Convergence patterns in the city of Seville. Sociolinguistic study of seseo

Convergence patterns in the city of Seville Sociolinguistic study of seseo Juana Santana Marrero Universidad de Sevilla Research carried out over the past few years in several Andalusian cities has revealed the processes of convergence towards conservative realisations that are occurring among the allophones of the phoneme /θs/. This time, using the Seville PRESEEA corpus, we have analysed the frequency of the seseante and dento-interdental pronunciation patterns that coexist in this city, and we have correlated them with linguistic, social, and individual parameters. Our data showed that this variation process was mainly influenced by closeness of [s] and [θ] in the immediate phonetic context, and by the educational attainment level, the age, and the sex of the respondents. The analysis per surveyed subjects indicated that most of them combined both variants. Keywords: seseo , PRESEEA-Seville, convergence patterns, sociolinguistics 1. Introduction The sociolinguistic studies carried out in recent years in various Andalusian cities have highlighted the variation between vernacular pronunciation patterns and convergent realisations with the standard. This is the case with consonants in the onset position (Villena 2012; Moya 2018). This time we will focus on the phonetic realisations corresponding to the orthographic segments <z + vowel> and <c + e, i> that have been recorded in the city of Seville in recent years in PRESEEA1cor- pus. Specifically, we intend to analyse what degree of variation exists between the seseante norm (i.e., zona [sóna] – ciclo [síklo]), the dialectal variant of the selected geographical area (Alvar et al. 1973, map 1705), and the dento-interdental pronunciation (i.e., zona [θóna] – ciclo [θíklo]), which is typical of the northern norm https://doi.org/10.1075/sic.20003.san |Published online: 8 March 2022 Spanish in Context 19:1 (2022), pp. 122–145. ISSN 1571-0718 |E‑ISSN 1571-0726 © John Benjamins Publishing Company 1. Proyecto para el estudio sociolingüístico del español de España y de América ‘Project for the Sociolinguistic Study of Spanish from Spain and America’. For further information, the website http://preseea.linguas.net/ can be consulted. and the standard. We will relate this alternation with linguistic variables (position in the word, number of syllables, stress, following vowel, closeness of [s] or [θ] in the phonetic context, and use in proper nouns) and social factors (educational attainment level, age, and sex). In addition, we will examine whether the distribution of the two variants is conditioned by their use in lexemes or recurrent word endings in the sample. Finally, we will also focus on the individual variation in each respondent. 2. Background Seseo is one of the phonetic results of the dental consonant /θs/, which is a product of the phonological reduction that occurred in southern areas of European Spanish after the readjustment of medieval sibilants (Lapesa 1991, 374–377). As an innovative or divergent realisation, the seseo forms part of the phonetic features in Andalusia, with a heterogeneous distribution on both the spatial and the social levels. Regarding the geolectal framework, without intending to make an exhaustive delimitation, the seseante norm is located between ceceante and distinción zones and covers “las tierras occidentales del sur de Badajoz, el sudoeste de Huelva, el norte de Sevilla, el sur de Córdoba y el norte de Málaga, junto con algunas localidades de Jaén y de Granada” (RAE and ASALE 2011, 190). Narbona et al. (1998, 170) specify that the city of Seville is included in this seseante area. Therefore, seseo is widely distributed in different parts of the Andalusian geography. On the other hand, sociolinguistic studies have demonstrated that this dialectal landscape is complex and heterogeneous, especially in urban areas (Villena 2012). This is principally because Andalusia has, over time, permeabilised pronunciation phenomena of central-northern Spanish, mainly those that are perceived as more prestigious phonetic realisations because they have been associated with the standard. In fact, the sociolinguistic research has shown that the alternation between the seseante and the convergent pattern is a characteristic that affects different Andalusian cities.2Thus, in Malaga (Villena 1997, 98–99; Villena 2001, 130–131) and Granada (Salvador 1980; Moya and García 1995; Moya and Sosiński 2015; Moya 2018) the dialectal realisations, seseante and ceceante, and the pronunciation patterns coming from the northern norm, have been coex- 2. Similar processes of convergence towards the northern norm have been detected in the Andalusian cities of Huelva and Jerez de la Frontera, where ceceo is the main local variant (Regan 2017; García-Amaya 2018). Convergence patterns in the city of Seville 123 isted for decades. Broadly speaking, this change is mainly led by subjects with high level of formal education, young people, and women.3 For the city of Seville, we have sociolinguistic analyses that can be distributed in two stages. In the first phase, when recordings from the seventies and eighties were studied, it was revealed that seseo was the most prominent dialectal variant, with upward progression as the educational attainment level of the respondents decreased (Sawoff 1980, 245–246; Lamíquiz and Carbonero 1987, 36–39; Carbonero 2003 [1985], 43). Therefore, it was suspected that changes were taking place in subjects with higher level of formal education, that pushed towards the dento-interdental allophone. Indeed, research in recent years on recordings from the beginning of the twenty-first century (Santana 2016; Santana 2016–2017; Santana 2017; Santana 2020a), reveals the presence of this phonetic realisation, with a different distribution among the speakers, which mainly depends on their educational attainment level: the degree of university education favours the convergence towards the prestigious variants of the northern norm and the scarce academic formation reinforces the dialectal realisations. Unlike what has been done so far, in this paper we present an integral analysis of all the recent recordings we have in the Andalusian capital, low, middle, and high levels, this time considering a wide range of linguistic factors. We start from the hypothesis that some of the latter, such as similar allophones in the closer phonetic context or the use of the variants in the same lexeme or word ending, can influence this alternation. On the other hand, we think that social variables, mainly education and sex, may also affect. It is also likely that the study by subjects will reveal an important degree of hesitation in the use of one or another variant among the respondents. If we look for an explanation to these convergence processes, we cannot forget that the Castilian variety has been the main reference for the conformation of the correct idiomatic speech model for centuries (Méndez 1999, 122–123). This characteristic is reflected in the beliefs of Europeans Spanish-speaking, who extensively think that the Castilian variety is the most prestigious dialect modality (Yraola 2014, 584–585; Cestero and Paredes 2018, 55; Hernández and Samper 2018, 186; Manjón-Cabeza 2018, 152; Méndez Guerrero 2018, 96). If we focus on Seville, Ropero and Pérez (1998, 283) point out that more than 60% of the respondents considered that the most prestigious speech model is the one used in Castile, followed by the Andalusian modality. We have documented similar data some decades later, when we surveyed the perception of a group of Sevillian university 3. This permeation has also been documented in Andalusian politicians who lived and worked in Madrid for years (Cruz 2020). This is another example of the incidence of this phenomenon in people with a high educational attainment level who were in contact with the northern norm and had a significant role in public life and media projection. 124 Juana Santana Marrero students (Santana 2018a, 121; Santana 2018b, 88–89; Santana 2020b, 83–85). In this case, 46% of respondents considered that the prestigious reference speech was in the Castilian variety and that it also projects a higher social status among those who use this dialectal pronunciation. It is worth noting how this recent data has special relevance for our research, because it gathers the opinions of young people from Seville, a group of individuals that can champion possible future changes in their community. Regarding to what has been said, it is not surprising that in the last decades, Andalusian cities have documented processes of convergence towards pronunciation features from the northern norm, which have a high social status. Concerning seseo, we must not forget that it is a pattern of southern prestigious pronunciation (Villena 2012, 65). Therefore, the two variants that we are discussing in this research constitute, in turn, patterns which are positively valued by speakers. According to Villena (2001, 129), this has caused the convergence towards the uses of the northern variety to follow a slower rhythm in the western area of Andalusia, because they come into conflict with the allophones of the regional norm that is spread from Seville. Hence the importance of knowing the current phase of this phonetic variation process in the Andalusian capital. 3. Methodology To carry out this research, we used the PRESEEA corpus of Seville, composed of 72 recordings that were compiled between 2009 and 2019 (Repede 2019a; Repede 2019b; Repede 2019c). Following the methodological guidelines of this project (Moreno 1996), the speech samples collect semi-targeted surveys in which a researcher directs the dialogue towards topics that are established in advance: housing, family, neighbourhood, travels and so on. With this technique you get an orality that is raw and without prior planning. The goal is to get the respondent to express himself/herself with speech that is as fluid and natural as possible. The subjects surveyed were classified according to the variables of educational attainment level (low: without studies, with primary education or, in the case of some young people, some courses of compulsory secondary education; middle: compulsory secondary education, vocational training or baccalaureate; and high: university studies), age (first generation: from 20 to 34 years; second generation: between 35 and 54 years; and third generation: 55 years old and over) and sex (12 men and 12 women per sociolect). The respondents have remained all or most of their lives in Seville. Therefore, it can be said that all of them are fully integrated in the life of the city. Convergence patterns in the city of Seville 125 Regarding the collection of the analysis units, we took the first 20 minutes of each recording. Thus, the sample consisted of a total of 1440 minutes. The fraction of time selected allowed us to make sure that the respondents go from the first minutes of the interview, where there may be some increase in the tension due to them being observed, to a phase in which their linguistic production tends to be more natural, without them being so aware of the recording. Initially, we thought that this might be an interesting aspect to take into account in the research, in case it might have influenced the respondents’ pronunciation, but, in general, it did not. The location of the variants was made through direct listening of the recordings. Although we carried out a careful and slow review to identify the allophones under study, there were cases in which we could not specify exactly what the subject said. This is because, along with the realisations in [s] and [θ] that we clearly identified, intermediate allophones were also given that do not fully conform to one or the other phonetic variant. This has been verified more precisely when recordings like ours have been analysed from the acoustic point of view (Lasarte 2012). For this reason, those cases in which it was impossible for us to delimit if it was a realisation in [s] or in [θ], we had to leave them out of the analysis. Along with the intermediate pronunciation of the allophone, we also had to rule out some occurrences in which the speed of the enunciation or, mainly, the overlap of the sound that we were interested in analysing with a background noise or with the speech of the interviewer prevented us from identifying the pronounced phonetic segment. Finally, the data was processed through version 26 of the SPSS programme and the Pearson X2test was used to identify the statistical significance of the variables considered (Preacher 2001). 4. Data analysis and results When analysing the whole sample, we could see that, as shown in Table 1, of the total of 8472 registered realisations, 52.2% belong to the dento-interdental variant, [θ], and 47.8% belong to seseo pronunciation, [s]. Therefore, a very close ratio between one variant and another was noted. Table 1. Global distribution of [s] and [θ] n % [s] 4049 47.8 [θ] 4423 52.2 Total 8472 126 Juana Santana Marrero The relevant fact is that our materials point towards an advance of the conservative pattern in recent decades. While data from Lamíquiz and Carbonero (1987, 36–39) showed approximatively 90.0% of seseo pronunciation, in the latest semicontrolled interviews that we have analysed, this phenomenon is accompanied by a decline in the pronunciation norm. The percentages between both allophones, however, are quite balanced, so we are facing a close fight between vernacular and convergent phonetic realisation. This is probably due to the fact that there are two prestigious pronunciation patterns, a southern and a national one. So, it is necessary to look carefully at the internal distribution of these two variants in close competition in our sample. This struggle between the two allophones was also reflected in the fluctuation that each subject demonstrated in their interventions. This factor allowed us to know the degree of linguistic individual variation presented by the individuals surveyed with respect to the phenomenon we are dealing with. Specifically, we observed two possibilities, as can be seen in Tables 2, 3, and 4: the subjects who chose only one of the variants4and those who used both. In the first case, it will be necessary to know which of the two was favoured by a greater number of respondents. In the second case, it will be relevant to know how many individuals had predominantly seseante or conservative preference indexes. Table 2. Subjects with one or two variants n % Subjects with one variant 20 27.8 Subjects with two variants 52 72.2 Total 72 Table 3. Subjects with one variant n % Only [s] 9 45.0 Only [θ] 11 55.0 Total 20 4. In this block we included the respondents who recorded only one example of the other realisation, which may have been a slip. Convergence patterns in the city of Seville 127 Table 4. Subjects with two variants n % Prefer [s] 29 55.8 Prefer [θ] 23 44.2 Total 52 It draws attention to the high level of individual variation that was detected among the subjects surveyed concerning these two allophones (Table 2), as a large majority, 72.2%, used the two forms in alternation. In turn, when they used only one of the phonetic realisations (Table 3), there was a slight increase in the respondents who only used [θ], 55.0%. This data reinforces the drive that the convergence with the northern pronunciation is receiving, since it is about people who do not have seseo in their linguistic competence. However, compared to what would be expected by the percentages we are recording, among the individuals who combined the two variants (Table 4), those who opted for seseo formed a larger group, 55.8%. These results should be interpreted along the lines of the fierce competition between the two allophones that are recorded throughout the corpus, with very balanced percentages in both cases, and with a very significant presence in a large number of individuals. 4.1 Linguistic factors Next, we will consider a selection of linguistic variables that have been taken into account in previous research (Moya and García 1995; Villena 2007; García-Amaya 2008; Santana 2016; Santana 2017; Santana 2020a; Regan 2017; Cruz 2020): position in the word (initial or middle: i.e., cinco, lanzada), number of syllables (bisyllabic, trisyllabic or polysyllabic: i.e., hizo, macetas, relacionar), stress (if the variant is in stressed or unstressed syllable, the latter pretonic or posttonic: i.e., quizás, coincidir, espacio), following vowel (i.e., empiezas, cerca, relación, hizo, azul), closeness of [s] or [θ] in previous or later syllables (i.e., esencia, precisa, organización), and use in proper nouns (i.e., Galicia). Finally, we will analyse the distribution of the two variants in syllabic sequences forming lexemes and word endings that were recurrent in our materials, to see if there were pronunciation patterns associated with these forms. The quantitative data can be seen in Table 5. On this occasion we are only working with the 52 subjects who used the two allophones in their interventions (totalling 6131 cases). 128 Juana Santana Marrero Table 5. Distribution of [s] and [θ] according to linguistic variables [s] [θ] X2Sig.n % n % Position Initial 427 49.1 443 50.9 2.524 0.112 Middle 2735 52.0 2526 48.0 Number of syllables Monosyllabic 8 66.7 4 33.3 7.139 0.068 Bisyllabic 981 49.5 999 50.5 Trisyllabic 1279 53.3 1122 46.7 Polysyllabic 893 51.4 844 48.6 Stress Stressed 1503 50.6 1469 49.4 2.139 0.128 Unstressed 1659 52.5 1500 47.5 Following vowel a 279 47.6 307 52.4 9.408 0.052 e 1168 52.6 1051 47.4 i 1518 51.1 1452 48.9 o 177 56.9 134 43.1 u 20 44.4 25 55.6 Closeness of [s] or [θ] No closeness of [s] or [θ] 2692 51.2 2568 48.8 80.355 0.000 Closeness of [s] 459 59.2 316 40.8 Closeness of [θ] 11 11.5 85 88.5 Proper name Proper name 95 43.4 124 56.6 When relating the linguistic variables with [s] and [θ] realisations, no statistical profitability was found in four of them: location in the word, number of syllables, stress, and vowel that follows the consonant segment that we study. This data indicated that the alternation of allophones was not only the most widespread among the respondents, but it was also quite established in their pronunciation habits, without there being, in principle, too many linguistic constraints that justify the use of one or the other. Therefore, we focus on those phenomena that did affect their distribution. Convergence patterns in the city of Seville 129 First, the presence of [s] or [θ] in the immediate, preceding or following phonetic context (the influence will be, at most, three syllables before or later, which may or may not belong to the same word), favoured the vernacular (i.e., ne[s]esito, evolu[s]ionar Sevilla, pases un ve[s]ino) or convergent (i.e., urbaniza[θ]iones, peatonalizar el [θ]entro, celebra[θ]iones) pronunciation respectively.5 Thus, there was an assimilation process, without being statistically relevant whether it was progressive or regressive. Similar results have been obtained in Granada (Moya and García 1995, 148–149) and Malaga (Villena 2007, 69–70). In the city of Seville, the influence of [θ] is perceived more decisively in the use of the conservative variant (88.5%), with almost 38 points of difference with respect to the cases without a close [s] or [θ] (48.8%). It is important to point out that this assimilation process of dento-interdental pronunciation maintained its solidity even in those subjects who were mostly seseantes (22/32, 68.8%). On the other hand, dissimilation acted negatively, since it clearly contributed to the descent of seseo when there was a [θ] in the immediate context (11.5%) and, although in a less forceful way, the northern realisation also decreased when there was a [s] close by (40.8%). With regard to the presence of the variants under study in the proper nouns, we could think that the univocal nature of this grammatical unit, as far as it serves to identify one or several of the elements of a same class, but not all, would favour standard pronunciation, which in turn is more directly associated with the orthographic representation. In this way, it would help to identify the reference more precisely. In fact, the results of the sample corroborated our initial hypothesis: the majority of the speakers who used the two variants in their interventions preferred the dento-interdental pronunciation in the designation of proper nouns (56.6%). However, the results varied significantly when we considered only the subjects who preferred the seseante pronunciation in their interventions. In this case, the percentage of the vernacular variant was more significant than that of the convergent (81/114, 71.1%). Therefore, we can conclude that the proper noun is a variable that, in general terms, favoured dento-interdental pronunciation, but it was not a determining factor in our corpus, since this variant presented a clear decline in individuals who promoted dialectal pronunciation more prominently. Lastly, we analysed the alternation of the two variants in a group of terms that were used repeatedly in our materials. For this we focused on the recurrence of lexemes, with or without affixes (i.e., especial-: especial/-es, especialmente, especialidad, and especialistas; i.e., azul), and word endings. Specifically, in the latter context we analysed the diminutives and the participles (with variations of gender 5. We only considered the influence of allophones that responded to a norm-based pronunciation (fran[s]esas) and not those that constitute a dialectal realisation ([s] erve[s]a). 130 Juana Santana Marrero Table 11. Distribution of [s] and [θ] according to educational attainment level and sex Low sociolect Middle sociolect High sociolect Males Females Males Females Males Females [s] n 1012 627 865 714 599 232 % 81.3 55.6 57.5 53.0 53.4 14.9 [θ] n 233 501 639 634 1093 1323 % 18.7 44.4 42.5 47.0 64.6 85.1 Subtotals 1245 1128 1504 1348 1692 1555 Low sociolect: X2=182.965 sig.=0.000; Middle sociolect: X2=5.945 sig.=0.015; High sociolect: X2=178.514 sig.=0.000 Men of the low sociolect were the main drivers of seseo (81.3%), distancing themselves a lot from the other two blocks: middle and high. These showed a downward trend as academic education increased. Men of the high sociolect also reinforced the conservative variant (64.6%), following the general tendency of the surveyed subjects with this level of attainment education. On the other hand, women of the high sociolect were the main promoters of the dento-interdental allophone (85.1%), at a considerable distance from the females of the middle level (38.1 percentage points), and even more from those of the low one (40.7 percentage points). Women in the initial and intermediate levels of formal education continued to be more inclined to maintain the local variant and their uses reflected a more balanced situation between the two pronunciation alternatives. Therefore, the females of the high sociolect were the main promoters of the phonetic change of convergence towards the northern realisation [θ]. Finally, in Table 12 we observe to what extent the age of men and women could influence the variation process that we analyse. Table 12. Distribution of [s] and [θ] according to age and sex 1st generation 2nd generation 3rd generation Males Females Males Females Males Females [s] n 924 266 861 514 691 793 % 62.5 20.6 60.1 35.6 45.2 61.1 [θ] n 554 1023 572 930 839 505 % 37.5 79.4 39.9 64.4 54.8 38.9 Subtotals 1478 1289 1433 1444 1530 1298 1st generation: X2=492.705 sig.=0.000; 2nd generation: X2=172.86 sig.=0.000; 3rd generation: X2=71.462 sig.=0.000 Convergence patterns in the city of Seville 137 The most outstanding fact, concerning the relation between age and sex, is that the opposition of the two variants was much more pronounced in the first two generations: women opted more significantly for the convergent pronunciation, especially the youngest ones (79.4%), while men decided on seseo to a greater extent, also mainly being of a younger age (62.5%). Therefore, the conflict between the two allophones takes on special relevance among the subjects of the first age group, which led this process of change. If we look at the analysis per individual (Table 13), we could think that if the dento-interdental allophone was fundamentally driven by respondents with a higher educational level, the subjects who used this pronunciation exclusively or mainly would principally belong to that social group. Table 13. Subjects with one or two variants according to educational attainment level Only [s] Prefer [s] Only [θ] Prefer [θ] Low 6 11 2 5 Middle 1 14 2 7 High 2 4 7 11 Subtotals 9 29 11 23 Totals 38 34 Our data confirms that, in effect, this is true: 66.3% of the respondents only used the conservative variant (7/11) and 47.8% of those who preferred that allophone (11/23) had university studies. On the other hand, just as the vernacular pattern was the majority option in the low sociolect, those subjects also used this allophone prominently as the only variant (6/9, 66.7%) or as the preferred phonetic realisation (11/29, 37.9%). On the other hand, the respondents of the middle educational level showed high indexes of individual variation, since most of them used the two variants (21/24, 87.5%), where the subjects who opted for the local seseante pattern prevailed (14/29, 48.3%). This confirms our initial hypothesis: the individuals of this sociolect are more noticeably divided between holding on to the local allophone or following the pattern that projects greater social consideration and that is used among speakers with higher formal education. As we already pointed out, the latter was the majority option among the youngest subjects. A further symptom of the degree of hesitation between both variants in this sociolect is that almost half of them (10/21, 47.6%) had poorly polarised use percentages between them, that is, that the majority allophone had less than 80% of the cases. This means that the alternation of [s] and [θ] for <z + vowel> and <c + e, i> was constant in their interventions, and they were not only isolated uses as a 138 Juana Santana Marrero result of a more punctual or sporadic hesitation. The profile of these respondents was preferentially second generation (5/10, 50.0%) and female (7/10, 70.0%). Next, data in Table 14 show the correlation between individual preferences and their age. Table 14. Subjects with one or two variants according to age Only [s] Prefer [s] Only [θ] Prefer [θ] 1st generation 5 6 5 8 2nd generation 2 11 4 7 3rd generation 2 12 2 8 Subtotals 9 29 11 23 Totals 38 34 As can be seen, it was the youngest respondents who showed the lowest rates of individual variation, since they are half of those who only used one of the variants (10/20, 50.0%). In addition, the majority of the subjects of this age group opted for the conservative realisation. Finally, based on data in Table 15, we are going to analyse the distribution of the surveyed subjects according to whether they preferred one or another pronunciation variant and its relationship with the variable sex. Table 15. Subjects with one or two variants and sex Only [s] Prefer [s] Only [θ] Prefer [θ] Males 8 14 3 11 Females 1 15 8 12 Subtotals 9 29 11 23 Totals 38 34 Our results showed that women led the group of individuals who only used the most normative and prestigious phonetic realisation (8/11, 72.7%), while, on the opposite side, men were the majority among the respondents who exclusively used the seseante norm (8/9, 88.9%). Convergence patterns in the city of Seville 139 5. Conclusions The results of the PRESEEA-Seville corpus indicate that today it is not possible to firmly state that this city is a seseante nucleus. On the contrary, it can be deduced from our materials that a convergence process is taking place towards the prestigious normative variant, the realisation of the orthographic segments <z+ vowel> and <c + e, i> in a dento-interdental allophone [θ]. The permeation of this northern phonetic realisation is already verified since we noticed that the latter slightly surpassed the vernacular pattern seseante when contrasting them. This fairly tight competition was also reflected in the individual variation, where a high index of hesitation between the two variants prevailed, since most respondents used both allophones in their interventions. This process of variation was influenced by linguistic and social factors. With regard to the linguistic variables, the assimilation with a [θ] in the close phonetic context especially favoured the dento-interdental variant and, to a lesser extent, when the allophone was used in a proper noun, although in this case it was not a decisive factor when the subjects were preferably seseantes. We also observed the association of each of these pronunciation patterns with certain lexemes and some word endings: seseo was preferred in especial-, vecinfacil-, nazareno/s, cena-, quizá(s), andaluc-, among other terms, and in the suffixes diminutive, –azo, -encia, -icionar, and –izar; and the convergent variant was more frequent in lexemes caz-, doce, bici(cleta), once, confianza, servicio/-s, piscina/-s, social-, among others. Therefore, there were some trends related to certain lexemes or suffixes, although we could not find any unquestionable rule. We think that this is an area that needs further research. Concerning social variables, our materials have allowed us to verify that a convergence process similar to that found in other Andalusian cities is taking place in the city of Seville, which is mainly driven by respondents with higher formal education, by young people, and by women. More specifically, young women of the high sociolect were the main promoters of this change, while young men of the low sociolect led the vernacular pronunciation. The use of both allophones showed more polarised positions between the upper and lower educational levels, in contrast to more balanced percentages in the middle sociolect. In this group, both the local pronunciation seseante and the northern dento-interdental were strongly represented, which highlights a close struggle between the two prestigious patterns alternating in our sample. In other words, Sevillian informants debated between remaining to the vernacular or accommodating to the northern pattern, the latter led by individuals with higher formal education. Once again, the dento-interdental allophone found greater follow-up among the respondents from the first generation, which might make that pronunciation pattern even- 140 Juana Santana Marrero tually progress in the years to come. Regarding the high rates of linguistic levels of individual variation registered in our corpus, in relation to the alternation we analysed, it should be noted that young subjects constituted the group that showed less indexes of hesitation, since they outperformed the other participants interviewed by using only one of the variants in their speech. On the other hand, the respondents of the middle level of formal education were those that most fluctuated between the two realisations. This is one more characteristic of the narrow competition that occurs in this population sector between the local seseante norm and the convergent pattern. Based on our data, it can be considered that we are facing a variation process and change in course that, while it has a great impact on young people, is likely to continue to advance and settle among the linguistic habits of Sevillians. Funding This research belongs to the project Patrones sociolingüísticos del español de Sevilla (PASOSSE), I+D Excellence project (reference FFI2015-68171-C5-3-P), financed by the Ministerio de Economía y Competitividad de España and by the Fondo Europeo de Desarrollo Regional . References Alvar, Manuel, Antonio Llorente, and Gregorio Salvador. 1973. Atlas lingüístico y etnográfico de Andalucía. VI. Madrid: CSIF. Carbonero, Pedro. 2003 [1985]. “Aspectos sociolingüísticos sobre la nivelación en el español actual.” In Estudios de sociolingüística andaluza, 39–46. Sevilla: PUS. Cestero, Ana M.ª, and Florentino Paredes. 2018. “Creencias y actitudes de los jóvenes universitarios del centro-norte de España hacia las variedades cultas del español.” Percepción de las variedades cultas del español: Creencias y actitudes de jóvenes universitarios hispanohablantes. Monográfico del Boletín de Filología LIII (2): 45–86. https://boletinfilologia.uchile.cl/index.php/BDF/article/view/51941/54406/ (accessed May 2019). https://doi.org/10.4067/S0718-93032018000200045 Cruz, Rocío. 2020. “Seseo, ceceo y distinción de /s/ y /θ/: el caso de los políticos andaluces en Madrid.” Nueva Revista de Filología Hispánica 68 (1): 137–174. https://doi.org/10.24201/nrfh.v68i1.3585 García-Amaya, Lorenzo J. 2008. “Variable Norms in the Production of /θ/ in Jerez de la Frontera, Spain.” In Indiana University Working Papers in Linguistics. VII. Gender in Language: Classic Questions, New Contexts, ed. by Jason F. Siegel, Traci C. Nagle, Amandine Lorente-Lapole, and Julie Auger, 49–71. Bloomington: IULC. Convergence patterns in the city of Seville 141 Hernández, Clara, and Marta Samper. 2018. “Creencias y actitudes de los jóvenes universitarios canarios hacia las variedades cultas del español.” Percepción de las variedades cultas del español: Creencias y actitudes de jóvenes universitarios hispanohablantes. Monográfico del Boletín de Filología LIII (2): 179–208. https:// boletinfilologia.uchile.cl/index.php/BDF/article/view/51945/54413 (accessed May 2019). https://doi.org/10.4067/S0718-93032018000200179 Lamíquiz, Vidal, and Pedro Carbonero. 1987. Perfil sociolingüístico del sevillano culto. Sevilla: IDR. Lapesa, Rafael. 1991. Historia de la lengua española, 8ª ed. Madrid: Gredos. Lasarte, M.ª Cruz. 2012. “Datos para la fundamentación empírica de la escisión fonemática de /s/ y /θ/ en Andalucía.” In Estudios sobre el español de Málaga. Pronunciación, vocabulario y sintaxis, ed. by Juan A. Villena, and Antonio M. Ávila, 129–166. Málaga: Sarriá. Manjón-Cabeza, Antonio. 2018. “Creencias y actitudes de los jóvenes granadinos hacia las variedades cultas del español.” Percepción de las variedades cultas del español: Creencias y actitudes de jóvenes universitarios hispanohablantes. Monográfico del Boletín de Filología LIII (2): 145–177. https://boletinfilologia.uchile.cl/index.php/BDF/article/view/51944 /54409 (accessed May 2019). https://doi.org/10.4067/S0718-93032018000200145 Méndez, Elena. 1999. “La norma idiomática del español: visión histórica.” Philologia Hispalensis 13: 109–132. https://doi.org/10.12795/PH.1999.v13.i01.08 Méndez Guerrero, Beatriz. 2018. “Creencias y actitudes de los jóvenes universitarios mallorquines hacia las variedades cultas del español.” Percepción de las variedades cultas del español: Creencias y actitudes de jóvenes universitarios hispanohablantes. Monográfico del Boletín de Filología LIII (2): 87–114. https://boletinfilologia.uchile.cl/index.php/BDF /article/view/51942/54407 (accessed May 2019). https://doi.org/10.4067/S0718-93032018000200087 Moreno, Francisco. 1996. “Metodología del ‘Proyecto para el Estudio Sociolingüístico del Español de España y de América’ (PRESEEA).” Lingüística 8: 257–287. Moya, Juan A. 2018. “Sobre el equilibrado reajuste de las hablas andaluzas.” Itinerarios 28: 35–66. http://itinerarios.uw.edu.pl/sobre-el-equilibrado-reajuste-de-las-hablas-andaluzas/ (accessed May 2019). Moya, Juan A., and Emilio J. García. 1995. El habla de Granada y sus barrios. Granada: Universidad de Granada. Moya, Juan A., and Marcin Sosiński. 2015. “La inserción social del cambio. La distinción s/θ en Granada. Análisis en tiempo aparente y en tiempo real.” Lingüística Española Actual 37 (1): 33–72. Narbona, Antonio, Rafael Cano, and Ramón Morillo. 1998. El español hablado en Andalucía. Barcelona: Ariel. Preacher, Kristopher J. 2001. Calculation for the Chi-Square Test: An Interactive Calculation Tool for Chi-Square Tests of Goodness of Fit and Independence [Computer software]. http://quantpsy.org (accessed May 2019). Real Academia Española, and Asociación de Academias de la Lengua Española (RAE and ASALE). 2011. Nueva gramática de la lengua española. Fonética y fonología. Madrid: Espasa. Regan, Brendan P. 2017. “A Study of ceceo variation in Western Andalusia (Huelva).” Studies in Hispanic and Lusophone Linguistics 10 (1): 119–160. https://doi.org/10.1515/shll-2017-0004 142 Juana Santana Marrero Repede, Doina. 2019a. El español hablado en Sevilla. Corpus PRESEEA-Sevilla. Vol. I Hablantes de instrucción alta. Sevilla: Editorial Universidad de Sevilla. https://doi.org/10.12795/9788447221752 Repede, Doina. 2019b. El español hablado en Sevilla. Corpus PRESEEA-Sevilla. Vol. II Hablantes de instrucción baja. Sevilla: Editorial Universidad de Sevilla. https://doi.org/10.12795/9788447221875 Repede, Doina. 2019c. El español hablado en Sevilla. Corpus PRESEEA-Sevilla. Vol. III Hablantes de instrucción media. Sevilla: Editorial Universidad de Sevilla. https://doi.org/10.12795/9788447221936 Ropero, Miguel, and Francisco J. Pérez. 1998. Análisis estadístico-sociológico de los comportamientos lingüísticos en la ciudad de Sevilla. Sociolingüística Andaluza 11. Sevilla: PUS. Salvador, Gregorio. 1980. “Niveles sociolingüísticos de seseo, ceceo y distinción en la ciudad de Granada.” Español Actual 37–38: 25–32. Santana, Juana. 2016. “Seseo, ceceo y distinción en el sociolecto alto de la ciudad de Sevilla: nuevos datos a partir de los materiales de PRESEEA.” Boletín de Filología LI (2): 255–280. http://www.boletinfilologia.uchile.cl/index.php/BDF/article/view/44879/46949 (accessed May 2019). Santana, Juana. 2016–2017. “Factores externos e internos influyentes en la variación de /θs/ en la ciudad de Sevilla.” Analecta Malacitana 34: 143–177. http://www.revistas.uma.es/index .php/analecta/article/view/5613/5262 (accessed May 2019). Santana, Juana. 2017. “Variación de las realizaciones de /θs/ en el sociolecto bajo de la ciudad de Sevilla: datos de PRESEEA-SE.” Procesos de variación y cambio en el español de España. Estudios sobre el corpus PRESEEA. Lingüística en la Red. Monográfico XV. http:// www.linred.es/monograficos_pdf/LR-monografico15-articulo2.pdf (accessed May 2019). Santana, Juana. 2018a. “Creencias y actitudes de los jóvenes universitarios sevillanos hacia las variedades cultas del español.” Percepción de las variedades cultas del español: Creencias y actitudes de jóvenes universitarios hispanohablantes. Monográfico del Boletín de Filología LIII (2): 115–144. https://boletinfilologia.uchile.cl/index.php/BDF/article/view/51943 /54408 (accessed May 2019). https://doi.org/10.4067/S0718-93032018000200115 Santana, Juana. 2018b. “Creencias y actitudes de jóvenes universitarios sevillanos hacia las variedades normativas del español de España: andaluza, canaria y castellana.” Pragmática Sociocultural / Sociocultural Pragmatics 6 (1): 71–97. https://www.degruyter.com/view /journals/soprag/6/1/article-p71.xml (accessed May 2019). Santana, Juana. 2020a. “Patrones sociolingüísticos de pronunciación en la ciudad de Sevilla: variación sociofonética de seseo y ceceo.” In Patrones sociolingüísticos del español hablado en la ciudad de Sevilla, ed. by Doina Repede, and Marta León-Castro, 169–207. Berna: Peter Lang. Santana, Juana. 2020b. “Percepción de las variedades andaluza y castellana de los jóvenes sevillanos: un análisis contrastivo.” Onomázein 50: 71–89. http://onomazein.letras.uc.cl /Articulos/N50/50_5.pdf (accessed December 2020). Sawoff, Adolf. 1980. “A Sociolinguistic Appraisal of the Sybilant Pronunciation in the City of Seville.” Festgabe für Norman Edison. Grazer Linguistische Studen 11–12: 238–262. Villena, Juan A. 1997. “Convergencia y divergencia dialectal en el continuo sociolingüístico andaluz: datos del vernáculo urbano malagueño.” Lingüística Española Actual 19 (1): 83–125. Villena, Juan A. 2001. La continuidad del cambio Lingüístico. Granada: Universidad. Convergence patterns in the city of Seville 143 Villena, Juan A. 2007. “Interacción de factores internos y externos en la explicación de la variación fonológica. Análisis multivariable del patrón de pronunciación no sibilante [θ] de la consonante fricativa coronal /θs/ en el español hablado en Málaga.” In Las hablas andaluzas y la enseñanza de la lengua. Actas de las XII Jornadas sobre la Enseñanza de la Lengua Española, ed. by Juan A. Moya, and Marcin Sosiński, 69–97. Granada: Universidad. Villena, Juan A. 2012. “Patrones sociolingüísticos del español de Andalucía.” In Estudios sobre el español de Málaga. Pronunciación, vocabulario y sintaxis, ed. by Juan A. Villena, and Antonio M. Ávila, 27–66. Málaga: Sarriá. Yraola, Aitor. 2014. “Actitudes lingüísticas en España.” In Bergen Language and Linguistic Studies 5. Actitudes lingüísticas de los hispanohablantes hacia el idioma español y sus variantes, ed. by Ana B. Chiquito, and Miguel Á. Quesada, 551–636. https://www .researchgate.net/publication/307174734_actitudes_linguisticas_de_los_hispanohablantes _hacia_el_idioma_espanol_y_sus_variantes/link/57fca1f208ae6ce92eb42e5a/download/ (accessed May 2019). https://doi.org/10.15845/bells.v5i0.685 Address for correspondence Juana Santana Marrero Departamento de Lengua Española, Lingüística y Teoría de la Literatura Facultad de Filología Universidad de Sevilla C/Palos de la Frontera S/N C. P. 41004 Sevilla España [email protected] Biographical note Juana Santana Marrero is currently a professor in Spanish language area at Seville University. She has also been a guest professor in Heidelberg, Buenos Aires and Athens universities. She is responsible for the Research Group Sociolingüística Andaluza: Estudio Sociolingüístico del Habla de Sevilla and her research activity is channeled through two international projects: Proyecto para el Estudio Sociolingüístico del Español de España y América (PRESEEA) and Proyecto de la Norma Culta Hispánica Juan M. Lope Blanch. She is a Correspondent Member of the Bolivian Academy of the Spanish Language. 144 Juana Santana Marrero Publication history Date received: 13 December 2019 Date accepted: 20 March 2020 Published online: 8 March 2022 Convergence patterns in the city of Seville 145