Taikomoji kalbotyra, 21: 77–103 eISSN 2029-8935
https://www.journals.vu.lt/taikomojikalbotyra DOI: https://doi.org/10.15388/Taikalbot.2024.21.6

First graders’ metalinguistic awareness of the societal prestige of the standard variety

Krisztina Fehér
University of Debrecen
Department of Hungarian Linguistics
feher.krisztina@arts.unideb.hu

Abstract. This paper discusses first graders’ linguistic biases in a Hungarian town where both the standard and the local varieties are used. Tests and interviews were conducted to determine whether the children who preferred the standard variety to the local dialect were able to verbalize their biases. In a preceding study conducted in the same town, kindergarteners of both genders preferred the standard variety but could not justify their linguistic choices. The current work showed that children became more judgmental and more aware of their biases in the first year of school. First graders commented on their preferences and followed the adult stereotype when describing the typical local speaker as “older, rural, male”. However, there was no evidence of gendered socialization in the children’s linguistic preferences. The qualitative change in the first graders’ metalinguistic awareness of the prestige of varieties highlights the impact of school on sociolinguistic acquisition.
Keywords: first graders, linguistic prestige, local variety, metalinguistic awareness, standard variety

Pirmosios klasės mokinių metalingvistinis sąmoningumas standartinės kalbos socialinio prestižo atžvilgiu

Santrauka. Šiame straipsnyje nagrinėjamos vieno Vengrijos miesto pirmosios klasės mokinių kalbinės nuostatos. Šiame mieste vartojama ir standartinė kalba, ir vietinė kalbos atmaina. Tiriant buvo atlikti testai ir surinkti interviu duomenys siekiant nustatyti, ar vaikai, kurie pirmenybę teikė standartinei atmainai, o ne vietinei tarmei, gebėjo paaiškinti savo pasirinkimą. Ankstesniame tame pačiame mieste atliktame tyrime abiejų lyčių darželinukai pirmenybę teikė standartinei atmainai, tačiau negalėjo pagrįsti savo kalbinio pasirinkimo. Šis tyrimas parodė, kad pirmaisiais mokyklos metais vaikai tampa sąmoningesni kalbinių atmainų atžvilgiu ir geriau suvokia savo pasirinkimą. Pirmosios klasės mokiniai gebėjo paaiškinti savo kalbinį pasirinkimą ir buvo linkę vadovautis suaugusiųjų stereotipais, o tipišką vietinį kalbėtoją apibūdino kaip „vyresnio amžiaus, kaimo gyventoją, vyrą“. Tačiau iš vaikų kalbinių nuostatų neišryškėjo jokių socializacijos pagal lytį požymių. Pirmosios klasės mokinių metalingvistinio suvokimo apie kalbos atmainų prestižą kokybinis pokytis išryškina mokyklos svarbą sociolingvistinei raidai.
Raktažodžiai: pirmosios klasės mokiniai, kalbinis prestižas, vietinė kalbos atmaina, metalingvistinis sąmoningumas, standartinė kalba

_________

Copyright © 2024 Krisztina Fehér. Published by Vilnius University Press.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

1. Introduction

Varieties/languages that carry societal prestige and have undergone standardization and codification are referred to as “standard”. Adults and adolescents usually prefer languages/varieties with a higher status in their society to those with a lower societal prestige (Preston 1998; Price, Tamburelli 2020). They comment on their status-based preferences, which are manifested in stereotypes (Maegaard 2010; Lippi-Green 2012). This behavior results from social learning in which schooling appears to be crucial (Cremona, Bates 1977; Day 1980; Kinzler, Shutts, Spelke 2012). During elementary school, the increasing use of the standard variants is a stable trend across societies (Chevrot, forthcoming). Schooling also suggests a reorganization of socio-indexical knowledge: their widening stylistic repertoire, i.e. the formal/standard variety of classroom materials and the hierarchical relationship with their teachers vs the informal/vernacular varieties of the influential peer group, gives children the opportunity to experience associations between social and linguistic features (Chevrot, forthcoming). Studies have addressed children’s linguistic biases around school entry, focusing on young children’s reactions to languages/varieties used in distinct domains (home vs school) and/or in ethnically or geographically diverse communities (Rosenthal 1977; Kinzler, DeJesus 2013; Cornips 2020; Morales, Cornips 2023; Paquette-Smith, Bucker, Johnson 2023). However, few studies have asked children around the age of school entry to evaluate speakers from their community where both the high-prestige standard variety and the low-prestige local dialect are spoken in all domains (Kaiser, Kasberger 2021). Subjective perceptions of the varieties of one’s own community may differ from those of dialects of other ethnic groups or regions. They may also be shaped by the exclusive or overlapping domains of use of the varieties.

The age at which status-based language preference emerges in children may also vary by context. In the Netherlands, many toddlers stopped speaking their regional language at home after entering preschool, even though it was used in their homes (Cornips 2020). Children speaking a transitional variety between the poles of the Bavarian-Standard continuum of Austrian German showed a clear status-based preference that seemed to set in after school entry, with literary language acquisition, at the age of 7-8 (Kaiser, Kasberger 2021). A speaker of Received Pronunciation was rated lower than a Welsh-accented one by 7-year-olds in the southwest of England, but this evaluation was reversed in 10-year-olds (Giles et al. 1983). Flemish 8–10-year-olds were able to discriminate between Belgian varieties of Dutch, and they seemed to recognize the standard variety as the model speech form. Considerable changes in attitudes in the Flemish context were reported in adolescence (De Vogelaer, Toye 2017). In terms of status, 5–6-year-olds from Illinois already preferred a northern-accented speaker to a southern-accented one, while their peers from Tennessee did not show any biases at that time. However, 9–10-year-olds from both areas rated the northern speaker higher (Kinzler, DeJesus 2013). Regardless of ethnicity and socioeconomic status, 3–6-year-old preschoolers from the northern part of the United States chose to receive a gift from a Standard English speaker rather than from an African American Vernacular English speaker (Rosenthal 1977). As for creole languages, before school entry, Hawaiian children preferred to receive a gift from a Creole English speaker than from a Standard English speaker, but they changed their behavior in the first year of school (Day 1980). In a post-colonial context, only the 5–10-year-olds from Xhosa-speaking families who attended school in English expressed a preference for English speakers over Xhosa speakers (Kinzler, Shutts, Spelke 2012). With respect to the varieties of English as an official lingua franca vs as a foreign language, neither Singaporean nor Korean 5-year-old kindergarteners preferred American English (considered “standard/native” by their teachers) to the local varieties of English, regardless of whether English was used in the kindergarten and whether they attended the kindergarten’s English classes (Choi et al. 2021).

Quantitatively, the development of status-based preferences has been found to be non-linear in some cases, but the trajectory of the process is not yet known. A status-based preference in Bavarian children increased significantly between the ages of 7–9 and leveled off slightly around age 10 (Kaiser, Kasberger 2021). In contrast, Italian children showed a considerable increase in their status-based preference between the first and fifth grades, with no decrease in preference for the standard variety (Cremona, Bates 1977). In terms of ethnic dialects, a temporary dip in preference for Standard English over African American Vernacular English occurred between the ages of 3 and 5 in both black and white children in the United States (Rosenthal 1977).

Qualitatively, young children may distinguish between the dimensions of “correctness”/“competence” and “pleasantness”/“social attractiveness” associated with varieties/speakers. They may also linguistically profile regional speakers, i.e. have thoughts about the social characteristics of people based on accent/dialect (on linguistic profiling, see Baugh 2003). Quantitative methods were used to test whether a given label, associated with one of those dimensions/characteristics, is used for a variety. In a choice task, 5–6-year-olds from Illinois chose a northern-accented speaker over a southern-accented one to be “smarter” and “nicer”, whereas 9–10-year-olds from the same area believed the northern speaker was “smarter” and the southern speaker was “nicer” (Kinzler, DeJesus 2013). In a questionnaire, Flemish 13–14-year-olds preferred the Belgian variety of Standard Dutch to the local dialect based on status, but their choice regarding social attractiveness was the opposite (De Vogelaer, Toye 2017). Similarly, Greek 11–12-year-olds perceived a lack of societal prestige in geographical varieties, but they acknowledged the social attractiveness and the personal reliability of regional speakers. They also associated the regional dialects with rural settings and informal contexts (Papazachariou et al. 2018).

The ability to verbalize linguistic biases represents a higher level in the process of sociolinguistic acquisition. The dimensions of evaluation are less clear in young children’s comments on their linguistic choices. Exploring children’s original labels requires a qualitative approach, which is difficult to apply (Ponizovsky-Bergelson et al. 2019). Interviews have been conducted with young children on various aspects of their language attitudes and linguistic beliefs (Szabó, Mátyus 2014; Stamou, Maroniti, Griva 2015; Choi et al. 2021), but rarely on their ability to justify their biases. From the age of 5, children seem to be able to make snap judgments but evaluations referring to the prestige of linguistic variables do not occur until preadolescence (Chevrot, forthcoming). However, the original labels used by children have not been systematically collected and analyzed, especially in a European context. 3–6-year-old white children attributed labels such as “nice(r)”, “good/better”, “bigger”, “clearer”, “gentle” to describe a speaker of Standard English, while they used labels such as “friend”, “cool”, “black”, “bigger”, “silly”, “bad” when describing a speaker of African American Vernacular English (Rosenthal 1977). Labels such as “better” or “nicer” also appeared in the responses of Hawaiian kindergarteners and first graders when commenting on a Standard English speaker compared to a Creole English speaker (Day 1980).

Women use more high-prestige variants than men of the same class/age (Trudgill 1972). However, findings are inconclusive as to how and when those gender differences develop during the acquisition of the linguistic prestige of varieties. Mothers in England used more non-standard variants when talking to their 2-year-old sons than to their daughters of the same age (Foulkes, Docherty, Watt 2005). While children produce more standard variants from age 6, findings are contradictory regarding gender differences. Style-shifting develops rapidly in elementary school, and it is more common in preadolescent girls than in boys of the same age (Chevrot, forthcoming). In terms of variety perception, 3–4-year-old girls outperformed boys in one of the subtests in discriminating dialects of British English (Jeffries 2019). 6- and 10-year-old Italian boys, as well as 10–11-year-old French boys with long-term local friends, produced more local variants than their female peers did (Barbu, Martin, Chevrot 2014; Cremona, Bates 1977). Yet, no significant gender difference was found in the preferences of Italian 6–10-year-olds (Cremona, Bates 1977). Compared to their female peers, Danish 4-, 6- and 8-year-old boys and male adolescents used more non-standard variants in their speech (Ladegaard 2003) and they were also more tolerant of the local variety (Ladegaard 1998). In contrast, Greek 11–12-year-olds’ attitudes toward the standard and other geographical varieties were not affected by gender (Papazachariou et al. 2018). As for ethnic dialects, 3–6-year-old white girls preferred Standard English to African American Vernacular English to a greater extent than white boys did, while the pattern was reversed for black children (Rosenthal 1977). Concerning post-colonial English in South Africa, no gender difference was found in the preferences of Xhosa- and English-speaking 5–10-year-old bilinguals for speakers of their first and second languages (Kinzler, Shutts, Spelke 2012).

2. The Hungarian context

In Hungary, it is compulsory for children to attend both kindergarten and elementary school. According to the regulations in force at the time of the study, children may graduate from kindergarten and be admitted to elementary school between the ages of 5–8, based on expert recommendations. While kindergarteners have less organized group activities and learn through play, in schools, children begin subject-based education, including literacy instruction, in 45-minute sessions of classes separated by 10-minute breaks. The time slots are signaled by the ringing of bells.

Hungarian teachers prefer using the standard variety and often discriminate against vernacular dialects (Parapatics 2020; Jánk 2020). The standard variety is generally rated higher than the regional varieties by Hungarian adolescents and adults (Sándor, Pléh, Langman 1998; Molnár 2022). Some language ideologies are transmitted from teachers to kindergarteners (Szabó, Mátyus 2014). Negative attitudes toward a non-local dialect and stereotypical features of Hungarian regional dialects have also been found among fifth and twelfth graders (Kaján, Deme 2020). However, little is known about the impact of schooling on students’ subjective reactions to standard vs local varieties in communities where bivarietal speakers use both varieties in overlapping domains.

The research site of the current work was a small town in northeastern Hungary where most residents are bivarietal, using a wide range of speech forms along a standard-regional continuum in all domains (cf. Tukacs-Károlyi 1990). The study was preceded by field observations in local schools and kindergartens, interviews with teachers and questionnaires with parents. The sociolinguistic background of the individuals in the local schools did not differ from that in the kindergartens. Linguistic insecurity was found among local parents and teachers, indicating a conflict between the attitudes toward the standard and local varieties. The informants considered the local dialect as “pleasant”, while the standard variety was labeled as “correct”. Teachers profiled the typical local speaker as older and rural. Parents were aware of the most salient vowel variables, i.e. (eː), (oː) and (øː), and they reported using them as well. Yet, the local variant of (eː) did not occur in the speech of the parents and teachers, while some local variants of (oː) and (øː) were produced. The local [iː] variant of (eː) appeared to be less used and more stigmatized than the local diphthongal variants of (oː) and (øː) variables. Beyond this background study, there has been no further research on the local adults’ attitudes. Nonetheless, the attitudes of the local teachers and parents were in line with the general attitude toward the dialect of this region: Hungarians have called the variety “ugly” and its speakers “peasant” (Kontra, 2003).

When local 14–18-year-olds were asked to choose between two candidates for a shop assistant position based on two audio recordings (not knowing that they were hearing the same voice twice), they hired the standard speaker and showed negative attitudes toward the local speaker (Hanyiszkó 2011). Females were less judgmental of the local dialect, suggesting gendered patterns of linguistic evaluation. They rated the standard speaker higher than the local dialect speaker on being kind, unselfish, helpful and emphatic. However, they had reversed responses for reliability, honesty, morality and humanity. They associated the local dialect with a rural environment and thought the standard speaker was smarter and younger. This confirms the standard order of indexicality in Europe where local dialects are devalued in the dimension of “competence” but ascribed some positive characteristics regarding “social attractiveness”.

In a recent study, local kindergarteners of both genders preferred the standard variety to the local dialect, but they could not yet justify their choice (Fehér 2020). Justification of the linguistic biases and gendered patterns of linguistic preference may develop later in local children. However, there are no studies that suggest how the acquisition of the societal prestige of the two varieties proceeds in this context.

3. Aims and questions

This paper is a continuation of a study that investigated kindergarteners from northeastern Hungary (Fehér 2020). Both studies used samples from a language and an age group that had not been previously studied in terms of the societal prestige of varieties. The present paper aims to investigate the further development of status-based preferences in first graders from the same community. Of particular interest is the development of their metalinguistic awareness of the prestige of varieties. “Metalinguistic awareness” is a fluid concept. In the current study, it refers to the children’s ability to verbalize their linguistic biases in linguistic terms. The focus of this paper is to determine the development of this ability, which represents a certain level of awareness, rather than to distinguish between unconscious and conscious biases in children (on the consciousness issue, see Kristiansen 2010).

The study aims to answer the research questions below:

(1) Do the Hungarian first graders show an increasing preference for the standard variety compared to the kindergarteners?

(2) Have they developed metalinguistic awareness of their preferences?

(3) If so, do Hungarian first graders associate the standard and local varieties with distinct characteristics, and do they do linguistic profiling?

(4) Do Hungarian first graders’ preferences become gendered in the first year of school?

4. Methods

Tests and interviews are well-established methods for eliciting linguistic preferences and language attitudes. In tests, participants are usually asked to rate speakers from audio recordings (guises) representing different languages/varieties. The match-guise technique (MGT) uses the same voice for all guises, while the verbal-guise technique (VGT) uses different voices (for a review, see Kircher, Zipp eds. 2022). In this work, both tests and interviews were conducted. To make the design child-friendly, the children were presented with a tale and asked to choose between two storytellers for a reward. Each child answered a few test questions followed by an open-ended question that introduced a mini-interview. The test was designed to determine the degree of status-based preferences in first graders. The mini-interview was designed to determine if the children could also verbalize their biases. Follow-up questions also allowed for the detection of verbal reactions associated with personality-based preferences.

This research on first graders is a replication of a recent study on kindergarteners from the same community (Fehér 2020), with two methodological modifications. First, while the kindergarteners were tested using VGT with control for voice differences (for details, see Fehér 2020), the first graders were tested using MGT for practical reasons. If the first graders had recognized that they were hearing the same voice twice when a distracting question (“Do you like tales?”) was inserted between the guises, VGT would also have been introduced to test another group of children. Second, no study has examined or controlled for the effect of the language variety of the instructor when conducting experiments on children’s linguistic biases. Instructors spoke only one of the varieties/languages tested (Rosenthal 1977; Day 1980; Kinzler, DeJesus 2013; Kaiser, Kasberger 2021). In the present study, the instructor used either the standard variety or the local dialect.

4.1. Procedure

88 first graders (age: M = 7;8, SD = 0;54, R = 6;5–8;10; girls: N = 48, boys: N = 40) were recruited at the same research site as in the previous study conducted with 37 kindergarteners (M = 6;3, SD = 0;6, R = 5;0–7;6) (Fehér 2020). Data was collected in a school setting, one year after the previous study, during the last month of the school year. Children from monolingual, bivarietal local families and with typical cognitive and linguistic development were invited to participate. Their age range partially overlapped with that of the kindergarteners due to the relevant school admission regulations. Among the participants, inadvertently, there were 12 first graders who had also participated in the previous study.

Research site approval and parental informed consent forms were signed and returned. Teachers and parents were informed that the children would listen to a tale and answer a few questions to check their linguistic skills. All the children were tested individually. The activities were audio-recorded and notes were also taken to register behavioral reactions. The research complied with the ethical standards in Hungary.

The local community most often uses speech forms from the middle of the standard-regional continuum. To test the preferences in this situation, the stimuli in the guises were taken from the poles of the continuum. A one-minute animal tale (Fehér 2020) was used as stimuli (Appendix 1). The script highlighted the most salient features of the two varieties. The main characters of the tale, the stork, (ɡoːjɒ) and the frog, (beːkɒ) were chosen to represent the (o:) and (e:) variables. The token numbers of the two characteristic types of vowel height variation, i.e. that of the rounded and unrounded long vowels, (o:), (ø:) and (e:), were balanced in the script. Other characteristic but less salient variables were also presented in the stimuli (for more details, see Fehér 2020).

The story was read aloud in both varieties by two female voices that were used in a balanced way. Each child was presented with the guises of the same voice only. The children were randomly assigned to the voices. Both speakers were older, monolingual, bivarietal, relatively non-mobile, long-term residents. Their occupations (teacher, secretary) and their working-class environment enabled them to provide samples of both varieties at the poles. The investigator was their relative. Having grown up in the town, she was also bivarietal. She checked the authenticity of the recordings. Moreover, the investigator was the instructor, and she used the standard variety in the first group of children (N = 44) and the local dialect in the second one (N = 44).

The children were randomly assigned to standard- or regional-instructed groups. The instructor went into their classroom and invited each child to come play with her in a separate room. While inviting them, she already used the variety that served as the variety of instruction. Before that, there was no communication between the instructor and the participants. The children were told that the guises belonged to two teachers who would tell them the same tale, one after the other, and they should listen carefully to decide whose present they wanted from the boxes. The choice of teacher characters was advantageous for detecting status-based preferences. Teachers are in charge, have higher status, and represent the school environment. However, it also carries a potential pitfall that should be considered when interpreting the data. The “voice of authority” may raise expectations for language use.

Slides of two sketches of a female teacher (Fehér 2020) were shown on a laptop screen. The only difference between the two figures was their coloring. Two identical boxes (Fehér 2020) were placed on a table. They had the same set of stickers inside. Each box had a label on the top referring to a teacher with the usual form of address used by children: her first name followed by néni ‘~aunt’: “Zsuzsi/Kati néni”. The names were not stereotypical. The audio guises (Fehér 2020) were played on a laptop through a speaker.

To control for non-linguistic biases, names and images were pseudo-randomly associated with the language varieties. A tale might be more interesting for the first time. Thus, the order in which the guises were played also varied pseudo-randomly. (When tested afterward, the order was not found to be relevant though.) These controls for the non-linguistic variables also ensured that neither the children who had just completed the task would be able to discuss their experiences with others nor the first graders who had participated in the previous study would be able to connect the two occasions in a way that would affect the outcome.

Each participant listened to both tale scripts once. Each child was then asked seven choice questions (Table 1). To keep the test simple for children, two basic adjectives, “good” and “beautiful”, were included in the choice questions and will be treated as prompt labels. If a child did not respond to a question, it was repeated. If they still did not answer, the reaction was registered as neutral. An additional “why” question (Rosenthal 1977; Day 1980) was added to start a mini-interview. The responses were followed up by further questions until they stopped responding. When finished, each child opened the box of the preferred teacher and chose a sticker. The investigator escorted each child back to the classroom where they continued their learning activities.

Table 1. Questions (Fehér 2020)

4.2. Data analysis

Responses to each choice question were encoded as nominal variables (“standard”/“regional”). If a child did not make a choice in response to a given question, the cells for both varieties were left empty. To measure the relative degree of the children’s preferences, sums of the encoded responses to Questions 1–7 were also calculated. For instance, if a child chose the standard guise four times, the regional guise two times and there was no choice once, it was encoded as “4” for the standard variety and “2” for the regional dialect in total. CogStat (Krajcsi 2021) was used for statistical analysis. The audio recordings of the mini-interviews were manually transcribed, annotated and encoded.

When the children provided linguistic examples in their comments or there were obvious cases of their using regional variants, words were tagged with [R]. Annotations on the guise (standard or regional) to which their responses referred were also added at the end of the given part: (S) or (R). Linguistic examples were italicized and ungrammatical forms or overgeneralized uses of local variants were marked with ∗. A few comments that were likely to refer to the images/names used in the design were filtered out such as that of Participant (hereafter abbreviated as P) 115 (girl, 8;3 [hereafter numbers with semicolons refer to years;months]): because her outfit is very beautiful. (This work was the continuation of a previous study; thus, numbers were assigned to the first graders starting with 110.) Elicitation turns were identified to distinguish between immediate and subsequent reactions. Each response to an eliciting question in the mini-interviews was counted as an elicitation turn and marked as →. Reactions that preceded the “why” question were considered responses to the first elicitation. Annotations on further information were marked between < > symbols (Table 2). Novel occurrences of evaluative labels and linguistic examples were encoded in each elicitation turn.

Table 2. Annotated transcript of the mini-interview with P 186 (boy, 7;7)

Words that were not evaluative in nature, but were used in an evaluative context that gave them positive or negative semantic poles, were also marked as labels (hereafter typed in small caps). For instance, in the comment of P 157 (girl, 8;1), That I think she said it in a deeper voice., the attributive deep as a critical remark about the regional guise was annotated as an evaluative label.

To avoid the encoding being driven by the word class of adjectives which is usually used for attributes, any attributive linguistic manifestation was encoded as a label and transformed into a form of adjective. For instance, the opinion of P 117 (girl, 7;2), I also liked her voice., was encoded as the label likable describing the direct reference “voice”. When the verb like did not have an identifiable direct reference as was the case in P 186’s comment (see Tables 2 and 3), its occurrence was not considered a label.

To explore the labels that are inductively provided by the children, the encoded labels were keywords that stayed as close to the interview script as possible. Labels were subordinated into higher-level themes afterward. Encoding was not keyword-based when the words used were unlikely to be real content words for the children. The label <empty> was used to mark a type of comment that occurred once in the data when P 139 (girl, 7;8) gave her explanation: Because she, like, lowers the pitch at the end. She probably took her teacher’s word without understanding the concept of “pitch”. The encoding procedure accurately followed the interview script in the case of attributive elements with negative words. For instance, when P 186 (see Tables 2 and 3) argued that it is not ugly speech (S), he did not explicitly state that the speech in the other guise was ugly. Thus, this part was encoded as the label not ugly for the standard variety rather than the label ugly for the local dialect. However, the continuation of the script, “not, like, funny”, was not encoded as not funny for the standard variety because in the preceding sentence the child had claimed the other guise to be the opposite (it is like… funny speech) and that part had been encoded as funny.

The semantic poles of the labels were also encoded, based on the context. For instance, the label long, given by P 181 (boy, 8;0), was encoded as “negative”: [I did not like it.] Zsuzsi néni told ... a longer tale. However, the label lengthened, provided by P 129 (boy, 7;6), as a favorable comment on the regional guise, was encoded as “positive”: she lengthens the letter. Anyway, I liked her tale too. Each label was associated with its direct reference (hereafter typed in small caps in brackets) as well. For example, in the above comment of P 117 (girl, 7;2), I also liked her voice., the encoded reference was [voice]. Direct references were always identified based on the children’s perspective, without further interpretation. For instance, since P 129 mentioned “lengthened letters” (see above), the reference was encoded as [letter] instead of [sound].

Linguistic examples given by the children were marked as sociolinguistic variables. For instance, in the comment of P 124 (girl, 7;7), Because Zsuzsi néni says the names of the animals like góulya <‘stork’> and bíka <‘frog’>, the variables (goːjɒ) [‘stork’] and (beːkɒ) [‘frog’] were encoded (Table 3).

Table 3. Encoded responses of P 186 (boy; 7;7) in the mini-interview (see script in Table 2)

Statistical analysis was carried out on the responses to the test questions, and a content analysis on the interview data. The results were also compared with those of the previous study on kindergarteners (Fehér 2020).

5. Results

The answers to the research questions are as follows:

1 First graders showed an increasing preference for the standard variety compared to kindergarteners.

2 First graders showed a developing metalinguistic awareness of their preferences.

3 The two language varieties had different characteristics, and their speakers had different linguistic profiles among first graders.

4 There was no evidence of gendered preferences in first graders’ linguistic choices. However, the profile of the typical regional speaker was gendered.

The results are presented in detail below.

5.1. Tests

For the sum of the seven choices per child, the first graders showed a highly significant preference for the standard variety over the local dialect (Figure 1). This is consistent with the results of the preceding study with kindergarteners who also favored the standard variety over the local dialect (Fehér 2020). Statistical analysis was also carried out on the 12 first graders who participated in the preceding study too, but the sample size was not sufficient to reach statistical power.

Figure 1. Number of choices in all test questions by variety. Gray lines show the matched pairs of the two varieties in individual responses. For statistical data, see Appendix 2

Like in the kindergarten group, there was no significant difference in linguistic preferences by gender among the first graders (Figure 2). The language variety of the instructor did not have a significant effect on the first graders’ choices (Figure 3). The kindergarteners were not instructed in the local dialect: thus, only the first graders who were instructed in the standard variety were compared with the kindergarteners. As a result, first graders showed a moderately greater preference for the standard variety, significant only at the < .005 level with a small effect size (Figure 4).

Figure 2. Number of choices of the standard variety in all test questions by gender. For statistical data, see Appendix 2

Because children may repeat their initial answers, responses to each question were also tested separately, paying particular attention to the first/last questions, which represent the immediate reaction and the choice of the present. Responses to each question were also tested by gender and the instructor’s language varieties. Except for Question 3, which showed differences by gender at the < .005 significance level, the outcome was consistent with the results above (see Appendices 3 and 4).

Figure 3. Number of choices of the standard variety in all test questions by the instructor’s language variety. For statistical data, see Appendix 2

Figure 4. Number of choices of the standard variety in all test questions by education level, when instructed in the standard variety. For statistical data, see Appendix 2

5.2. Mini-interviews

There were significant differences in the mini-interviews across the educational levels. Most kindergarteners made no comments (Fehér 2020), whereas many first graders justified their choices.

5.2.1. Labels of evaluation

At the individual level, 1-2 label types were identified in the comments of 11% of the kindergarteners (Md = 0; Q1 = 0; Q3 = 0) (Fehér 2020), while 1-6 labels occurred in the mini-interviews of 81% of the first graders (Md = 2; Q1 = 1; Q3 = 3). No pattern was identified in the number of labels by the choice of present, the instructor’s language variety and gender.

In the case of the kindergarteners, few labels occurred, and they were all based on the adjectives used in the choice questions: (not) beautiful and good/bad (Fehér 2020). In first grade, the beautiful/ugly and (not) good labels remained the most robust (57%), but 48 original label types also appeared with low(er) token frequencies (TTR =.31) (Table 4). Among them, 23 label types appeared only once.

Table 4. Evaluative labels in the mini-interviews

Numbers without parentheses refer to all elicitation turns. Numbers in parentheses indicate the first elicitation turn.

In the first graders’ comments, labels were used more often for the standard variety than for the local dialect. When the numbers of labels for the varieties were compared by the choice of present, a characteristic pattern emerged. 16 children who chose the local speaker’s present used labels (Md = 1.5; Q1 = 0; Q3 = 2) exclusively to describe the regional guise. In contrast, 72 first graders requested the present of the standard speaker. While they used positive labels (Md = 1; Q1 = 1; Q3 = 2) to describe their preferred guise, they also used some negative labels (Md = 0; Q1 = 0; Q3 = 1) to comment on the other guise.

The label types were also distinctive to each variety. First graders used more types but fewer tokens to comment on the regional guise than on the standard guise (TTR = .39 vs .25). The labels beautiful and good covered 57% of the tokens of the standard variety. In contrast, in the case of the local dialect, eight label types out of 30 accounted for a similar percentage of tokens, i.e. deep (1), funny (2), weird (3), different (4), long (5), peasant-like (6), old-fashioned (7), and stretched (8):

(1) ’Cause it [her voice] was deeper. Well, no [I did not like it]. (R); P 146 (boy, 7;6)

(2) Zsuzsi néni’s one [voice] was funny (R); P 167 (boy, 7;7)

(3) That she speaks, like, weird. (R); P 135 (girl, 7;4)

(4) Well, that um.... she spoke differently from Kati néni (R); P 157 (girl, 8;1)

(5) that the tale... got longer (R); P 166 (boy, 7;10)

(6) Yes, like ... a bit peasant-like, like, interestingly very much [peasant-like]. (R); P 139 (girl; 7;8)

(7) a word was pronounced, like, old-fashioned (R); P 189 (boy; 7;10)

(8) ’Cause Kati néni stretches the letters (R); P 125 (girl; 7;7)

The distribution of labels by variety was even more distinctive in the first elicitation turn providing 51% of the total of label tokens (62% for the standard guise; 36% for the regional guise) (Table 4). In the first elicitation turn, the first graders also used more label types but fewer tokens to comment on the regional guise than on the standard guise (TTR = .59 vs .23). However, the number of tokens of the leading label for the standard variety, beautiful, which was one of the prompt labels, surpassed the total number of label tokens for the local dialect. While the prompt label tokens, beautiful and good, covered 70% of the tokens used for the standard guise, almost all of the labels used for the regional guise were original.

While the semantic poles of the labels used for the standard variety were all positive, those for the local dialect were negative in 55% of the cases in the first elicitation turn and in 73% of the cases overall (Table 4). Only four label types out of 48 were used for both varieties in the first elicitation turn and in total: beautiful, good, likable and interesting, with a dominance of describing the standard guise (Table 4). There were only two labels that were used exclusively for the local variety and had more tokens for the positive semantic pole than for the negative one: funny and long (Table 4).

5.2.2. Themes of evaluation

The labels could be arranged into seven themes of evaluation (hereafter between quotation marks). The words beautiful and good appeared in the test questions. They were also the most popular and almost exclusively used adjectives in the mini-interviews with the kindergarteners (Fehér 2020). Thus, the related labels such as beautiful, not ugly, good, not good and bad were treated separately so that they were aggregated into the theme “[beautiful/good]” (36% in total; 60% and 1% of the labels for the standard and local varieties, respectively). Beyond this, six original themes were identified (Table 5).

Table 5. Evaluative themes in the mini-interviews

The second leading theme was “expressivity” (20% in total; 13% and 30% of the labels for the standard and regional varieties, respectively). This was consistent with the frame of the experiment, i.e. a performance of storytelling. The performance of the standard guise was evaluated as likable in the first place, while that of the regional guise was funny. The standard guise was also interesting, elegant, emphasized, enthusiastic, exciting and not long. In contrast, the regional guise was rather long, old-fashioned, folktale-like, exaggerated and colorful.

The second most popular theme differed by variety. In the case of the standard guise, the theme “personality” was present the most (11%) after “expressivity” and labels such as nice and likable occurred. In the case of the regional guise, however, labels in the theme of “normality” (22%) and “acoustics” (22%) were used. As for “normality”, the regional guise appeared to be weird, different, foreign, interesting, not Hungarian and unfamiliar. The theme “acoustics” was the third most popular (10%) in the case of the standard variety. The corresponding labels were specific to each guise.

The theme “acoustics” was further elaborated in the case of the local variety. The standard variety was associated with the labels soft (9) and thin (10). In contrast, the local dialect was described with the labels deep (11), lengthened/stretched/slow (12), loud (13) and hoarse (14):

(9) Also, that her voice was more beautiful, softer (S); P 153 (girl; 7;4)

(10) That her voice is a bit thinner. (S); P 119 (girl, 8;0)

(11) Well, [Kati néni’s voice was] a bit deeper (R); P 189 (boy; 7;10)

(12) That she was, like, stretching the sentences. That as if it went, like, slower. (R); P 123 (boy, 7;3)

(13) That Kati néni spoke, like, louder. Something ... with some loudspeaker. Or I don’t know with what. (R); P 185 (boy, 6;6)

(14) [I noticed] that she had, like, a hoarse voice. (R); P 195 (girl, 8;0)

Some labels such as loud/deep (15), soft/girlish vs deep/boyish (16) in the theme “acoustics” were associated with each other, and the theme seemed to be related to gender in general:

(15) when the ... Zsuzsi néni told [the tale] then [she said it] like... like... like like ... like loud that like... Yes, also... also like that, you know, like that, like... like, that like <saying this in deep voice>. She said it like deep when she started. (R); P 136 (boy, 7;3)

(16) ’Cause she [Zsuzsi néni] read it softly, and she did it li[ke]... and Kati néni did it like boys <embarrassed laugh>. Well, that she said it um... in a very deep voice, and the girls didn’t... didn’t... say it in a deep way, and she didn’t say it softly like Zsuzsi néni, and the girls always... um... say it softly, but sometimes so do boys. (S) vs (R); P 163 (girl, 7;6)

First graders also had beliefs about the background of the speaker of each variety (16% of the label tokens for the regional guise and 2% for the standard guise). Based on the theme “demographics”, the profile of the local speaker was more complete than that of the standard one. The speaker associated with the standard guise was young (17) and girlish (18). In contrast, the person behind the regional guise was more likely to be peasant-like (19), old (20) and boyish/mannish (21):

(17) Becaaause... because she is younger? (S); P 128 (girl; 8;3)

(18) and she didn’t say it softly like Zsuzsi néni, and the girls always... um... say it softly (R) vs (S); P 163 (girl, 7;6)

(19) she just speaks like peasants (R); P 149 (boy, 7;6)

(20) she is also, like, old (R); P 137 (girl, 8;2)

(21) but, like, ... well..., like, it was as if a bácsi <a word for an adult male in children’s language> told it [the tale]. (R); P 191 (boy; 7;10)

5.2.3. Collocations of evaluation

The kindergarteners associated the few labels they mentioned at all with [speech] and [voice] (Fehér 2020). The first graders’ associations were more diverse. The most frequently labeled direct references were [storytelling] (38%) (22) and [voice] (23%) (23), probably reflecting the framing of Questions 3 and 5. Beyond them, a language-related reference, [speech] (24), was the most often mentioned (15%) (Table 6). Other linguistic references such as [language], [pronunciation], [reading], [letter], [sentence], [word] and [sound] rarely occurred:

(22) ’Cause her storytelling was a bit better. (S); P 155 (girl, 7;10)

(23) Well... um ’cause I think um her voice is more beautiful. (S); P 172 (girl, 7;3)

(24) Also, she spoke, like, weeeird. (R); P 139 (girl, 7;8)

From the children’s perspective, the storytelling, the voice and the speech were all beautiful in the standard guise, while the regional guise could rather be described as funny/weird+[storytelling], deep+[voice], peasant-like/weird+[speech]. The tale was good in the standard guise, while it was long in the regional one. The person behind the standard guise is likable in the first place, but, behind the regional guise, she is funny (Table 6).

Table 6. Label+reference collocations in the mini-interviews

5.2.4. Examples of evaluation

Linguistic examples were rare. Only 11% of the kindergarteners provided 1–2 words in the previous study (Fehér 2020), and 22% of the first graders gave 1–3 words in the current study. The examples all belonged to the local dialect and represented the two most characteristic vowel variables, (eː) (25), (o:) (26), as well as (ol) (27), which is related to the latter:

(25) She said bíka for the frog. (R); P 110 (girl, 8;6)

(26) For example, wheeeen when [she said] jóu, jóu <‘alright, alright’>. Well, I can give you more [examples], that, when also the ... she said the stork too. Góulya <‘frog’>.” (R); P 140 (girl, 8;3)

(27) Well... well... she said it like ... like vó <‘was’>… like ∗vút <‘was’>. (R); P 137 (girl, 8;2)

They all were tokens that occurred in the guises, with three exceptions: words containing two regular and one overgeneralized use of the local variant of the (eː) variable. Four children used overgeneralized forms of the local vowel variants in words that occurred in the guises.

5.2.5. Pragmatics of evaluation

Some children got confused when justifying their choices. The labels (not) long appeared only when the children were instructed in the local variety, while other similar labels such as stretched, slow and lengthened occurred regardless. The participants believed that the tale in the regional guise was longer, while they also thought that both took the same amount of time:

(28) Well, but I would like to choose Kati néni because um because um I like the... those tales that take a bit longer time, and, well, <confused> both of them took the same time, but... <confused> (R); P 171 (girl, 7;11)

(29) Zsuzsi néni told ... a longer tale. She read a longer tale. Yes, a bit [longer]. <confused> One sentence [longer]. <confused>; (R); P 181 (boy, 8;0)

The laughs of five children and the gossipy whisper of one child were also recorded during and after listening to the regional guise. The non-verbal reactions coincided with the labels weird, funny (three children), peasant-like (two children), boyish (one child) and with the linguistic examples of the [i:] local variant of the (e:) variable (four children):

(30) <whispering> Zsuzsi néni spoke, like, weird. That peasant-like. (R); P 164 (girl; 7;6)

Laughing and whispering occurred regardless of the instructor’s language variety. Yet, at least on one occasion, the fact that the interviewer (I) used the local dialect led to the uncertainty of P 165 (boy, 7;9). Meanwhile, he also asked May I say...? when he realized the weight of his negative judgment (peasant-like) in this context:

(31) P: <laughing out loud when listening to the guise, and saying:> Bíka[R] <‘frog’>.

<conversation continues after listening to the guises>

I: What did you laugh so much at?

P: Bíka[R].

I: Is that funny?

P: <laughing> Wha’, isn’t it?

I: I don’t know[R]. I’m askin’[R].

P: <stops laughing> She [Zsuzsi néni] told it [the tale] in a weird way.

<conversation continues after answering the test questions>

I: And why[R] did you choose Kati néni?

P: ’Cause she told the poem <sic!> more beautifully.

I: I see. Why[R] did she tell the tale[R] more beautifully[R]? What was[R] more beautiful?

P: <getting confused> ‘Cause she... May I say that she [Zsuzsi néni] told it [the tale] peas...ant-like?

I: You should say [R] what you think. <smiling>

P: Um... she told it [the tale] in a peasant-like way.

I: What do you mean by “peasant-like”?

P: Well, like... that like... like bíka <‘frog’> and like... um... she told it [the tale], like, not good, like...

I: What is not good[R]?

P: Well, that, like,... say the bíka <‘frog’>, like,.. she lengthened it, like, bíííka <‘frog’>.

A label was explicitly associated with education by P 183 (girl, 7;5):

(32) P: Well, [she told the tale] not so Hungarian as Zsuzsi néni did.

I: What do you mean by “not so Hungarian”?

P: Well, we haven’t learned that yet, but....

6. Discussion

This paper addressed a few research gaps in developmental sociolinguistics. The study presented evidence for the emergence of the metalinguistic awareness of the societal prestige of the standard variety in bivarietal children after school entry. It applied child-friendly methods focusing on the mini-interviews preceded by a test in which the evaluation of the standard variety was compared to that of the local dialect. The research was conducted in a community where speech forms between the high-status standard variety and the low-status local dialect are used in all domains. It was a replication of a previous work with kindergarteners (Fehér 2020) to result in a trend study that allowed for a cross-educational comparison. The samples were taken from Hungarian-speaking children, an understudied population in the field.

In the quantitative analysis, a preference for the standard variety was found in Hungarian first graders. The kindergarteners already showed status-based preference (Fehér 2020), but this bias became moderately but significantly stronger in the first graders. Since this trend has also been reported in other studies using different experimental methods and examining children from different backgrounds, the question arises whether the degree of status-based preference in Hungarian children will continue to increase over time, as reported in the case of Italian students (Cremona, Bates 1977), or whether it is a temporary peak that will later level off, as seen in the case of Bavarian children (Kaiser, Kasberger 2021). Nevertheless, previous research has provided quantitative evidence for children’s increasing preference for the standard variety after school entry in different environments (Day 1980; Kaiser, Kasberger 2021), yet, the qualitative aspect has been overlooked. This study also provided a systematic qualitative analysis and showed that the major change in the first year of school among Hungarian children was qualitative rather than quantitative: in contrast to kindergarten children (Fehér 2020), many first graders were already able to justify their linguistic choices, indicating an emerging metalinguistic awareness of the social prestige of varieties.

The differences between the local elementary schools and the kindergartens were the formality of the education and the age of the children, both of which may account for the change observed in the Hungarian first graders. Instead of the playful group activities of kindergarten, first graders had classes organized into regular, subject-based sessions, including literacy instruction. The school setting created a more formal context and increased both the authority of the teachers and the explicit metalinguistic information provided through literacy instruction. The choice of a teacher figure in the guises was intended to target status-based preference, so we can only speculate whether the linguistic expectations toward another figure would be different. Because the age ranges of the educational groups overlapped and did not allow for a powerful statistical analysis by age, the relative weight of age and school setting in the development cannot be estimated. Organized instruction alone may not lead to considerable changes if children are too young to recognize differences between varieties or associate linguistic features with socio-indexical meanings: no preference for the high-prestige American variety was found in Korean-speaking kindergarteners as young as 5, regardless of whether they received afternoon English classes (Choi et al. 2021). However, the language of instruction changed the attitudes of both Hawaiian and African children after school entry (Day 1980; Kinzler, Shutts, Spelke 2012). A clear preference for the standard variety of Austrian German was set in with literary education and its further growth correlated with age in Bavarian-speaking children (Kaiser, Kasberger 2021).

The early stage of the development of the Hungarian first graders’ metalinguistic knowledge may be reflected in the comments with no linguistic references, in some overgeneralized forms of the local dialect and in the uncertainty/confusion in their behavior. A few participants used labels to describe the regional guise while recognizing the inconsistency between the stimuli presented and their comments. Interestingly, 13 tokens of five label types from 11 children indicated this subjective perception: long, stretched, slow, lengthened for the local dialect vs not long for the standard variety. Nine children could not associate this perceived feature with the diphthongs of the local variety, but two cases (Kati néni/she stretches/lengthens the letter[s]) indicate that it was likely related to them, representing an adult-like pattern of which first graders were not yet fully aware. Hungarian adults have never been studied in this regard, but according to the international literature, adults often label accents as “drawl” in reference to their diphthongs (Preston 1999). The label deep for the local variety relates either to the voice of a male as the typical regional speaker or to a so-far overlooked acoustic feature of the dialect. Some labels such as <empty> or colorful (Hungarian ízes, a folk linguistic term often used by adults for regional accents) probably had little or no meaning for the child participants. One child referred to the school when asked to explain the meaning of the label not Hungarian she had used. These labels are more likely to be the signs of the children’s automatic and mimetic alignment with the words of their parents and teachers than real labels. The inconsistent use of the direct references of the labels resulting in numerous types of label+reference collocations, the high proportion of use of prompt references from the test questions and the few linguistic examples provided by the children may also indicate the early stage of developing metalinguistic awareness of the prestige of varieties. It should be noted, however, that more diverse and less clear responses are to be expected in interviews than in tests.

Despite the early stage of the children’s ability to verbalize their linguistic biases, 44% of the labels in the mini-interviews already went beyond the level of snap judgments. The themes “[beautiful/good]” and “expressivity” can still be considered simple and immediate reactions associated with the prompt labels and references of the test questions, whereas the other themes, “acoustics”, “normality”, “demographics”, “personality” and “intelligibility”, point more to adult-like judgments based on societal prestige and stereotypes, especially those of the regional dialects (69% vs 27% of the labels used for the local vs standard varieties). In some cases, the pragmatics of the subjective reactions also indicate a higher level of awareness: some children showed embarrassed laughter, whispering and politeness when introducing negative labels such as peasant-like into the conversation. A developing metalinguistic knowledge of sociolinguistic prestige is also reflected in the fact that there were children, though not too many, who could already give linguistic examples when explaining their choices. The adult-like labels/examples appeared earlier in the Hungarian children than expected. The audio guises, which represent the two poles of the standard-regional continuum, may play a role in this outcome. Previous literature suggests that evaluation based on social values of linguistic forms emerges only during preadolescence (Chevrot, forthcoming). However, comparisons with other studies need to be made with caution, especially in the case of qualitative data. With regard to children’s abilities to verbalize linguistic biases, 5–9-year-olds have rarely been studied. Moreover, the interview methods are more varied than the experimental designs.

The Hungarian first graders’ verbal evaluations were adult-like in their critical approach to the local dialect. The semantic poles of the labels for the standard variety were all positive, whereas the local dialect was labeled negatively in 73% of the cases. The test was designed to detect the dimension of “correctness/competence” of the standard variety/speaker, but labels associated with the “pleasantness/social attractiveness” of the local dialect/speaker could also appear in the follow-up questions. Nevertheless, only a few tokens of two label types used in the interviews, likable (2) and the positive uses of funny (4), were found to be potential markers of the latter dimension. The higher lexical diversity of the labels for the local dialect (especially in the first elicitation turn) indicates that the children perceived the standard guise as “normal” (unmarked) and the regional guise as “deviant” (marked). This perception can also be seen in their linguistic examples which all referred to the most salient variants of the local dialect and were consistent with the findings for adults in both the previous research (Tukacs-Károlyi 1990) and the background study of this paper. Hungarian first graders seem to learn about the negative features of the local dialect rather than the positive features of the standard variety. This is also reflected in the number of original labels. In the case of the regional guise, almost all the labels were novel, giving the variety a diverse profile: deep [voice], funny, weird, different, long, peasant-like, old-fashioned, etc. However, more than half of the label tokens for the standard variety were the prompt labels: the two basic adjectives, beautiful and good that were included in the test questions. In the first elicitation turn, their proportion was even higher (70%). It also shows the developing judgmental behavior of the participants that the children who chose the standard guise did not necessarily mention the features of the standard guise when explaining their choices of present, but they always mentioned some negative traits of the local dialect.

As in the case of the kindergarteners (Fehér 2020), no evidence of gendered socialization was found in the linguistic preferences of the first graders. This contrasts with some previous work (Rosenthal 1977), but is consistent with many others (Cremona, Bates 1977; Kaiser, Kasberger 2021), all of which used different methods in diverse environments. Most importantly, however, when the findings are compared with the pattern observed among adolescents from the same town and elsewhere, the children did not yet follow the adult behavior. Local female adolescents were less judgmental of the regional dialect than their male peers (Hanyiszkó 2011). Gendered linguistic biases may develop later than the first year of school in the Hungarian context. Interestingly, however, a gendered feature that was not reported in local adolescents (Hanyiszkó 2011) and adults (the background study of the current work) emerged when the children profiled the regional speaker. Hungarian first graders believed that the typical local speaker was male (boyish/mannish) and the person behind the standard guise was female (girlish). They did this regardless of the fact that they heard female voices in both guises and were instructed by a female. We can only speculate whether the results would be different if male voices were used and/or if a male were the instructor. When 9–10-year-old Austrian bivarietal children were presented with both female and male voices (the gender of the instructor was not reported), they had higher linguistic expectations for a male doctor than for a female one (Kaiser, Kasberger 2021), whereas adolescents and adults typically show the opposite pattern (Trudgill 1972).

The study revealed further evidence of linguistic profiling. Surprisingly, the profile of the regional speaker was almost complete. The first graders had ideas not only about the gender of the typical local speaker but also about their age and geography/education. They believed that the local speaker was old, old-fashioned, folktale-like, great-grandma-like, and the standard speaker was young. Further studies could determine whether this would change if the children were presented with a less conservative regional guise. Like 11–12-year-old Greek students when asked in a questionnaire (Papazachariou et al. 2018), Hungarian 6–9-year-olds in the current study associated the typical local speaker with a rural setting and a lower level of education, as indicated by the label peasant-like. The character of the local speaker was also old and rural in an experimental study on the attitudes of local adolescents (Hanyiszkó 2011), as well as in the interviews with local teachers in the background study of the present paper. Interestingly, the first graders followed the stereotypical concept of NORM (“non-mobile, older, rural, male”) speakers (Chambers, Trudgill 1980) in their linguistic profiling of regional speakers. The children characterized the speakers of the local variety as ORM (“older, rural, male”) speakers. This is also consistent with the adult pattern (Dailey-O’Cain 1999; Diercks 2002; Rensink 1999). Only the “non-mobile” feature was not found. Children as young as the participants have little or no experience and concept of commuters/travelers. Apart from that, the linguistic profile of the typical regional speaker in Hungarian first graders was already fully adult-like.

7. Conclusions

Previous findings have suggested that children’s preference for the standard variety increases after school entry, but judgments referring to the societal prestige of varieties do not occur until preadolescence. This paper demonstrates that the major change in Hungarian first graders’ preference for the standard variety is qualitative rather than quantitative. The kindergarteners studied already showed a status-based preference, but they could not justify their choice. The first graders, however, could verbalize their biases, while their status-based preference grew moderately. Many of the evaluative labels used by the first graders were more than snap judgments. The children already followed the adult pattern when they linguistically profiled the regional speakers as “older”, “rural” and “male”. The difference between the kindergartens and schools studied was the (in)formality of the learning activities rather than the teachers’ language use. This highlights the impact of formal instruction on the metalinguistic awareness of the prestige of varieties.

References

Barbu, S., N. Martin, J-P. Chevrot 2014. The maintenance of regional dialects: a matter of gender? Boys, but not girls, use local varieties in relation to their friends’ nativeness and local identity. Frontiers in Psychology 5, 1251.

Baugh, J. 2003. Linguistic profiling. Black Linguistics: Language, Society, and Politics in Africa and the Americas. S. Makoni, G. Smitherman, A. F. Ball, A. K. Spears (eds.). London and New York NY: Routledge. 155–163.

Chambers, J. K., P. Trudgill 1980. Dialectology. Cambridge Textbooks in Linguistics. New York NY: Cambridge University Press.

Chevrot, J-P., forthcoming. Acquiring sociolinguistic variation in a first language. Toward a Socialization-based Framework. Routledge Handbook of Variationist Sociolinguistics. Y. Asahi, A. D’Arcy, P. Kerswill (eds.). London: Routledge.

Choi, N., H-J. Cho, S. Kang, H. Ahn 2021. Korean children’s attitudes toward varieties of English: The role of age and English learning environment. Languages 6(3), 133.

Cornips, L. 2020. The impact of preschool attendance on children’s bidialectism in the Netherlands: Why toddlers may stop speaking a regional language (Limburgish) at home. Language in Society 49(3), 333–355.

Cremona, C., E. Bates 1977. The development of attitudes toward dialect in Italian children. Journal of Psycholinguistic Research 6(3), 223–232.

Day, R. 1980. The development of linguistic attitudes and preferences. TESOL Quarterly 14(1), 27–37.

Dailey-O’Cain, J. 1999. The perception of post-unification German regional speech. Handbook of Perceptual Dialectology. Vol. 1. D. Preston (ed.). Amsterdam and Philadelphia PA: John Benjamins. 227–242.

De Vogelaer, G., J. Toye 2017. Acquiring attitudes towards varieties of Dutch. Acquiring Sociolinguistic Variation. Studies in Language Variation 20. G. De Vogelaer, M. Katerbow (eds.). Amsterdam and Philadelphia PA: John Benjamins. 117–154.

Diercks, W. 2002. Mental maps: Linguistic-geographic concepts. Handbook of Perceptual Dialectology. Vol. 2. D. Preston, D. Long (eds.). Amsterdam and Philadelphia PA: John Benjamins. 51–70.

Fehér, K. 2020. Status-based preference of varieties in bidialectal kindergarteners: an experimental study. Argumentum 16, 147–172.

Foulkes, P., G. Docherty, D. Watt 2005. Phonological variation in child-directed speech. Language 81(1), 177–206.

Giles, H., C. Harrison, C. Creber, P. Smith, N. Freeman 1983. Developmental and contextual aspects of children’s language attitudes. Language and Communication 3(2), 141–146.

Hanyiszkó, K. 2011. Nyelvváltozatok és attitűdök avagy egy speciális ügynökvizsgálat tanulságai. [Language varieties and attitudes: Lessons from a special Matched-Guise Experiment.] BA thesis. Unpublished. Debrecen: University of Debrecen. https://mnytud.arts.unideb.hu/szakdolgozat/1688/hanyiszko_k_1688.pdf

Jánk, I. 2020. Linguistic discrimination and linguistic bias in pedagogical evaluations of Hungarian teachers. Hungarian as a Pluricentric Language in Language and Literature. I. Vančo, R. Muhr, I. Kozmács, M. Huber (eds.). Frankfurt am Main: Peter Lang. 211–221.

Jeffries, E. 2019. Preschool children’s categorization of speakers by regional accent. Language Variation and Change 31(3), 329–352.

Kaiser, I., G. Kasberger 2021. Children’s sociolinguistic preferences: The acquisition of language attitudes within the Austrian standard-dialect continuum. Sociolinguistic Variation and Language Acquisition across the Lifespan. A. Ghimenton, A. Nardy, J-P. Chevrot (eds.). Amsterdam: John Benjamins, 130–160.

Kaján, A., A. Deme 2020. Nyelvi attitűdvizsgálat iskolások körében. [Language attitude test among fifth- and twelfth-grade students.] Anyanyelv-pedagógia 13(3), 23–48. 10.21030/anyp.2020.3.2

Krajcsi, A. 2021. CogStat: An automatic analysis statistical software (Version 2.1.0). [Computer software.] Retrieved from https://www.cogstat.org

Kinzler, K., J. DeJesus 2013. Northern = smart and Southern = nice: The development of accent attitudes in the United States. Quarterly Journal of Experimental Psychology 66(6), 1146–1158.

Kinzler, K., K. Shutts, E. Spelke 2012. Language-based social preferences among children in South Africa. Language Learning and Development 8(3), 215–232.

Kircher, R., L. Zipp (eds.) 2022. Research Methods in Language Attitudes. Cambridge: Cambridge University Press.

Kontra, M. 2003. A szép magyar beszéd és a csúnya [The beautiful Hungarian speech, and the ugly]. Nyelv és társadalom a rendszerváltáskori Magyarországon [Language and society in Hungary during the change of regime]. M. Kontra (ed.). Budapest: Osiris, 240–255.

Kristiansen, T. 2010. Attitudes, ideology and awareness. The SAGE Handbook of Sociolinguistics. R. Wodak, B. Johnston, P. Kerswill (eds.). Los Angeles, London, New Delhi, Singapore, Washington DC: SAGE Publications, 265–278.

Ladegaard, H. 1998. Boys, girls, language and identity: language attitudes and linguistic behaviour in a rural community in Denmark. International Journal of Applied Linguistics 8(1), 3–25.

Ladegaard, H. 2003. Gender differences in young children’s speech: The acquisition of sociolinguistic competence. International Journal of Applied Linguistics 13(2): 222–233.

Lippi-Green, R. 2012. English with an Accent: Language, Ideology, and Discrimination in the United States. Second edition. London and New York NY: Routledge.

Maegaard, M. 2010. Linguistic practice and stereotypes among Copenhagen adolescents. Multilingual Urban Scandinavia: New Linguistic Practices. P. Quist, B. Svendsen (eds.). Bristol: Multilingual Matters. 189–206.

Molnár, M. 2022. A nyelvjáráshoz fűződő földrajzi és nyelvi szempontú attitűdvizsgálat pécsi gimnazisták körében. [Investigating attitudes from a geographic and linguistic point of view toward regional varieties in students at a high school in Pécs.] Modern Geográfia 17(1), 87–112.

Morales, G., L. Cornips 2023. Minoritising a regional language in multilingual preschools in Dutch Limburg: Teachers’ and toddlers’ choices between Dutch and Limburgish. International Journal of Bilingualism 27(3), 275–292.

Papazachariou, D., A. Fterniati, A. Archakis, V. Tsami 2018. Metapragmatic stereotypes about geographical diversity in Greece: Evidence from elementary school pupils’ responses to mass culture texts. Journal of Greek Linguistics 18(1), 124–154.

Paquette-Smith, M., H. Bucker, E. Johnson 2023. How sociolinguistic factors shape children’s subjective impressions of teacher quality. Quarterly Journal of Experimental Psychology 76(3), 485–496. doi:10.1177/17470218221094312

Parapatics, A. 2020. Teachers’ attitudes toward regional dialects in Hungary. Didacticae 8, 59–77.

Ponizovsky-Bergelson, Y., Y. Dayan, N. Wahle, D. Roer-Strier 2019. A qualitative interview with children: What encourages or inhibits young children’s participation? International Journal of Qualitative Methods 18.

Preston, D. 1998. They speak really bad English down South and in New York City. Language Myths. L. Bauer, P. Trudgill (eds.). London: Penguin Books, 139–149.

Preston, D. 1999. Attitude approach to the perception of regional variety. Handbook of Perceptual Dialectology. Vol. 1. D. Preston (ed.). Amsterdam and Philadelphia PA: John Benjamins. 359–373.

Rensink, W. 1999. Informant classification of dialects. Handbook of Perceptual Dialectology. Vol. 1. D. Preston (ed.). Amsterdam and Philadelphia PA: John Benjamins. 3–7.

Rosenthal, M. 1977. The Magic Boxes: Children and Black English. Urbana IL: University of Illinois, ERIC Clearinghouse on Early Childhood Education and ERIC Clearinghouse on Languages and Linguistics.

Sándor, K., Cs. Pléh, J. Langman 1998. Egy magyarországi ügynökvizsgálat tanulságai: A nyelvváltozatok hatása a személypercepcióra. Valóság: Társadalomtudományi Közlöny 1998(8), 27–40.

Stamou, A., K. Maroniti, E. Griva 2015. Young children talk about their popular cartoon and TV heroes’ speech styles: Media reception and language attitudes. Language Awareness 24(3), 216–232.

Szabó, T. P., K. Mátyus 2014. Óvodások és kisiskolások nyelvi ideológiái [Kindergarteners’ and primary school children’s language ideologies]. Nyelv – társadalom – kultúra: Interkulturális és multikulturális perspektívák.[Language – society – culture: Intercultural and multicultural perspectives.] Vol 1–2. M. Ladányi, Zs. Vladár, É. Hrenek (eds.). Budapest: MANYE and Tinta, 215–221.

Trudgill, P. 1972. Sex, covert prestige and linguistic change in the urban British English of Norwich. Language in Society 1, 179–195.

Tukacs-Károlyi, M. 1990. Az ingázók és nem ingázók beszédének sajátosságai Hajdúhadházon [Features of commuters’ and non-commuters’ speech in Hajdúhadház]. Élőnyelvi tanulmányok [Studies on present-day language]. L. Balogh, M. Kontra (eds.). Budapest: Hungarian Academy of Sciences, Institute of Linguistics, 200–208.

Submitted June 2024

Accepted October 2024

Appendices

Appendix 1. Stimuli. Underlined bold letters show vowel variation, regular bold letters refer to other variations. The ‿ symbol between words was applied for long consonants only (Fehér 2020)

Appendix 2. Number of variety choices in all test questions

Appendix 3. Choices of the standard variety by questions

Two-tailed binomial test. Participants without choice were excluded.

Appendix 4. Choices by questions and groups