4 Findings and Discussion
4.1 Findings by Research Instruments
4.1.1 Findings from the Pre-Task Interview
The questions of the pre-task interview aimed at collecting general data about the 12 Transylvanian-Hungarian multilingual participants with a special focus on the habits and preferences of language use. In addition to answering the interview questions, they were invited to use two additional support materials to evaluate their own language mastery and context of language use. The Grosjean-grid (Grosjean, 2011) was employed as a starting point in the evaluation of own language mastery along the dimensions of frequency of use and fluency. Block’s matrix (Block, 2003) was employed as a background for delineating the contexts of language use. The findings from the pre-task interview are presented according to the categories that emerged from the data analysis. The presentation of the findings linked to each category is followed by a short discussion of how the phenomenon captured by the category relates to the study and to existing theories. Within the discussion of each category, examples from the participants’ accounts are given. They contain the L1 Hungarian version, the language in which the interviews were conducted, and its English translation in square brackets. The following six provisional categories (a-f) were identified in the data collected in the pre-task interviews:
a. Awareness of being able to understand and use multiple languages;
b. The facilitating role of cognates in learning and using typologically close languages,
c. L2 Romanian as a supporter language for non-native languages;
d. Use of language governed by external factors, such as place and activity and internal factors, such as perceived usefulness;
e. Use of languages across the four skills is not homogeneous; and
f. Translation as a regular linguistic assignment in the participants’
educational context.
For this segment of data, these categories were termed core categories. However, relative to the entire amount of collected data, they were only inceptive and temporary categories, and they were subjected to further analysis in the subsequent steps of data evaluation. It is to be mentioned here that the seemingly large number of core categories at
this stage of data analysis is not in antagonism with some authors’ suggestions to keep a reduced number of categories while working towards theory development. Albeit GTM methods textbooks (Charmaz, 2006; Glaser, 1978a) advise researchers to narrow down the number of core categories to one or two, this suggestion is aimed at the more advanced stages of the data analysis, when the contours of the substantive theory are taking shape (Urquhart, 2013).
The two main properties of the categories derived from the participants’ accounts in the pre-task interviews were inceptiveness and temporariness. They were inceptive because they were derived in a relatively early stage of the study and because they acted like some kind of scaffolding structures which I used to support my further inquiry. They were also temporary, because they were taken forward in the data analysis and subjected to constant comparison with the categories developed in the other phases of data analysis.
Some words have to be said also about the length of the category labels. I intentionally settled on more detailed descriptions to name the categories, knowing that at a later stage they would be re-examined analytically, and on these occasions their descriptions would also undergo changes; that is, they would be replaced by more analytic ones.
4.1.1.1 Awareness of Being Able to Understand and Use Multiple Languages All 12 Transylvanian-Hungarian multilingual participants had a developed sense of what it means to learn, know and use several languages on different levels of fluency and with different levels of frequency. They reflected on how they were learning languages, they compared the languages they knew and use, mentioned strategies relating to both language learning and use verified by own practice (“this-works” belief statements), demonstrating a high level of awareness. In Excerpt (2) P.F. mentions English and German as being similar, infers the positive transfer from English towards German, and compares German to English in terms of the presence of the definite articles:
(2)
Amikor elkezdtem németet tanulni, már két éve tanultam az angolt és mondták, hogy lesznek átfedések, de aztán rájöttem én is, hogy mennyire hasonló a két nyelv. A némettel azért haladtam lassabban, mert ott vannak a der, die, das-ok.
[When I started to learn German, I had been learning English for two years, and I was told that there would be overlaps, but then I myself realised how similar the two languages are. I progressed with German slower because of der, die, das.]
In Excerpt (3) B.Cs. talks about the discovery of being able to say basic things in two of his non-native languages and the feeling of joy attached to the ability use several languages.
(3)
P.E: Milyen előnye van annak, hogy több nyelvet tudsz?
B.Cs.: Szerintem az, hogy el tudok mondani egy valamit, egy dolgot több nyelven. Egyszer, régebben rájöttem, hogy el tudok mondani egy rövid mondatot angolul és franciául is. Azt hiszem, az volt, hogy Engem Csabának hívnak.
Azután már mindent próbáltam elmondani négy nyelven.
P.E: Ez igen. Most is csinálod ezt?
B.Cs.: Persze. És amikor sikerül hibátlanul, boldoggá tesz.
[P.E: What are the advantages of knowing several languages?
B.Cs.: In my opinion, the fact that I can say something, a thing in more languages. Once, a long time ago, I discovered that I could say a short sentence in both English and French. I think it was My name is Csaba. After that, I tried to say everything in four languages.
P.E: Awesome. Do you still do this?
B.Cs.: Of course. And it makes me happy when I get it right.]
The increased sensitivity towards languages, the insight into how different non-native languages are similar or different and the evaluation that positive transfer operates between two non-native languages are all traits most characteristic to multilingual individuals (Jessner, 2006, 2014). This group of traits is known as metalinguistic awareness in specialist literature, having “been identified as (…) the key factor of multilingual learning” (Jessner, 2014, p. 175, emphasis in the original). Metalinguistic awareness has been defined as the “ability to focus attention on language as an object in itself or to think abstractly about language and, consequently, to play with or manipulate language” (Jessner, 2006, p. 42). It has been shown to be directly proportional with the number of languages known by an individual and to positively influence the ability to monitor language use. Metalinguistic awareness deserves special attention in the inquiry of how languages are relied on and how non-native words are accessed by multilinguals because it influences directly multilinguals’ ability to keep track of the cognitive-linguistic processes and to take appropriate measures to deal with possible breakdowns (Jessner, 2006, 2014).
Closely connected to metalinguistic awareness is the subcategory which I termed
“this-works” beliefs, where “this works” is an in vivo code. In the analysis of the pre-task interviews, an emergent characteristic of the participants’ awareness of being able to understand and use multiple languages was their well-contoured beliefs about how language learning should take place and what some viable lexical problem-solving strategies are.
In Excerpt (4) K.J. talks about how he proceeds when he has to learn verb tenses in two non-native languages, L3 English and L4 French.
(4)
Ez úgy működik, hogy az ember, például, már megtanulta angolul a present perfect-et és akkor jön a francia passé composé. Ami, ugye nagyon olyan, mint a present perfect. Megjegyzem, hogy mi a különbség a kettő között és mindig a különbségeket jegyzem meg. A többit már nem, mert már egyszer megjegyeztem.
[This works when, for instance, one has already learnt the present perfect in English and then comes the French passé composé. Which, well, is very much like present perfect. I commit to memory what is different between the two, and I always memorize the differences. I do not memorize the rest because I already know it.]
From the perspective of the research interests of the present study, beliefs are important because of their power to influence behaviour and language use. Beliefs are
“propositions about the world that are held true” (Abdi & Asadi, 2015, p. 104) and are central components of human behaviour and learning. According to Pajares (1992), a “belief is based on evaluation and judgement” (p. 313). Because beliefs arise out of judgements, and judgements are the results of the subjective feeling that certain cognitive activities are easy or difficult, that is of processing fluency, a direct relationship can be established between processing fluency and beliefs (Greifeneder et al., 2010; Kurilla & Westerman, 2008;
Topolinski, 2013; Unkelbach & Greifeneder, 2013a; Wanke, 2013). This relationship, as it is explained later, is relevant from the standpoint of reliance on languages and strategies employed to access non-native words in linguistic problem solving situations.
4.1.1.2 The Facilitating Role of Cognates in Learning and Using Typologically Close Languages
A second category that emerged from the analysis of the pre-task interview data was the important role ascribed by the participants to the words that are similar in both form and meaning in two or more languages. They mentioned the importance of cognates in language learning and use. Harnessing the potential of cognates seems to be one of the strategies developed under the aegis of efficiency, or the principle of least effort (Zipf, 1949). In Excerpt (5) K.J. talks about how he makes use of cognates to retain some animal’s names across languages.
(5)
Az állatok neveit tudom megjegyezni, speciálisan azért, mert sok állatnak ugyanaz a neve több nyelven is. A zsiráf például és a tigris ilyenek. A zsiráf románul girafǎ, franciául giraffe, és még latinul is tudom a nevét: Giraffa.
[I can memorize the names of the animals, especially because many animals are called the same way in several languages. The giraffe and the tiger are like this.
The giraffe, in Romanian is girafǎ, in French is giraffe, and I know even its Latin name: Giraffa.]
Interestingly, the participants highlighted the positive aspect of cognates versus their less positive aspect, that is, the interference they can sometimes cause. This seemed to be due to the fact that the favour they do is much beyond the hindrance they cause. Excerpt (6) illustrates a case in which H.B. refers to the pronunciation of cognates:
(6)
Hát, néha rosszul ejtem ki, mert hasonlóan írják, de aztán azonnal rájövök, hogy rosszat mondtam és kijavítom magam.
[Well, sometimes I pronounce it wrongly because it is written similarly, but then I realize that I said something wrong and I correct myself.]
Excerpt (6) also hints at the possible existence of a heightened tolerance towards errors in multilinguals.
The influence and role of cognates have been extensively documented by second language lexical research, especially by cross-lexical influence studies, and cognates have been proved to act as strong links between languages, known as the cognate facilitation effect (Dijkstra et al., 1999; M. Gibson & Hufeisen, 2003; Ó Laoire & Singleton, 2009; Odlin, 1989; Voga & Grainger, 2007). In Ecke’s (2015) and Ecke and Hall’s (2003) vision, for instance, cognates are at the base of the parasitic relationships which establish between the lexical units of an individual’s already known language(s) and the lexical units of a language that is newly learnt.
For the present discussion on multilinguals’ reliance on languages and retrieval of non-native target lexical units, cognates are important because the Transylvanian-Hungarian multilingual participants’ linguistic repertoires comprised typologically related languages, known to share a considerable number of words with phonetic and semantic similarity (Kondrak, 2013; Peeters et al., 2013), thus potentially acting as support languages for each other (H. Wang & Sitbon, 2014).
4.1.1.3 L2 Romanian as a Support Language for Non-Native Languages
The category of L2 Romanian as a facilitator or a support language for the non-native languages emerged from the participants’ answers to questions relating to learning foreign languages in formal contexts. It was not surprising that L2 Romanian played an important role in the learning processes of those six Transylvanian-Hungarian participants who were attending L2 Romanian mainstream education. However, I expected to find L2 Romanian having a lesser relevance in the case of those participants who were attending L1 Hungarian minority-language education. The participants from the first group mentioned L2 Romanian as a scaffolding language employed in the foreign language classroom. That is, during the French, English, German and Latin lessons, the teachers gave the explanations in L2 Romanian and student comprehension of vocabulary was achieved and negotiated in L2 Romanian. Excerpts (7), (8), (9) and (10) are data extracts from the pre-task interviews with participants from the group who took part in the L2 Romanian mainstream education. The participants’ initials are given in parentheses at the end of each excerpt in Hungarian.
(7)
Románul tanulok minden nyelvet. Mondjuk, ez nem baj, mert jól tudok románul.
(Sz.E.)
[I learn each language in Romanian. Well, it doesn’t matter because I am very good at Romanian.]
(8)
Ez egy román iskola. Nem is magyarázhatnának más nyelven. Csak a franciatanár magyaráz néha franciául is. (B.L.)
[This is a Romanian school. They couldn’t explain things in any other language.
Only the French teacher explains sometimes in French.]
(9)
Soha nem is tanultam angolt magyarul, vagyis az angoltanár mindig román volt.
(P.L.)
[I have never learnt English through Hungarian; that is, the English teacher was always Romanian.]
The participants from the L1 Hungarian minority school mentioned L2 Romanian especially in connection with “being prepared for the exams” (B.Cs.), and further learning.
The participants’ language teachers were all native speakers of Hungarian; thus, during the foreign language lessons L1 Hungarian was used as a scaffolding language for the foreign languages taught at school. In the pre-task interview, only one participant, in the data extract in Excerpt (10), remarked that she preferred to use the set of Romanian-English and English-Romanian bilingual dictionaries because of “containing more words” (S.S.).
(10)
A román szótárral keresem a szavakat általában. Azért, mert a magyar szótárban nincs meg minden szó. A románban mindent megtalálok.
[I usually look up the words in the Romanian dictionary. Because the Hungarian dictionary does not contain all the words. In the Romanian one, I find everything.]
Employing a scaffolding language in the FL classroom is well-documented in the applied linguistics research literature. However, the focus of the existing research has been mainly on how the students’ L1 contributes to the efficiency of the teaching of a foreign language (Bhooth, Azman, & Kemboja, 2014; Butzkamm & Caldwell, 2009; Storch &
Wigglesworth, 2003). Employing a scaffolding language other than the L1 in a foreign language task performance is related to the psycholinguistic area of language choice characteristic to people fluent in multiple languages (Levine, 2011). In the present study, from the perspective of the discussion of multilinguals’ reliance on languages and access of non-native target-words, the category of L2 as a support language bears importance because it can be one of the factors which shape an individual’s dependence upon languages for “cognitive support” (Storch & Wigglesworth, 2003, p. 760).
The dimensions of this category—the strong presence of the L2 Romanian as a scaffolding language for foreign languages in one context (L1 Hungarians in the L2 Romanian mainstream education) and its more nuanced presence in the other context (L1 Hungarians in the L1 Hungarian minority education)—were considered patterns with both common and diverging properties. The relationship between these two dimensions of the category is one of partial similarity and can be conceived of in the form of a continuum along which the presence of the L2 Romanian as a support language is an increasing progression between the two ends of the continuum: L2 Romanian in mainstream education and L2 Romanian in the L1 Hungarian minority education. The existence of these two dimensions indicates that the use of an L2 in educational settings as the medium of instruction impacts on the students’ patterns of language use, a fact already supported by sociolinguistic empirical evidence (Gregory & Williams, 2000; Martin-Jones & Jones, 2000; Versluys, 2008).
These pieces of evidence, however, only identify the influence that the language of education
has on the students’ language use, but do not offer explanations about why this influence manifests itself.
The category of L2 Romanian as a support language for non-native languages underwent changes in the subsequent stages of data analysis: its range became narrowed down to support language for the L3 English, and it became one of the dimensions of the category reliance on languages, the category developed in the stage of the think-aloud protocol analysis.
4.1.1.4 Use of Language Governed by External Factors, Such as Place and Activity and Internal Factors, Such as Perceived Usefulness
This category is grounded in the participants’ accounts of the contexts in which they used their different languages and the purposes the languages fulfilled. The dimensions of the category are as follows:
• the languages the participants knew and used,
• the context in which particular languages were used,
• the purposes fulfilled by each language, and
• the perceived practicality of each language.
To represent visually this quite complex category, it is useful to employ Block’s (2003, p. 34) context-matrix developed for describing language-learning scenarios. In the pre-task interview, I adopted a simplified version of the matrix and adapted it to explore the situations in which languages were used. Figure 19 shows such a completed matrix.
Block’s (2003) matrix has four quadrants, and each represents the combination of two contexts: school and community. Languages can be placed within the four quadrants depending on the contexts in which they are used. The section +School/+Community is the place for the language(s) used both at school and in the larger community, including home.
The quadrant + Community/- School is the one for the language(s) used in the larger community but not at school. The quadrant –School/-Community is the place for the language(s) that are not used either at school or in the larger community, such as a self-taught language. The quadrant -Community/+School can include the languages that are learnt at school and are not used in the community.
Figure 19 shows that participant S.S. placed L1 Hungarian and L2 Romanian in the +School/+Community quadrant, and L3 English, L4 German, and L5 Latin in the – Community/+School. The distribution of languages in the case of all 12 participants was similar.
L1 Hungarian and L2 Romanian were placed in the +Community/+School quadrant.
In some cases, L3 English was also included in this quadrant as a language present in the larger community because it was the language in which the participants were exchanging written messages on the internet. While L3 English was placed on occasions both in quadrant +School/+Community and quadrant +School/-Community, the other languages L4 French or L4 German, and L5 Latin were typically placed in the quadrant – Community/+
Figure 19. A Completed Language-context Matrix.
The sign + represents the presence of the language (in the community and/or at school), the sign – represents the lack of the language (in the community and/or school). The hungarian words mean:
ISKOLA=SCHOOL
KÖZÖSSÉG=COMMUNITY ANGOL=ENGLISH
FRANCIA=FRENCH LATIN=LATIN
MAGYAR=HUNGARIAN ROMÁN=ROMANIAN
The Hungarian sentence MELYIK TERÜLETEN HASZNÁLOD A NYELVEKET, AMELYEKET ISMERSZ? means WHERE DO YOU USE THE LANGUAGES YOU KNOW?
School as languages used exclusively at school. These latter languages, especially L5 Latin, were perceived by the participants as being rather subject matters or spheres of knowledge, to be learnt.
The category of use of language governed by external factors, such as place and activity and internal factors, such as perceived usefulness connects with the functionalist approach to language, particularly that of Nuyts’ (1989). In this sense, language is action, and the goal of language in these situations is to “contribute to the performance of these actions” (Nuyts, 1989, p. 92). In these actions, language is present either with its interpersonal role (Hymes, 1969) or with its intrapersonal role (Fernyhough, 2016; Harman, 1975). In the case of multilingual language users, the use of different languages in different interpersonal contexts has been shown to be in strong connection with the language user’s social identity (Bayer, 1990; Gogolin et al., 2013; Versluys, 2008). It has been also shown that in interpersonal contexts the employment of one or another language is very much a function of socio-culturally rooted conventions, such as politeness, solidarity, prestige, and deference (Versluys, 2008). Additionally, research on how the employment of several languages manifests itself has documented the fluent and flexible ways in which multilinguals are able to juggle and mix their languages. These aspects of multiple language use in varied contexts have been termed “translanguaging” (García & Wei, 2014; Mazzaferro, 2018), and
“polyanguaging” (Jorgensen et al., 2011b).
Excerpt (11), in which participant H.B. talks about his home languages, illustrates a context in which two languages coexist and their employment is governed by politeness. The participant does not mention it, but it is a tacit socio-cultural norm to switch to the language spoken by all those present in the immediate surrounding of the speakers even when there is one person who does not take part in the conversation and does not understand or speak the language employed by the speakers.
(11)
Otthon mamámmal románul beszélek, mert ő ért magyarul, de nem beszél.
Anyukámmal, ha nincs ott mama, akkor magyarul beszélek, de ha mama is ott van, akkor románul beszélünk. Apukám tud magyarul egy kicsit, de vele is románul beszélek.
[At home, with my grandmother I speak Romanian because she understands Hungarian but does not speak it. With my mother, if grandmother is not there, I speak Hungarian, but if grandmother is there, too, we speak Romanian. My father knows Hungarian a bit, but I speak Romanian with him, too.]
Besides context, the language use of the participants proved to be influenced by the nature of the activity to be performed. Participant S.B., in Excerpt (12), reports about using L1 Hungarian for basic arithmetic operations while solving arithmetic problems in L2 Romanian.
(12)
Azt hiszem, lassabb lennék, ha románul kellene számoljak. Amikor hangosan számolok órán, akkor persze, hogy románul számolok. Magamban mindig magyarul számolok. A szorzó táblát is magyarul tudom.
[I guess I would be slower if I had to do the operations in Romanian. When I do the operations aloud in the lesson, of course, I do them in Romanian. When I do them using my inside voice, I always use Hungarian. I also know the multiplication table in Hungarian.]
Several studies on the numerical cognition of bilinguals and second-language learners have examined language-reliance in arithmetic problem solving from the perspective of language proficiency and have found that proficiency in a second language is one of the crucial factors which determines which language is used to perform arithmetic calculations (Van Rinsveld et al., 2015, 2017; Y. Wang et al., 2007). The same phenomenon has not been investigated yet in the case of multilinguals, and there are also questions to be answered in relation to the case of balanced bilinguals or individuals who are equally proficient in their two or even three languages.
The last dimension that emerged in connection with the category of use of language governed by external factors, such as place and activity and internal factors, such as perceived usefulness is perceived practicality of a language in certain situations. This dimension was identified mostly in connection to L1 Hungarian, and L3 English. For instance, participant P.L. speaks about L3 English in Excerpt (14) as being “very useful” in playing cooperative or multiplayer online video games in which he could pair up with players who also spoke English.
(13)
Call of Duty-t szoktam játszani és nagyon hasznos ilyenkor az angol, mert egy csomóan játsszák ezt a játékot és majdnem mindenki tud angolul. Sokszor volt már olyan, hogy angolul beszélt a párom és angolul beszéltem végig vele.
[I usually play Call of Duty, and English is very useful in these situations because there are many who play this game and almost everybody knows English. It happened many times that my partner spoke English and I spoke English with him throughout.]
Within the practicality dimension, an example of limited usefulness emerged in connection to L1 Hungarian. In Excerpt (14), F.L. speaks about the usefulness of her L1 Hungarian and about the awareness of not speaking the Hungarian considered mainstream.
(14)
Sajnos a magyarnak csak itthon tudom hasznát venni és Magyarországon.
Vannak magyarországi ismerőseim és voltam már többször Magyarországon. Ott gyorsan észreveszik, hogy nem magyarországi magyart beszélek.
[Unfortunately, I can use Hungarian only here and in Hungary. I have Hungarian acquaintances, and I have been to Hungary several times. There people quickly notice that I do not speak the Hungarian spoken in Hungary.]
It can be noticed that among the four dimensions of the category use of language governed by external factors, such as place and activity and internal factors, such as perceived usefulness there are overlaps and that these dimensions are interconnected. For instance, in the case of Excerpt (14), the participant considered L1 Hungarian useful and she also connected it to two places: Transylvania and Hungary. Then, Excerpts (11) and (12) are illustrations of L1 Hungarian and L2 Romanian used at home and also for carrying out a particular activity. It is obvious that these overlaps and interconnectedness are due to the situated nature of cognition and language use: we use language to carry out certain activities in certain places, and we perceive the practicality of a language as a degree of its characteristic of being useful in completing an activity (Levine, 2011). Especially because of its perceived practicality trait, the category of use of language governed by external factors, such as place and activity and internal factors, such as perceived usefulness connects with the Levine’s (2011) theory of language choice in task performance, which sees the ability to choose between languages as heightened “awareness of which language to use, with who, when, and why” (p. 4).
Attention has to be drawn to the fact that the category of use of language governed by external factors, such as place and activity and internal factors, such as perceived usefulness has clear links with the category L2 Romanian as a supporter language for non-native languages, because L2 Romanian in that category appears as a language employed for well-defined purposes. In the process of data analysis and interpretation, however, I considered the latter category an independent one because of the several situations in which the participants mentioned L2 as a support language in the process of learning foreign languages.
4.1.1.5 Use of Languages Across the Receptive and Expressive Language Modalities is not Homogeneous
This category can be considered the extension of the previous category, languages used with different purposes and in different context, because it takes language use a step forward in the direction of the receptive and expressive language modalities traditionally considered when referring to mastery of a language.
In the data analysis, several language-use patterns surfaced for the participants’
different languages. The participants’ L1 Hungarian and L2 Romanian were used in varied situations and across the receptive modalities of reading, listening and watching moving pictures, and the expressive modalities of speaking and writing. English, the language that the participants learned as an L3, was also put to use across the receptive language modalities of reading, listening and watching moving pictures, and the expressive modalities of speaking and writing. However, its position emerged as being clearly delineated by issues relative to classroom-learning. Participant B.A., in Excerpt (15) talks about the L3 English employed at school, and how he considers the use of L3 English different from L4 German.
(15)
Az angolt jóformán az iskolában használom. Ja, igen és próbáltam olvasni angolul, néha. A Harry Potter-t, például. De mégsem olyan, mint a német, amit beszélek az ismerőseimmel. A filmekből és a zenékből tanultam sok (angol) szót.
Néha használom őket, de nincs kivel. Az iskolában nem beszélünk angolul, csak felelünk.
[I use English mostly at school. Well, and I tried to read in English sometimes.
Harry Potter, for example. But, still, it is different from German, which I speak with my acquaintances. I learned many (English) words from films and songs.
Sometimes I use them, but there are no people around to speak with. At school we do not speak in English, we only recite the lesson.]
In Excerpt (15), B.A. also mentions reading, listening and watching content in L3 English as language modalities used as quotidian activities.
Reading in L3 English outside the school was mentioned by another five participants, and writing in L3 English outside the school was mentioned by seven participants.
Furthermore, listening and watching films in L3 English outside the school were mentioned by all the 12 participants. The only language modality relative to L3 English, which was confined almost exclusively to school (with few exceptions, see excerpt 13), was speaking.
Based on the participants’ reports, French and German, which the participants learned as an L4, were used across the receptive language modalities of reading and