16. PhD revisited: Meaning in English1The chapter presents
the overall results of a doctoral study (Rindal, 2013) from the
University of Oslo. This is an article-based thesis with three published
articles (Rindal, 2010, 2014; Rindal & Piercy, 2013). The doctoral
thesis in its entirety – with theoretical, methodological and empirical details
– can be found here: https://www.duo.uio.no/handle/10852/65151
L2 attitudes, choices and pronunciation in Norway
- Side: 335-355
- DOI: https://doi.org/10.18261/978-82-15-03074-6-2019-17
- Publisert på Idunn: 2019-06-18
- Publisert: 2019-06-18
- Creative Commons (CC BY 4.0)
This chapter reports a doctoral study (Rindal, 2013) that investigated L2 practices in a English language teaching (ELT) context using a theoretical and methodological framework at the intersection of linguistics and education. The results suggest that Norwegian learners can express local and individual identity through English. The chapter discusses the implications of such L2 social meanings for the teaching of English in Norway, and presents suggestions for future research.Keywords: pronunciation, language attitudes, social practice, English as a global language
In an increasingly globalised world, English has become the foremost global language of communication. English is not only the result of a more economically, culturally and professionally interconnected world; it also provides the transnational contact language that enables these connections. Because of this function as a transnational lingua franca, between one and two billion people are learning English around the world (Crystal, 2003; Graddol, 2006), and the status of English in the world is increasingly characterised by those who use it as a second or later language, rather than by its native speakers (Kirkpatrick, 2010a). English is not merely spread around the world, but appropriated by speakers in different communities. The global development of English is thus characterised by variation and diversity, as more and more speakers negotiate English in their local environments. Speakers enact English to meet communicative and situational demands – in this way, the globalised English is being constantly localised through its use.
In Norway, as in most other European countries, English has traditionally been labelled a foreign language, acknowledged for international travel, business and education, and taught at scheduled hours in the classroom. However, following increased out-of-school exposure and English language proficiency, English no longer feels foreign to Norwegians. Young Norwegians are frequent users of entertainment and social media, from which they are exposed daily to English. They travel on holiday or to study in countries with both native and non-native speakers of English. Almost half of the younger population (age 25–34) undertake higher education (OECD, 2011) where a considerable amount of written material and lectures are given in English (Ljosland, 2008; Schwach & Dalseng, 2011). English language competence is increasingly considered a basic skill. And yet, Norwegians do not qualify as speakers of English as a second language as these are often described in the literature (e.g. Graddol, 2006); they are neither speakers of New Englishes in postcolonial countries nor immigrants to a native-English-speaking country, and English is not an official language in Norway. For young Norwegians, then, English is neither a first, second or foreign language.
It was in this social reality, where English is characterised by diversity and variability globally, and without any clear status to its L2 speakers locally, that the participants of the present doctoral study practised English as L2. The primary research question was:
In order to access potential social meaning in the English of Norwegian adolescents, the study investigated different aspects of their L2 behaviour: their pronunciation of English, their L2 choices, and their social evaluations of the target language.
In the sociolinguistics of style (Coupland, 2007; Eckert & Rickford, 2001), language is a practice – an activity of creating social meaning. Speakers pick from a range of linguistic resources to generate new meanings or reorganise old meanings. Stylistic practice concerns “adapting linguistic variables available out in the larger world to the construction of social meaning on a local level” (Eckert, 2004, p. 44). In order for language use to be meaningful, speakers must have a shared idea of the potential meanings that are associated with linguistic forms. Such meanings can be accessed by investigating language attitudes (Garrett, 2010; Irvine & Gal, 2000).
A much-used definition of attitude for language attitude research purposes is taken from Sarnoff (1970, p. 279): “a disposition to react favourably or unfavourably to a class of objects”. As a “disposition”, attitudes are not directly observable, and can be difficult to access. They must be inferred from more obvious processes, such as statements, language behaviour, or reactions to other people’s language behaviour. Language attitudes are therefore often elicited using a matched-guise test, which is designed to elicit spontaneous reactions, as was also the case in the doctoral study reported here.
In the social sciences, there is considerable agreement that language attitudes are dynamically related to language behaviour, but the link between them is not a direct one (Garrett, 2010; Kristiansen, Garrett, & Coupland, 2005); a positive disposition towards certain speakers does not mean that one necessarily wishes to behave like them. In order to predict behaviour, attitudes are probably best investigated at the right level of specificity (Ajzen & Fishbein, 1980). Therefore, in addition to investigating evaluations of native accents of English, this doctoral study also accessed language choices related to participants’ own use of English pronunciation.
The global spread and local appropriation of English has elicited a growing body of research that describes and discusses the new and developing Englishes around the world. Most of this research has been conducted in postcolonial settings where English has official language status alongside one or more national languages and bilingualism has led to creole or indigenised Englishes (e.g. Kachru, Kachru, & Nelson, 2006; Kirkpatrick, 2010b). Furthermore, an emerging research trend has been the investigation of English used as a lingua franca by non-native speakers (e.g. Jenkins, 2006; Seidlhofer, 2010). However, very little research has focused on environments where L2 speakers are proficient enough to use English as part of their linguistic and identity repertoire, but where English does not have status as an official second language or is used as a necessary language of communication. It is this sort of L2 environment that was the context for the study presented here.
Prior to the doctoral study, an increasing number of studies had investigated attitudes of non-native speakers towards varieties of English (e.g. Dalton-Puffer, Kaltenboeck, & Smit, 1997; Evans & Imai, 2011; Ladegaard, 1998; van der Haagen, 1998). Among the non-native speakers in European studies, Standard Southern British English (henceforth SSBE) had been considered superior to both General American (henceforth GenAm) and so-called “non-standard” varieties (Dalton-Puffer et al., 1997; Ladegaard, 1998; van der Haagen, 1998). Learners in Europe had also considered SSBE the most attractive model of pronunciation (Ladegaard, 1998; van der Haagen, 1998). However, scholars had suggested avoiding the so-called “standard” accents such as SSBE or GenAm as models for learners (e.g. Bex, 2008; Dürmüller, 2008) because these accents are “imbued with the cultures of their origin” and following these norms might mean “implicitly signalling that they wish to convey the values of such native speakers” (Bex, 2008, p. 233).
However, although scholars were questioning native-speaker standards, there had been little research into alternatives for English language learners. Research related to social aspects of language learning in the Second Language Acquisition (SLA) field seemed to focus mostly on how the learner’s identity develops (or refrains from developing) when taking on another’s language (e.g. Norton & Toohey, 2011), not how they create meaning with language. However, following the developing status of English globally and locally, it became relevant to investigate the meaningful use of linguistic resources by L2 speakers of English, which was the endeavour embarked upon by this study. Nonetheless, SLA research had shown that learning is enhanced when the teacher engages the learners’ identities, encouraging educators to take social context into account when developing curricula and instructional design (Menard-Warwick, 2005). This educational concern from SLA was also central to the doctoral study, giving it an interdisciplinary nature with a foot in both sociolinguistics and language education.
Although this study was conducted within the field of English didactics and clearly had educational concerns, sociolinguistic methods seemed the most appropriate for the purpose of investigating social meanings of language, and were therefore predominant in the methodological approach. As what was most likely the first sociolinguistic investigation into L2 practices in a Norwegian ELT context, the study was exploratory, using both quantitative and qualitative methods and analyses. The project was a mixed-methods study (as defined in e.g. Johnson & Onwuegbuzie, 2004), which pragmatically used the methods necessary to meet its research objectives, and in which the purpose of using mixed methods was complementarity and expansion.
Ninety-seven students from four different classes at four different upper secondary schools in Oslo participated in the study. Similar data was collected at two different stages, allowing for the methodology to be developed further after the first stage. Data set 1 comprised 23 students from one school (School S), while Data set 2 comprised 74 students from three different schools (Schools A, B and C). All students (58 females and 39 males) were approximately 17 years old and in their second year of upper secondary school (Vg2). All the students in Data set 1 and 90% of the students in Data set 2 reported having attended Norwegian school from year 1, which means that most of them would have learnt English in Norwegian society and would have had English as a school subject for 11 years at the time of the study.
The research design consisted of methods which were appropriate for the different parts of the study; L2 pronunciation, attitudes and reported choices – as well as for the study as a whole; social meaning in the L2 of this specific context. Table 16.1 gives an overview of the methods and analyses used in the study.
|Quantitative data||Qualitative data|
|Audio recordings of speech in two situations |
Data set 1: N=23
Data set 2: N=70
Data set 1: N=21
Data set 2: N=70
|Matched-guise test (MGT) |
Data set 1: 21
Data set 2: 70
Data set 1: 5
Data set 2: 7
Data set 1: N=21
Data set 2: N=70
Note. Two students were absent when the questionnaire in Data set 1 was administered. Two audio recordings with a total of four students from Data set 2 could not be analysed because they were inaudible, and four students (not the same students) were absent when the MGT and questionnaire was administered. The quantitative and qualitative data come from the same questionnaire.
The participants were recorded while reading a word list and while having a casual conversation in English with a classmate of their choice for approximately 10 minutes.
In Data set 1, four different pronunciation variables were chosen for examination; the quality of the vowels in the lexical sets LOT and GOAT, the presence or absence of non-prevocalic /r/, and the presence of intervocalic /t/ voicing. In Data set 2, another three variables were added; the quality of the vowel in the lexical set BATH, the presence or absence of post-coronal /j/, and the realisation of voiceless th. The first six variables are salient distinguishing differences between the reference accents of GenAm and SSBE (e.g. Trudgill, 1986), and the seventh (th) is pronounced as a dental fricative /θ/ by speakers of both GenAm and SSBE, but sometimes replaced with /t/ by learners of English, since most of these, Norwegians included, do not have dental fricatives in their L1 (e.g. Cruttenden, 2008).
The variables were analysed using auditory analysis. The participants’ realisations of phonological variables were classified with reference to GenAm or SSBE accents. The analysis in Data set 2 was more detailed than in Data set 1, combining realisations that defied categorisation into one of the two main variants into an “Other” category, and adding a fourth category for Intervocalic /t/. In total, 10,864 tokens were analysed. Paired sampled t-tests were performed to investigate differences in pronunciation across the situations word-list reading and conversation speech.
A matched-guise test (MGT) was used to investigate participants’ reactions to different English varieties. In a matched-guise test (Lambert, Hodgson, Gardner, & Fillenbaum, 1960) recordings of two or more varieties produced by a single speaker are evaluated by listeners who are under the impression that they are listening to different speakers, thus allowing for evaluation of accents when individual voice attributes are the same. A total of 4 recordings in Data set 1 and 11 recordings in Data set 2 of speakers with four different accents were used as stimulus: the “standard” accents SSBE and GenAm, as well as the “non-standard” accents Scottish English and Leeds English. The students who participated in the study were asked to evaluate the stimulus recordings by filling out evaluation forms with semantically labelled scales. The scales were chosen from those employed in previous and comparable studies (e.g. Ladegaard, 1998; van der Haagen, 1998), but were moderated and supplemented after pilot tests and conversations with comparable raters. Students were asked to evaluate a set of person-related qualities, as well as three language-related qualities (see Table 16.2). Students were asked to rate stimulus speakers on scales from 1 (low) to 5 (high), reporting their first impression.
A principal components analysis (PCA) was performed in order to investigate whether the evaluations patterned into dimensions as in previous language attitude research based on the students’ ratings. A PCA with Varimax rotation was therefore performed on the ratings given to each individual quality on the evaluation form. A one-way repeated measures analysis of variance (ANOVA with Bonferroni-adjusted post-hoc comparisons) was then conducted to compare the attitude dimensions that emerged from the PCA.
Questionnaire and interviews
In a questionnaire administered after the audio recordings and the matched-guise test, the students were asked to report which pronunciation they aimed towards when they spoke English. For Data set 1, the students were asked to report either American English, British English, Other or I don’t care as their “accent aim” in a closed-end question. The results from this data set called for two additional accent aims to choose from, and in Data set 2 students were also given the alternatives Norwegian and “Neutral”. Norwegian was included as an alternative because the results from Data set 1 suggested that students had an idea of such an accent. “Neutral” was given as an alternative following results from Data set 1, as well as an additional pilot study where students reported that instead of aiming towards any of the well-known native accents, they preferred a geographically unrecognisable English accent. In both data sets, in an open-ended question in the questionnaire students who had chosen British English were asked why they had not chosen American English, and vice versa, in order to explore the differences between them.
The students’ reported accent aims were counted, and a Kruskal-Wallis test was performed in order to investigate the effect of accent aim on pronunciation, i.e. whether the students were using the pronunciation they reported to target. A thematic analysis was conducted on their reported reasons to aim towards either a British English or an American English accent; 26 answers by American English aimers (henceforth AmE aimers) and 21 answers by British English aimers (henceforth BrE aimers) produced a total of 66 reasons, which were grouped and labelled according to content.
Furthermore, in order to explore and discuss the rationale for the choice of target accents in social context, focus-group interviews with students were held at each of the four schools (two interviews at School S, one interview at each of the Schools A–C). For Data set 2, interviews were held with students who had reported a “Neutral” accent aim, since the questionnaire did not focus on this accent choice. A detailed analysis of the interviews is beyond the scope of this chapter, but quotes are included to illuminate the quantitative data. While questionnaires were administered and responded to in English, interviews were held in Norwegian and quotes from these are translated into English by the author.
The results from the matched-guise test showed that SSBE was evaluated as the accent with highest linguistic quality and the most status and competence qualities compared to any other native-speaker accent. However, American English was more popular than British English as target accent, and some students reported to avoid native accents as targets altogether. Furthermore, results from the auditory analysis showed that GenAm phonological variants largely dominated students’ pronunciation.
Attitudes towards native-speaker accents of English
The principal components analysis (PCA) generated three categories of attitudinal dimensions. Two of them are prevalent in language attitude research, namely Status and Competence and Social Attractiveness, while Linguistic Quality is a common category in evaluational research in non-native speaker contexts. Table 16.2 shows how the qualities from the matched-guise test (MGT) were loaded into these three components.
The matched-guise test (MGT) in both data sets gave the same result: SSBE was rated significantly higher than GenAm for Status and Competence and for Linguistic Quality (p<0.001 with an ANOVA with Bonferroni-adjusted post-hoc comparisons). This means that the students found speakers of SSBE to be more educated, formal, ambitious, etc. than speakers of GenAm, and the SSBE accent more intelligible and aesthetic and a better model of pronunciation than GenAm. In Data set 2, which included Leeds and Scottish English accents, GenAm was rated similarly to these. There were no significant differences in the evaluations between any of the accents for Social Attractiveness, which means that speakers were found equally attractive, pleasant, etc. regardless of accent.
Table 16.3 gives an overview of the reported accent aims. 11/21 (52%) students in Data set 1 and 23/70 (33%) students in Data set 2 chose British as their accent aim. 8/21 (38%) students in Data set 1 and 30/70 (43%) students in Data set 2 reported American as their English pronunciation target. Nobody chose the option Norwegian. One student from each data set reported an Other aim, and one participant in Data set 2, here also in the Other category, circled both British and American and wrote It can be useful to learn both. “Neutral” was only given as an alternative in Data set 2; 11/70 (16%) students chose this alternative.
Students who had chosen either British English (BrE) or American English (AmE) as their accent aim were asked why they did not choose the other accent as aim. A thematic analysis of their answers produced a total of 66 reasons (Table 16.4). Italics refer to quotes from the questionnaire and interviews.
|Reason||AmE aim||BrE aim||Description||Example|
|Access||16||2||Accent feels more natural, easier, more accessible||Because I find American English easier to pronunciate and it's talked more on TV, so it's the language I hear the most.|
|Linguistic quality||4||12||Accent more aesthetic (n=15) or intelligible (n=1)||I think British English sounds better|
|Status and Competence||6 (neg)||10||Accent associated with/not associated with e.g. formality, class, intelligence, education||because British English is more classy, and sounds nicer and less vulgar I feel that American English is more relaxed, and not VERY Formal.|
|Validity||0||4||Accent is the real and original English||Well, I do think British English is and was the original English while American just “happened”.|
|Authenticity/Markedness||4||0||Accent sounds right or appropriate and not fake/Accent is more neutral||[British English is harder], and if you don't quite get it right, you would just sound stupid, and people would think you are trying hard to be something you are not.|
|Other||2||3||Audience or social attractiveness reasons||because I think American english [is] not so boring such as british|
Note: Examples are copied verbatim from the questionnaire. (neg) refers to reasons not to choose this accent as aim.
A few main patterns are identifiable in the reasons given for accent aims. First, the most common reason to aim towards an American English accent was accessibility. Two of the AmE aimers who gave Access as reason pointed out that British English nevertheless sounded better. Another two AmE aimers in this category explicitly reported TV to be the reason for accessibility. Second, the most common reason to aim towards a BrE accent was aesthetics. One of the BrE aimers in this category pointed out that AmE was easier, but a ugly language. Third, another common reason for aiming towards BrE was that this accent was associated with status and competence qualities such as education, formality or class. These associations to British English were also a common reason not to choose this accent (cf. “(neg)” in Table 16.4), but rather the less formal American English. Finally, in addition to reasons related to status and competence and linguistic quality (mostly in favour of BrE), and accessibility (mostly related to AmE), there was also a matter of markedness: British English is considered the original English, but L2 speakers of this accent might easily sound like they are trying too hard, and American English is the more neutral choice.
Table 16.5 shows the overall results for the production of the six variables that are known to distinguish GenAm and SSBE for both data sets. The production results show that for almost all the variables, the variant also found in GenAm English is favoured. The speakers are mostly rhotic, they tap or voice Intervocalic /t/, use [æ] for BATH, [ɑ] for LOT, and have a back onset for GOAT. One exception to this pattern is Post-coronal /j/, which is present in the majority of tokens. The most frequent realisation of Voiceless th (as in e.g. thing, thought) was a voiceless fricative [θ] (76%), but there was a large minority of variants pronounced with a stop [t] (17%).
|Phonological variable||Examples||GenAm variants||SSBE variants||Number of tokens|
|Variant||Data set 1||Data set 2||Variant||Data set 1||Data set 2||Data set 1||Data set 2|
|Intervocalic /t/||little, atom||[ɾ]||50%||66%||[t]||49%||32%||582||396|
|Post-coronal /j/||new, student||Ø||n/a||19%||[j]||n/a||81%||n/a||123|
Note: Other variants than those found in GenAm or SSBE have been excluded from the table. n/a = not applicable.
However, although there is a preference for GenAm variants, Table 16.5 also shows that there is considerable variation in the pronunciation of each variable; no single variable is pronounced entirely like GenAm or like SSBE. This variation existed for each individual speaker; all participants used variants found in both GenAm and SSBE, no one had an entirely native-like pronunciation. Patterns are found in this variation in the pronunciation across word-list reading and conversation speech; especially related to Non-prevocalic /r/, Intervocalic /t/, LOT, and GOAT. The word list elicited more rhotic variants, more [t], more [ɒ], and more [oʊ]. This is not a pattern of specific accents; the variants that are preferred in the word list compared to in the conversations are both SSBE and GenAm variants. Rather, these word-list preferences might be accounted for by the effect of orthography; in Norwegian, an orthographic r is always pronounced, t is never voiced, o is mostly pronounced similarly to [ɒ], and there is no back-closing diphthong with a mid-central starting position such as [əʊ].
In order to investigate the pronunciation of the phonological variables related to the participants’ desired pronunciation, production results were calculated for the three main accent aim groups, namely BrE aimer, AmE aimers and Neutral aimers. For all variables, AmE aimers produced more GenAm variants than BrE aimers and vice-versa. These differences were significant (p<0.05) for all variables except Post-coronal (j). Voiceless th was included to investigate possible patterns of L1 influence, but participants seemed to be treating this variable similarly regardless of accent aim. On average, Neutral aimers’ L2 production was in between BrE aimers and AmE aimers; they produced more GenAm variants than BrE aimers, but less than AmE aimers. However, the production results for each individual participant show that Neutral aimers do not pattern together – some use many GenAm variants, some use few. This would indicate that they are not trying to do something in between GenAm and SSBE, but rather that they are not trying to do either.
Discussion: Contributions to the english didactics field
The doctoral study presented here explored three areas of English language practices among Norwegian 17-year-old students; their attitudes towards native accents of English (responses to practice), their choices of target accent (reported practice), and their pronunciation of phonological variables (actual practice). The results formed a continuum of these three practices, which were paired with the methods used to investigate them. This continuum of results and methodologies is shown in Figure 16.1, and will be referred to below in the discussion of empirical, methodological and theoretical contributions. Quotes from the interviews will be used throughout the discussion to illustrate some of the claims, and will be presented with student codes consisting of a letter referring to Schools S and A–C, and a number referring to a student in this school.
Empirical contributions: L2 identity vs. exposure and competence
At one end of the continuum in Figure 16.1, there is a discrepancy between the participants’ responses to practice – the attitudes elicited from the evaluation tests, and their reported practice – the accent aims given in the questionnaire. While SSBE was considered the most prestigious accent in the matched-guise test, this was not the majority’s preferred accent aim. Even though speakers of SSBE were judged positively by the Norwegian adolescents, this did not necessarily mean that they wished to behave like speakers of SSBE. Native accents of English, and perhaps especially SSBE, carry with them social meanings that are transferred to their speakers. In the case of L2 speakers, the social meanings were perhaps even strengthened. The formality function assigned to SSBE was attributed to its speakers, and in the case of the L2 speakers, this function was related to school and ELT, and might signal a student who was trying too hard (cf. Table 16.4 above). GenAm might therefore have been a more preferred accent aim, seeing as it was not only more linguistically accessible due to more extensive exposure, but also more socially accessible because it was not associated with formality and school, as SSBE was.
The students’ reasoning behind accent choices showed that they could use GenAm or SSBE to project their desired identities – to show who they are (or who they are not), based on the meanings attached to these accents, which they shared with peers and which were elicited in the matched-guise test. Some students also argued that they wished to draw on the function of both native-speaker accents, because they were appropriate for different purposes:
A09 If you’re going to sit in class and impress the teacher a little then you should maybe try more towards the English – the British
S19 When we hang out with friends […] we don’t want to use the British English we try to learn at school, we would rather do what we think is cool.
Conversely, even though these L2 speakers might have agreed on the functions and meanings of SSBE and GenAm, they might still have avoided both accents as target pronunciation because they were not right for them – a native accent did not reflect who they are:
C21 If I suddenly should have started speaking British then that would just be weird because I don’t live in Great Britain, I’m not a Brit, and I’m not influenced by British culture, like, at all, so that would, like, change parts of the identity
I think many of us want to be neutral because I want to, like, be thought of as someone who actually knows the language […] I don’t want to, like, be thought of as an American or a Brit.
On the other end of the continuum (see Figure 16.1), there was a discrepancy between reported practice – accent aims, and actual practice – the realisation of phonological variables. The participants’ L2 speech was considerably influenced by GenAm pronunciation. Although there was a correlation between the reported accent aims and actual pronunciation, the extent of GenAm-influenced pronunciation vastly exceeded what would be expected from the reported accent aims. This discrepancy suggests that language use is not necessarily always an intentional choice, perhaps particularly in an L2. Other variables that might affect pronunciation are competence and exposure. For instance, some participants reported that they had “settled” for an American accent because British forms were inaccessible.
Furthermore, limited competence might in part explain why most of the participants used more than one variant per variable; they might not have been able to imitate a native accent entirely. Conversely, it could be that variation in pronunciation was deliberate: even though the participants reported a native accent as target, they might not have wished to unconditionally imitate this accent, in an attempt to avoid putting on another’s identity. This was one reason behind some of the participants reporting a “neutral” English accent aim altogether, attempting to blend some of the available resources into an “undetectable” accent:
B03 A cleaner form of English, that there isn’t anything dialect-distinctive about it […] you sort of have to find that thing in between
C02 I speak, in a way, what I’ve learnt and what I’ve picked up, a little here and a little there
Notwithstanding considerable variation for each individual speaker, what many participants have “picked up”, regardless of accent aim, seems to be mostly GenAm resources. Although sociolinguists traditionally have been reluctant to accept that television has any effect on language behaviour (Chambers, 1998), this could work differently with an L2. The results in this doctoral study strongly suggest that television is a major medium of English language exposure and a significant source of linguistic resources influencing learners’ pronunciation.
The L2 practices among the Norwegian adolescents – responses to practice, reported practice and actual practice – echo the transitional ELT community in which they are members. In a context where the status of English is hybrid, fluid, in transition from one place to a yet unknown other, speakers’ use of English, too, is characterised by hybridity, negotiation and variability. The variability is used in systematic – and meaningful – ways. This social sensibility and linguistic aptitude can only mean that speaking English is a social practice (Eckert & Rickford, 2001). As participants in social practice, these Norwegian students, regardless of the ELT context, were not primarily learners of a foreign language; they did not attempt to access the cultures and opportunities of others (as in SLA research such as e.g. Norton & Toohey, 2011). The empirical contribution of this thesis suggests that L2 use of English among these Norwegian students is not an interlanguage or an “incomplete” learner language, but that they are speakers of English as a second language, in line with Graddol’s (1997) description of this as “part of the speaker’s identity repertoire” (p. 11).
Methodological and theoretical contributions: sociolinguistic approach and mixed methods design
The main contribution related to methodology and theory was the (successful) application of sociolinguistic inquiry in an L2 context. This doctoral study shows that it makes sense to describe and analyse L2 use of English in Norway related to standard norms of English by finding and distinguishing a set of phonological features used across a group of speakers. By doing so, the study has included Norwegian adolescent L2 speakers of English in English sociolinguistics.
The discussions related to the main objective of this doctoral thesis – the social meanings of the L2 – arose from the combination of methods and pertaining results in Figure 16.1. The combination and integration of qualitative and quantitative approaches enhanced the validity of the results, as the different data sources and analyses both supported and added complexity to the findings. The mixed-methods approach thus accessed social meanings related to the ways in which Norwegian adolescents used English in a way that would not have been possible with qualitative or quantitative methods alone.
Implications for teaching English
Based on the discussions of the contributions of this doctoral thesis, it seems clear that presenting native accents of English as “correct” and asking adolescent students to imitate these in the English L2 classroom is problematic, as also suggested by scholars within a framework of English as an international language (Bex, 2008; Dürmüller, 2008; Jenkins, 2006). For some students, imitating a native speaker implies not just putting on an accent, but also putting on an identity. Instead, SSBE and GenAm are very useful as reference accents to present as examples or to which students and teachers can compare own and other people’s accents, but they should not be presented as targets that students are expected to reach. Not only is native-like attainment near impossible (Derwing & Munro, 2005), it is also unnecessary. If accent is taught in the English L2 classroom, focus should be on helping students develop an accent of English that shows who they are. This would involve listening to and identifying available (native and non-native) accents of English, reflecting on the functions and meanings they carry with them, and assessing whether they are appropriate for the students’ purposes and contexts of communication. Engaging the learners’ identities in this way can enhance learning, as suggested by SLA scholars (Menard-Warwick, 2005).
As English continues to spread globally and be appropriated locally, the diversity and hybridity of English increases (Mufwene, 2010). This view of English as L2 (and language in general) as diverse and highly variable can be presented to students in Norway, rather than presenting only native-speaker cultures. If students learn how English is used as a lingua franca, i.e. how linguistic properties and norms are co-constructed and established in order to achieve a communicative goal, they will be able to adapt their English language to different purposes, whether that is appropriating the local accent in a native-speaking country to which they move, or making themselves understood when speaking to less competent non-native speakers of English with first languages other than Norwegian. This is not just relevant in Norway, but also in other contexts where the primary goal of learning English is to use it to communicate with native and non-native speakers for a wide variety of purposes.
Suggestions for future research
Part of the rationale for the doctoral study reported here was the lack of research on the meaningful use of linguistic resources by L2 speakers. As the status of English continues to develop and the diversity of the English language and English-speaking communities continues to increase, there is still a need for research into meaningful L2 practices by non-native speakers. This study shows that there is reason to study L2 English in Norwegian ELT further using theoretical approaches and methodologies related to language attitudes and stylistic practice.
In relation to responses to practice (cf. Figure 16.1), future research should focus on attitudes towards non-native speaker pronunciation rather than native-speaker accents. Considering the deliberate avoidance of native-speaker accents among Norwegian adolescents and the patterned variation in their pronunciation that mirrors the transitional status of English in Norway, it is relevant to treat this pronunciation in itself as the object of investigation. There is some research emerging on this topic (see e.g. Beinhoff, 2013; Hendriks, van Meurs, & de Groot, 2015; Nejjari, Gerritsen, van der Haagen, & Korzilius, 2012), but in Norway this has mainly received attention in MA theses (Haukland, 2016; Hordnes, 2013). The students in this doctoral study explained the variability in their L2 production with a desire to use a neutral accent of English. Of course, there is no such thing as a neutral accent. Further research is needed into how these local English language practices are perceived by Norwegians themselves as well as by native and non-native speakers of English. What functions and meanings are associated with Norwegian-accented English?
Related to reported practice, there is a need for more research into the nuances of L2 choices. This doctoral study asked participants to report their target accent in a closed-ended question in a questionnaire, taking for granted that it would be possible to anticipate participants’ full range of potential answers. Although alternatives to this questionnaire item were developed through two data sets and an additional pilot study, the discussion has shown that accent choice is more complex than circling one of six accent alternatives. Language choices are probably best explored through more open-ended questions and preferably in interviews, with or without stimulated recall (Calderhead, 1981), so that participants’ reflections can be discussed. What kind of pronunciation do Norwegian students want to use if they do not need to conform to variety labels?
Related to actual practice, further research is needed into L2 production across situations with different audiences in order to investigate whether Norwegians actually do adapt their pronunciation to different contexts and purposes, or if this is only an intention. Furthermore, what happens to their pronunciation when they use English to interact with speakers of other first languages, i.e. when they use English as a lingua franca? Such research should consider the range of language components, not just phonological variables. In particular, the discussions in the present study call for investigations into the individual’s L2 strategies and styles. In addition to investigations into intra-individual variation (how individual speakers pronounce the same variables with different variants), there is also a need for further research on intra-variant variation (how different variants are pronounced differently) in L2 production, in order to describe the local appropriation of English in more detail. In short, what does it sound like when Norwegians speak English?
Through the application of sociolinguistic inquiry in an ELT context, this doctoral study has accessed a vast field at the intersection between sociolinguistics and English language education in dire need of investigation. Hopefully, the inferences and discussions presented here will encourage scholars and students to take on the challenge.
|1||The chapter presents the overall results of a doctoral study (Rindal, 2013) from the University of Oslo. This is an article-based thesis with three published articles (Rindal, 2010, 2014; Rindal & Piercy, 2013). The doctoral thesis in its entirety – with theoretical, methodological and empirical details – can be found here: https://www.duo.uio.no/handle/10852/65151|