On Vocalic* Variations in Portuguese Dialects

Under the rubric of standard Weinrich (1954, p. 396) includes, among other notions, those of ‘socially acceptable’ or ‘average’ or ‘typical’. For Portuguese linguists the main criterion for a model to represent standard language has always been the speech modality of the cultured or urban people. There is no agreement, however, over whether the speech of Lisbon or that of Coimbra should set the norm. In an article reviewing the position of early Portuguese linguists regarding the interpretation of what is the Portuguese padrão (standard), Boléo (1946) expresses that Gonçalves Viana (1892) and José Pedro Machado (1940) shared the conviction that it is the speech of the cultured parlance of Lisbon that is considered as the standard, as opposed to that of Coimbra, which was the choice of Oliveira Guimarães (1927) as well as Boléo (1946, p. 11-12). The notion that the speech of Coimbra represented the best model of speech in Portugal had been postulated and defended already in the seventeenth century. Vasconcelos (ed. 1970, p. 54) quotes a passage by Dom Francisco Manuel de Mello in Apólogos dialogaes (ed. 1721, 26162) which reads as follows:


Context
Under the rubric of standard Weinrich (1954, p. 396) includes, among other notions, those of 'socially acceptable' or 'average' or 'typical'. For Portuguese linguists the main criterion for a model to represent standard language has always been the speech modality of the cultured or urban people. There is no agreement, however, over whether the speech of Lisbon or that of Coimbra should set the norm. In an article reviewing the position of early Portuguese linguists regarding the interpretation of what is the Portuguese padrão (standard), Boléo (1946) expresses that Gonçalves Viana (1892) and José Pedro Machado (1940) shared the conviction that it is the speech of the cultured parlance of Lisbon that is considered as the standard, as opposed to that of Coimbra, which was the choice of Oliveira Guimarães (1927) as well as Boléo (1946, p. 11-12).
The notion that the speech of Coimbra represented the best model of speech in Portugal had been postulated and defended already in the seventeenth century. Vasconcelos (ed. 1970, p. 54) quotes a passage by Dom Francisco Manuel de Mello in Apólogos dialogaes (ed. 1721, 261-62) which reads as follows: se cá entre nós dissessemos se fallava mais elegante em Coimbra, que em outra parte, não mentiriamos, sendo alli o coração e alma das sciencias que se ensinam e aprendem.
Vasconcelos, however, questions this view when stating: "Chacun prèche pour son saint, selon ses prédilections naturelles" (p. 54). Boléo's argument supporting his preference for the speech of Coimbra to be accepted as the norm for the standard language is based on the fact that the speech of the capital does not necessarily reflect a norm for the best pronunciation. He points out that, contrary to the linguistic situation in Spain, France and England, where the role of the capital was unquestioned in the spread of a language norm, other countries, such as Italy and Germany do not exhibit the best speech characteristics based on the speech of their capitals. He remarks: Ora no estado actual dos nossos conhecimentos, se é legítimo defender, como norma ortoépica, a pronúncia da capital, por motivos de ordem política e social, não é menos defender a de Coimbra, por motivos sobretudo lingüísticos (1946, p. 12).
It is not within our scope to deal more extensively with the reasons, controversies and theories that guided Portuguese linguists to have common or opposing viewpoints in regard to whether Coimbra or Lisbon represents the standard language. It is important to recognize that, in spite of any argument regarding location, what prevails, and is shared by the opinions (Cuesta and Mendes da Luz, 1980;Vasconcelos, 1970;Boléo, 1942;Viana, 1891) on this subject is that the language of the cultured people is the main criterion used in the selection of a norm representing the standard language. In regard to this view Vasconcelos (ed. 1970, p. 15) writes: Jusqu'au XII e siècle, la langue portugaise n'existait que comme langue parlée; au moins on ne connaît pas, antérieurment à cette époque, de documents entiérement écrits en portugais. Naturellement, les classes les plus instruits, comme celle des clercs, employent, par opposition au peuple, des formes considérées comme plus choisies; mais ce ne fut que lorsqu'on commença à écrire le portugais, et surtout depuis que cette langue devint proprement littéraire (XIII e et XIV e siècles), que l'on peut distinguer clairement deux courants, l'un populaire, et l'autre érudite.
For Vasconcelos, it is this literary language that exhibits distinct features from the language used in the north of Portugal (Minho), the centre (Beira) and the south. It is clear that there is a norm for the standard language and it is against this that other speeches (within the country) exhibit variations. We believe that there is no valid reason for assuming a value judgement suggesting that one region speaks 'better' than another. This is an oversimplification.
In this study we are not guided by the presuposition of a 'better' speech when outlining the phonetic characteristics of Standard Portuguese, but our intent is to establish a model of standard language against which to evaluate the phonetic characteristics of the regional variants, and that will bear upon our research.

Dialect Variations of Portuguese
Portuguese on the whole represents a linguistic community with a distinctly unified pattern of speech. Only the speech of Miranda, Riodonor and Guadramil, in the northeast of Portugal, digress considerably from the norm; they are more akin to Leonese than to Portuguese, 1 otherwise, the diffferences from one region to another are not acute: "não há, como em muitos outros países, duas normas linguísticas que se oponham e influam uma sobre a outra. (Helmut Lüdtke, 1954, p. 216). This being the case, for example, in Italy or in Germany. Given that the differing features found in these regions often do not map on to predictable boundaries, it is difficult to speak of dialects of Portuguese in the strict sense of this term. In referring to the linguistic situation in Portugal, Barbosa (1983, p. 23) affirms: On prend même le pretexte d'une telle uniformité pour affirmer que le portugais ne connaît pas de <<dialectes>>, mais seulement des <<parlers régionaux>> … Nous dironts toutefois que, si l'on prend le mot <<dialecte>> dans le sens qui est le sien lorsque l'on parle de <<dialecte de New York>>, de <<dialecte de Boston>>, etc., il y a des dialectes portugais sur le continent européen… Within this general pattern, however, there are regional varieties with a number of phonetic traits of some importance. On the basis of such differences, Portuguese linguists have proposed a number of division of the Portuguese spoken in mainland Portugal. Each linguist, however, has applied his or her own criterion of the terms dialecto, subdialecto, codialecto, variedade, or falar (Vasconcelos, 1893codialecto, variedade, or falar (Vasconcelos, -18971901;Maria Helena Santos Silva, 1959-1962;Pilar Vázquez Cuesta and M. Albertina Mendes da Luz, 1980;Lindley Cintra, 1983).

Vocalic dialect variations in Portuguese: Mainland Portugal
The region of Beira Baixa and Alentejo which comprises the urban centres of Castelo Branco and Portalegre is characterized by a profound alternation of the whole vocalic system, especially that of the stressed vowels. The most distinct feature of this zone is the change of /u/ > [ü], frequently registered in some localities in Portalegre, Moção, Nisa, Castelo Branco, Idanha-a-Nova, Oleiros, and Covilhã. Lüdtke (1957, p. 106-110) studied the vowel system of this region, zone 2, and saw a chain reaction as follows: e > E > A > ç > o > u > ü. And he establishes two vocalic systems both with fronted /e E o/. The [ü] pronunciation is not confined to this area, but is widespread in zone 3, in the area known as Barlavento do Algarve (Cintra, 1983, p. 157-58). 7 Leite de Vasconcelos (ed. 1970:83) saw it as an area feature and delimitates it as follows: SIGNUM: Estud. Ling., Londrina, n. 5, p. 77-106, dez. 2002 6 The Portuguese of the North has neutralized the distinction /v/ and /b/ in favour of the latter, as in Galician and Castilian; has kept the distinction between the apico-alveolars /s z/ and the reflexes of the medieval [ts] and [dz] as predorso-dentals /s z/; has maintained the old [tЀ], grapheme ch. Standard Portuguese, as is known, has one pair of sibilants, predorso-dental /s/ and /z/, having merged the alveolars with these. 7 The same geographic delimitation for [ü] is also reiterated by Cuesta and Mendes da Luz (1980:61), and in regards to the intensity of this vowel they conclude that "A palatalização do u pode ser mais ou menos intensa e não se dá em posição átona: maduro (ü) mas madurar (u). Others also remark that the French [ü] is a feature of the speech of Vila do Bispo (Nunes, 1902, p. 34;Lüdtke, 1957, p. 95-112;and Maia, 1975-78, p. 58). ... de Fundão et Sertã (Beira Baixa) jusqu'à Portalegre (alto Alentejo), et qui comprend quelques territoires de l'Estrémadure (Alvaiázere, Paialvo), l'u de la langue littéraire devint ü: rüa, müro, düas. Ensuite le phénomène apparait dans l'Algarve (Barlavento), où je l'ai observé à Lagos et à Vilado-Bispo: üm, lüa, viüva. Il est probable que l'ü existe aussi dans certains endroits compris entre ces zones. Hammarström (1953, p. 51-52 and passim) on the other hand, disputes the délimitation of this vowel in western Algarve and he emphasizes that its occurrence is not "si limitée que les dialectologues semblent l'indiquer, mais peut être constatée dans une grande partie de la province" (p. 146). He outlined the following chain reaction in the vocalic development (1953, p. 73 Cintra (1983) questions Hammarström's delimitation of [ü] in this speech, and he remarks: "...é uma fonte a manejar com cuidado, por razões metodológicas" (p. 157, note 59). He confirms, though, that the presence of [ü] is, however, an important linguistic factor setting apart the western Algarve from eastern Algarve (p. 137, note 33). 8 The Portuguese of the North has maintained the diphthong [ow] also pronounced as [aw] (in the south [ow] is monophthongized with different results). The diphthong /ow/ is kept in the North as [ow], pronounced in some areas as [aw]. Vasconcelos speaks of the existence of a variant [öw], -also occurring in the Azorean speech-in north of Trás-os-Montes, and part of Entre Douro e Minho, and possibly in part of Beira (Esquisse). For Cintra (1983, p. 43) has noted [öw] as typical of the north and centre of Trás-os-Montes.
Diphthongization of both /o/ and /e/, which occurs after labials and velars respectively, appears in other speech areas of mainland Portugal, in Mértola near Algarve and in Alandroal, southwest of Elvas (Dámaso Alonso, 1962, p. 39-40).
Similar phenomena are recorded in the speech of some of the islands in the Azores, mainly in the central group, Terceira, S. Jorge, Graciosa, and less extensively on the eastern and western groups, as well as in the speech of the islands of Madeira and Porto Santo. In addition, an important phonetic characteristic of the speech of Madeira is the formation of diphthongs as a result of the detachment of the glides [w] and [j]. Rogers (1946, p. 244-246) divides this development into two categories: a) the diphthong is formed independently of the vowels or consonants in contact; b) the diphthong results from the contact of the stressed vowel with the consonant. The consonants affecting the release of the glide are the velars /k/ or /g/, the bilabial consonants /p b m/, or / (Rogers, 1948, p. 3). Other examples registered in Porto Santo also appear in the north of Portugal (Monteiro, 1948-50, p. 93). b) Zone 2 is characterized by a profound alternation of the whole vocalic system, especially that of the stressed vowels. The most distinct feature of this Zone is the palatalization of /u/ as [ü] and Cintra (1983, p. 155) proposes its isogloss as the limit of dialect demarcation. 13 There is also 13 The palatalization of /u/ is frequently registered in some localities in Portalegre, Moção, Nisa (especially in this area), Castelo Branco, Idanha-a-Nova, Oleiros, Covilhã. In reference to the presence of [ü] in this region, Boléo (1969-71:875-76) informs that this vowel is found "do Minho ao Alentejo, e não apenas na Beira Baixa, como geralmente se supõe." palatalization, to a lesser or greater degree, according to locality, of stressed /a/ when in contact with a consonant or a semivowel or when, in the preceding syllable, one of the vowels i and u, or their glide counterparts is found. Some examples of this palatalization in this region, and also some cases registered in Algarve are identified when the syllable preceding the stressed /a/ has /i/ or /u/, or /ej/ or /ow/ as in the following words: buraco, aguilhada, giada, felicidade, cidade, vinagre, buraco, animal, buscar, ficar, fumar (Sabugal, Fundão, Sobral, Oleiros, Sertã, Bemquerença, Monsanto and Idanha-a-Nova), geada, alguidar, curral (Portalegre and Castelo Branco); or in barraca, amarelado, azulado, canastra in areas of Algarve (Dámaso Alonso, 1962, p. 150;Maia, 1975-78, p. 46;Boléo and Silva, 1961, p. 100). In other examples the palatalization of /a/ seems to be caused by the influence of the preceding palatal phoneme as in balhar, relampejar heard in Monsanto, Beira Baixa (Dámaso Alonso, 1962, p. 151), bacalhau, calhau registered in Algarve (Maia, 1975-78, p. 46). This palatalization is also recorded in the north, Minhoto, as in the words: buraco (also heard with [e]), embala, jarro (Santos Silva, 1961, p. 309). Boléo and Silva (1961) write: uma linha vertical divide as zonas em que se verifica a passagem do a oral a e, e a do a nasal a e, correspondendo a zona ocidental à transformação sofrida pelo a oral. Cintra (1983, p. 133, note 28) disagrees with this delimitation of / a/ as [ϯ] and he writes: o traço 'geral' 5 do minhoto -aparecimento de e aberto, correspondendo a a tónico aberto do português-padrão, só nas palavras bacalheu e bureco (e por vezes em pestenas), e não, de um modo geral, nas palavras com a tónico aberto...é um fenómeno que se regista esporadicamente em regiões muito afastadas do Minho.
We have heard this phenomenon in some sporadic cases in the speech of Terceira (Azores), and a number of examples pertaining to the pronunciation of /a/ as [ϯ] in this speech (Ribeirinha, S. Mateus) is also provided by Borba e Maia in her unpublished thesis presented at the University of Lisbon 1965 (p. 3-5). Moreover, there is labialization of /e/, and to some degree that of the diphthong /ow/. 14 c) Zone 3 also palatalizes /u/ (Cintra, 1983, p. 157-158) as in Zone 2. The palatalization of /u/ in this region had already been noted by Leite de Vasconcelos (ed. 1970, p. 83). 15 This geographic delimitation of the palatal /u/ is also reiterated by Cuesta and Mendes da Luz (1980, p. 61). Regarding the intensity of this vowel they conclude that "A palatalização do u pode ser mais ou menos intensa e não se dá em posição átona: maduro (ü) mas madurar (u). Nunes (1902, p. 34) referred to a French [ü] in Vila do Bispo. Also registered here by Lüdtke (1957, p. 95-112) and Maia (1975-78, p. 58). 16 As well, the stressed /a/ velarizes and is pronounced as [ѐ] (as [ѐ] in areas of the Algarve (Cintra, 1983, p. 158;Hammarström, 1953, p. 137-9;143 and passim;Lüdtke, 1957, p. 95-112;Cuesta and Mendes da Luz Cuesta, 1980, p. 62). In an attempt to explain the distribution of velar /a/, Hammarström says: SIGNUM: Estud. Ling., Londrina, n. 5, p. 77-106, dez. 2002 14 As already Indicated, Lüdtke (1957:106-110) has studied the vowel system of this region, and he establishes two vocalic systems: one for the area Póvoa e Meadas and Montalvão, south of Tagus; and the other for the area of Castelo Branco and Idanha-a-Nova, south of Tagus. 15 See above. 16 For other references regarding the geographic delimitation of ü, see above.
Hammarström, who studied this speech, gives the following chain reaction: (ei > e) > ϯ > a > ѐ >o > u > ü (p. 160). 22 A profound change in the vocalic system, with features similar to those noticed here is found in the speech of S. Miguel (Azores), and sporadically manifested in the speech of the other islands in the Azores: Flores, Corvo, Graciosa, Terceira, Pico and Santa Maria (Blayer, 1992). The palatalization of /u/, a prominent feature of the speech of São Miguel, is registered, though less extensively in the islands of Flores, Corvo, Graciosa, Santa Maria, Terceira, and Pico; and is found in Porto Santo and Madeira. Particularly in Machico and Porto-Moniz (Vasconcelos, 1970, p. 130), and on the "southern part of the island along the coast eastward from Câmara dos Lobos to Machico, including both of these places" (Rogers, 1946, p. 241-42). In this speech, [ü] is also heard when the stressed vowel is followed by /l/ as in desculpe, última (Rogers, 1946, p. 242). The Madeiran [ü] has been described as the Swedish /u/ (Viana, 1883, p. 34, note 2 to p. 33). Rogers (1946, p. 241) notes that this vowel is "slightly less 22 The stressed /i/ has also been labialized (in pedidos, fitas) in Sagres and Faro, and compared to the Swedish /i/ (Hammarström, 1953, p. 47, 99, 126-127). Hammarströn describes the /i/ as follows: "Notons que notre impression auditive d'une voyelle labialisée est peu sûre, parce qu'il est possible de prononcer, sans arrondir les lèvres, un son très proche de l'y suédois: il s'agit sans doute d'une sorte d'i postérieur assez ouvert." Lüdtke (1952, p. 479) describes the /i/ of Loulé-Faro-Olhão as a central or velar unrounded vowel and "identique à l'i perçu par nous comme labialisé." Maia (1975-78, p. 54) has found in Vila do Bispo and Alte that the pronunciation of stressed /i/ "...algumas vezes, apresenta uma realização bastante aberta, dando a impressão acústica de uma vogal intermédia entre e e i." In the speech of the island of Porto Santo (about twenty-seven miles northeast of Madeira) the stressed oral or nasal /i/ is often heard as oral or nasal [öj] as in bonito, meio dia, domingo, assim. (Rogers, 1948:2). In Madeiran speech, the stressed /i/ is also pronounced as the diphthong /oj/ and the first element appears to have a timbre "between i and ö" (Rogers 1946:239). Moreover, Rogers mentions that the pronunciation of the stressed /i/ is not consistent, not even when pronounced by the same individual, and he remarks as well that this phonetic trait is one of the "most striking features" of this speech (p. 240). Vasconcelos (1970:130) had already referred to the peculiarity of the Madeiran stressed /i/, and he noted: "Dans toute l'île de Madère, il y a un i spécial que M. Gonçalves Vianna (Essai de Phonétique, p. 6) compare à l'y polonais, et que je note par ï, ex. navïo (presque navêio), rïo." forward than the ü of German müde or the u of French pur." Within the same areas where /u/ is heard as [ü], the nasal counterpart is also registered with a labialized timbre (Rogers, 1946, p. 242).
The back pronunciation of /a/, heard on all islands with various degrees of velarization (as [a], as [å] or as [ѐ]), is heard as [å] in the insular speech of Porto Santo, and in Madeira its pronunciation ranges "from Standard á almost to open ó" (Rogers, 1946, p. 242). The pronunciation of /o/ as [u], widely heard on the speech of all Azorean islands, not just on S. Miguel is also a charatectistic of the speech of Madeira (Vasconcelos, 1970, p. 130;Rogers, 1946, p. 248).

2.2.Vocalic Tendencies in the Speech of the Azores
The vowel system of the speech of the Azores Islands has several important tendencies which are generally most prominent on São Miguel, and sporadically interspersed throughout the other islands. It is in São Miguel that the vocalic system has undergone the most advanced stage of development. The linguist, Leite de Vasconcelos (1925), who studied the speech of Arrifes in São Miguel remarked: O dialecto dos Arrifes, no concelho de Ponta Delgada, representa a fase mais evolucionada, que conheço, do português, com excepção dos crioulos, como o dialecto de Melgaço é o mais arcaico de todos. (p. 49) As we know today, it is not only the speech of this area that differs from the Standard language, but that of São Miguel as whole. This speech has the fronted palatals /ü/ and /ö/. Its front members have opened and the center and back members show closing as well as some of the palatalization in an overall chain reaction. Most interesting, the spoken language displays in the articulation of the vowels an array of variety which unables us to see the intermediate stages that take place. The western group of islands (as well as sporadic examples collected in other islands) also have /ü/ and /ö/, but otherwise their system has not fully developed and it has not reached the degree of evolution that exists in São Miguel.
The change affecting the vocalic system of São Miguel, somewhat parallel to that of mainland zone 2 as identified by Lüdtke, may be schematized as follows: An analysis of vowel fronting for the other islands 24 shows that the change affecting asymmetry of the vocalic systems 25 of the insular speeches involve the following patterns: 26 I. The conversion of falling diphthongs to monophthongs and then SIGNUM: Estud. Ling., Londrina, n. 5, p. 77-106, dez. 2002 23 Some phonological characteristics of the speech of the island of São Miguel, such as the presence of the palatal vowels /ü/ and /ö/, have been ascribed to Celtic influence. As we saw above, similar traits are clearly apparent in other speeches (Corvo, Flores, Sta. Maria and the Central islands). However, it is at its most advanced in São Miguel. There are some interesting parallels between the vowel system of São Miguel (including the presence of /ü/ and /ö/), and that of some regional variants of Portuguese. It is very easy, of course, to think of the influence of early settlement from regions of Portugal where the system is prevalent. This cannot be excluded, but it is better not to be categorical as one has little knowledge of the history of these developments, or of the antiquity that these have on the Continent. We are still far from being able to explain why Portuguese varieties (such as the speech of São Miguel) have the characteristics that they do. Based on the evidence that we have (in particular that we see all different stages that lead to the formation of the system of São Miguel, and also the mixed stages in the vocalic system of other islands), it seems reasonable to claim that what had been attributed to Celtic influence may actually indicate that we are dealing with the internal evolution of the language. 24 I am not delimiting the vocalic sound system for the entire insular speeches, but referring to specific speech areas as detailed in Blayer (1992). 25 The nasal vowels counterparts will not be included here. The developments recorded in S. Miguel agree in part with those presented by Martinet (1955). Martinet's account of the phonological changes in the speech of São Miguel are based on Rogers' article (1948). It must be emphasized, however, that today we have more data which provide a more clearer idea of the changes taking place in the vowel system of this island.
Parallel developments, especially the fronting of mid and high-back or mid-high fronted vowels can also be described as part of the tendencies affecting some speech areas of Peninsular Portuguese, as discusses above.
A further comparison of the speech of São Miguel with that of the other islands produces other important parallels, but overall these do not profoundly affect the vocalic system, as it occurs in São Miguel, since they appear to be at a beginning or intermediate stage. Some developments may be cathegorized as follows: (a) labialization of the stressed velar back vowels, as well as the mid fronted /e/; (b) As we indicated above, the stressed /a/ velarizes in the speech of S. Miguel, and is pronounced as [ѐ] or with intermediate stages between [a] and [å]. Similar pronunciation generally occurs in the speech of all the other islands. 27 The opening also affects the front vowels /e/ and /ϯ/.
Finally, an important feature which distinguishes the speech of the Central group -Terceira, São Jorge, Pico and Fayal-is the release of the glides [w] or [j] before stressed vowels, and this diphthongisation is, according to our research and also Rogers' study (1949, p. 49) and Maia's thesis (1965), a general characteristic of the speech of Terceira. The other groups (Eastern and Western) also have sporadic cases of this diphthongisation. The release of [w] occurs after any consonant. It appears, due to the number of the examples, that it was common after the bilabials /p b m/, and the velars /k g/, and it may have begun with this context and afterwards it spread in contact with other consonants such as the labio-dentals /f v/, the dento-alveolars /t d n l r s z/, and the palatal /ѭ/. 28 The release of [j] occurs in particular after the dento alveolars /t d n l r s z/, the bilabials /p b m/, the velars /k g/, the palatals /ѭ ? ْ/, and the labio-dental /v/. It is also registered after the consonantal groups tr, br, gr. 29

Concluding Remarks
Looking at these speeches synchronically, it appears that some of the vocalic tendencies featured in the speeches of the regional varieties in mainland Portuguese are clearly apparent in the insular speeches of the SIGNUM: Estud. Ling., Londrina, n. 5, p. 77-106, dez. 200227 See Blayer (1992. 28 For details see Blayer (1992). 29 For details see Blayer (1992).
Azores and the Madeira islands. There are some interesting parallels between these vowel systems.
Based on the evidence that we have, in particular the finding that we see all stages that lead to the formation of the system of São Miguel (Azores) as well as in the southern continental Portuguese -zone 2 and 3-and all mixed stages in the vocalic system of the various regions in the insular and continental speeches, it seems reasonable to claim that we are dealing with a Romance phenomenon which clearly deserves careful analytical examination in future research in this area. The challenge is of course to extend the study to a detailed and refined analysis of phonological environments conditioning the sound change.