as in caixa , or velar [w], as in cenoura , in a decreasing diphthong. In Brazilian Portuguese, this process presents different constraints depending on the nature of the glide and the following phonological context: the suppression of velar glide tends to be a categorical rule in an informal context, including writing ( ARAUJO; BORGES, 2019 ; GONÇALVES; AMARAL, 2014 ; TOLEDO, 2013 ; HAUPT; SEARA, 2012 ; CRISTOFOLINI, 2011 ). Non-salient linguistic features, such as the monophthong, do not present differences in judgment that interfere with the lexicality task ( FREITAG; SOUZA, 2019 ). The palatal glide is internally constrained by the following phonological context. It is a semi-categorical rule independent of formality for contexts in which the following syllable presents the palatal feature, as in caixa, beijo , and is restricted by contexts in which the following syllable is initiated by stops, such as in leito, caibo . The pattern is stable in all regions of the country, without social or dialectal effects.

In contrast, the palatal diphthongization takes place in a stressed syllable trapped by an alveolar/postalveolar fricative ( NEVINS, 2012 ; CALLOU; MILK; MORAES, 2012). The process, even covered in the prescriptive grammar, presents dialectal, but not social restrictions.

The -R deletion in external coda in verbs is a categorical rule in spoken Brazilian Portuguese, even in situations of higher formality and monitoring, without association with social differences. The deletion in monomorphemic contexts and medial position in the word presents dialectal and social restrictions: the Northeastern region differs from the Southeastern and Southern regions because of -R deletion in non-verbs, but in this region, the -R deletion in medial position presents social differences (CALLOU; MORAES; LEITE, 1998; MONARETTO, 2000 ; CALLOU; SERRA; CUNHA, 2015).

There is dialectal and social association in the deletion of -d in -ndo segment: While in some regions, such as in the countryside of São Paulo, it is restricted to verbs (FERREIRA; TENANI; GONÇALVES, 2012), in others, such as in Rio de Janeiro and the Northeastern region, it affects the system as a whole ( MOLLICA; MATTOS, 1992 ; NASCIMENTO; ARAÚJO; CARVALHO, 2013; FREITAG; CARDOSO; PINHEIRO, 2018), and in other regions, it is not even reported, such as in Southern Brazil. The process is related to low education or lower formality and is restricted to situations of higher formality, such as in reading aloud (CARDOSO; PINHEIRO; SILVA, 2019).

The simplification of clusters is a step in a leniency process that can result in the deletion of the second segment in complex CCV attacks formed by a plosive and a liquid ( CRISTOFARO-SILVA, 2002 ). In Brazilian Portuguese, it takes place with alveolar flap, as in pró/pɾ/io > pró/p_/io, li/vɾ/o > li/v_/o. The tautosyllabic context presents a late acquisition in Portuguese; but, in this case, the non-production of CCV syllables is counterbalanced by lengthening ( MIRANDA; CRISTÓFARO-SILVA, 2011 ). When entering elementary school, the typical student has already acquired this linguistic feature. The persistence of simplified cluster implementation is a variable process, and its recurrence is associated with specific social profiles, and subject to social stigmatization (MOLLICA; PAIVA, 2011).

The orthographic context of “lh” presents a variable realization by a palatal lateral approximant /ʎ/, a palatalized alveolar lateral /l^j/, or a vocalized glide /y/. The vocalized realization is socially and dialectically delimited, associated with low education and the rural region ( BRANDÃO, 2007 ). It is found in the school space and transposed to writing ( COAN; FREITAG, 2010 ; FREITAG, 2011 ).

Along with retroflex realization, rhoticism is possibly the most stereotypical sociolinguistic feature in spoken Brazilian Portuguese: The neutralization of the distinction between lateral approximant and flap in complex attacks, as in f/l/anela ~ f/ɾ/anela, or syllabic coda, as in a/l/moço ~ a/ɾ/moço, is the object of aware manipulation of “ caipira ” caricatures, such as the character Jeca Tatu and Chico Bento ( COAN; FREITAG, 2010 ; COSTA, 2013 ; FREITAG et al ., 2020). This feature is a socio-dialectal marker of Brazilian Portuguese present in the speech of distant regions of large urban centers, and of less educated and older people.

The observation of reading aloud can allow the identification of the value judgment attributed to variable linguistic features, contributing to the scientific construction of sociolinguistic awareness ( Chart 3 ). On the other hand, the presence of variable features in reading is a clue of the access to the lexical route and, therefore, of the automaticity in decoding, an essential skill for comprehension, which can be verified with the success rate in a reading assessment.

Chart 3
– Distribution of linguistic features regarding productivity and frequency

Process	Example		The standard in the community	Feature		Standard
				dialectal	social
monophthongization -ow		cen/ow/ra
		cen/o_/ra	√			√
coda R deletion (verbs)		cant/aR/
		cant/a_/	√			√
diphthongization in stressed syllabic coda S		arr/oz/		√
		arr/ojz/	√			√
monophthongization -aj, -ej		c/aj/xa
		c/a_/xa	√			√
coda R deletion (monomorphemic)		so/R/vete	√	√		√
		so/_/vete
d- deletion in-ndo		canta/ndo/		√		√
		canta/_no/	√
simplification of cluster		pe/dɾa/	√		√	√
		pe/d_a/
vocalization of lateral palatal approximant		pa/ʎa/ço	√	√	√	√
		pa/ya/ço
rhoticism		f/l/or	√	√	√	√
		f/ɾ/or

Source: Author’s elaboration (2021).

Method

The procedures of observation and sample composition are detailed in Freitag and Sá (2019) , which aimed to identify the reading profile of students in a 3rd grade class of a public school in Sergipe; here, the necessary procedures for the sociolinguistic description are summarized. Twenty-one students, aged between 7 and 11, all regular students in a 3^rdgrade class in 2017, participated in the survey in four types of data collection: short interviews; naming picture tasks; reading aloud; and reading comprehension test. The first three types of data collected reflect a stylistic context of greater or lesser monitoring of speech: short interview (less formal) oral activity (naming picture task, intermediate) reading aloud (more formal).

The texts for reading aloud and the figures for the naming picture task were chosen according to the words that could also allow the presence of the same linguistic phenomena. The reading comprehension test (RAT) similar to the ANA, published by Editora Moderna (2016), unlike the previous collections, was collectively applied to 18 students⁶ . All the linguistic documentation was individually recorded in a school environment and submitted to orthographic transcription with audio aligned in Elan software (WITTENBURG et al ., 2006).

All possible contexts of variable phenomena occurrences in short interviews, naming picture tasks, and reading aloud were identified and controlled for the rate of the standard variant frequency in the community per student and univariate analysis by variable phenomenon. The RCA correct answers were individually computed and analyzed in a linear regression model regarding the frequency rate of the standard variant in the community at the extremes of the stylistic monitoring continuum (reading aloud and short interviews), and then meta-analysis. The graphic visualization of the results was developed with ggstatsplot package ( PATIL; POWELL, 2018 ) for R.

Results and discussion

To explain the results, three aspects were considered: is there an association between the occurrence of the variant and the stylistic continuum? (observation of the percentage and the result of Pearson’s chi-square test); how strong is this association? (Cramer’s V coefficient) and what is the pattern? (direction of the percentages). From these aspects, the rates of occurrence of variant assumed as standard in the community for each one process present four distinct patterns:

• - categorical occurrence of the standard variant in the less formal stylistic contexts and variable occurrence of the non-standard variant in the more formal situation ( image 1 );

• - low recurrence of the standard variant in the less formal stylistic contexts and high recurrence in the more formal situation ( image 2 );

  

  Image 2
  – Distribution of linguistic features with predominance of the non-standard in speech
  

• - high recurrence of the standard variant in the less formal stylistic contexts and low recurrence in the more formal situation ( image 3 );

  

  Image 3
  – Distribution of linguistic features with stereotype behavior
  

• - stability in all stylistic situations ( image 4 ).

  

  Image 4
  – Distribution of linguistic features with indicator behavior.
  

In less formal situations, the monophthongized realization of -ow, the coda -R deletion in verbs, and the diphthongization in final stressed syllable ended by -S express as a categorical rule for the first two and semi-categorical for the last phenomenon.

The variant assumed as non-standard in the community occurs only in reading aloud, which suggests that it is not necessarily an effect of stylistic monitoring, but a reflection of transparency in graphophonemic conversion that signals the activation of the phonological route and not the lexical route. Because they are not socially marked, the features would not be filtered by sociolinguistic awareness in the situation of higher formality; therefore, it is expected that the proficient reader has already the information that the graphophonemic relationship is not transparent in the visual memory of the words in which this feature occurs: one writes “cenoura”, but reads “cenora”.

In reading aloud, these features present a weak association (as -ow monophthongization) or an insignificant one (coda -R deletion in verbs and diphthongization). The behavior of these variables could suggest a distribution of markers, considering their distribution in the stylistic monitoring continuum. However, since it is a sample of readers in early reading learning, it is possible that sociolinguistic awareness has not been activated yet. The conservation of the variant assumed as non-standard based on the effects of frequency may indicate the convergence of the effects of prescriptive orientations and the transparency of the system, which would explain the non-monophthongization and the conservation of coda -R in verbs, but not the diphthongization, which has a prescriptive orientation against the linguistic variant assumed as standard (cf. BECHARA, 2009 ). On the other hand, morphological awareness could explain the conservation of -R in verbs, as Guimarães’s (2005) results suggest. An investigation with proficient readers (at more advanced levels of instruction) could provide more consistent cues: if the pattern is maintained, there is a stylistic monitoring effect (and, in this case, the feature may behave as markers), if the pattern in reading is the same as that of the interviews and naming tasks, there is an effect of the development of sociolinguistic awareness (and, in this case, the features would have categorical behavior, not a variable rule anymore).

The deletion of -d in segments -ndo and the simplification of clusters with complex attack are high recurrent processes in less formal situations and low recurrent in the more formal ones. The pattern of distribution suggests the effect of the transparency of the orthographic system, but in a different direction from that of the previous variables: the visual memory of the words with these segments include two realizations, one for speech, in which the variant with deletion (-d or -R) predominates, and one of reading, with the realization of the segment. Unlike the previous variables, the association between the variants is strong and significant in all contexts of stylistic monitoring. This result reinforces the cues of sociolinguistic studies of production that, as far as social evaluation is concerned, the traces behave as markers: sensitive to stylistic monitoring.

The vocalization of the palatal lateral is a stereotype in Brazilian Portuguese ( MADUREIRA, 1999 ; CYRANKA; RONCARATI, 2010 ) and this perception of stereotype seems to be captured by the students’ sociolinguistic awareness in the sample. Although it is the sociolinguistic feature with the weakest association among the eight selected, the behavior of the frequency of the standard variant is sensitive to the stylistic context, with decreasing recurrence due to increased formality, but still with occurrence in reading. This means that the visual memory of the words with this segment comprises two forms, the standard and the non-standard, and the non-standard form is constrained by social monitoring (evidenced by the reduced recurrence in orality situations, either in the short interview or in the naming picture task, and by the restriction in the reading aloud). The pattern is different from the one recognized as a marker because, in addition to the stylistic sensitivity which is present in the markers, there is a restriction on the occurrence in speech. Due to being consciously manipulated and targeted by evaluative meta-comments, whose most emblematic example is the poem “Pronominais” by Oswald de Andrade, this linguistic feature is considered a stereotype ( COAN; FREITAG, 2010 ). Nevertheless, much stronger than vocalization is rhoticism: in the sample, the non-standard variant was not found in the situation of reading aloud. This result, while corroborating production studies, points to the development of sociolinguistic awareness denoted by sensibility to the stylistic context and the selection/lacking of linguistic features.

The pattern of the coda -R deletion as a general trend points to the non-sensibility to stylistic situations, with a relative constant rate of occurrence in the case of internal coda -R deletion, an indicator behavior, which requires an analysis with other reader profiles to resolve the effect of monitoring or the development of sociolinguistic awareness.

Considering the pattern of the controlled linguistic features in the reading aloud, the results indicate that the sociolinguistic awareness contributes to the monitoring of the occurrence of the non-standard variant as the suppression of d- in -ndo and the simplification of complex attacks, which are sensitive to the stylistic context (markers); the coda -R deletion in names does not seem to be sensitive to stylistic monitoring (indicator); and the vocalization of the palatal lateral approximant is highly sensitive, with a restricted distribution in situations of less monitoring and strong restriction in reading. Concerning the other contexts, due to their proximity to prescriptivism and the transparency of the orthographic system, results with other schooling profiles are requested for more reliability in delimiting a social evaluation.

Image 5
– Distribution of linguistic features regarding productivity and recurrence

The occurrence of sociolinguistic features provides clues to sociolinguistic awareness, which affects reading by activating the routes: the occurrence of the feature taken as the pattern in reading aloud signals the activation of the lexical route, with more precision and speed than that the reader who uses the phonological route. The reader with greater accuracy and speed in decoding can expend more cognitive effort in the comprehension and thus have better performance reading comprehension tests, as the official assessments and the RAT applied in this study. The occurrence of standard linguistic features in reading aloud can be a predictive factor of performance in a reading comprehension test. To test this hypothesis, for each sociolinguistic feature, the rate of the standard variant per student and per less formal (short interviews) and more formal (reading aloud task) stylistic situations was computed.

Image 6
– Regression model of the effects of sociolinguistic features and the performance in the reading comprehension test

The linear models for predicting performance in the reading comprehension test and the use rate of the standard variant of each of the sociolinguistic features in the short interviews were not statistically significant (Figure 9, left). The linear models to predict the performance in the reading comprehension test and the rate of use of the standard variant of each of the sociolinguistic feature controlled in the reading aloud (Figure 9, right), except the monophthongization of -ow, were significant, with a great positive effect of using the standard variant in the TCL score, such as the deletion of -d in -ndo (β = 598). 23, p < .01), substantial effect, as the diphthongization, (β = 748.54, p < .01) and monophthongization of -aj, -ej (β = 936.64, p < .01), and medium effect, as the palatal lateral vocalization (β = 328). 12, p < .05) and -R deletion in internal coda (β = 1033.69, p < .05), in verbs (β = 415.99, p < .05), and in complex attack clusters (β = 532.94, p < .05).

Meta-analysis is a strong evidence that sociolinguistic awareness, measured by the rate of occurrence of the standard variant in the community in the stylistic situation of the reading aloud task, has a strong effect on reading performance, measured by the score of the reading comprehension test applied to the same sample of students.

Overall, this result indicates that listening to the student reading aloud can help to identify the reading comprehension capacity. It seems a simple procedure, which does not require specific resources, only a basic instruction for the teacher of the early grades to be aware of sociolinguistic variation. The recommendation for sociolinguistically sensitive early training is also proposed by Barrera and Maluf (2004) and Guimarães (2005) , however, the Brazilian pedagogical reality has not incorporated the approach of linguistic variation in the early grades yet ( FREITAG; SÁ, 2019 ), perhaps due to a lack of detail more focused on the pedagogical practice based on descriptive sociolinguistic studies. The experiences in L2 can be a starting point: Van Compernolle and Williams (2013) argue that sociolinguistic awareness can be developed by instruction; controlled use of linguistic variation can be developed by the students as they internalize their sociolinguistic knowledge as a cognitive tool. The pedagogical recommendations for L2 (but we can extend them to L1 early grades) involve exposing learners to socio-stylistically differentiated linguistic variants, explaining their socio-stylistic values, and providing students with the opportunity to use the variants in communicative tasks.

Schwalm (1985) states that the development of sociolinguistic awareness is much more a problem for speakers of an L2 than for native speakers of a language, since the first group demands much more sociocultural adjustments in relation to the use of a language. Nevertheless, the process of transition to school environment may be an experience that demands the recognition of students, especially when the linguistic variety of the community is more distant than what it is assumed as a prestige variety. After all, the school is a culture other than students’ daily life, and for being successful, they need to learn the values and the variable linguistic features that are characteristic of that culture. Pendergast and Kaplan (2015) reinforce the importance of better training teachers to observe cultural and linguistic effects on reading aloud practices in the early grades. The adoption of these measures may contribute to the reduction of the asymmetry in the results of reading assessments.

Conclusion

The investigation of the pattern of the same linguistic feature in different situations of stylistic monitoring by the same speakers allows us to contribute to the improvement of the scale of social evaluation of sociolinguistic variables, and, since it has been conducted in the school environment and involves a test of reading comprehension as a parameter, it also allows us to make some observations about the development of sociolinguistic awareness and reading performance.

Negative stereotypes, such as vocalization and rhoticism, although present in speech (short interviews), do not transpose into situations of higher stylistic monitoring, such as the appointment and reading tasks. Variables with marker behavior, such as the deletion of -d in -ndo and the simplification of complex attacks have also been blocked in reading aloud.

Monophthongization, diphthongization, and coda -R deletion (especially verbs), linguistic features that are traditionally considered as indicators, in this study, presented a marker behavior, sensitive to the context of stylistic monitoring, but in an opposite direction from what was expected: the non-standard variant is more recurrent in the context of higher monitoring, the reading aloud. In this case, the effect can be associated with the lack of transparency between the orthographic form and the standard realization in the community; as it occurs in early reading, students access the phonological rather than the lexical route.

The results show that the students who had the highest scores of standard variables in the reading aloud task were those who performed better in a reading comprehension test. The transposition of the standard variant of vernacular speech to reading aloud suggests automaticity in the lexical decoding process and consequent cognitive capacity availability for processing.

It may sound obvious, but an empirical observation, such as performance in official evaluations, shows that reality is very distant from the theoretical assumptions. Reversing this situation requires, on the one hand, that descriptive results from sociolinguistic studies be resorted to during the initial education of first language teachers, contributing to the identification of variable features of speech that are transferred to reading and their social value; and, on the other hand, the observation of variable features of speech that are transferred to reading can contribute to the refinement of sociolinguistic descriptions regarding the behavior of marker, stereotypes, and indicators: While negative stereotypes present stable and demarcated patterns, markers seem to work in two directions: one in which the standard variant is recurrent in speech, but does not move to reading (a non-standard variant emerges), and the other in which the non-standard variant is recurrent in speech, and does not move to reading either (this is the most prototypical behavior associated with stereotypes). The diversification of stylistic contexts for the observation of linguistic variation has gained space in the current research agenda (cf. VIEIRA; LIMA, 2019 ), with results that particularize patterns in different genres and textual types. The inclusion of reading tasks and the documentation of linguistic production in the school environment to identify variable features of speech that move to reading can contribute to the development of a diagnosis protocol for reading in a classroom situation.

Acknowledgements

This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001, selected by call POSGRAP/COPGD/UFS 06/2020.

REFERÊNCIAS

• ARAUJO, A. S.; BORGES, D. K. V. Atitudes linguísticas de estudantes universitários: o fenômeno da monotongação em foco. Tabuleiro de Letras , Salvador, v.12, p.97-113, 2019.
• BARRERA, S. D.; MALUF, M. R. Variação lingüística e alfabetização: um estudo com crianças da primeira série do ensino fundamental. Psicologia Escolar e Educacional , Campinas, v.8, n.1, p.35-46, 2004.
• BECHARA, E. Moderna gramática portuguesa . Rio de Janeiro: Nova Fronteira, 2009.
• BELL, A. Language style as audience design. Language in society , Cambridge, v.13, n.2, p.145-204, 1984.
• BRANDÃO, S. F. Um estudo variacionista sobre a lateral palatal. Letras de Hoje , Porto Alegre, v.42, n.3, p.89-99, 2007.
• CALLOU, D.; LEITE, Y.; MORAES, J. A ditongação no português do Brasil: estudo de dois casos. Vivacité et diversité de la variation linguistique , [ s.l .], v.3, p.95-102, 2012.
• CALLOU, D.; MORAES, J.; LEITE, Y. Apagamento do R final no dialeto carioca: um estudo em tempo aparente e em tempo real. DELTA: Documentação de Estudos em Lingüística Teórica e Aplicada, São Paulo, v.14, nesp., 1998. Disponível em: https://doi.org/10.1590/S0102-44501998000300006 Acesso em: 05 ago. 2021
  » https://doi.org/10.1590/S0102-44501998000300006
• CALLOU, D.; SERRA, C.; CUNHA, C. Mudança em curso no português brasileiro: o apagamento do R no dialeto nordestino. Revista da ABRALIN , [ s.l .], v.14, n.1, p.195-219, 2015.
• CARDOSO, P. B.; PINHEIRO, B. F. M.; SILVA, L. S. Variação do segmento /d/ no contexto -ndo: efeitos prosódicos e de leitura. Leitura , [ s.l .], v.2, n.63, p.174-191, 2019.
• COAN, M.; FREITAG, R. M. K. Sociolinguística variacionista: pressupostos teórico-metodológicos e propostas de ensino. Domínios de Lingu@gem , Uberlândia, v.4, n.2, p.173-194, 2010.
• COLTHEART, M.; RASTLE, K. Serial processing in reading aloud: evidence for dual-route models of reading. Journal of Experimental Psychology: human perception and performance, Arlington, v.20, n.6, p.1197-1211, 1994.
• COSTA, L. T. da. Fenômenos variáveis e variantes líquidas produzidas no ataque complexo. Acta Scientiarum, Language and Culture , Maringá, v.35, n.2, p.179-186, 2013.
• CRISTÓFARO-SILVA, T. Branching onsets in Brazilian Portuguese. Revista de Estudos da Linguagem , Belo Horizonte, v.10, n.1, p.91-107, 2002.
• CRISTOFOLINI, C. Estudo da monotongação de [ow] no falar florianopolitano: perspectiva acústica e sociolinguística. Revista da ABRALIN , [ s.l .], v.10, n.1, p.205-229, 2011.
• CYRANKA, L. F. de M.; RONCARATI, C. Atitudes lingüísticas: uma pesquisa em escolas públicas de Juiz de Fora (MG–BRASIL). Diacrítica , Braga, v.1, n.24, p.211-231, 2010.
• ECKERT, P. Variation and the indexical field. Journal of Sociolinguistics , Chichester, v.12, n.4, p.453-476, 2008.
• EHRI, L. C. Orthographic mapping in the acquisition of sight word reading, spelling memory, and vocabulary learning. Scientific Studies of Reading , Philadelphia, v.18, n.1, p.5-21, 2014.
• EHRI, L. C. Learning to read words: theory, findings, and issues. Scientific Studies of reading , Philadelphia, v.9, n.2, p.167-188, 2005.
• FERREIRA, J. S.; TENANI, L. E.; GONÇALVES, S. C. L. O morfema de gerúndio ‘ndo’ no Português Brasileiro: análise fonológica e sociolinguística. Letras & Letras , Uberlândia, p.167-188, 2012.
• FREITAG, R. M. K. Reparos na leitura em voz alta como pistas de consciência sociolinguística. DELTA: Documentação de Estudos em Linguística Teórica e Aplicada, São Paulo, v.36, n.2, 2020. Disponível em: https://doi.org/10.1590/1678-460X2020360206 Acesso em: 05 ago. 2021.
  » https://doi.org/10.1590/1678-460X2020360206
• FREITAG, R. M. K. Uso, crença e atitudes na variação na primeira pessoa do plural no Português Brasileiro. DELTA: Documentação de Estudos em Linguística Teórica e Aplicada, São Paulo, v.32, n.4, p.889-917, 2016.
• FREITAG, R. M. K. Implicações da variação na alfabetização: a lateral palatal e seus correspondentes grafêmicos. Leitura , [ s.l .], v.2, n.46, p.37-56, 2011.
• FREITAG, R. M. K.; CARDOSO, P. B.; PINHEIRO, B. F. M. Saliência na conservação de /d/ no segmento -ndo: efeitos sociais e estilísticos. Gragoatá , Niterói, v.23, n.46, p.654-678, 2018.
• FREITAG, R. M. K.; SÁ, J. J. de S. Leitura em voz alta, variação linguística e o sucesso na aprendizagem inicial da leitura. Ilha do Desterro , Florianópolis, v.72, n.3, p.41-62, 2019.
• FREITAG, R. M. K.; SOUZA, V. R. A. 2019. Discriminação de palavras e efeitos da variação linguística. In: SYMPOSIUM IN INFORMATION AND HUMAN LANGUAGE TECHNOLOGY AND COLLOCATES EVENTS, 12., 2019, Salvador. Proceedings . Salvador: SBC, 2019. p.345-353
• FREITAG, R. M. K. et al. O uso da língua para a discriminação. A Cor das Letras , Feira de Santana, v.21, n.1, p.185-207, 2020.
• GONÇALVES, G. F.; AMARAL, V. S. Produções orais e escritas dos ditongos [aj],[ej] e [ow]: dados de São José do Norte/RS. Diadorim: revista de estudos linguísticos e literários, Rio de Janeiro, v.14, p.127-154, 2014.
• GÖRSKI, E. M.; FREITAG, R. M. K. O papel da sociolinguística na formação dos professores de língua portuguesa como língua materna. In: MARTINS, M. A.; TAVARES, M. A. (org.). Contribuições da Sociolinguística e da Linguística histórica para o ensino de língua portuguesa . Natal: EdUFRN, 2013. p.11-52.
• GUIMARÃES, S. R. K. Influência da variação lingüística e da consciência morfossintática no desempenho em leitura e escrita. Interação em Psicologia , Curitiba, v.9, n.2, p.261-271, 2005.
• HAUPT, C.; SEARA, I. C. Caracterização acústica do fenômeno de monotongação dos ditongos [aj, ej, oj] no falar florianopolitano. Revista Linguagem & Ensino , Pelotas, v.15, n.1, p.263-290, 2012.
• KRISTIANSEN, T. Two standards: one for the media and one for the school. Language Awareness , Clevedon, v.10, n.1, p.9-24, 2001.
• LABOV, W. Language in the inner city: studies in the black English vernacular. Pennsylvania: University of Pennsylvania Press, 1972a.
• LABOV, W. Sociolinguistic patterns . Pennsylvania: University of Pennsylvania Press, 1972b.
• LEIRIA, L. L. A ditongação variável em sílabas tônicas finais travadas por -S. Organon , Porto Alegre, v.14, n.28-29, p.133-141, 2000.
• MACHADO, A. P. G.; FREITAG, R. M. K. Pistas dos processos de decodificação que levam à compreensão da leitura. Letras de hoje , Porto Alegre, v.54, n.2, p.132-145, 2019.
• MACHADO, A. P. G. Variação linguística e leitura: fenômenos variáveis da fala na leitura em voz alta. A Cor das Letras , Feira de Santana, v.19, n.4, p.196-218, 2018.
• MADUREIRA, E. D. Reanálise de alguns aspectos da vocalização da lateral palatal no português. Revista de Estudos da Linguagem , Belo Horizonte, v.8, n.1, p.125-145, 1999.
• MCKENZIE, R. M. The sociolinguistics of variety identification and categorisation: free classification of varieties of spoken English amongst non-linguist listeners. Language Awareness , Clevedon, v.24, n.2, p.150-168, 2015.
• MIRANDA, I. C. C.; CRISTÓFARO-SILVA, T. Aquisição de encontros consonantais tautossilábicos: uma abordagem multirrepresentacional. Revista Linguíʃtica , Rio de Janeiro, v.7, n.1, p.14-31, 2011.
• MOLLICA, M. C. Enfoques de pesquisa sobre a relação língua e sociedade. Veredas-Revista de Estudos Linguísticos , Juiz de Fora, v.5, n.1, p.7-19, 2001.
• MOLLICA, M. C.; MATTOS, P. B. Pela conjugação das abordagens variacionista e difusionista. Revista de estudos da linguagem , Belo Horizonte, v.1, n.1, p.53-64, 1992.
• MOLLICA, M. C.; PAIVA, M. da C. Restrições estruturais atuando na relação entre [l] ~ [ɾ] e [ɾ] ~ [0] em grupos consonantais em português. Boletim da Abralin , Campinas, n.11, p.181-190, 1991.
• MONARETTO, V. N. de O. O apagamento da vibrante posvocálica nas capitais do sul do Brasil. Letras de hoje , Porto Alegre, v.1, n.35, p.275-284, 2000.
• NASCIMENTO, K. R. S.; ARAÚJO, A. A.; CARVALHO, W. J. A. A redução do gerúndio no falar popular de Fortaleza: um olhar variacionista. Veredas-Revista de Estudos Linguísticos , Juiz de Fora, v.17, n.2, p.398-413, 2013.
• NEVINS, A. Vowel lenition and fortition in Brazilian Portuguese. Letras de Hoje , Porto Alegre, v.47, n.3, p.228-233, 2012.
• NICOLOSI, L.; HARRYMAN, E.; KRESHECK, J. Terminology of communication disorders: speech-language-hearing. Philadelphia: Lippincott Williams & Wilkins, 2004.
• PAIVA, M. da C.; SCHERRE, M. M. P. Retrospectiva sociolingüística: contribuições do PEUL. DELTA: Documentação e Estudos em Linguística Teórica e Aplicada, São Paulo, v.15, n.3, p.201-232, 1999.
• PAIVA, M. da C.; SILVA, V. L. P. Cumprindo uma pauta de trabalho: contribuicões recentes do Peul. Alfa: Revista de Lingüística, São Paulo, v.56, n.3, p.739-771, 2012.
• PATIL, I.; POWELL, C. ggstatsplot: “ggplot2” Based Plots with Statistical Details. [ S.l .]: ggstatsplot, 2018.
• PENDERGAST, L. L.; KAPLAN, A. Instructional context and student motivation, learning, and development: commentary and implications for school psychologists. School Psychology International , London, v.36, n 6, p.638-647, 2015.
• PIPER, D. Language awareness in Nova Scotia schools: an exploratory study. Brock Education Journal , St. Catharines, v.13, n.1, p.61-71, 2003.
• SCHWALM, D. E. Degree of difficulty in basic writing courses: insights from the oral proficiency interview testing program. College English , Urbana, v.47, n.6, p.629-640, 1985.
• SILVERSTEIN, M. Indexical order and the dialectics of sociolinguistic life. Language & communication , Oxford, v.23, n.3-4, p.193-229, 2003.
• SQUIRES, L. Processing grammatical differences: perceiving versus noticing. In: BABEL, Anna (ed.). Awareness and control in sociolinguistic research . Cambridge: Cambridge University Press, 2016. p.80-103.
• TOLEDO, E. E. Estudo em tempo real da monotongação do ditongo decrescente/ej/em amostra de Porto Alegre. Letrônica , Porto Alegre, v.6, n.1, p.94-107, 2013.
• TRUDGILL, P. Dialects in contact . Oxford: Blackwell,1986.
• VAN COMPERNOLLE, R. A.; WILLIAMS, L. The effect of instruction on language learners’ sociolinguistic awareness: an empirical study with pedagogical implications. System , [ s.l. ], v.41, n.2, p.298-306, 2013.
• VIEIRA, S. R.; LIMA, M. D. A. de O. Variação, gêneros textuais e ensino de Português: da norma culta à norma-padrão. Rio de Janeiro: Letras UFRJ, 2019.
• WITTENBURG, P. et al. Elan: a professional framework for multimodality research. Proceedings of LREC , p. 1556-1559, 2006.

• 1
  Version of this text was presented at 48 New Ways of Analyzing Variation (NWAV), 2019, in Eugene, Oregon, USA.
• 2
  In addition to these features, the authors reported having controlled the following aspects: lexical variation (catá (=pegar); tacá (=jogar, atirar); buchuda (=grávida); tropicá (=tropeçar); caçá (=procurar); abano (=ventilador); pexera (=faca); coisá (= fazer alguma coisa); mor (= para) e mais (=com)), collapsed [a] in a few words (gulha/guia, bacaxi, rancá, sustô, marelo, ranhá, cabô, inda), insertion of [a] in some words (atropeçô, alembrandu, avoar), sentence constructions considered atypical (mor deli tropcá i caí; eu venhu mais o meu amigu; a bruxa tava durminu mais u gatu; mandô eli num mais robá; não se lembru; aguá as pranta; você ponhava; ela se acordô; meu pai vai demorá de ir; ela vai dá um castigu neli), contractions, agglutinations, abbreviations, and other variations related to the phonetic context and/or to prosodic aspects of the general rhythm of speech (ca/cas (com a/as), cu/cus (com o/os), cê (você), ni (em)), and other frequent variations (árvri/árvi, tamém, ingual).
• 3
  Original - pronúncia ou omissão do “r” nos infinitivos verbais (p. ex.: pronunciar /passear/ ou /passiá/, /cantar/ ou /cantá/, /comprar/ ou /comprá/) [queda do -R em coda em verbos]; - pronúncia ou omissão do “u” final em verbos no pretérito perfeito, terceira pessoa do singular (p. ex.: pronunciar /dançou/ ou /dançô/, /falou/ ou /falô/, /mostrou/ ou /mostrô/) [monotongação de -ow]; - pronúncia ou omissão do “s” final em verbos no pretérito perfeito, primeira pessoa do plural (p. ex.: pronunciar /compramos/ ou /compramo/, /perdemos/ ou /perdemo/, /brincamos/ ou /brincamo/); - pronúncia ou omissão da nasalização do ditongo nasal átono em verbos no pretérito, terceira pessoa do plural (p. ex.: pronunciar /cantaram/ ou /cantaro/, /deram/ ou /deru/). ( GUIMARÃES, 2005 , p.265).
• 4
  Original: “ Quanto aos ditongos, cumpre notar: ai, ei, e ou, na pronúncia, devem guardar, na pronúncia cultivada, sua integridade, não se exagerando o valor do i ou do u, nem os eliminando, como o faz a modalidade distensa: caixa, queijo, ouro. [...]. Normalmente ditongamos, pelo acréscimo de um i, as vogais tônicas finais seguidas de um -z ou -s. Assim não fazemos a diferença entre pás, pais e paz; mas e mais .” ( BECHARA, 2009 , p.60).
• 5
  The process of monophtongization, for example, occurs in a different way in -aj, -ej as the -R final drop is different in nouns and verbs, among other aspects ( MOLLICA, 2001 ).
• 6
  The National Assessment of Alphabetization (ANA) is a census large-scale assessment applied each two years by INEP/MEC to public school students who are enrolled in 2 grade in elementary school. The reference framework of Portuguese Language comprises two structuring axes (reading and writing) and 12 skills. The tests are not available, and there are no commented items either in the Devolutivas Portal (where Prova Brasil items can be found, for example). Considering the skills of the reference framework, similar tests are developed in order to prepare and pre-assess students. The RCT adopted in this study comprises 1 multiple choice item and 3 written production. All skills from the reference framework are present in the test. Apart from skill 1, assessed only in item 1, the other skills are assessed in two items. The exercises were graded with scores attributed according to the reference framework. Details of the complete study may be found in Freitag and Sá (2019) .

Publication Dates

• Publication in this collection
  29 Oct 2021
• Date of issue
  2021

History

• Received
  6 Oct 2019
• Accepted
  10 May 2020