Sociolinguistic accommodation by Ecuadorian migrants in Granada (Spain)

4 Abstract This chapter shows the processes of linguistic accommodation of Ecuadorian migrants in a Spanish-speaking community that has very different linguistic features from the Ecuadorian community. Two linguistic features are studied using quantitative methods. Ecuadorian migrants in Granada maintain very strongly the features of Andean Spanish. This is evident in the treatment of the /-s/ in coda, which continues to have typically Andean characteristics, very different from the treatment given to the same sound by those born in Granada. In the case of the diminutive morphemes, we can no longer be so categorical because a sociolinguistic accommodation is observed, as there are cases of the morpheme -illo, which is unusual in Ecuador. The scarce convergence is influenced by the fact that the Andean features studied coincide with the prestigious features of the Spanish of central-northern Spain, with which the emigrants have contact both through the media and through their penetration of the Andalusian variety, which leads to the stigmatisation of vernacular features. Keywords: sociolinguistics, accommodation, Ecuadorians, Andalusian variety of Spanish, Andean variety of Spanish, diminutives, -s in coda position 4.1 Introduction 4.1.1 Preamble Steiner (1997, p. 78) reminds us that “Trees have roots; men and women have legs. With which to traverse the barbed-wire idiocy of frontiers”. It is natural for people to go from one territory to another, and it has been since the dawn of humanity to the present day. Linguists’ interest in the consequences of the recent contact made by groups of humans with the vernaculars already used in certain territories is, however, more modern. It was sociolinguistics that sparked the systematic study of the relations between languages and migratory movements. García Marcos (2002) in an annotated bibliographical compilation points to Weinreich (1953), Cohen (1956), De Mauro (1963), and, of course, Fishman (1964, 1965), among others, 4 Sociolinguistic accommodation by Ecuadorian migrants in Granada (Spain) Antonio Manjón-Cabeza Cruz DOI: 10.4324/9781003130703-4 10.4324/9781003130703-4 Sociolinguistic accommodation by Ecuadorian migrants in Granada 73 Sociolinguistic accommodation by Ecuadorian migrants in Granada as precursors. Sociolinguistics considers migrations to involve processes that modify the linguistic repertoire of the new host community, and also to involve linguistic mechanisms whereby the migrant community participates in the community in which they have come to live. 4.1.2 Spanish, a language of migrants One long chapter in the history of migration is that of peoples that spoke and still speak a language that we call Spanish. In order to characterise the history and the present of this language we cannot forget that it has been, and still is, a language of migrants. As Moreno Fernández (2013, p. 67) points out, “la lengua española le debe su configuración interna y externa al hecho de haber migrado, con sus hablantes, a lo largo de muchos siglos y a lo ancho de una geografía intercontinental” (the Spanish language owes its internal and external configuration to the fact that it has migrated, with its speakers, over many centuries and across an intercontinental geography). The ancient colonisations of large areas of the Iberian Peninsula, the Canary Islands, and a large part of America, although linked to military conquests, can be characterised as migratory movements. The process of colonisation entailed the establishment of a large number of settlers in the territories that the Crown was conquering. In addition to members of the army, the church, and the administration, it should be borne in mind that the vast majority of the colonists could be considered economic migrants. Since the migratory movements of Spanish speakers have not ceased since Spanish came into being, a fixed beginning for the contemporary migration of Spanish-speaking groups cannot be identified. However, a beginning can be proposed: it took place about 75 years ago, after the global geopolitical changes brought about by World War II. These movements have run parallel in America and Europe. From the Americas, modern migrations of Latin Americans have taken place in three main directions. The most massive one has been towards the north, towards the prosperous territory of the USA. The migrations of Cubans, Mexicans, Puerto Ricans, Central Americans, and many other groups from the middle of the last century to the present are well known. The linguistic consequences of these migrations have interested many scholars and occupy a central place in many collective works and reference works, such as Roca and Lipski (1993), Roca (2000), Ortiz López and Lacorte (2005), López Morales (2008), Otheguy and Zentella (2012), Fuller (2013), Escobar and Potowski (2015), Márquez Reiter and Martín Rojo (2015), and Potowski (2016, 2018). Internal movements between the territories of different countries and between rural and urban areas within the same country are also very important. As an example of linguistic studies on this subtype of migratory movement, one may recall the works of Martín Butragueño (2004, 2009) on Mexico City, Caravedo and Klee (2012) on Lima, or Fernández Mallat (2018) on Bolivian migrants in Chile. 74 Antonio Manjón-Cabeza Cruz Many Latin Americans have migrated to Europe, though with less demographic intensity. According to the United Nations (2019), the number of Latin American migrants in the USA is 22,124,003, while in Europe the total number is 3,534,003, of which 2,258,316 are in Spain. These movements have attracted less scientific attention (with the exception of the migration of Latin Americans to Spain), although there have been some recent studies such as those by Pellegrino (2004) on general sociological aspects, Calvi (2011) and Calvi Uberti-Bona (2020) on Hispanics in Italy, and Patiño-Santos and Márquez Reiter (2019) on Latin Americans in London. From Europe, the modern migration of Spanish speakers has also taken place in three directions. First, there was the migration, during the first Francoist period, of European Spanish speakers to America: Canarians to Venezuela, Galicians to Argentina, Andalusians to Brazil, etc. According to Palazón (1991–92) some 40,000 Spaniards emigrated to American territory every year from 1945 to 1959. Approximately two million Spanish speakers, according to Vilar Ramírez (2000), left their homeland in the 1960s and 1970s to settle mainly in France, Germany, and Switzerland. While the Canary Islands and Galicia provided the largest contingent of modern Spanish migration to the Americas, Andalusia was the main focus of Spanish emigration to Europe, although all areas were represented. The linguistic history of these migrants, especially in their relationship with Germany, has been the subject of attention for various scholars, such as Schmid (1994), Vilar Sánchez (1995), Lüdi (1998), Jiménez (2000), or Higueras and Fuentes (2017). Similar to the American process, there has been internal emigration from rural regions to industrialised areas. Millions of workers, especially from the south of the Iberian peninsula, settled in areas of central and northern Spain, and some of these areas are bilingual. The Basque Country and, above all, Catalonia were the recipients of most migrants. The linguistic and social consequences of these migratory movements have interested scholars such as Martín Butragueño (1992), who studies Andalusians in Madrid, although more attention has been paid to Andalusian migrants in Catalonia, as shown by the works of Báez de Aguilar (1997, 2000) or a chapter in Narbona, Cano, and Morillo (2011) dedicated to the same issue. In this chapter, the authors conclude that emigrants of Andalusian origin abandon the stigmatised linguistic features that betray their status as emigrants with a low social status (rotacism of /-l/, ceceo, heheo, the loss of /-s/ in coda position). On the other hand, the youngest speakers reach rates close to 100% active proficiency in Catalan. 4.1.3 Latin American migration to Spain This section is dedicated to studying the characteristics of one of the previously mentioned migratory movements: Latin Americans to Europe. If we ignore, as is logical, the continued presence of Latin American elites in Europe, then it is a relatively modern phenomenon, because, until the 1970s, the path was in the opposite direction: European emigration towards America. Spain, as Pellegrino (2004) observes, has been the preferred country of settlement for these migrants. Historical relationships, sometimes the family Sociolinguistic accommodation by Ecuadorian migrants in Granada 75 relationships of descendants of Spanish emigrants to America, in some cases the possibility of dual nationality, and undoubtedly the shared language explain this preference. Latin Americans, along with other communities (especially Moroccans and Romanians), converted Spain in a very short time from a country of emigrants to a destination of immigration. The socio-historical conditioning factors for this change are diverse. We cannot forget Spain’s economic progress, since, according to the Spanish Chamber of Commerce and Eurostat ‒ quoted by Europa Press (2020) ‒ the GDP (gross domestic product) per capita went from 1,010 euros in 1975 to 11,596 in 1995 and 26,432 in 2019. Material progress is coupled with the accession to the European Community (1986) and the recognition of nondiscriminatory rights such as free education and health care for the entire population, including newcomers. Data for 2020 from the National Institute of Statistics (INE) indicate 5,423,198 registered persons of foreign origin, of which 1,505,956 are from Latin American countries.1 Latin American immigration to Spain has been the object of sociolinguistic studies centred on contacts between American and European varieties of Spanish. These studies, which are becoming more and more numerous, have frequently been published in the journal Lengua y Migración/Language and Migration, issued by the University of Alcalá. We cannot fail to mention the general proposals of Moreno Hernández (2009, 2013). When it comes to specific areas, it has been the Latin American communities in Madrid that have been most studied. We can cite studies by Palacios Alcaine (2007), Sancho (2010, 2013a, 2013b, 2014, 2020), Molina Martos (2010a, 2010b), and Lara (2018). All these works focus on Ecuadorian migrants, to which we can also add Peralta (2014) on Dominicans in Madrid and Sáez Rivera (2014) writing from a more multilingual perspective. Less numerous have been the studies on Latin Americans in other communities, such as those by Rodríguez and Vázquez (2017) on Galicia, or Corona and Unamuno (2008), Bonomi (2010), and Trenchs et al. (2013) on migrant communities in Barcelona, Catalonia. We must remember that these are bilingual areas and scholars prefer to study the relationship of migrants with official single, local languages. They focus on studying the process of construction of the linguistic and cultural identity of Latin American migrants, especially young people. Andalusia also has been the subject of studies, although less numerous than those on Madrid. Von Essen (2016, 2020) has investigated the phonic accommodation of Argentines in Málaga. The sociolinguistic situation of Ecuadorians in Granada has been studied by: Sosinski (2018), who compares phraseology used by migrants and the vernacular; Manjón-Cabeza (2018b), who focuses on variation in diminutives; and Fernández de Molina (2019), who studies taboos in the speech of migrants in relation to employment. 4.1.4 Granada and the Andalusian variety among other Hispanic varieties The city of Granada is no stranger to immigrants. According to Fernández de Molina (2019, p. 11), in 2016 the city, not counting the rural areas, had 243,758 76 Antonio Manjón-Cabeza Cruz inhabitants, of which 9.5% were foreigners. Ecuadorians (1,063) were among the most numerous Latin American migrants. Ecuadorians represented 4.77% of the foreigners in the census of 2016. This is a group with special qualitative importance, thanks to their active associations, which managed, for example, to convince Granada City Council to inaugurate a square called “Plaza de la República de Ecuador” on 11 November 2017. When Ecuadorian migrants, mostly from the Sierra, arrive in Granada, they are confronted with a linguistic variety with characteristics very different from their own. They also encounter a variety of Spanish that is considered the European standard, i.e. that of the centre-north of the peninsula, exemplified in Madrid. The characteristics of the Andalusian dialect have been thoroughly analysed, described in dialectological studies, and summarised in general books such as Narbona, Cano, and Morillo (2011), or widely dealt with in many dialectological works as a whole, such as Alvar (1996), and thus they are sufficiently well-known by specialists. But the situation described in dialectological studies does not correspond to reality, because Andalusia, as sociolinguists such as Villena (2006) and Moya (2018) have highlighted, is undergoing a process of convergence with the northern variety, at different speeds depending on the town or city. The most dramatic example may be the historical increase in the distinction of /s/ vs. /θ/, as opposed to vernacular solutions (seseo and ceceo). This change has been studied by Moya and Sosinski (2015) for the city of Granada. These authors point out the distinction percentages shown in Figure 4.1. In Granada, this loss of the vernacular variety is relatively advanced when it comes to several linguistic phenomena. Migrants encounter a sociolinguistic 10% 10% 46.71% 52.50% 55% 79.40% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 1933 1953 1980 1992 1995 2014 Figure 4.1 Evolution of the s/θ distinction in Granada over the last 80 years (adapted from Moya and Sosinski 2015). Sociolinguistic accommodation by Ecuadorian migrants in Granada 77 stratification that we can describe as abrupt ‒ Hernández Campoy and Villena (2009) ‒ since speakers with high levels of education tend to converge with the northern variety, while lower levels are more faithful to the local variants. This process, together with other older language policy issues, has led to a decline in the prestige of the Andalusian variety, as shown by studies on attitudes and beliefs. Proof of these considerations is the negative ratings given to Andalusian by the people of Madrid, as reflected in data compiled by Yraola (2014), or the people of Toledo, as can be seen in Crespo and Manjón-Cabeza (1996) or Manjón-Cabeza (2000). These negative ratings are similar to those given by Andalusians themselves, as shown by Santana (2018a, 2018b) for Sevillians or Manjón-Cabeza (2018a, 2020) for people from Granada. It is therefore a very paradoxical situation since migrants who are going to settle in neighbourhoods with low or medium-to-low sociolinguistic levels2 find themselves surrounded by an Andalusian variety with different characteristics. Furthermore, the Ecuadorian vernacular often coincides with the prestigious variety used in the north of the Iberian peninsula and with the sociolinguistic variants preferred by the higher strata of Granada society, strata which will not form part of their social environment. This is not the place to make an exhaustive list of the sociolinguistic landscape of Granada, but, by way of orientation, we must comment on some of the most divergent features between the two varieties that come into contact there: Andean and Andalusian.3 Migrants find two local phonetic features common to all social groups in Granada because they are still resistant to convergence and are quite distant from the Ecuadorian variety. We refer here firstly to the generalised loss of the /-s/, which – according to Tejada (2015, p. 83) – is missing at an almost categorical 94.6% of the time in final position. The other common phenomenon is a consequence of the previous one: so-called vowel projection, or opening of vowels, which usually affects not only the final vowel but all the other vowels in the word (vowel harmony), so that one usually hears [‘dᴐ ‘mᴂ.nᴐ] for dos manos (two hands). Ecuadorians from the Sierra, as studies on this Andean subvariety remind us (ranging from the oldest ones like Toscano (1953) or Boyd-Bowman (1953) to some somewhat more recent ones like Haboud and De la Vega (2008), passing through Quilis (1988) and Canfield (1988)), maintain the /-s/, so that its pronunciation will be [‘dos ‘ma .n os], coinciding with the northern Iberian variety. However, the usual thing is that the Andalusian features are not so generalised in the vernacular because they are undergoing a process of convergence that causes sociolinguistic jumps between social groups. There are several, but we can highlight – due to their different treatment by Ecuadorians – the following: a. The sibilants in syllable onsets. In Granada, there are four sounds for which Ecuadorians have one. Thus, Ecuadorians, for what is written as solución, will say [so .l u.’sjon], while people from Granada, as Moya and Sosinski (2015) have analysed and quantified in a recent corpus, may opt for the prestigious northern variant [so .l u.’θjon], which is the one that is currently prevailing 78 Antonio Manjón-Cabeza Cruz (79.4% of cases), especially at higher social levels and among women. Alternatively, some opt for a similar performance to the American one, which in Granada is in decline and is usually called seseo: [so .l u.’sjon] (13.1% of cases); or for the ceceante variant, which is very deprecated: [θo.lu.’θjon] (5.7% of cases); or else for the even less prestigious variant heheante: [ho .l u.’hjon], so stigmatized that it does not appear in our interviews, conducted with the knowledge of the respondents, not with hidden microphones. b. While in the Ecuadorian Sierra, as Aleza and Utrilla (2010, p. 36) recall, “se han mantenido soluciones del «castellano viejo» como la oposición entre líquidas en posición implosiva”, in Granada there is a tendency to confuse /-l/ and /-ɾ/ in coda position, together with a high percentage of those consonants being lossed. For example, for /-l/ López Moreno (2018) reports 66.86% maintenance, 17.82% rhotacisation or realisation as /-ɾ/, and 15.33% loss. In the case of /-ɾ/, Fernández de Molina (2018) records 62.4% maintenance, 31% elision, and several minority solutions such as assimilations and lambdacism or pronunciation as /-l/. Both authors identify subjects’ education as being significant in the distribution of the variants since higher educational levels promote maintenance. Although there are many other phenomena that we will not comment on, the important thing to note is that Ecuadorian immigrants settled in Granada encounter a strongly stratified variety of Spanish, in which their neighbours (typically of low socio-cultural levels), employ variants that are very far from the prestigious variants of the high levels, which are those of central and northern Spain, and variants that often coincide with those of Andean Spanish. This situation regarding sociolinguistic accommodation by Andean people in Andalusia raises many questions: will immigrants abandon the divergent features and linguistically assimilate with their neighbours? Will they maintain the convergent features with the Spanish of the centre-north, with which they have contact through the media? Is the evolutionary direction of the changes homogeneous or will they differ according to the specific feature? Will there be social factors that influence this process? The following pages are dedicated to trying to answer these questions. 4.2 Methodology 4.2.1 Preamble This is a variationist study in which, through an analysis of variations in linguistic features in two corpora – one of Ecuadorian immigrants and the other of Granadians’ vernaculars – we can infer situations of convergence or divergence between the varieties. To find answers to the questions posed earlier, we have selected a generalised Granada feature of a phonetic nature, the loss of the /-s/ in coda position, as well as a Granada feature that exhibits greater social differences of a grammatical nature: diminutive suffixes. Sociolinguistic accommodation by Ecuadorian migrants in Granada 79 4.2.2 Corpus of migrants For the data on Ecuadorians, we have used the ECUGRA corpus (an oral corpus for the social study of Ecuadorian migrants in Granada) (forthcoming from Alvarado and Manjón-Cabeza). The corpus consists of 30 interviews with Ecuadorian migrants conducted between November 2016 and March 2017, totalling 13 hours and 30 minutes of recordings. All the respondents are residents of Granada. The sampling was non-probabilistic, i.e. it was done following the snowball technique, and always in the subjects’ usual place of leisure, in an outdoor court for ecuavolley (three-man volleyball), which serves as a meeting and recreation place at weekends. Since the interviewers went to the volleyball court, the interviewees were able to relax, as they were in a place with positive connotations, and the intention was to avoid their feeling uncomfortable in front of unknown interviewers. In total, there were 21 men and nine women. The lower number of women is mainly explained by the fact that they felt more inhibited when it came to participating in the interviews. We also identified two age groups: those in their thirties and those in their forties and above. It should be clarified that this is a trivial division because the ages are very homogeneous, and there is little dispersion: the minimum is 33 years, the maximum is 60 years, and the average age is 41.97 while the standard deviation σ = 7.34. There were also two levels of education applicable: primary education and secondary education. The basic data of the corpus are presented in Table 4.1. Table 4.1 contains no other personal characteristics that could be important when studying sociolinguistic accommodation, such as how many years they have lived in Spain, because in our case it is a very homogeneous characteristic, as the minimum is 12 years, the maximum is 18, the average is 15.1 and the standard deviation σ = 1.54. Only two of those surveyed were from the Ecuadorian coast (Guayaquil and Milagro), while 28 were from different Andean areas (e.g. Loja and Ambato), although most were from Quito and the surrounding area. Likewise, as a relevant characteristic, we must report that 15 interviews were carried out by Ecuadorians and 15 by Spaniards, a strategy which we believe to be useful in order to see the degree of accommodation according to the vernacular variety of the interviewer. The age or gender of the interviewers has not been considered because all Ecuadorian interviewers were male, while Spaniard Table 4.1 Distribution of the 30 Ecuadorian speakers Gender Age/level of education Primary Secondary Men 30–40 years 7 4 > 40 years 4 2 Women 30–40 years 2 6 > 40 years 1 4 80 Antonio Manjón-Cabeza Cruz interviewers were divided between men and women, so possible differences due to interviewer gender are hidden under differences due to interviewer origin. Also, the age of the interviewers was fairly homogeneous. We did not seek to conduct a special interview to look for particular phenomena, but rather opted for a semi-structured interview so that it could be compared with the corpus of Granada’s vernacular speakers. 4.2.3 Corpus of vernacular speakers The Granada sample is based on the PRESEEA corpus (Project for the Sociolinguistic Study of Spanish in Spain and America) of Granada, published by Moya (coord.) (2007, 2008, 2009). It is an oral corpus, obtained on the basis of a sampling by uniform fixing quotas. The distribution of speakers is shown in Table 4.2. 4.3 The /-s/ in coda position 4.3.1 Extraction of examples and coding To study the pronunciation of the /-s/ in coda position in Ecuadorian migrants in Granada, we extracted from the ECUGRA corpus all the cases found between minutes 10 and 12 of each of the 30 recorded surveys, i.e. a total of 60 minutes. Thus, in a fragment like the one in example (4.1), we recorded ten cases: (4.1) Vivíamos en el carro/ ¿no? no podíamos alquilar una casa porque no teníamos permiso legal en España/ desde/ desde el día/ desde el año dos mil llegamos a tener el permiso legal en España. [ECUGRA-H11-022]4 We obtained 1,170 examples, with an average of 39 examples per subject and a standard deviation σ = 13.12. Once the examples had been recorded, we proceeded to codify them, following recent studies, especially Tejada (2015), who studied six minutes of each of the PRESEEA-Granada subjects. In this way, the dependent variable is the /-s/, which initially presented four variants: maintenance, aspiration, elision, and assimilation with the next consonant. Table 4.2 Distribution of the 54 speakers in the PRESEEA (Granada) corpus Gender Age/level of education Primary Secondary University Men 19–34 years 3 3 3 34–54 years 3 3 3 ≥ 55 years 3 3 3 Women 19–34 years 3 3 3 34–54 years 3 3 3 ≥ 55 years 3 3 3 Sociolinguistic accommodation by Ecuadorian migrants in Granada 87 Ecuadorian migrants are found to use morphemes that are not found in their home variety or in the prestigious European variety. In contrast to the generalised loss of /-s/, they now find themselves with a very complex sociolinguistic distribution, since processes of convergence with the centre-north are taking place and a not inconsiderable penetration of -ito from the Granada vernacular can be observed. 4.4.2 Examples and coding We proceeded to extract all the examples from the ECUGRA corpus, using the same methodology as Manjón-Cabeza (2012, 2016, 2018b). In this process, as in other studies on diminutives, we were presented with several problems. The first of these has to do with lexicalisations and the second with double diminutives. We did not write down the historical lexicalisations, such as rodilla or bocadillo. There are two special cases of lexicalisation, which deserve separate consideration. These are bonito and chiquito. In the case of bonito we noted it at the beginning because we were trying to compare it with the frequent examples of reanalysis in bonico used by people of Granada, such as the one in example (4.5): (4.5) Sobre todo cuando (risas) quería adelgazar aprovechaba la cuesta/ que se te quedaba un tipillo mu(y) bonico. [GRAN-M13-005] This is a case of lexicalisation that shows pseudo-suffixal alternation since bonico/bonito have lost any semantic relationship with their root, bueno. We found ten cases of bonita/o in the ECUGRA corpus, but, as we did not find any cases of bonico in Ecuadorian migrants, we chose to eliminate the examples with bonita/o since we could not check whether there was alternation as in the Granada vernacular. A more difficult problem is that of chiquillo/chiquito, which is lexicalised in some American countries, such as Venezuela (Malaver, 2012, 2017). We have chosen to keep the nine cases found for two reasons: firstly, because their root chica/o is often seen in the corpus; and secondly, because there is chiquitochiquitito/chiquillo-chiquitillo alternation in Ecuadorian migrants, as can be seen in examples (4.6) and (4.7): (4.6) Y usted que me considera a mí// dice/ bueno/ dice un morenito chiquitito/ digo bueno/ la estatura/ el color de la piel/ eso no juega// lo que juega es acá la cabeza. [ECUGRA H22-011] (4.7) Yo tengo tres niños y los traje// yo vivo aquí desde el año dos mil/ traje a mis chiquitillos/ uno de tres/ cinco años// y una de nueve años. [ECUGRA H22-019] 88 Antonio Manjón-Cabeza Cruz Another problem in the count is with the cases of double diminutives of the type chico>chiquito>chiquitillo. In these cases, they have been considered as belonging to the group that contained the suffix that appeared last, i.e. -illo. The coding took into account the dependent variable which is the different appreciative morpheme, so that, in principle, it is a variable with three variants: diminutives with -ico, such as that in (4.8); with -illo, such as that in (4.9); and with -ito, such as that which can be seen in (4.10): (4.8) Me agrada a mí/ también a mi mujer/ venir para acá// estar con ellos un momentico/ reírnos y para la casa. [ECUGRA-H11-004] (4.9) He estado parado/ también/ pero he hecho cosillas/ por ahí. [ECUGRA-H22-019] (4.10) Si alguien cumple años// nos reunimos/ y nos vamos a las tapitas españolas. [ECUGRA-M12-012] We have considered two independent linguistic factors and four independent social variables. The linguistic independent variables have been: a. Type of root: nouns (to which we add pronouns), adjectives (together with verbal participles), and adverbs and adverbial locutions. Examples of the three variants appear in (4.11), (4.12), and (4.13), respectively. (4.11) Hombre/ siempre hay por ahí/ algún problemilla/ pero no// nada serio. [ECUGRA-H21-30] (4.12) Mira/ estaban jovencitos/ mira/ le digo/ si no estudias/ anda a trabajar. [ECUGRA-H21-24] (4.13) Si pudiera ascender un poquito// podría estudiar y ascender. [ECUGRA-M12-06] b. Final phoneme of the root. Given the statistical dispersion of the variable, we have grouped examples into two sets: roots that finish with unvoiced sounds, such as un poquito in example (4.14), and roots that end with voiced sounds, such as extranjerito in example (4.15): (4.14) Que seamos un poquito/ reconocidas/ las empleadas de hogar. [ECUGRA-M22-13] Sociolinguistic accommodation by Ecuadorian migrants in Granada 89 (4.15) Ahá/ extranjerito/ así me decía con el dedo/ extranjerito decía/ ya me vienes a engañar. [ECUGRA-H21-24] The independent social variables are the same as for the /-s/ study: sex, age, level of education, and the nationality of the interviewer. 4.4.3 Results and discussion Once the ten cases of bonito are eliminated, we have 161 examples, of which 30 end in -illo5 (18.63%); 130 in -ito (80.75%); and only one in -ico (momentico), which we show in example (4.8). This is the morpheme -ico applied to a base finishing in /t/, so we cannot be sure that it is a case of accommodation to the Granada vernacular since it is a common solution in many Spanish American countries, including Ecuador. Two facts are revealed to us after this basic count. The first is that there is a certain degree of accommodation by Ecuadorian speakers to the vernacular of Granada as there are a significant number of cases of -illo; the second is that migrants uttered far fewer diminutives than their Granada neighbours. Although, as we discussed in Manjón-Cabeza (2016), the number of diminutives seems to be an idiosyncratic feature, if we look at Table 4.8, where we compare the primary data in people from Granada and Ecuadorians – including cases of bonito/ bonico – the differences are obvious. People from Granada employ almost four times more diminutives than their new Ecuadorian-born neighbours. These data are similar to those found by Sancho (2015, p. 161) for Madrid, although the difference is not as pronounced in Madrid as it is in Granada. In fact, of our 30 subjects, five did not use any diminutives at all in their interview. They were respondents H11-002, H11-003, M22010, H11-014, and H22-027. It is to be noted that younger, more educated men dominate here in not using diminutives. It is noteworthy that four of them were interviewed by an Ecuadorian, which hints that what is important in inhibiting the use of diminutives is not the nationality of the interviewer since we would expect more empathy for an Ecuadorian than for a Spaniard. The data on the low use of diminutives is striking, given that various studies indicate that a feature of Andean Spanish is a greater use of diminutives than in Europe, as Toscano (1953), Reynoso (2001), and Palacios Alcaine (2006, p. 194) have pointed out. Table 4.8 Number of diminutives used by migrants and people from Granada n Corpus duration Diminutives per minute One diminutive occurs every… Granadians 1,296 2,140 min 0.61 1.65 min Migrants 171 810 min 0.21 4.74 min 90 Antonio Manjón-Cabeza Cruz Although the diminutive counts are based on different corpora, the tenor of the interviews, with their high communicative tension, influenced the scarce production of diminutives by migrants who might feel uncomfortable when observed by people from outside their community, whether they are Ecuadorian or Spanish researchers. We have now proceeded to try to delimit which linguistic or social factors influence the alternation between morphemes. As there is only one case of -ico termination, the dependent variable has the variants -illo and -ito, so, for the subsequent analyses, the total number of diminutives is 160. As in the case of /-s/, we proceeded to analyse the data with Rbrul. Firstly, we carried out an exploratory analysis using the step-up and step-down method of fixed effects. Both models coincided, indicating a good fit. Both the upward and downward analyses determined that, unlike /-s/, they are significant linguistic and social variables. The best exploratory model considers the following factors: sex (p-value = 0.00271), interviewer (p-value = 0.0456), age (p-value = 0.0381), and phoneme in the base (p-value = 0.0152). These data reveal that the variation seems to be due to sociolinguistic processes of accommodation, since, together with linguistic factors, social factors appear. In order to refine the analysis, we carried out a mixed-effects logistic regression with a random subject factor. The main results are shown in Tables 4.9 and 4.10, where the reference value was the use of -ita/o. When the individual factor is involved, we observe some changes with respect to the previous exploratory analysis of fixed effects. On the one hand, the word class variable to which the diminutive is applied is included in the model (p-value = 0.00566); on the other hand, the nationality of the interviewer becomes marginally significant (p-value = 0.0629). Table 4.9 Factors favouring -ito in Ecuadorian migrants One-level analysis of responses with predictor(s): subject [random] and phoneme in the base (0.000968) + base (0.00566) + gender (0.00872) + interviewer (0.0629) + age (0.209) + level (0.869) Factor Log-odds Tokens 1/1 + 2 Weight Phoneme in the base Voiced 1.344 38 0.895 0.793 Unvoiced –1.344 122 0.787 0.207 Base Adverbs 1.560 70 0.843 0.826 Nouns –0.687 63 0.810 0.335 Adjectives –0.872 27 0.741 0.295 Gender Women 2.083 33 0.970 0.889 Men –2.083 127 0.772 0.111 Interviewer Ecuadorian 1.059 68 0.897 0.742 Spanish –1.059 92 0.750 0.258 Sociolinguistic accommodation by Ecuadorian migrants in Granada 91 It should be noted that it is very common that, by involving the individual factor in the analysis, social factors are altered. We do not believe that this is the case, because it can be observed that sex continues to be a fundamental factor and the nationality of the interviewer, although it weakens, is not completely nullified. The results shown in Table 4.9 tell us that the use of -illa/o (remember that this is a symptom of accommodation by migrants to Granada Spanish) is favoured when the final consonants in the base are voiceless. In this case, we believe that lexical diffusion phenomena are involved, because, of the 30 cases that use -illo, 11 are poquillo/un poquillo, like the one in example (4.16): (4.16) Pues con la gente compañera pues estás/ tiés que estar un poquillo a gusto. [ECUGRA-H11-016] Table 4.10 Individual variation in use of -ito Subject Intercept Tokens 1/1 + 2 Weight H11-022 3.062 20 1 0.96 H21-024 2.498 13 0.923 0.931 H22-026 1.229 3 1 0.792 H12-025 1.223 10 1 0.791 H11-004 0.806 4 1 0.714 H22-019 0.395 8 0.5 0.623 M22-007 0.119 7 1 0.557 M11-029 0.104 4 1 0.553 H21-001 0.052 11 0.909 0.54 H21-018 0.051 3 0.667 0.54 M22-013 0.049 5 1 0.539 M12-012 0.017 3 1 0.531 M12-006 0.002 3 1 0.527 M12-008 0.002 2 1 0.527 H12-015 –0.036 7 0.857 0.518 H22-028 –0.418 1 0 0.423 H22-005 –0.596 11 0.818 0.38 H11-023 –0.672 6 0.5 0.363 H22-011 –0.878 6 0.667 0.316 M21-021 –0.932 9 0.889 0.305 H11-009 –0.99 9 0.889 0.293 H12-020 –1.19 4 0.5 0.253 H12-017 –1.268 2 0.5 0.239 H21-030 –2.492 6 0 0.084 H11-016 –2.838 3 0 0.061 Misc. 1: n = 160; df = 9; intercept = 4.178; proportion = 0.812; centred input prob = 0.985. Misc 2: log.likelihood = –54.123; AIC = 126.247; AICc = 127.447; Dxy = 0.863; R2 fixed = 0.461; R2 random = 0.278; R2 total = 0.739. 92 Antonio Manjón-Cabeza Cruz As part of this same type of phenomenon we can group the three cases in which the morpheme -illa appears linked to a word unknown in Ecuadorian Spanish, chapuza,6 as in the example (4.17): (4.17) También eh// también he tenido// sí/ he tenido chapucillas// así he trabajado para particulares. [ECUGRA-H22-028] Additionally, noun bases and, above all, adjectives favour the presence of -illo, as can be seen in example (4.18) where it is applied to the adjective chiquito: (4.18) Salíamos a vender una cabeza de ganado a los comerciantes […]/ mi padre yo chiquitillo por debajo de las piernas de él. [ECUGRA-H12-015] The distribution of solutions according to gender leaves no room for doubt: it is men who favour the use of the morpheme -illo, while women have a clear tendency to reject it since they only used it in one case, which is also doubtful. The following is an example of -illo, uttered by an Ecuadorian woman: (4.19) Le digo/ vete al coño// digo [chi]quillo/ ya no es/ es/es/es por mi salud/ porque pasar por las quimioterapias/ por la radioterapia. [ECUGRA-M21-021] This single example is special, since we found it in an interview with an Ecuadorian migrant woman conducted by a Spanish female interviewer, in a context imbued with little communicative tension. This can be deduced from the rude expression vete al coño, and surely also because a painful experience is being narrated: breast cancer. A very local appellation (quillo), almost lexicalised, arises. It seems, on first analysis, that all the data point towards a greater degree of accommodation by men to the linguistic characteristics of their host community, while women show a clear tendency to maintain their vernacular linguistic characteristics. We believe that this first impression needs to be qualified, however, since in the host community it is also usual to use the ending -ito, so the choice of women may be influenced both by their vernacular and by the variety of Spanish spoken in the centre of the peninsula, which is considered the most prestigious among European ones. However, there is no doubt that male accommodation to the Andalusian variety has taken place due to the use of -illo. One might think that men interact more with locals outside of their households, but, at least from the data available to us, we cannot affirm this. Both men and women were working outside the home, and in some cases, the interviews showed couples in which the man was unemployed while the woman was working. We also believe that the use of one or another morpheme is influenced by the nationality of the interviewers. Although there is an influence caused by the individual factor, if we analyse Sociolinguistic accommodation by Ecuadorian migrants in Granada 93 the data in Table 4.9, we observe the different choices of the speakers caused by adapting to the interlocutor, in this case, the interviewers. It is clear that, although -ito constitutes the majority of cases, if the interviewers are Spanish the appearance of the Andalusian -illo rises, so that the percentage of -ito is 75% of cases, while if the interviewer is Ecuadorian the percentage of -ito increases to 89.7%. The influence of the variety spoken by the interviewer shows how Ecuadorian migrants adapt to the new environment in which they find themselves. Likewise, although we only have one dubious case of -illo amongst the Ecuadorian women, it is still symptomatic that it is found in an interview conducted by a Spanish woman. We should also not disregard the cases of -illo that appear with Ecuadorian interviewers. These seem to reflect the fact that the influence of the Andalusian variety is beginning to take root in these new granadinos of Ecuadorian origin. We have observed that, when considering the individual factor, some changes in significant factors occur, so it is necessary, as we did with the phonetic variable /-s/, to present Table 4.10, which shows the results of the random factor analysis for variation in diminutive morphemes. The reference value is the use of -ito. From the analysis set out in Table 4.10 we can highlight the following: a. The individual factor is not as determining in the variation of the appreciative morphemes as it was in the case of the variation of /-s/. We can verify this because the fixed R2 is very high (0.461), while the variation explained by the random factor is also high (0.278), although less than that of the fixed factors. With the sum of the two, we obtain a very high total R2 = 0.739. b. Groupings of individuals can be established if we observe the distribution of speakers at the upper end (tendency to maintain -ito), central, or lower end (tendency to use -illo): 1. There are categorical speakers, i.e. they show no variation in the interviews. Ten interviewees only use -ito, while three interviewees only use -illo. If we extend the spectrum of -illo a little more, the seven subjects for whom at least 50% of diminutives use -illo/a are interviewed by Spaniards. However, of the ten that only use -ito, there are six interviewed by Ecuadorians and four by Spaniards. 2. If we divide the group into thirds, among the first eight speakers, i.e. users of -ito, we find, contrary to expectations, six interviewed by Spaniards and two by Ecuadorians. Furthermore, some of them are characterised by the use of a good number of diminutives, such as H11022, with 20, and H21-024, with 13. In the lower third, that is, relative exponents of -illo, we find the same proportion: six interviewed by Spaniards and two by Ecuadorians. The latter is less surprising because it can be attributed to the accommodation of these interviewees to the local interlocutor. 3. There is no influence of age or educational level factors on extreme speakers. Neither do the years they spent in Spain have any influence. 94 Antonio Manjón-Cabeza Cruz The distribution of citizens of Ecuadorian origin with regard to the gender factor is, at first glance, paradoxical. Among those who prefer -ito, we find six men and two women. These two women also appear in seventh and eighth place. Among the subjects in the lower third, users of -illo, there are seven men and only one woman, M21-021, which is the only case of -illo that we saw in the example (4.19). Most of the women are in the middle third. This does not mean that they present variation: they are categorical in their use of -ito, but they use few diminutives and do not have as much weight as those who use more. Although it may seem strange, if we look at the distribution of subjects in Table 4.10, we find that the Ecuadorians at the extremes are men interviewed by Spaniards. Women tend to occupy intermediate positions, with little use of diminutives. It seems that Ecuadorian men tend to accommodate the Spanish interviewer with two strategies: the first one is represented by the first third and consists of using appreciative diminutives in abundance, for which they use -ito; the second strategy, evident in the last third of subjects, is to use diminutives in -illo. In short, the rejection by migrants of the local -ico morpheme is evident, since we only have one example, almost certainly shared with Andean Spanish. The preference for -ito is also clear, helped both by its almost exclusive use in Ecuadorian Spanish and by the manifest prestige of the use of this morpheme in the Spanish of the centre of the peninsula and its penetration into the Spanish of Granada. Nevertheless, there is a not inconsiderable percentage of -illo among Ecuadorian migrants: 18.63%. This percentage is not uniformly distributed since it varies according to the nationality of the interviewers and the sex of the interviewees. It seems that accommodation can be expected in the percentage of -illo morphemes (25%) that appear when the interviewer is from Andalusia, but the penetration of this morpheme is significant when the interviewer is from Ecuador (10.3%). Without a doubt, these percentages show a gradual process of sociolinguistic accommodation to a new variety, so we observe both short-term accommodation (with the local interlocutor) and long-term accommodation (among Ecuadorians), as analysed by Trudgill (1986) in a pioneering work. As regards the obvious influence of the sex factor in the use of one or another suffix, the data is striking: there is only one case of -illo in women. If we follow Trudgill (1983), it would seem that this is the result of the performance of prestige patterns, to which women are usually more predisposed. 4.5 Conclusions To answer the questions posed at the end of the introductory section, we have studied two linguistic features using quantitative methods. For this purpose, we have used two comparable corpora, one from Granada’s vernaculars, the PRESEEA corpus from Granada, and the other from Ecuadorian migrants living in Granada: the ECUGRA corpus. We wondered whether migrants tended to abandon the divergent features with respect to their neighbours and accommodate themselves linguistically to them. Sociolinguistic accommodation by Ecuadorian migrants in Granada 95 In this case, the answer is generally negative. Ecuadorian migrants in Granada very strongly maintain the features of Andean Spanish. This is evident in their treatment of the /-s/ which still has some typically Andean characteristics, very different from the treatment that the vernaculars give to the same sound. In the case of diminutive morphemes, we can no longer be so categorical because a sociolinguistic accommodation is observed as cases of the morpheme -illo appear, mainly amongst men. As an answer to the second question, we believe that this general lack of convergence is greatly influenced by the fact that the Andean features studied (maintenance of the /-s/ and the majority use of the appreciative suffix -ito/a) coincide with the prestigious features of the Spanish of central-northern Spain, with which migrants have contact both through the media and through its increasingly evident penetration into the Andalusian variety, which entails the stigmatisation of vernacular features. From the preceding two paragraphs, the answer to the third question can be inferred. The evidence makes us choose the second option, that is to say, the phenomena studied, one of a phonetic nature and another grammatical, do not receive the same treatment. We have already seen that the phonetics of the respondents do not converge with the vernacular, while the suffixes used show a certain tendency to converge. In this process of convergence, we believe that the phenomenon of lexical diffusion is fundamental since migrants use the suffix -illo/a with very few lexemes, which are repeated. Above all, the adverbial locution un poquillo and the use of the suffix with a word unknown in Ecuador, chapucilla, stand out. If there are no processes of accommodation in a phenomenon, as is the case with /-s/, we cannot speak of social factors, but we can answer the initial question regarding the possible existence of social factors if we stick to the appreciative suffixes. The answer, in this case, must be in the affirmative. The incipient use of -illo shows short-term accommodation processes since it tends to be used more frequently when the interviewer is a speaker of the vernacular, although we also observe long-term accommodation because it also appears when the interviewer is Ecuadorian. Along with the social factor of the type of interviewer, the sex of the respondent is fundamental. Women show an almost categorical absence of the Andalusian -illo/a, which, however, appears in men. We believe that this situation of different rates of accommodation is due to the fact, repeatedly documented in various sociolinguistic situations, that women are more sensitive to patterns of prestige, so they always use the Andean -ito because it coincides with the most prestigious suffix in Spain which is the predominant one in the variety used in the centre and north of the peninsula. To conclude, we must say that we have tried to explain the paradoxical sociolinguistic situation of Ecuadorian migrants who, when they arrive in Granada, find themselves surrounded by a variety of Spanish in which their neighbours, of generally low socio-cultural levels, employ variants that are sometimes very different from those used by people from Granada from higher classes or the mass media. These higher-level variants are sometimes used by Andean migrants in a vernacular way. 96 Antonio Manjón-Cabeza Cruz It is almost certain that investigating these bidialectal contexts from a sociolinguistic point of view will help us to better understand the processes of convergence between varieties of the same language, both because of the accommodation of migrants and because of the changes in the variety used by the host community. 4.6 Acknowledgements This work has received funding from the Paler-Gr project (Proyecto Para el Estudio Lingüístico de la Adecuación de los Emigrantes Residentes en Granada), reference: B-HUM-268-UGR18, of the ERDF Operational Programme Andalucía 2014–2020. I would like to thank the members of the project “Agenda 2050. El español de Granada: Procesos de variación y cambio espacial y sociales”, reference: PID2019-104982GB-C53 PROJECT/SRA (State Research Agency/10.13039/501100011033). Notes 1 The mismatch between of the INE data with those of the UN (2019) are explained because the INE does not consider migrants who have acquired Spanish nationality. 2 Ecuadorians prefer to settle in the Zaidin-Vergeles and Norte-Beiro neighborhoods, which according to 2018 AEAT (State Tax Administration Agency) statistics are the neighborhoods with the lowest disposable income in the city, with 18,198 and 17,759 euros of income per capita, respectively, compared to neighborhoods such as Centro (28,539 euros), Realejo (28,245 euros), or San Antón (27,430 euros), which occupy the highest positions. 3 It must always be borne in mind that we are dealing with the same language and that dialectal differences do not, in any case, endanger intercomprehension. 4 In this case, the coding is ECUGRA = corpus of Ecuadorian migrants (the corpus of vernaculars is coded as GRAN), H = man, 1 = first age level, 1 = primary education, and 022 is the number of the subjects. The number of the subjects is relevant because 1 to 15 were interviewed by Ecuadorians and 16 to 30 by Spaniards. 5 The pronunciation of the sound represented by <ll> in Spanish can be very varied. In Granada, the central palatal pronunciation /ʝ/ is exclusive, sometimes with a slight friction or rehilamiento, reminiscent of English Joe. In the case of Ecuadorian migrants, we have not found any palatal lateral pronunciation /ʎ/, which is detected in the Ecuadorian highlands due to the influence of Quechua. The most common pronunciation is the same as in people from Granada: central palatal with slight rehilamiento, although we have also found cases of a tendency to vocalization /j/, also common in some varieties of American Spanish. Phonetic accommodation, therefore, does not seem to represent added difficulties. 6 On the Ecuadorian coast it is usually cachuelo, while in the mountains chauche is used more. References AEAT (2018). Memorias y estadísticas, https://www .agenciatributaria .es /AEAT .internet /Inicio /La _Agencia _Tributaria /Memorias _y _estadisticas _tributarias /Memorias _y _estadisticas _tributarias .shtml