2.1. Metalinguistic concepts: the differentiation between linguistic attitudes and perceptions

Following the sociolinguistic parameters, at the end of last century studies that take into account the metalinguistic ideas of Spanish speakers began to emerge: metalinguistic utterances are linguistic statements that refer to something linguistic (Gauger, 1976, p. 44). However, metalinguistic research on Andalusian varieties focuses exclusively on the attitudinal aspects[7] and thus restricts the term metalinguistic consciousness to the attitudinal aspect, although Schlieben-Lange (1991, p. 25) points out that metalinguistic consciousness, despite being interrelated with attitude, forms another part of cognitive consciousness. Interesting regarding this contribution seems to be the idea that metalinguistic consciousness leads the speakers to identify themselves with a certain linguistic variety and helps them in the social categorization of speakers in different communicative situations (Scharloth, 2005, p. 15). However, it should not be forgotten that the comments of speakers about their linguistic consciousness do not always have to be conscious (Gauger, 1976, p. 55). Therefore, several linguists (Krefeld & Pustka, 2010; Preston, 1999) prefer to differentiate the metalinguistic consciousness from the supposedly more conscious metalinguistic knowledge. Consequently, this contribution understands the metalinguistic consciousness as the basis of all concepts of metalinguistic knowledge. Therefore, the term is subdivided into linguistic attitude and metalinguistic knowledge. The aspect of metalinguistic knowledge, not restricted to the attitudinal aspect, is very little investigated within Spanish linguistics, as Serrano Morales (2001) points out: “In the Hispanic context, however, such works are practically non-existent” (p. 2). For the Andalusian varieties, it even represents a desideratum.[8]

In general, the scientific history of the metalinguistic perspective of the speakers themselves, despite having a modest beginning in the past,[9] is relatively recent. For the first time, Preston (1986, 1989) and Niedzielski and Preston (2000) focused on the research of speakers’ ideas about linguistic variation. According to Preston (2005, p. 149), the goal of the so-called perceptual dialectology is to clarify the cognitive knowledge of speakers about linguistic variation through the analysis of conscious and unconscious reactions to perceived spoken language. However, adaptations of the perceptual dialectology in the Hispanic domain are scarce:[10] “While there is growing interest in the investigation of the perception of dialect variation in the United States, such research related to Spanish has not been very productive” (Díaz-Campos & Navarro-Galisteo, 2009, p. 181). The perceptual variety linguistics, then, broadens the theoretical basis of perceptual dialectology. While perceptual dialectology focuses on diatopic variation and, following the English term dialect in the American tradition, on the speakers’ linguistic attitudes as an element of sociolinguistics, perceptual variety linguistics extends this range by social and situational variations (see Krefeld & Pustka, 2010, p. 10).

Krefeld (2005, p. 162) is the first to use the term Perzeptive Varietätenlinguistik (perceptual variety linguistics), which seeks to clasp the entire variation system from the speakers’ point of view.[11] Here, it is necessary to demonstrate the relationship between metalinguistic knowledge and concrete speech, which are mutually conditioned by a circular process between linguistic production and perception (Postlep, 2012, p. 89). Following Pustka (2007, p. 18), we believed that speakers modify their linguistic competence through perceptions throughout their life. This is why it is very important, besides from investigating linguistic production, to analyze the speakers’ perception, which precedes production from childhood: “Die Perzeption geht der Produktion fast immer voraus” (Krefeld & Pustka, 2014, p. 12). Chomsky (1980) already indicates that it is not only necessary to investigate the linguistic production of the speakers, but also their linguistic knowledge. In this sense, he speaks of linguistic and mental representations of an ideal speaker who knows how to categorize grammatical elements (p. 5). Within the linguistics of variation it is a commonplace that a homogeneous speech community with ideal speakers (Chomsky, 1980, p. 25) does not exist. Accordingly, Pustka (2007, pp. 9–11; 2008, pp. 214–215) and Krefeld and Pustka (2010, pp. 10–16) include mental representations of the linguistic variation in the part of speakers’ linguistic knowledge, representations that are subdivided into an abstract variational knowledge that a speaker can manifest without an external stimulus and another concrete metalinguistic knowledge about variation that is based on direct perceptions. Along with the Chomskian competence then, according to Krefeld and Pustka (2010, p. 11), there is also knowledge about the variation in a given language.

2.2. About the necessity to distinguish between abstract cognitive concepts and concrete perceptions

Although they exert a mutual influence, it is important to distinguish between the terms of concept and perception, since they constitute two fundamental theoretical pillars of the perceptive linguistics of variation. As already mentioned, abstract representations or concepts are part of the metalinguistic knowledge and are also present without concrete stimuli. Therefore, they are part of the permanent metalinguistic knowledge and belong to the langue. Contrary to abstract concepts, concrete perceptions belong to the parole. Consequently, we distinguish between abstract representations and representations based on concrete perceptions (Krefeld & Pustka, 2010, p. 14).

However, it should be mentioned that all representations, including abstract ones, are based on concrete perceptions and in a constant process of change through each communicative experience. Additionally, concrete perceptions are controlled by abstract representations, whereby both are reciprocal (Krefeld & Pustka, 2010, p. 14). This relationship does not prevent the speakers from maintaining strong stereotypes about certain varieties, due to extra-linguistic influence and the maintenance force of these in the abstract representations. As a result of this possible ignorance, it has often been shown that some speakers emphasize in their mental maps dialectal frontiers represented without stimuli that in the concrete perception no longer exist (Glose & Pustka, 2014). Other methodological problems are attached to the previous: what one wants to investigate is the metalinguistic knowledge of the speakers or the langue, but only the concrete production of the speakers or the parole can be analyzed in cartoons, imitations, surveys on their variational concepts or experiments with voice tests (see Krefeld & Pustka, 2010, p. 15). And since there is no direct access to the permanent metalinguistic knowledge of the speakers, there is always the risk of answering survey questions in accordance with a supposed social desirability (Postlep, 2012, p. 91). Apart from that, there are differences in linguistic knowledge not only at the individual level, where it is evident (Krefeld & Pustka, 2010, p. 15), but also depending on the metalinguistic tradition in certain communities (Sinner, 2014, p. 131). Closely linked to the metalinguistic tradition in a speech community is also the division between self- and hetero-representations: while metalinguistic concepts about the community itself are defined as self-representations, metalinguistic ideas about other linguistic communities unrelated to the speaker’s own would be hetero-representations (Krefeld & Pustka, 2010, p. 15).

While perceptual variety linguistics intends to include all the variational aspects of the diasystem, it simultaneously approaches, like perceptual dialectology, the linguistic attitudes of speakers, which are part of the theoretical model of perceptual variety linguistics—initially introduced by Pustka (2007) as “culture”. However, there is some debate about the position of linguistic attitudes within the scheme: while Krefeld and Pustka (2010, p. 12) separate the attitudes from representations, although they point to a certain relation between them, Postlep (2010a, pp. 55–57) denies dividing concepts into a cognitive and an affective component. Drawing on the ideas of Kabatek (1996, p. 42) and Berruto (2002, p. 353) claiming that a separation between the cognitive structure and the affective component of metalinguistic knowledge is impossible since linguistic attitudes also obtain a conative component that influences linguistic production, Postlep links metalinguistic representations and linguistic attitudes (pp. 55–56). In the present article, a division is maintained between attitudes and cognitive knowledge; the acts of delimiting a certain dialectal zone or naming linguistic features by themselves seem to us simply cognitive even though, like Krefeld and Pustka (2010), we do not deny that they can also have an attitudinal aspect, for example when the linguistic traits lead to positive or negative comments, which we call following Anders (2010)Triggermerkmale (linguistic triggertraits).

2.3. Methodology

For the investigation of the Andalusian varieties from the perspective of the perceptual variety linguistics we have created a corpus based on a direct survey that is divided into two parts.[12] To create the survey we have used methodological elements that have already been applied on other perceptive studies (Postlep, 2010a; Preston (Ed.), 1999; Pustka, 2007; Stoeckle, 2014), since the methodological diversity of perceptual variety linguistics should be treated as complementary (see Krefeld & Pustka, 2010, p. 15).

The first part of our contribution focuses on the associations that arise between the respondents about the Andalusian varieties and various phonetic features: here, we distinguish between the metalinguistic knowledge, which is demonstrated in mental maps, and the linguistic attitudes of the respondents. This first part of the survey follows the methodical ideas of the perceptual dialectology in asking the speakers to draw regions of different Andalusian speeches on a blank map of Andalusia, to characterize the linguistic zones they highlight and to point out the typical phonic features of each specific variational zone. The latter is the most important aspect in the survey of abstract representations, since it deals with the dimension of specific linguistic knowledge or the inhaltsbezogene Dimension (Anders, 2010, p. 164). This is where the respondents describe the linguistic phenomena that help to identify a specific variety, thus making explicit their idea of salient phenomena of each Andalusian variety. Unlike Postlep (2010b, p. 70), the comments of the speakers are not filtered to extract only the most universal and scientifically valid aspects. Regardless of the extreme complexity of the comments and the greater difficulty in categorizing comments (Anders, 2010, p. 272), this contribution tries to collect strictly emic data[13] and to do a scientific categorization, which is performed in the analysis through a linguistic classification adapted from Anders (2010, pp. 268–275) and Pustka (2007, p. 231). This classification consists of five different groups in terms of phonic, lexical, and morpho-syntactic aspects as well as descriptive and evaluative expressions that are further subdivided into 24 subgroups. Although in this contribution we are solely interested in the phonic aspects and the descriptive or evaluative expressions of this classification, the challenge of this kind of classification of linguistic associations made by naïve speakers is still that it has to satisfy the demands of an empirical work and, at the same time, should reflect the inexperience of the utterances made by the speakers (Anders, 2010, p. 267). The last two parts of the survey of abstract representations deal with the subjective difference of the idiolect of the speakers with other Andalusian speakers, the “degree-of-difference”, to put it in Niedzielski’s and Preston’s terms (Niedzielski & Preston, 2000, p. 77), and with the linguistic attitude of the respondents to the different Andalusian varieties. The last aspect of the questionnaire is similar to what Preston (1999) calls “correct and pleasant” (p. XXXIV). However, instead of giving grades from 1 to 10 to the sympathy and antipathy of the different Andalusian varieties, respondents have to decide on sympathy or antipathy towards the other Andalusian varieties on a binary nominal scale. In this respect, the questionnaire rather resembles Hundt’s pilot study (2010, p. 218) on the linguistic consciousness of German speakers. Respondents can choose on a binary scale between sympathy and antipathy from a variety or leave the boxes blank. Likewise, respondents have the option to rate linguistic zones as correct or incorrect.

The second part of the survey is dedicated to the perception. Speakers listened to audio samples and decided from which Andalusian variational area the speakers are and to what social strata they belong. In total, there are 14 different audio samples: twelve of the speakers are from western Andalusia and have different levels of education. The main focus in the collection of these 14 stimuli is on a heterogeneous presence of phonic features of different Andalusian speech communities: in some of the stimuli one can perceive seseo, in other ceceo or heheo,[14] etc. The audio samples last from 20 to 45 seconds. Unlike Postlep (2010a, p. 103), each respondent had to listen to all 14 stimuli, which are made up of spontaneous, rather than practiced, speech acts by Andalusian speakers. We follow Brade’s (2010, p. 443) and Williams, Garrett, and Coupland’s (1999, p. 346) contributions illuminating that the spontaneous speech of a speaker is much more authentic than some sentences being read. The attention of the respondents is higher because the task becomes more interesting if it is a free talk. Moreover, the difficulty in filiation to a given variety is easier than in a read stimulus (Brade, 2010, p. 445).

Answers of a total of 30 participants[15] constitute the corpus, which does not seek to achieve sufficient representativeness for a quantitative work, but to offer tendencies for the speech community. This number of speakers approaches the participants in Anders (2010) in each speech community of her study on the German of Saxony and can guarantee the participation of a wide base of Jerezano speakers without exceeding the exploration capacity of a single interviewer, as Stoeckle (2014, p. 102) already admits. The surveys were conducted at different places in Jerez in 2015 and lasted between 40 and 75 minutes per speaker. In order to guarantee diverse social parameters, not only participants of different generations or gender have been interviewed, but also the respondents’ level of education was taken into account. Thirteen speakers are college graduates and 11 at least have the ESO (secondary education). Six participants are less educated or illiterate. As the test does not reach an absolute representativeness of the local speech community, we aim at increasing the representativeness by taking into account extra-linguistic parameters and creating a heterogeneous corpus based on these. Especially the integration of speakers with lower education is new within the research of the perceptual variety linguistics that, in many cases, continues to find its participants exclusively among the young students (see Hundt, 2010; Preston, 1989; Purschke, 2011; Pustka, 2007). Only in a few cases—mainly for a better understanding of the qualitative results—we will quantify the data, which only indicate trends, but absolutely not a representative scientific saturation (see Kuckartz, 2014, p. 84).

3.1. Abstract cognitive representations of the phonic variation

On average, speakers divided the blank maps in the draw‑a‑map task (Preston, 1999) into 7 different speech zones. The conception of the dialectal area of Andalusia ranges from a minimum of 3 (Figure 1) to a maximum of 16 (Figure 2) different speech areas, which refers to an enormous capacity of differentiation. Most respondents (15) structure the Andalusian dialect area in 6 to 9 zones, as the average of 7.0 mentions reflects. The average of divisions varies slightly according to the extra-linguistic factors. The most determinant factor seems to be age: while the younger generation (6.0) and the more advanced one (6.4) identify a similar number of Andalusian varieties, the intermediate generation highlights an average of 8.7 dialect sectors. This result reaffirms Hansen’s (2012) conclusion on the influence of the generational factor on dialectal knowledge: very young speakers emphasize less areas of divergent speech (p. 69).

Figure 1. Minimum of selective markings in the map (speaker R): Sevillian, Jerezano, Granadino.

Figure 2. Maximum of selective markings in the map (speaker B): Sevillian, Jerezano, Gaditano, Onubense, Almeriense, Cordoban, Jienense, Malagueño, Lepero, Astigitano, Llanito, Sierra de Cádiz, Sanluqueño, Chipionense.

The way of marking the points is selective (Anders, 2010, p. 197) in almost all cases (Figures 1 and 2). This means that the respondents have only highlighted areas of Andalusian varieties that can stand out from other areas, but for them the Andalusian dialect does not form a continuum. This is not surprising since, according to Anders (cf. 2010, p. 197), this type of map structure seems to be the most practiced in these types of mental representations of speakers who are not specialized in linguistics. Preston’s (1989) data points in the same direction. Contrary to the selective way of mapping, the exhaustive way (Hundt, 2010, p. 180) allows us to infer that the respondent believes that the Andalusian dialectal area is a continuous one. However, only the respondent N does a comprehensive marking of the map (Figure 3).

Figure 3. Exhaustive marking in the map (speaker N): Cordobese, Onubense, Sevillian, Rural Sevillian, Gaditano, Sierra de Cádiz, Ubriqueño, Jerezano, Almeriense, Jienense, Grandino, Malageño.

Among the respondents who selectively separate the Andalusian varieties, the attention they give to provincial capitals (Cádiz, Seville, Córdoba, Granada, and, to a lesser extent, Málaga and Jaén) is more striking than that for Jerez and the rural areas around the Sierra de Cádiz. Twenty-four of 30 respondents take the main urban nucleus of the Andalusian provinces as a point of reference in the reproduction of the cognitive map. In highlighting provincial capitals, respondents point out the importance they exert within their cognitive space and underline their importance in the Andalusian reality, as already stressed by Bustos Tovar (2013): “En el caso de Andalucía, las áreas dialectales que se aglutinan en torno a ciertos núcleos, tienen una gran estabilidad y no sufren la agresión de variantes focalizadas en otros núcleos” (p. 36). According to Lynch (1965/2001, p. 62), cities function as an optical point of reference in general. As in the blank map also the names of some cities appear as points of orientation, the attention given to the urban centers of Andalusia by the speakers is not surprising. In his pilot study, Hundt (2010) observes that speakers draw dialect zones around important cities (p. 193). Thus, the spatial categorization of the autonomous community of Andalusia has a strong daily relevance in the distribution of the cognitive space and influences the consciousness of the Andalusian speakers not only on a political-administrative level, but also regarding questions related to the use of language. The political provinces, and especially the capitals of these provinces, constitute the cognitive space of the Andalusian speakers, which is preceded by the categorization of the variational space of the Andalusian varieties, as Christen (2010) also points out for Switzerland.

With regard to the speech communities of Seville, Jerez, and rural Cádiz, which are the ones that are important to this contribution, we must conclude that the two major urban centers of the region are the most prominent in the corpus. Moreover, more than half of the respondents pointed to the Sierra de Cádiz or different villages within this province of Cádiz as a zone of divergent variety to their own. Also, more than a quarter of the respondents name the small speech communities of other regions of Cádiz (e.g., Bajo Guadalquivir with Sanlúcar de Barrameda, etc.). It is noteworthy that some respondents (4) differentiate between the capital of Seville and the speech communities of Sevillian small towns (e.g., Dos Hermanas or Écija), thus indicating an urban/rural division also in the province of Seville. This fact is reflected in the scientific literature on the speech community of Seville (Carbonero Cano, 1982, p. 26). Consequently, the geographical proximity of rural localities to their own speech community influences cognitive representations, at least in those of older generations. The younger generation appears to be much more focused on urban points, including those in the east of Andalusia (Granada, Córdoba, Málaga), leaving out of their maps many times the rural areas highlighted by older generations, and focusing on urban points of larger demographics within the autonomous community.

Linguistic features that have frequently been mentioned in scientific linguistic concepts, such as “ceceo” (B) or “seseo” (Y), underline a great metalinguistic knowledge of the respondents. In total, there are 121 divergent linguistic phenomena that have been described by Jerezano speakers. Of these phenomena nearly two thirds are phonic features, which highlights the phonic character of the Andalusian dialectal peculiarities already underlined by Narbona Jiménez, Cano Aguilar, and Velarde-Pérez (2011, p. 165). The phonic features are divided into three subcategories of vowel, consonantal, and prosodic features, which in turn differ in several characteristics. The subcategories of phonological change in consonants (e.g., “ceceo” for respondent B) and vowels (e.g., “alargamiento vocálico” for respondent H) together account for half of all phonic-related mentions. Apart from these subgroups, a further three account for a quarter of all references: consonant quality, single consonant, and negative evaluative expressions.

For respondents, the phonic elements are also the most prominent linguistic traits of their own variety. Within these, the salient phonological phenomenon of Jerez is ceceo, a feature that is very abundant in linguistic characterizations, as for example in “cecean” (Q), “ceceo” (B) or “s = z” (γ). Within the descriptions of ceceo there are no extra-linguistic differences, since independently of the educational level, gender, or age, the phonic feature stands out and all describe it with a specific linguistic term, as in “ceceo” (O) or “cecean” (N), or with a more plastic vision, as for example in “cambian la s por la c; ejemplo caza (casa)” (D), in “sustituyen la letra s por la z” (P), or in “cambian la s por la z” (Ñ). In addition, some indicate that typical of their own speech community are the relaxed pronunciations of the /t͡ʃ/ and /x/: “aspiración de la j → h” (X), “sh” (O), “musho” (ß) o “koshe” (Z) (Table 1).

The variety of Seville seems to be different from that of Jerez: according to the speakers, in the Sevillian variety there is a phonological neutralization in favor of seseo instead of ceceo. This linguistic feature is, with 16 comments, salient for the respondents. Although men indicate the phonic feature more frequently, the great extra-linguistic difference with respect to the naming of seseo for Seville is the educational level of the speakers. While low-educated speakers do not refer to phenomena by virtue of the technical term, but using the actual sound [s], as in “s” (O) or “hablan con s” (G), high-educated people often employ the appropriate linguistic term, for example in “sesear” (Q), “seseo” (X), or “sesean” (N). A certain prestige of the Sevillian variety in the opinion of the respondents is only partially noticed in the comments made. Although Seville’s variety is defined as “muy fino” (E) or even in a self-referential manner as “el habla es más fina” (V), there are also negative references indicating that the variety is seen as “chulita” (Z) (Table 2).

Unlike the capital of Seville and the own speech community, the participants frequently value the rural area of the Sierra de Cádiz. These comments are always negative, as in “demasiado vulgar” (α), or ironic, as in “refinado jejeje” (Z), and usually made by college students. Aside from prosodic signs, as in “hablan cantando” (D) or “melodía al hablar” (G), the notes on the use of [h] instead of /s/ abound, for example “pronuncian s como j (Puerto Jerrano)” (N), “s = h” (γ) or “hablan mucho con la j y z (Jervejita)” (T). There are participants who emphasize the ceceo of the speakers of the rural area and there are others that emphasize the heheo, but only a few use in their descriptions the appropriate scientific term, “ceceo” (B) or “heheo” (β) (Table 3).

Therefore, Jerez and the rural areas of the western provinces of Cádiz and Seville, although separated by isoglosses put by the participants, constitute a ceceante zone. This naïve view separates a seseante area, linked to the speech communities of the urban centers with political-administrative functions at the provincial level (Málaga, capital of Seville, capital of Cádiz), from another clearly rural ceceante area plus their own speech community. They think that they have more phonic features in common with rural areas, such as ceceo, use of the [ʃ] fricative and aspiration of the /x/. This difference between the prominent spaces is also manifested in the degree of difference that the speakers themselves establish between Seville, Jerez, and the rural areas of Cádiz. While the eastern areas of Andalusia (e.g., Granada or Almería) are rated at an average of 4.0, participants split the western part into two different areas: a part is made up of the cities of Seville and Málaga, adding some of the southeast regions of the province of Huelva. The zones in this part receive the placement of “more different than similar” with an average of more than 2.9. Finally, they point to the nearest part of Jerez, some rural areas of the province of Seville and the rural regions of the province of Cádiz including the capital. For Jerez respondents, these areas are “similar” (1.8–2.4) to their own speech community. On the one hand, middle-aged speakers perceive the rural areas very close to their own and establish a greater distance between their own speech community and those of Seville and Cádiz. On the other hand, educated and younger speaker see a slightly approach to Cádiz’s variety, but not to the urban variety of Seville. Gender differences offer an image similar to the generational ones, manifesting a closer proximity to the ceceante zones (Sierra de Cádiz) and greater distance to the seseante regions (Cádiz, Seville) by the women surveyed, but a greater proximity of the men to the regions with a salient seseo.

With regard to the linguistic attitude, it is necessary to mention that Jerezano speakers evaluate very positively the sympathy of the closer western Andalusian varieties. Along with the variety of Jerez, the rural ones of western Andalusia are qualified as being sympathetic. However, the urban variety of Seville is, according to the speakers, also rather sympathetic than unfriendly. The only sociolinguistic differences are in the handling of the variety of urban Seville, whose habla seems to attract the sympathy of the most prepared men and younger generations. The second generation, the women and averagely-educated respondents show more antipathy towards the Sevillian variety. However, only in the case of women does antipathy outweigh sympathy. This observation underscores Kontra’s (2002) statements about divided subjective opinions of speakers who live outside provincial capitals towards major urban centers. Regarding the accuracy of the Andalusian varieties, it should be noted that the rural points of the province of Cádiz and some of the province of Seville receive the worst appraisal by the Jerezano speakers, a fact that reflects a possible stigma of the rural Andalusian varieties. The capital of Cádiz and Jerez itself receive slightly more positive than negative evaluations. The vast majority of participants considers the variety of the urban speech community of Seville as a bit more correct than incorrect. This fact shows a certain prestige of the Sevillian habla for the Jerezano speakers on the one hand. On the other, however, more than a quarter considers the variety of Seville to be incorrect.

3.2. Concrete perceptions of the phonic variation

The Jerezano speakers have a great metalinguistic knowledge, which is also demonstrated in the part of the concrete perceptions. Although second-generation participants and those with an average level of education appear to be the respondents who correctly track the localization of voice sample tests rather than the rest, we must refute the hypothesis that, compared to young people, older people locate other speakers better. While perceptual studies in anglophone language communities have demonstrated the validity of this hypothesis (Torgersen, 2012, p. 91), it is important to emphasize that young people also have a high linguistic consciousness in the local speech community being examined here. It is striking that there are also no major differences in the accuracy of the perception of the level of education of the participants in the corpus, although it seems certain that lower educated people use less adequate linguistic terms than those with a college degree or have an average level of education.

The linguistic features perceived in greater numbers have been those of the phonological neutralization of /s/ and /θ/, associated with divergent dialectal zones in western Andalusia. The defonologización in favor of ceceo is always located in a self-referential way, that is, within the own speech community of Jerez, or within the rural varieties of Cádiz: therefore, the phonic feature of ceceo is associated mainly with the own speech of the respondents or with the near varieties of the rural region, highlighted in the degree of difference by their closeness. Many times the separation of both areas in the concrete perception, in the part of the simple representations still well distinguished, is very complicated for the Jerezano speakers: if they perceive a ceceante speaker they doubt if it is of their own speech community or from the rural areas of the province of Cádiz, or even from the rural area of ​​the province of Seville, for example in the stimuli (a), (c), (f), (g), (l), and (m). This result strengthens, on the one hand, Ropero Núñez’s (2001) conclusion for the metalinguistic consciousness of Sevillian speakers that ceceo is located more in the rural speech communities of Andalusia: “El ceceo en cambio todavía se suele asociar con el habla rural” (p. 39). But the Jerezano speakers associate a perceived ceceo of a stimulus not only with the varieties of the rural areas of Cádiz and Seville, but also with their own variety. This fact, on the other hand, emphasizes a remarkable rurbanity of the speech community of Jerez (Harjus, 2018), since, unlike in the part of the representations, the respondents do not perceive any difference between both dialect zones.

On the contrary, the phonological neutralization in favor of seseo, as perceived in stimuli (b), (d), (e), or (k), is never associated with the speech community of Jerez, but always with the varieties of the provincial capitals of western Andalusia, especially with those of Seville and Cádiz, and, although rarely, with the speech community of Málaga and the zone of the Cordoban variety in eastern Andalusia. Therefore, the respondents divide the western Andalusian communicative space into two zones: the varieties of the large urban conglomerates, Seville, Cádiz, and Málaga, are associated with the phonic feature of seseo, while the ones spoken in the rural areas of the provinces of Cádiz and Seville along with the variety of Jerez are linked to the phenomenon of ceceo.

Therefore, according to the conscience of the respondents from Jerez, there is a multidimensionality of linguistic communication in the dialectal space of western Andalusia, since they distinguish between an urban and a rural area. Piredda (2014) analyzes the same phenomenon in a study on the perception of Sardinian speakers, where she summarizes that their respondents cannot exactly locate the stimulus but can always distinguish perfectly between Sardinian rural and urban areas (pp. 75–79). In contrast to Sardinian speakers, who do not mention specific linguistic markers (Piredda, 2014, p. 79), the Andalusian participants of this contribution justify their locations in the urban area (capitals of Seville, Cádiz, Málaga, or Córdoba) or rural areas of Cádiz or Seville, including its own speech community, with the distinctive features of seseo or ceceo respectively. Above all, the absolute and total rejection of seseo for the local speech community indicates, as in the part of the representations, that this phonic feature is perceived as a deviation in the Jerezano variety. The fact that Boughton (2010) indicates that respondents in Paris note small linguistic differences in Alsace-Lorraine variety, but do not know how to delimit the geographical origin of perceived speakers (p. 107), underlines the enormous metalinguistic knowledge that Jerezano speakers have about their linguistic variety in general, and of the phonological neutralization of the /s/ and the /θ/ in particular, since they can distinguish perfectly between these salient linguistic features.

With regard to the perception of social factors, it should be noted that seseo, as a phenomenon perceived from a stimulus, leads the speakers—much more frequently than it is the case with a salient ceceo—to locate the stimuli speakers within a higher educational level. On the contrary, ceceo perceived in a voice sample is often associated with the variety of speakers with a low level of education. This result seems to confirm, for the Jerezano speech community, Ropero Núñez’s ideas about the Sevillian one: that seseo has more prestige than ceceo and it is highly socially accepted throughout western Andalusia (2001, p. 39).

Within the Jerez speech community, seseo also seems to be prototypical for more cultured speakers, and with this it can be concluded that it is also highly accepted, since there is a correlation between the assignment of a certain social character and the acceptance of the perceived linguistic variable (see Krefeld, 2010, p. 171). The results of linguistic attitudes about the variety of urban Seville demonstrate this as well. However, it must be reiterated that, despite the supposedly higher social status of seseo versus ceceo, this phonic feature is never associated with the local speech community itself. Seseo, which is accepted regardless, may not be part of the local variety, since it cannot be from there: “no puede ser de aquí” (α). Therefore a stimulus with the salient phonic phenomenon of seseo is never located within the Jerezano speech community.

In addition, seseo is not always associated with speakers of more educated levels. While only three respondents believe that the speaker from stimulus (d) has a college degree, almost a third of all respondents (9) perceive a speaker with little education. And even among the social categorizations of the speakers of stimuli (b), (e), and (k), which are mostly characterized as college students, there are up to one sixth of the perceptions of speakers without secondary education. Not only does this fact demonstrate that the individual qualification of phonic features can vary considerably (Anders, 2010, p. 354), but also that the use of seseo is not always prestigious.

Then, it is necessary to reject an alleged stigmatization of ceceo in Jerez, often associated to it in other Andalusian speech communities, such as Seville (Ropero Núñez, 2001) or Granada (Moya Corral & García Wiedemann, 1995): the social perceptions of the speakers of the stimuli (c), (f), and (m) as speakers with the highest level of education clearly demonstrate that ceceo forms part of the prestigious phonic forms of the Jerezano variety. However, it is striking that among low-educated respondents almost nobody recognizes a high-cultured speaker: uneducated speakers do not associate a high level of education with a salient ceceo. On the contrary, the rest of the respondents does not hesitate to locate the ceceo speakers between the higher educational levels. And although it is not possible to rule out the possibility that certain high social categorizations of the ceceante stimuli (c, f, m) and the low social caracterization of the seseante stimulus (d) can be motivated by protection of the own variety itself, as Brade (2010) supposes for self-consciousness of Caribbean varieties (p. 452), this result underscores the persistence of the old stereotype of the use of ceceo among less educated speakers. Yet, among the higher educated Jerezano speakers, this stereotype seems to be antiquated and the social perceptions on high educational levels show the concealed prestige that ceceo receives even among the most educated speakers of the local speech community.

Finally, it is indispensable to verify the importance of the content of a stimulus and the narration style of a speaker in the voice sample for the social perception. Brade (2010, p. 445) already emphasizes the problem that respondents can turn their attention in unread stimuli to the style of the narrative: this way we can explain the positive comments about the speaker of stimulus (b), who speaks fluently (“habla fluido”; N), the positive ones about the stimulus (m), “por la forma del comentario” (F), or the negative ones about the speaker of the voice sample which is described as relatively uncultivated by the way of expressing itself (“por la forma de expresarse”; O). Additionally, some respondents turn to the content of what is expressed in the stimuli. As already mentioned in the methodology part, the decision to use stimuli taken from free conversations has the advantage of obtaining a maximum of spontaneity and authenticity, but they invite the speaker to look at the content: “[Es] liegt die Annahme nahe, dass der Inhalt des Textes besonders großen Einfluss auf die Bewertung nimmt” (Brade, 2010, p. 443). In the part of the concrete perceptions, only some participants express openly that they base their social categorization of the speaker in the content narrated in the voice sample, as for example in “contenido básico” (N) for the stimulus (i) or “no tiene inquietudes” (Q) for the stimulus (g). Nevertheless, it is evident that often the combination of the narrative style and the content expressed in the stimulus forms the main basis for the social categorization of the perceived speaker: in this way, the perceived salient linguistic features of the speakers of the stimuli (a) and (c), mainly ceceo and the fricativization of the affricate, are the same. However, once respondents identify a low-educated speaker in stimulus (a), they recognize a speaker with a university grant in stimuli (c). This result demonstrates the utmost importance of what a speaker expresses and the least importance of which salient variables they realize: therefore, it can be emphasized as a result of this perceptive inquiry that the content of a statement is more important in the social categorization of a speaker than the realization of certain linguistic features. Therefore, the ceceante realization is not at all a phonological feature linked to the lower social strata within the Jerezano speech community.