SYNCOPE OF NON-FINAL POSTONIC VOWELS IN THE INLAND OF PARANÁ: AN AUTOSEGMENTAL AND VARIATIONAL ANALYSIS FROM ALIB DATA

ABSTRACT

In this paper, we aim to describe and analyze the phenomenon of syncope of postonic vowel in non-final context, based on data collected in interviews conducted by the project Linguistic Atlas of Brazil (ALiB), in sixteen points of inquiry located in the inland of Paraná, with a total of 64 informants. Thereby, we evaluate the frequency and possible linguistic and extralinguistic factors that can condition syncope/maintenance of non-final postonic vowel of proparoxytones in eleven lexical items: lâmpada (‘light bulb’), elétrico (‘electric’), fósforo (‘match’), pólvora (‘gunpower’), abóbora (‘pumpkin’), árvore (‘tree’), sábado (‘Saturday’), número (‘number’), fígado (‘liver’), vômito (‘vomit’) and hóspede (‘guest’) that are part of the Phonetic-Phonological Questionnaire (QFF in Portuguese) of the ALiB Questionnaire (ALiB PROJECT NATIONAL COMMITTEE, 2001). The investigation rests on the perspective of Pluridimensional Dialectology (THUN, 1998) and in light of the autosegmental phonological model for the syllable, in analysis of the syllabic pattern of Brazilian Portuguese (COLLISCHONN, 1996; BISOL, 1999). We found out that the phonological context of postonic vowels and adjacent segments may interfere in the conditioning of syncope, while the extralinguistic variables are not productive for this process in our corpus .

Linguistic Atlas of Brazil; Syncope; Proparoxytones; Syllable; Pluridimensional Dialectology

RESUMO

Este estudo busca descrever e analisar o fenômeno da síncope da vogal postônica não final no falar paranaense interiorano, com base em dados coletados em entrevistas realizadas pelo Projeto Atlas Linguístico do Brasil (ALiB), em dezesseis pontos de inquérito situados no interior do Paraná, totalizando 64 informantes. Desse modo, foram avaliadas a frequência e possíveis fatores linguísticos e extralinguísticos que podem condicionar a síncope/manutenção da vogal postônica não final das proparoxítonas em onze itens lexicais: lâmpada, elétrico, fósforo, pólvora, abóbora, árvore, sábado, número, fígado, vômito, hóspede , que constam do Questionário Fonético-Fonológico (QFF) do Questionário do ALiB (COMITÊ NACIONAL DO PROJETO ALiB, 2001). A investigação fundamenta-se na perspectiva da Dialetologia Pluridimensional (THUN, 1998) e à luz do modelo fonológico autossegmental para a sílaba, em análise do padrão silábico do português brasileiro (COLLISCHONN, 1996; BISOL, 1999). Constatamos que o contexto fonológico das vogais postônicas e segmentos adjacentes podem interferir no condicionamento da síncope, enquanto as variáveis extralinguísticas não se mostraram producentes para esse processo em nosso corpus .

Atlas Linguístico do Brasil; Síncope; Proparoxítonas; Sílaba; Dialetologia Pluridimensional

Introduction

Proparoxytones, a less productive accentual pattern in the context of Brazilian Portuguese (hereinafter BP), tend to a phonic reduction in the segment following the tonic syllable, which undergoes the syncope of non-final unstressed vowel, turning words into paroxytones. As demonstrated by Silva Neto (1956 [1938]), the tendency to reduce proparoxytones was already common in Vulgar Latin, and was documented in Appendix Probi . This phenomenon went through several phases of Portuguese and, currently, can be observed in popular speech.

Ilari (2004) documents that many divergent forms have evolved in different ways in Romance languages and have originated from Classical Latin (proparoxytones) and from Vulgar Latin (paroxytone words), namely: óculos (‘eyeglasses’) and olho (‘eye’), artículo (‘article’) and artelho (‘articulation’), partícula (‘particle’) and partilha (‘sharing’), coágulo (‘clot’) and coalho (‘rennet’). In relation to the transition from Latin to Portuguese, Ali (1964) postulates that some phonemes within words have disappeared, as in: dedo ( digitu ) (‘finger’), verde ( viride ) (‘green’) etc. Coutinho (1976 , p. 32, author’s emphasis, our translation) highlights the characteristic of Vulgar Latin as a “tendency to avoid proparoxytone words: masclus (masculus), domnus (dominus), caldus (calidus), fricda (frigida), virdis (viridis) ”¹ , which underwent reduction. The philologist already asserted that the sonority of the tonic vowel could be reflected in the final syllable or extended to the nearest consonant, sustaining the permanence of some words in the proparoxytone form.

Câmara Jr. (1985, p. 35, author’s emphasis, our translation), in turn, ratifies what the aforementioned scholars affirm: “Portuguese words stressed in the antepenultimate syllable rarely come from the evolution in Vulgar Latin (e.g. pêssego (‘peach’), from Latin persi ( cum ).”² According to the linguist, most of the proparoxytones come from Classical Latin borrowings, processed primarily from the sixteenth century, among them, some Greek words adapted to the structure of Classical Latin.

The author also explains that, even in the standard language, there is a tendency to reduce and, in popular speech, this phenomenon operates through “[...] suppression of the phonic segment between the stressed vowel and the final vowel (e.g.: Petrópis for the toponym Petrópolis ; exérço instead of exército (‘army’); glóbo replacing glóbulo (‘globule’)”³ (CÂMARA JR., 1985, p. 35, author’s emphasis, our translation).

What the above mentioned linguists and philologists defend clarifies the fact that the reduction of proparoxytones is not a linguistic process restricted solely to Portuguese spoken in Brazil by poorly educated speakers, but it comes from a historical process, which is part of the evolution of the language, in stable variation, as Gomes (2011) also observed.

Therefore, we outline the general goal of this research: to describe the syncope process of one or more non-final postonic segments in proparoxytones, analyzing the syllabic pattern of BP based on an autosegmental phonological model for the syllable and in the light of the Pluridimensional Dialectology, from oral data collected by the Linguistic Atlas of Brazil Project (hereafter ALiB) in the inland of Paraná.

In order to do this, we aim to: i) present the frequency of the maintenance of proparoxytones and syncopated variants in the inland of Paraná; ii) demonstrate, under the autosegmental phonological model for the syllable, which postonic syllabic structures benefit or inhibit syncope in proparoxytone forms; iii) examine the possible extralinguistic factors (gender and age group) and locality that influence the incidence of the phenomenon; iv) contribute to studies about phonetic and phonological facts concerning BP, in order to eliminate possible linguistic prejudice.

From the objectives outlined above, the research questions that lead this work emerge:

1. Is the use of the syncopated form of proparoxytones productive even by speakers living in urban environments?

2. Can social factors, such as gender and age group, favor the process of deletion in the informants’ speech in the inland of Paraná?

3. Is the geographical distribution of syncope of proparoxytones able to demonstrate the vitality of the process in any locality of Paraná?

After a brief summary of the linguistic-historical panorama regarding the reduction of proparoxytones, we proceed to the literature review that documents this phenomenon in regional speech studies.

The reduction of proparoxytones in regional studies

Research on regional speeches, conducted by dialectologists, has demonstrated that, in addition to the recording of the reduction of proparoxytones in Vulgar Latin, such phenomenon takes place in the vocabulary of speakers living in several areas of Brazil.

Amaral (1982 [1920]), in relation to the caipira dialect, finds out that the speakers presented a tendency to reduce the forms that are not recurrent in linguistic uses. In proparoxytones,⁵ the tendency is to suppress the vowel of the penultimate syllable and even the whole syllable, making the word a paroxytone ( rídico ⁵ = ridículo (‘ridiculous’), legite = legítimo (‘lawful’/‘legal’), cosca = cócega (‘tickle’), musga = música (‘music’).

Regarding the popular speech in Alagoas and Pernambuco, Marroquim (2008) comments that words, when changing from Latin to Portuguese, have transgressed the phonetic rule of stressed vowel conservation, moving it forward, as in: límitem > limíte ; océanum > oceáno ; íntegrum > intéiro ; cátedra > cadéira . In the study of the popular Northeastern speech, in the chapter dealing with phonetics, the researcher exemplifies that, with the variants fôrgo for fôlego (‘breath’) and córgo for córrego (‘stream’).

The same dialectologist stated that, following the Principle of Least Effort, the Romans avoided proparoxytones by deleting unstressed vowels after the tonic ones. These words were listed in Appendix Probi and Marroquim (2008) compared them to the phenomenon that occurs in the speech of caboclos: sábado (‘Saturday’) not sabo etc.

In the work O linguajar carioca , Antenor Nascentes (1953 [1922], p. 37, author’s emphasis, our translation) observes that similarly to what happened in Vulgar Latin change to Portuguese, postonic vowels in proparoxytones also underwent syncope, e.g. “ arvore-arve , passaro-passo , polvora-porva , marmore-marme , Alvaro-Arvo , abobora-abobra (or aborba )”⁶ .

Other dialectological and variational works record the phenomenon of reduction of proparoxytones in addition to the above mentioned fundamental works. In the next section, we present recent research, taking into account linguistic and extralinguistic conditionings that may influence this process.

Reduction of proparoxytones in other variational studies

The reduction of proparoxytones was documented by Araújo (2012) in a corpus consisting of 11 lexical items of the Phonetic-Phonological Questionnaire – PPQ (Questionário Fonético-Fonológico – QFF) (ALiB PROJECT NATIONAL COMMITTEE, 2001), in interviews with 200 informants, in 25 Brazilian capitals participating in the ALiB project.

The extralinguistic variables take into account age group (band I - 18 to 30 year-olds and band II - 45 to 60 year-olds), gender (male and female) and schooling (elementary and university level), since four profiles of informants with higher education are added to the inquiries of the ALiB project in each capital. The author did not work with the possible linguistic conditioning for the phenomenon.

For the statistical analysis, the GoldVarb X program was utilized. In a second round of data, 94% of the occurrences of non-suppression of the postonic and 6% of occurrences of deletion were found. In the statistical analysis by region, the researcher observed that the favoring of syncope occurred primarily in Florianópolis and Boa Vista. Belo Horizonte, São Paulo, Rio de Janeiro, Curitiba, Recife, Salvador, João Pessoa, Fortaleza, Natal, Manaus, Macapá, Belém and Porto Velho did not present speakers that favored deletion. In short, Araújo (2012) concludes that the variable schooling (elementary level) favored syncope, as well as locality and age group (band II).

Castro (2008) conducted an investigation about the reduction of proparoxytones in 65 localities that form the corpus of Atlas Linguístico do Paraná – ALPR (‘Linguistic Atlas of Paraná’, AGUILERA, 1994 ). Castro (2008) analyzed the questions whose answers should be the words árvore (‘tree’) , útero (‘uterus’) , eucalipto (‘eucalyptus’) , eclipse (‘eclipse’) , amigdalas (‘tonsils’) , relâmpago (‘lightning’) , glândula (‘gland’) and estômago (‘stomach’). Eclipse and eucalipto were selected, according to the author, because when the speaker pronounces them, he/she usually inserts an epenthetic vowel /i/, producing eclipise and eucalipito.

The variant árvore was recorded in all localities, and in 93% of the responses, paroxytones ( arve , arvre , auve ) were obtained. The reduction of proparoxytones prevailed in the variants árvore , útero , eucalipto and eclipse . In the items amígdalas and glândula , paroxytones and proparoxytones came up in an even way, whereas the reduced form for amígdalas was less frequent.

In general, the aforementioned researcher concludes that paroxytone forms predominate in the ALPR (except for estômago , which favors the maintenance of the proparoxitone form). Other extralinguistic variables chosen by the author were schooling and gender of the informant, whose numbers did not evince great differences.

Starting from the study on Fortaleza speech, Aragão (2000) conducted a study named “Social Dialects of Ceará”, through interviews, spontaneous conversations and interaction between doctor and patient. Data were analyzed according to the variables: a) gender; b) age group (10-11, 14-15, and 18-25 year-olds); c) schooling (former primary, gymnasium, high school), in order to describe and analyze proparoxytones in the speech in Fortaleza, correlating linguistic and sociolinguistic contexts.

In the analysis of social factors, minor differences were verified regarding age group, possibly due to the short age interval between them, as suggested by the author. As for the gender factor, the dialectologist found a small difference between women (71.42% of reductions) and men (69.76%). Regarding schooling, results did not support the idea that education might have influence on the maintenance of proparoxytones, since the students with a higher level of education registered a major percentage of proparoxytones reduction.

In a study based on Variational Sociolinguistics, Amaral (2002) utilizes a corpus of 40 informants from the rural area of São José do Norte, located in Rio Grande do Sul, to investigate the syncope of proparoxytones in order to thoroughly examine linguistic and social factors that favor/disfavor the phenomenon. As for that, the author analyzed the process by means of a quantitative analysis, using the VARBRUL program package.

Among the independent variables applied to verify the deletion or maintenance of the non-final postonic syllable, the one that stood out was the subsequent phonological context – proparoxytones that exhibit /r/ or /l/ in the resyllabification, generating a consonantal group in C + /l/ or /r/, as in petla ( pétala ) (‘petal’) or arvri ( árvore ) (‘tree’). When analyzing the vowel articulation features, the researcher detected that syllables with labial /o/ and /u/ favor syncope, while coronal /e/ and /i/ do not have as much weight. The results concerning previous syllable structure demonstrated that light syllables (with CV structure – pétala ) favor reduction, while closed CVC structures ( véspera – ‘eve’) are unfavorable. The place of articulation of the segment previous to the target vowel was examined as “preceding phonological context”, revealing that the velar consonant /k/ constitutes a more favorable environment for deletion, followed by a labial consonant in onset context, as in: óclus – óculos (‘glasses’).

Regarding social variables, schooling proved to determine the reduction of proparoxytones. In relation to the type of interview, syncope was more frequent in informal than formal contexts; as for gender, the results demonstrated a slight difference, somewhat more significant for men, who delete more. Age group was more significant in determining deletion between older and younger speakers, while the intermediate age groups avoided reduction.

After presenting the panorama of the studies regarding the reduction of proparoxytones, based on oral data, we proceed to the discussion of the theories that underlie our analysis.

Theoretical contribution - Autosegmental phonological model - syllable treatment

From the 1970s, the syllable acquired phonological status in linguistics studies, says Collischonn (1996) . According to the mentioned scholar, based on the theory proposed by Selkirk (1982) , the arboreal representation of the syllable is organized as follows:

Figure 1
– Internal syllable structure

The syllable consists of elements called syllabic constituents, namely: onset (O) and rhyme (R). Rhyme, therefore, is formed by a nucleus (N) and a coda (C). All categories can be empty except the nucleus.

Syllables can also be categorized as light or heavy. The way a syllable is formed determines the syllabic weight. The syllables that present rhymes consisting of solely one vowel (solely the nucleus) are light, and those with nucleus and coda - vowel and consonant or vowel and glide - are heavy.

Lima (2008) comments on the phonotactic constraints based on the syllabic structure, exemplifying with the restriction of the obstruents in coda position in BP. Speakers tend to avoid this segment in a syllabic coda by inserting an epenthetic vowel /i/, forming the CV pattern. Phonotactic constraints may limit the positions of the segments in the syllable structure, involving the onset, nucleus, or coda. Processes involving rhyme, that is, nucleus and coda, are more frequent, according to the author.

The template, according to Selkirk (1982) , is universal for all languages. It is up to each language to take advantage of each part of the template, according to particular restrictions. It is a distinctive characteristic of languages, Collischonn (1996 , p. 101-102, our translation) affirms that “languages differ in terms of the number of segments allowed in each syllabic constituent [...]”, and also adds: “[...] the mold is a general statement about the possible structure of syllables in a given language”⁷ .

In the case of Portuguese, if there is more than one consonant in the left of the syllable rhyme, a complex onset occurs, which, according to Bisol (1999) , admits a maximum of two elements, while the nucleus always consists of a vowel and, in coda position, a maximum of two segments are listed. According to Lima (2008) , solely /r/ and /l/, sound or nasal consonants (besides glides), are permitted in the BP to occupy a coda position. In the following item, the syllabic constituents of the onset and the phonological process of syncope are detailed.

The phenomenon of syncope based on an autosegmental phonological model for the syllable

Syncope consists of the deletion of a phoneme or phonemes within a word. According to Quednau (2002 , p. 79, our translation), there were systematic processes of syncope in the evolution of Lusitanian Romance: “1) syncope of the postonic vowel of Latin proparoxytones, with reduction of the word to become a paroxytone and possible subsequent evolution of the resulting consonant group as in apicula> apicla> abelha (‘bee’) [...]; 2) syncope of sound consonant between vowels, as in mala> maa> má (‘bad’) [...]”⁸ . In a more recent period, the researcher continues, the “stop syncope as the first member of the consonant group in erudite words, as in excepção> exceção (‘exception’) [...]”⁹ still recorded. In this study, we are interested in the deletion of one or more segments in the postonic syllable(s), reducing them to paroxytones, following the phonotactic rules of the language, which determine the positions of each segment in one syllable. Thus, as Amaral (2002, p , p. 102, our translation) explains, the consonantal group that results from the process must “[…] constitute a well-formed onset (ár.vo.re> ar.vre (‘tree’), re.lâm.pa.go> re.lam.po) (‘lightning’) or a well-formed coda (pé.ro.la> per.la) (‘pearl’)”¹⁰ . For this, the syncope triggers the phonological processes of assimilation, resyllabification and restructuring of metric feet, clarifies Lima (2008) .

In BP, the onset position allows up to two segments by the system, which can be stop or fricative labial consonants (first position), followed by lateral liquid or vibrant liquid (second position), forming the obstruent + liquid sequence, in the examples: br, cr, cl, tl etc. When forming a well-formed onset, the segments obey the Sonority Sequencing Principle (SSP), according to Clements (1990 apudAMARAL, 2002 ), arranged according to the scale:

(More sonorous) vowels> glides> liquid> nasal> obstruents (Less sonorous)

The nucleus, in BP, must be constituted by segments of higher sonority (vowels), while its edges are formed of the segments of lower sonority, that is, increasing from onset to nucleus, and decreasing from nucleus to coda, as shown in the diagram:

Figure 2
– Scheme according to degrees of sonority

Thus, sequences of segments in nt (nasal> obstruent) are never in the onset position, as they violate the sound sequence (nasal is louder than obstruent and, in an onset position, the movement should increase). Amaral (2002) postulates that the speakers present this internalized phonatory rule, so it is possible to predict syncope in proparoxytones, the insertion and deletion of certain segments in the words. Lima (2008) explains that the phonological processes occur in the rhyme, therefore, the deletion of the postonic vowel occurs in this syllabic position; when the nucleus is erased, the onset segment can be incorporated into the next syllable or undergo syncope as well.

Lima (2008) postulates that, in the resyllabification process, the floating consonantal segment of the postonic syllable moves to the onset position of the following syllable, or to the coda of the preceding stressed syllable. Consequently, the syllabic foot undergoes a restructuring process, making paroxytone the previously proparoxytone word. The assimilation occurs when the segments assimilate characteristics of the precedent (progressive assimilation) or subsequent (regressive assimilation) segment, listing as examples “ cócega > cóska ” (‘tickle’) for the former, and “ físico> fisco ” (‘physical’) for the latter. In this way, it is inferred that the characteristics of the segments that make up the syllabic structure and its adjacent ones interfere in the resyllabification after the deletion of the postonic vowel, a process that is examined in this work in the variety spoken in the state of Paraná.

In the following item, the studies that support the theory of Pluridimensional Dialectology are reviewed, since the extralinguistic and social factors, as well as the locality in which the phenomena are documented, may also interfere in the phonological processes registered in popular BP.

Pluridimensional Dialectology

Dialectology, as a discipline, has the task of identifying, describing and situating the uses in which a language is diversified, according to the spatial, chronological and sociocultural distribution ( CARDOSO, 2002 ). Through the method of Linguistic Geography, in predetermined spaces, the testimonies of dialectal reality are systematically collected and mapped. Essentially, according to Contini and Tuaillon (1996 , p. 7, our translation), “Dialectology aims essentially at studying geolinguistic variation”¹¹ . In its genesis, it was characterized by the monodimensional aspect of analysis of the linguistic variation phenomenon, by the primacy conferred at the diatopic (spatial) level as the object of investigation. It began with the great national Atlases of Georg Wenker (1881) in Germany, and it was systematized with the publication of the Linguistic Atlas of France by Gilliéron and Edmont (1902-1910). From these studies, several others were developed in Europe and America, following the one-dimensional orientation.

Social, historical and cultural transformations, the change in the degree of human mobility, formerly characterized by isolation and sedentarism; the evolution of the media, human migration from rural to urban environments, among other factors, influenced the dynamics of the language. With the beginning of the sociolinguistic studies in the 60s, by William Labov, progressively, Dialectology began to include social factors. Thus, Pluridimensional Dialectology, according to Thun (1998 , p. 369, our translation) “is characterized by the enlargement of the field of observation and by a work in more developed productivity”¹² .

In addition to the diatopic dimension, other parameters that condition the variation, such as social factors, external to the language, are considered, opening up ways to analyze linguistic facts according to diasexual, diagenerational and diastratic variables, among others. The first one refers to the way in which male and female speakers utilize language; the second one concerns the influence of age group in the linguistic uses; the latter refers to the degree of schooling and its impact on linguistic registers ( CARDOSO, 2010 ).

Immersed in this new panorama of studies that guides geolinguistic studies, in 1996 national ALiB project¹³ arose, which, among other objectives, aims to “[…] describe, on the basis of systematically collected empirical data, the linguistic reality of the country, in respect of Portuguese, providing up-to-date linguistic data not solely of diatopic diversity, but also of diagenerational, diastratic, diageneric and diaphasic variation”¹⁴ (CARDOSO et al ., 2014a, p. 23, our translation). The first results of the project referring to the interviews conducted in 25 Brazilian capitals were published in 2014 (CARDOSO et al ., 2014a, 2014b).

For our study, given the ALiB methodology for surveys in inland localities, we solely verified diasexual, diagerational and diatopic variables. In the item below, the methodological procedures that guided our analysis are outlined.

Methodological procedures

The corpus of this research consists of data extracted from the database of the ALiB project, whose collection is conducted in loco and follows the assumptions of Pluridimensional Dialectology ( THUN, 1998 ), which associates traditional Dialectology with Sociolinguistics. The interviews, transcripts and revisions were conducted by the Paraná team of the ALiB project, under the coordination of Dr. Vanderci de Andrade Aguilera.

We selected eleven lexical items, namely: lâmpada ( lamp ) , elétrico ( electric ) , fósforo ( match ) , pólvora ( gunpowder ) , abóbora ( pumpkin ) , árvore ( tree ) , sábado ( saturday ) , número ( number ) , fígado ( liver ) , vômito ( vomit ) , hóspede ( host ) of the Phonetic-Phonological Questionnaire (QFF) of the ALiB Questionnaire (ALiB PROJECT NATIONAL COMMITTEE, 2001). We selected all the proparoxytone items discussed in the QFF of this project, which, in turn, rests on other geolinguistic corpora studies, such as the published linguistic atlases, in order to facilitate the intercomparison of linguistic phenomena in several studies. The answers given throughout the interview were recorded, either in the first or second response, as in other parts of the recording.

We emphasize that the QFF presents directed questions, eliciting the linguistic environment for the study of proparoxytone words, providing short and direct answers, besides being a moment of the interview in which the speaker pays more attention to his/her speech, in a continuum from a less formal to a more formal level. In the following section, extracted from the transcripts of the ALiB project, we present an interview model:

INQ.- What is the name of the thing that is put in fireworks so that they burst?

INF.- Pólvora (‘gunpowder’).

The sample of this research consisted of 64 informants, distributed in 16 localities in the inland of the state of Paraná: Nova Londrina, Londrina, Terra Boa, Umuarama, Tomazina, Campo Mourão, Cândido de Abreu, Piraí do Sul, Toledo, Adrianópolis, São Miguel do Iguaçu, Imbituva, Guarapuava, Morretes, Lapa and Barracão.

Informants are stratified according to the social variables presented in Table 1 .

Table 1
– Informants’ profile

Informant	Schooling	Age Group	Gender
01	Elementary	I (18-30 year-olds)	Male
02	Elementary	I (18-30 year-olds)	Female
03	Elementary	II (50-65 year-olds)	Male
04	Elementary	II (50-65 year-olds)	Female

Source

In order to observe the syncope or maintenance of the non-final postonic vowels, we collected the items by listening to the interviews and subsequent tabulation of the data. With the spreadsheet set up, we quantified the data in percentage and absolute numbers.

Then, for the items that presented a significant syncope index, we outlined models of representations of the syllabic restructuring, analyzing the syllabic pattern of BP based on an autosegmental phonological model for the syllable.

Finally, the quantitative analysis of the extralinguistic variables gender and age group, and the distribution in the sixteen points of inquiry surveyed were contemplated by means of the organization of a comparative table of the productivity of proparoxytonic realizations and syncopated forms.

Data analysis

Quantitative analysis of maintenance and/or syncope of the postonic vowels

Most of the items under study were isolated from prosodic contexts larger than the phonological word, which is a characteristic of the ALiB Phonetic-Phonological Questionnaire, in order to observe whether postonic syncope is related to the formation and shape of the foot.

The syncope processes verified consist of the deletion of the non-final postonic vowel, which sometimes culminates in the deletion of other segments. Thus, as cases of syncope, we consider the deletion of non-final postonic vowel and syncope of non-final postonic vowel allied to the deletion of consonant in the onset position of the last syllable. In others, we include the other phenomena observed during the collection, but which do not fit the analyzed process, such as the apocope. Finally, as non-answer, we grouped the cases in which the questioner did not pose the question and in which the informants were not able to respond or preferred other items that were not pertinent to the study, as in question 127 of the QFF, in which the informants recorded the verbal forms ( vomitô, gomitô, vomitá, vomitado ).

Table 1 lists the data obtained in absolute and percentage numbers.

Table 1
– Results obtained in the perceptual analysis of the data

Variants	Total records						Non-answer
	Maintenance		Syncope		Others		Total
	Total	%	Total	%	Total	%
Lâmpada	62	98.41	1	1.59	-	-	1
Elétrico	58	95.08	2	3.28	1	1.64	4
Fósforo	58	82.86	10	14.28	2	2.86	-
Pólvora	53	86.89	8	13.11	-	-	4
Abóbora	58	77.4	16	21.4	1	1.3	-
Árvore	60	83.3	12	16.7	-	-	1
Sábado	64	98.5	1	1.5	-	-	-
Número	64	100	-	-	-	-	-
Fígado	58	95.08	3	4.92	-	-	3
Vômito	43	81.13	-	-	10	18.87	13
Hóspede	55	96.5	2	3.5	-	-	7
	633	90.17	55	7.83	14	2	33

Source

The pattern that emerges from Table 1 indicates that the reduction of proparoxytones in relation to productivity evinced the preference in words like: abóbora (16 occurrences), árvore (12), fósforo (10) and pólvora (8), with less occurrences in: fígado , hóspede and elétrico , and the hápax legomena (a record) of the stigmatized variants for lâmpada and sábado . The maintenance of the proparoxytone was hegemonic for the items número and vômito , making up 100% of the occurrences, indicating a possible unfavorable context for syncope, which will be analyzed below.

In the study by Castro (2008) , reduction was also predominant in variants for árvore . The results of the research conducted by the aforementioned researcher on the corpus of the ALPR (1994) in the inland of Paraná indicated the predominance of the reduced forms; in contrast, the current linguistic scenario, based on the ALiB data in places in the inland of the state of Paraná,¹⁵ demonstrates the primacy of maintaining proparoxytone forms. The percentages of maintenance of standard variants (90.2%) and of syncopated forms (7.7%) were obtained, numerical dimensions close to the percentage of the research in capitals with the ALiB corpus, conducted by Araújo (2012) .

In the following section, we analyze the process of ressyllabification, based on the analysis of the syllabic pattern of BP according to the autosegmental phonological model for syllables ( COLLISCHONN, 1996 ; BISOL, 1999 ).

Examination of syllabic restructuring for syncopated variants

We proceeded to the analysis of the phenomenon according to the order of the items listed in Table 1 .

a) Reduction of the proparoxytone in lâmpada:

We recorded solely one case of reduction of the proparoxytone lâmpada (‘lamp’) in the corpus , possibly an idiosyncrasy of the speaker, since, when verifying the whole interview, we could observe his/her tendency to delete word final segments.

The unstressed non-final dorsal postonic vowel /a/ undergoes syncope, therefore, the bilabial stop consonant (obstruent) /p/ becomes a floating segment and attempts ressylabification in the final syllable, in onset position, but this complex onset in /pd/ is not allowed in BP. Therefore, the alveolar stop consonant (also obstruent) /d/ of the final syllable is deleted and the /p/ segment is resyllabified, forming a restructured syllable according to the syllabic pattern of BP: lâm.pa.da> lâm.pØ.da > lâm.pØ.Øa> lâm.pa. In fact, Lima (2008 , p. 93, author’s emphasis, our translation) certifies in his research regarding the following phonological context that obstruents and nasals in a complex onset position do not favor the process, since “[…] these segments are not licensed for second position of a complex onset. Thus, in our data, when a vowel was deleted, the obstruents were deleted along with the postonic vowel ( relâmpago > relampu ; lâmpada (‘lamp’) > lampa )” ¹⁶ .

The author complements that the favorable context in this position is the one formed by the vibrant liquid vowel (more propitious) or the lateral liquid vowel, a statement that was verified throughout the present study.

b) Reduction of the proparoxytone in elétrico:

Solely two occurrences of the reduced forms for the proparoxytone elétrico in the corpus were obtained. After the deletion of the coronal postonic vowel /i/, the onset on obstruent /t/, followed by a vibrant liquid /r/ remains a floating segment, but can not be added to the posterior syllable in onset position, since this position is not allowed according to the rules governing the syllabic mold of BP; obstruent /k/ is syncopated and the last syllable is formed by complex onset, in /tr/: elé.tri.co > elé.trØ.co > elé.trØ.Øo > elé.tro.

c) Reduction of the proparoxytone in fósforo:

The standard form fósforo underwent syncope of the unstressed non-final postonic labial vowel /o/, and obstruent /f/ became a floating segment, forming a complex onset /fr/ in the subsequent syllable: fós.fo.ro (‘match’ )> fós.fØ.ro> fós.fro , with three occurrences. On the other hand, six records were obtained for the reduced variant fósfo , adding the process to the deletion of the vibrant liquid consonant; with this, in the resyllabification, the floating consonant /f/ forms a simple onset in the last syllable: fós.fo.ro> fós.fØ.Øo> fós.fo . Also, an informant recorded the syncopated variant fósso , which reveals the addition of the syncope phenomenon to the phonological process of total progressive assimilation of the contiguous alveolar fricative consonant /s/ by the labiodental consonant /f/, given the familiarity of the sonorous features of the two fricative segments: fós.fo.ro> fós.fØ.Øo> fós.fo> fós.so.

d) Reduction of the proparoxytone in pólvora:

The proparoxytone variant pólvora went through syncope in the non-final postonic labial vowel /o/, making the obstruent /v/ a floating segment, which is integrated with the vibrant liquid consonant /r/ in the following syllable, producing a new syllable according to the BP syllabic template: pól.vo.ra> pól.vØ.ra> pol.vra, in solely one of the records. The speakers from Paraná favored, as a syncopated variant, the reduced form pórva , with seven occurrences. The rhotacism of /l/ for /r/ through regressive assimilation of the vibrant feature and the subsequent deletion of the consonant /r/: pól.vo.ra> pól.vØ.ra> *pór.vra> pór.vØa> pór.va were added to the process of syncope of the labial vowel /o/. In fact, Câmara Jr. (1985) emphasizes the variation between the use of the consonant /r/ or /l/, with preference for the first one, indicating cases of free variation documented in the literature, for frecha rather than flecha (‘arrow’). In fact, the data reveal this inclination by speakers from Paraná towards the use of /r/ (in pórva ) rather than /l/ (in pólva ) as a syncopated variant.

e) Reduction of the proparoxytone in abóbora:

In the work O linguajar carioca (The carioca language), Antenor Nascentes (1953) emphasizes the tendency for the reduction of abóbora to abóbra , which is similar to the process that occurred in Vulgar Latin. In the process, the labial non-final postonic vowel /o/ was deleted and the obstruent /b/ became a floating segment, going through the process of resyllabification to the following syllable. Thus, the syllable forms a complex onset /br/, in a well-formed construction, permitted by BP phonotactics: abó.bo.ra> abó.bØ.ra> abó.bra. This phonological process proved to be productive in our corpus , adding up to fifteen occurrences among the recorded cases of syncope.

We highlight that the following phonological context, formed by a vibrant liquid, has been documented in Lima (2008) and Amaral (2002) as a favorable environment for syncope. In this way, the simple onset becomes a complex one, by means of the stop /b/ in the first position, followed by the vibrant liquid /r/. The complex onsets formed by the vibrant in the second syllabic position are more productive than those formed by the lateral (tl, dl, for example), as reinforced by the researchers.

In addition to the syncope of the postonic vowel, a case of syncope of the non-final postonic vowel added to the regressive metathesis was recorded, according to Hora, Telles and Monaretto (2007), in: abó.bo.ra> abó.bØ.ra> abó.bra> abró.ba. It represents a phenomenon of reordering segments in the phonological sequence, which is not as recurrent as assimilation or deletion, as explained by Hora, Telles and Monaretto (2007). Ali (1964) highlights that metathesis was productive in Old Portuguese, privileged by the syllabic context formed by a vibrant. Because it is not within the scope of this research, we refrain from delving deeper into the study of this phenomenon.

f) Reduction of the proparoxytone in árvore:

The standard form árvore , in turn, became arve following the syncope of the non-final labial vowel /o/ and the deletion of the vibrant /r/, located in the onset position of the final syllable. The speakers preferred the paroxytone word formed by this phenomenon, which constitutes a simple onset by the fricative /v/, with eight occurrences: ár.vo.re> ár.vØ.Øe> ár.ve. Nascentes (1953) documents this reduced variant in Old Portuguese.

Two syncopated variants alve were recorded, resulting from the syncope of the postonic vowel and the liquid /r/, in addition to the exchange of vibrant /r/ for lateral /l/, possibly caused by hypercorrection. Azambuja (2012) explains that hypercorrection is a mistake made by the speaker who wants to censor certain linguistic traits through the imitation of characteristics which, in his/her own view, constitute the standard form, because of an ideological construct.

Two forms were also obtained with syncope of labial vowel /o/ and resyllabification of floating fricative /v/, generating a complex onset in the following syllable: ár.vo.re> ár.vØ.re> ár.vre. Indeed, in the research conducted in the inland of Paraná, Castro (2008) and, similarly, Ramos and Tenani (2009) , in the variety spoken in the Northwest of São Paulo State, already acknowledged the reduced variants árve and árvre.

g) Reduction of the proparoxytone in sábado:

Marroquim (2008) documented that the caboclos in Alagoas and Pernambuco, by means of the Principle of Least Effort, similarly to the Romans in relation to Vulgar Latin, avoided the use of proparoxytones, e.g.: sábado > sábo . One can consider this process, since a syllabic structure CV.CV with a simple onset was formed, also simplifying the duration of emission of the word as a whole, reducing the number of syllables. Lima (2008) agrees with this prerogative, in relation to the simplification of the word, by the decrease of the number of syllables added to the simple syllabic structuring.

After the syncope of the dorsal postonic low vowel /a/, obstruent /b/ became a floating segment and lost its syllabic nucleus. Resyllabification of this segment in onset position in the following syllable is forbidden by the phonotactics of BP, as it would form the cluster /bd/. Consequently, alveolar stop segment /d/ also undergoes syncope and bilabial /b/ can assume the onset position in the following syllabic structure: sá.ba.do> sá.bØ.do> sá.bØ.Øo> sá.bo.

We assume that the reason why there is solely one record of reduction in this word in our corpus is the familiarity speakers have with this expression in their daily life, attested by the absence of non-answer to the question. Because it is an expression that can be easily seen in calendars and personal organizers, the speaker possibly tends to maintain its standard written form.

h) Reduction of the proparoxytone in número:

Maintenance of the proparoxytone for the word número has proved to be hegemonic in this study. We justify the non-realization of deletion in this word with the configuration of the syllable after the hypothetical syncope of its non-final postonic vowel, since the deletion of coronal vowel /e/ would turn bilabial nasal consonant /m/ into a floating segment, producing a consequent resyllabification and formation of a prohibited syllable onset /mr/.

However, in his research, Amaral (2002) attested the nasal consonant in onset position in número and such phenomenon can be verified in European Portuguese, according to Silva (2014 , p. 90, our translation), who makes reference to them (the most vibrant nasal in the onset) as “anomalous consonant groups occurring in postonic position” ¹⁷ . However, in general, speakers of BP tend to avoid reduction of non-final postonic vowels when the vowel is preceded by a nasal consonant and followed by a vibrant liquid consonant in the subsequent syllable onset. Another possible explanation for the maintenance of número in our corpus is the speakers’ familiarity with this expression, reinforced by the absence of non-answers to the question.

i) Reduction of the proparoxytone in fígado:

Three occurrences of the reduced variant figo were recorded. The canonical form fígado went through syncope of the postonic low vowel /a/; obstruent /g/ had its resyllabification in the following syllable restricted, as the onset /gd/ is not permitted according to the sonority scale, consequently forcing the deletion of obstruent /d/. Obstruent /g/ takes the simple onset position in the resyllabification, restructured as a light syllable formed by CV: fí.ga.do> fí.gØ.do> fí.gØ.Øo> fí.go. If the Principle of Least Effort is taken into account, according to Lima’s (2008) analysis, the syncopated word in question was simplified in relation to the number of syllables and its structuring in CV.CV.

j) Reduction of the proparoxytone in vômito:

The word vômito did not present cases of syncope in its postonic segments. The explanation for the absence of deletion in these words is the syllable configuration after the hypothetic deletion of its non-final postonic vowel. Deletion of coronal vowel /i/ would make bilabial nasal /m/ a floating consonant; the resyllabification of these segments and the formation of a complex onset in the following syllable in /mt/ would violate the phonotactically motivated rules of BP, and the sonority scale (nasal > obstruent - the nasal is more sonorous than the obstruent /t/).

In fact, Lemle (1978) cites elétrico and vômito as words not prone to suppression. Amaral (2002 , p. 108, our translation) claims that “[...] proparoxytones more resistant to syncope have /i/ as a postonic vowel and the surrounding contexts do not represent well-formed consonant groups”¹⁸ . Such prerogative may extend to the analysis of the word vômito [ mt ]. In the present study, the forms vômi and vumíto were recorded; in the first, the deletion of the last syllable makes the word a paroxytone, in the latter, the displacement of stress, probably in analogy to the verb vomitar , also forms a paroxytone. Both examples maintain well-formed consonant groups, not transgressing the analysis made by the researcher.

k) Reduction of the paroxytone in hóspede:

Solely two reductions of proparoxytone hóspede were pronounced by the speakers from Paraná. Coronal non-final postonic vowel /e/ suffered deletion and obstruent /p/ became a floating segment; in the resyllabification in the onset of the subsequent syllable, it would form a cluster forbidden by BP phonotactics: /pd/ (obstruent-obstruent). The coronal vowel of the last syllable, which constitutes the nucleus of the rhyme, remains, and obstruent /d/ undergoes syncope, so that a syllable in CV is formed, allowed by the phonological rules governing the system: hós.pe.de > hós.pØ.de> hos.pØ.Øe> hós.pe.

Regarding the analysis of syncope, we found that the most productive reduced variants were formed after the deletion of the postonic vowel and consequent deletion of the consonant in the following syllable in thirty-one cases. The phenomenon of the deletion of non-final postonic vowel proved to be efficient in twenty-four reduction occurrences.

In addition to the analysis according to the autosegmental phonological model for the syllable, we observed the extralinguistic factors that may influence certain processes. Thus, in the following section, we investigate whether variables such as gender, age group and locality motivate the syncope or maintenance phenomena.

Analysis according to extralinguistic variables

Sociolinguistic research generally highlights “a greater feminine awareness of the social status of linguistic forms” ¹⁹ ( PAIVA, 2007 , p. 35, our translation). Therefore, in general, women are more sensitive to the more prestigious variants and tend to utilize them more than men.

Data from Chaves (2011) , Head (1986) , Amaral (2002) , Lima (2008) and Castro (2008) indicate that male informants have a higher rate of syncope application. In contrast, the research conducted by Aragão (2000) demonstrates that women registered more non-standard forms.

In fact, as expected, women tended to maintain proparoxytones, accounting for 330 maintenance records and solely 21 cases of syncope, while men presented 303 occurrences of the standard form and 34 syncopated variants (as shown in Table 2 ).

Table 2
– Maintenance or syncope according to the variable gender

Gender	Total of occurrences	Maintenance		Syncope
		Total	%	Total	%
Female	351	330	94.02	21	5.98
Male	337	303	89.9	34	10.1

Source:

Percentage data reveal that the gender variable was somewhat more prevalent in relation to the maintenance of the proparoxytone, the most socially prestigious form, accounting for 94.02% of the responses given by women, against 89.9% by men. The rate of registration of stigmatized variants was higher among men, accounting for 10.1% of the responses, while females registered 5.98% of syncopated variants. Albeit the difference does not seem to be significant, we found that the data predisposes the interpretation of a tendency among women to speak favoring prestigious forms, according to Castro (2008 , p. 117, our translation).

[...] there is a minimal superiority of women in the use of prestigious (proparoxytone) forms, and a somewhat lower number in the use of stigmatized forms (paroxytones), which is in harmony with previous research findings that demonstrate greater adhesion of female speakers to forms of prestige. ²⁰

On the other hand, the variable age group allows one to observe indications of the stage in which a certain variant is found in a linguistic system, thus, the relation between the age of the speakers within the sample and the production of change may demonstrate signs that this phenomenon is in a process of stable variation or change in progress.

The data presented in Table 3 demonstrate that age group is an extralinguistic variable which is not very significant for the application or retention of the syncope phenomenon, since the difference between maintenance and the percentage of reduced variants in Bands 1 and 2 is around just over 3%. Similarly, Chaves (2011) and Lima (2008) affirmed that age group is not responsible for conditioning the realization of the observed phenomenon.

Table 3
– Maintenance or syncope according to the variable age group

Age group	Number of occurrences	Maintenance		Syncope
		Total	%	Total	%
Band 1	336	303	90.18	33	9.82
Band 2	352	330	93.75	22	6.25

Source,

Maintenance of proparoxytones prevailed in both age groups, recorded in 90.18% of valid responses for Age group 1 and in 93.75% of the responses for Age group 2. Syncope presented little productivity, with 9.82% in Age group 1 and 6.25% in Age group 2. Age group 1 subtly favors the least prestigious form in comparison to Band 2, which is in accordance with the research by Araújo (2012) and Amaral (2002) , in which individuals of the second age group favored syncope. Given the small difference in percentage between the bands, it is not possible to assure whether it would be a process of stable variation or change in progress.

In Table 4 , the possible influence of locality on either maintenance or syncope was measured.

Table 4
– Maintenance or syncope according to the variable locality

Locality	Total occurrences	Maintenance		Syncope
		Total	%	Total	%
Nova Londrina	43	39	90.7	4	9.3
Londrina	42	41	97.6	1	2.4
Terra Boa	43	41	95.35	2	4.65
Umuarama	44	39	88.64	5	11.36
Tomazina	45	39	86.67	6	13.33
Campo Mourão	41	41	100	-	-
Cândido de Abreu	41	38	92.69	3	7.31
Piraí do Sul	43	42	97.67	1	2.33
Toledo	41	36	87.8	5	12.2
Adrianópolis	41	37	90.25	4	9.75
São Miguel do Iguaçu	45	43	95.56	2	4.44
Imbituva	45	42	93.33	3	6.67
Guarapuava	46	40	86.96	6	13.04
Morretes	42	42	100	-	-
Lapa	45	39	86.67	6	13.33
Barracão	41	34	82.93	7	17.07

Source,

We can infer from Table 4 that the syncopated variants obtained the highest record by the speakers from Barracão, adding up to seven occurrences (17.07%) of valid answers for the research in that location. We also note that, at this point of inquiry, the highest non-answer rate was obtained. This city is located in the Southwest of Paraná, on the triple border between Argentina, Santa Catarina and Paraná, with solely 9,735 inhabitants ²¹ (IBGE, 2010). Given its “small town” character, one could suggest a relative influence of location on the characteristics of the speech of the barraconenses .

Subsequently, in other localities, such as Lapa, Guarapuava and Tomazina, six occurrences of syncope were obtained, and in Toledo and Umuarama, five. Each one is located in a region in the inland of Paraná. The lowest reduction rates were recorded in São Miguel do Iguaçu and Terra Boa, with two occurrences each, and in Londrina and Piraí do Sul, with one occurrence of the stigmatized form. Our attention was drawn to the absolute record of the maintenance of proparoxytones by the residents of Campo Mourão and Morretes, with the hegemony of the standard variant.

When we contrast our data, with its predominance of prestigious forms, to the results obtained by Castro (2008) , which presented a majority of stigmatized variants, we can suggest a social change brought about by urbanization and manifested in the speech of the inhabitants of Paraná.

Final remarks

As pointed out by Câmara Jr. (1985), the syllable in which the syncopated nucleus is found is a non-final atonic. Still, in response to the initial research questions, we verified that the results indicate that, in general, the maintenance of proparoxytones was mostly present in the urban speech of speakers from the inland of Paraná, in comparison to other studies. When we contrast our research with the work of Castro (2008) , which, on the contrary, indicated the strength of the non-standard variant, we were able to verify that the geographic reach of prestige forms in the inland of Paraná was clearly evidenced, revealing the vicissitudes through which the language runs. A conjecture to explain the formation of this new linguistic panorama is reflected in the fact that society has undergone many changes, such as intense urbanization and schooling, culminating in the modification of social references that have secured a place in speech.

Regarding the syncope of non-final postonic vowels, as a phonological process, we determined its conditioning to the maintenance of the same syllabic positions of the floating segments that, by means of resyllabification, occupied the syllable onset position, forming a syllabic template that respects the structure of the system and follows specific formation conditions. In accordance with the processes of evolution in the formation of Portuguese, reduction was performed preferably after the postonic vowel and the consonant in the final syllable were deleted.

It is possible to affirm that the characteristics of the segments that form the syllabic structure and its adjacent segments can interfere in resyllabification, which explains the preference for deletion of certain segments of postonic syllables by the speakers. The following phonological context, with a liquid in the onset position, which, after the phenomenon, becomes the second consonant of a complex onset, favored the deletion process, provided that the pattern of the BP syllabic template was respected. In this context, obstruents also conditioned syncope, except in cases in which vowel deletion would incur an onset not permitted by the system.

Regarding the observed social variables, factors such as gender and age group proved to be fruitless quantitatively as determinants of the processes under investigation, albeit women and Band 2 preferably chose the standard canonical form, with a small differential percentage in comparison to men and Band 1. As for the localities, there was a prevalence of the prestigious form in the speech of speakers from Paraná, and in Barracão, in the Southwest, the highest rate of non-standard variants was recorded.

With this work, we hope to contribute to the revival of studies on phonological processes and geolinguistic investigations of variation, thus promoting respect for the multiple forms of expression, extending such deference to the speakers themselves.

Acknowledgements

We thank CAPES and the Araucária Foundation (CAPES Agreement/Araucária Foundation CP 11/2014) and CAPES DS (Portaria CAPES, nº 76, of 04/14/2010), for the scholarships that enabled the accomplishment of this study and other studies concerning linguistic description. We also express gratitude towards the Scientific Committee of the ALiB project, as well as the informants inquired for making their database available.

REFERÊNCIAS

• AGUILERA, V. de A. Atlas Lingüístico do Paraná . Curitiba: Imprensa Oficial do Estado, 1994.
• AMARAL, A. O dialeto caipira: gramática, vocabulário. São Paulo: Hucitec/INL, 1982 [1920].
• AMARAL, M. P. do. A síncope em proparoxítonas: uma regra variável. In: BISOL, L.; BRESCANCINI, C. (org.). Fonologia e variação: recortes do português brasileiro. Porto Alegre: EDIPUCRS, 2002. p. 99- 126.
• ALI, M. S. Gramática histórica da língua portuguesa . São Paulo: Melhoramentos, 1964.
• ARAGÃO, M. do S. S. As palavras proparoxítonas no falar de Fortaleza. Acta Semiotica et Linguistica , São Paulo, v.8, p.61-88, 2000.
• ARAÚJO, A. A. de. A redução das proparoxítonas a partir dos dados do Projeto Atlas Linguístico do Brasil. Revista (CON)TEXTOS Linguísticos, Vitória. v.6, n.7, p.7-19, 2012.
• AZAMBUJA, E. B. Fatos de “hipercorreção” como um observatório do funcionamento ideológico e da memória discursiva da/na língua. In: ENCONTRO DE ESTUDOS DE LINGUAGEM, 4., 2011. Anais [...], Pouso Alegre: ENELIN, 2012. p. 1-9.
• BISOL, L. A sílaba e seus constituintes. In: NEVES, M. H. M. (org.). Gramática do português Falado. Campinas: Editora Humanista: FFLCH/USP, 1999. p. 701-742.
• CÂMARA JR. J. M. História e estrutura da língua portuguesa . 4. ed. Rio de Janeiro: Padrão, 1985.
• CARDOSO, S. A. M. et al. Atlas Linguístico do Brasil . Londrina: EDUEL, 2014a. v.1.
• CARDOSO, S. A. M. et al. Atlas Linguístico do Brasil . Londrina: EDUEL, 2014b. v.2.
• CARDOSO, S. A. M. Geolinguística: tradição e modernidade. São Paulo: Parábola Editorial, 2010.
• CARDOSO, S. A. M. A Geolingüística no terceiro milênio: monodimensional ou pluridimensional? Revista GELNE , Fortaleza, v.4, n.2, p.1-16, 2002. Disponível em: https://periodicos.ufrn.br/gelne/article/download/9088/6442/0 Acesso em: 22 jan. 2019.
  » https://periodicos.ufrn.br/gelne/article/download/9088/6442/0
• CASTRO, V. S.. A redução das proparoxítonas no português popular do Brasil: estudo com base em dados do Atlas Linguístico do Paraná. Estudos Linguísticos , São Paulo, v.37, n.2, p. 113-121, maio-ago. 2008.
• CHAVES, R. G. A redução de proparoxítonos na fala do Sul do Brasil. 2011. 173f. Dissertação (Mestrado em Letras) - Pontifícia Universidade Católica do Rio Grande do Sul, Porto Alegre, 2011. Disponível em: http://fonapli.paginas.ufsc.br/files/2013/03/Disserta%C3%A7%C3%A3o-RaquelChaves.pdf Acesso em: 10 fev. 2019.
  » http://fonapli.paginas.ufsc.br/files/2013/03/Disserta%C3%A7%C3%A3o-RaquelChaves.pdf
• COLLISCHONN, G. A sílaba em português. In: BISOL, L. (org.). Introdução a estudos de fonologia do português brasileiro . Porto Alegre: EDIPUCRS, 1996. p. 99-131.
• COMITÊ NACIONAL DO PROJETO ALiB. Atlas Lingüístico do Brasil: Questionários 2001. Londrina: Eduel, 2001.
• CONTINI, M.; TUAILLON, G. Introduction. In: ATLAS LINGUISTIQUE ROMAN [ALiR]. Présentation . Roma: Istituto Poligrafico e Zecca dello Stato, 1996. v.1. p. 1-18.
• COUTINHO, I. de L. Pontos de gramática histórica . 7. ed. rev. Rio de Janeiro: Ao Livro Técnico, 1976.
• FERREIRA, A. B. de H. Dicionário Aurélio da língua portuguesa . 5 ed. Curitiba: Positivo, 2010.
• GILLIÉRON, J.; EDMONT, E. Atlas Linguistique de la France . Paris: Honoré Champion, 1902-1910.
• GOMES, D. K. A síncope das vogais postônicas não-finais: variação na fala popular urbana do Rio de Janeiro. Diadorim: Revista de Estudos Linguísticos e Literários, Rio de Janeiro, n.8, p. 81-102, 2011.
• HEAD, B. F. O destino de palavras proparoxítonas na linguagem popular. In: ENCONTRO DE VARIAÇÃO LINGÜÍSTICA E BILINGÜISMO NA REGIÃO SUL, 4., 1986, Porto Alegre. Anais [...] Porto Alegre: UFRS, 1986. p. 38-56.
• HORA, D. da; TELLES, S.; MONARETTO, V. N. O. Português brasileiro?: uma língua de metátese?. Letras de Hoje , Porto Alegre, v.42, n.2, p.178-196, set. 2007.
• ILARI, R. Lingüística românica . 3. ed. São Paulo: Ática, 2004.
• LEMLE, M. Heterogeneidade dialetal: um apelo à pesquisa. Linguística e ensino do vernáculo , Rio de Janeiro, v. 53/54, p.60-94, abr./set. 1978.
• LIMA, G. de O. O efeito da síncope nas proparoxítonas: análise fonológica e variacionista com dados do sudoeste goiano. 2008. Dissertação (Mestrado em Linguística) - Programa de Pós-Graduação em Linguística, Universidade Federal de Uberlândia, Uberlândia, 2008.
• MARROQUIM, M. A língua do Nordeste: Alagoas e Pernambuco. 4. ed. Maceió: EDUFAL, 2008.
• NASCENTES, A. O Linguajar Carioca . 2. ed. Rio de Janeiro: Organizações Simões, 1953 [1922].
• PAIVA, M. da C. Transcrições de dados linguísticos. In: MOLLICA, M. C. Introdução à Sociolinguística: o tratamento da variação. 3. Ed. São Paulo: Contexto, 2007. p. 101-116.
• QUEDNAU, L. R. A síncope e seus efeitos em latim e em português arcaico. In: BISOL, L.; BRESCANCINI, C. (org.). Fonologia e variação: recortes do português brasileiro. Porto Alegre: EDIPUCRS, 2002. v. 1. p. 79-98.
• RAMOS, A. P.; TENANI, L. E. Análise métrica do apagamento das vogais postônicas não finais no dialeto do noroeste paulista. Estudos Linguísticos, São Paulo, v.38, n.1, p.21-34, jan.-abr. 2009. Disponível em: http://www.gel.org.br/estudoslinguisticos/volumes/38/EL_V38N1_02.pdf Acesso em: 13 fev. 2017.
  » http://www.gel.org.br/estudoslinguisticos/volumes/38/EL_V38N1_02.pdf
• SELKIRK, E. The syllable. In: HULST, H.; SMITH, N. The structure of phonological representation . Dordrecht: Foris, 1982. v. 2. p. 337-383.
• SILVA, T. C. Fonética e fonologia do português: roteiro de estudos e guia de exercícios. 10. ed. São Paulo: Contexto, 2014.
• SILVA NETO, S. da. Fontes do latim vulgar . Rio de Janeiro: Acadêmica, 1956 [1938].
• THUN, H. La géographie linguistique romane à la fin du XXe siècle. In: ENGLEBERT, A.; PIERRARD, M.; ROSIER, L.; Van RAEMDONCK, D . Actes do XXIIe Congrès International de Linquistique et de Philologie Romanes. Bruxelas: Max Niemeyer Verlag, 1998. p. 367-409.
• WENKER, G. Sprachatlas von Nord - und Mitteldeutschland: Auf Grund von systematisch mit Hülfe der Volksschullehrer gesammelten Material aus circa 30000 Orten. Abtheilung I Lieferung 1. Strasbourg/London: Karl J. Trübner, 1881.

• 1
  In the original: “ tendência a evitar as palavras proparoxítonas: masclus (masculus), domnus (dominus), caldus (calidus), fricda (frigida), virdis (viridis) ” ( COUTINHO, 1976 , p. 32).
• 2
  In the original: “ [...] os vocábulos portugueses de acentuação na antepenúltima sílaba raramente provêm da evolução no latim vulgar (um exemplo é pêssego, do lat. persi(cum). ” (CÂMARA JR., 1985, p. 35).
• 3
  In the original: “[...] supressão do segmento fônico compreendido entre a vogal acentuada e a vogal final (ex.: Petrópis para o topônimo Petrópolis; exérço em vez de exército; glóbo substituindo glóbulo)” (CÂMARA JR., 1985, p. 35).
• 4
  How proparoxytone words are also known.
• 5
  It is worth pointing out that Dicionário Aurélio documents ridico as a syncopated form of ridículo , marking it as a “ brasileirismo familiar mineiro ” (‘familiar brasilianism from Minas Gerais’), meaning “stingy” ( FERREIRA, 2010 ).
• 6
  In the original: “arvore-arve, passaro-passo, polvora-porva, marmore-marme, Alvaro-Arvo, abobora-abobra (ou aborba)” ( NASCENTES, 1953 [1922], p. 37).
• 7
  In the original: “as línguas diferem quanto ao número de segmentos permitido em cada constituinte silábico [...]”, e ainda completa: “[...] o molde é uma afirmação geral a respeito da estrutura possível de sílabas numa determinada língua ” ( COLLISCHONN, 1996 , p. 101-102).
• 8
  In the original: “ 1) síncope da vogal postônica dos proparoxítonos latinos, com redução do vocábulo a paroxítono e possível evolução posterior do grupo consonântico resultante, como em apícula> apicla> abelha [...] ; 2) síncope de consoante sonora entre vogais, como em mala> maa> má[…] ” ( QUEDNAU, 2002, p , p. 79).
• 9
  In the original: “ síncope da oclusiva como primeiro membro de grupo consonântico, em vocábulos eruditos, como em excepção>exceção [...]” ( QUEDNAU, 2002, p , p. 79).
• 10
  In the original: “ [...] constituir um ataque bem-formado (ár.vo.re > ar.vre, re.lâm.pa.go > re.lam.po) ou uma coda bem-formada (pé.ro.la > per.la) ” ( AMARAL, 2002, p , p.102).
• 11
  In the original: “ la dialectologie a pour tache essentielle d’étudier la variation geolinguistique ” ( CONTINI; TUAILLON, 1996, p , p. 7).
• 12
  In the original: “ caracteriza pelo alargamento do campo de observação e por um trabalho em produtividade mais desenvolvido ” ( THUN, 1998, p , p. 369).
• 13
  To know more about Project ALiB, access: < https://alib.ufba.br/> .
• 14
  In the original: “ descrever, com base em dados empíricos, sistematicamente coletados, a realidade linguística do país, no que tange à língua portuguesa, fornecendo dados linguísticos atualizados não só da diversidade diatópica, mas também da variação diageracional, diastrática, diagenérica e diafásica ” (CARDOSO et al. , 2014a, p. 23).
• 15
  Among the places of ALiB inquiry in the interior of Paraná, seven points coincide with ALPR: Adrianópolis, Barracão, Campo Mourão, Guarapuava, Lapa, Londrina, Umuarama.
• 16
  In the original: “[...] estes segmentos não estão licenciados para segunda posição de um ataque complexo. Assim, em nossos dados, quando uma vogal era apagada, as obstruintes apagaram junto com a vogal postônica (relâmpa g o > relampu; lâmpa d a > lampa) ” ( LIMA, 2008, p , p. 93).
• 17
  In the original: “ grupos consonantais anômalos ocorrendo em posição postônica ” ( SILVA, 2014, p , p. 90).
• 18
  In the original: “[...] as proparoxítonas mais resistentes à síncope têm o /i/ como vogal postônica e os contextos circundantes não representam grupos consonantais bem-formados ” ( AMARAL, 2002, p , p. 108).
• 19
  In the original: “ uma maior consciência feminina do status social das formas linguísticas ” ( PAIVA, 2007, p , p. 35).
• 20
  In the original: “[...] observa-se uma mínima superioridade das mulheres no uso das formas de prestígio (proparoxítonas), e um número ligeiramente menor no uso das formas estigmatizadas (paroxítonas), o que se harmoniza com constatações de pesquisas anteriores que apontam maior adesão de falantes do sexo feminino às formas de prestígio ” ( CASTRO, 2008, p , p. 117).
• 21
  Available at: https://cidades.ibge.gov.br/brasil/pr/barracao/panorama . Access on: Feb. 2019.

Publication Dates

• Publication in this collection
  20 Dec 2019
• Date of issue
  Sep-Dec 2019

History

• Received
  10 Apr 2018
• Accepted
  06 Jan 2019