Editor

JLLT edited by Thomas Tinnefeld
Journal of Linguistics and Language Teaching
Volume 6 (2015) Issue 1


A Study in the Perception of Native and Non-Native Englishes by German Learners

Julia Davydova (Mannheim, Germany)

Abstract
While advocating an integrative approach to the study of language attitudes, the present study explores perceptions of two native and two non-native varieties by German learners of English. The native varieties chosen for the elicitation of attitudes include standard British English and mainstream American English. The non-native varieties targeted in the study are Indian English, a second-language variety, and German English, a foreign-language variety. Exploiting the method mix consisting of a survey, a verbal guise test, and sociolinguistic interviews, the study brings forth converging evidence consistent with the foregoing research. More specifically, it shows that learners evaluate the standard variety of British English as the one showing high levels of prestige and status. In contrast, mainstream American English is perceived as highly socially attractive. It is argued that the social context (formal vs. informal) guides the acquisition of learners’ evaluations of different native Englishes. On a more general level, however, the native-speaker varieties receive much more favourable ratings than the non-native Englishes. This finding is indicative of “an inferiority complex” (Tan & Castelli 2013), a phenomenon whereby non-native speakers exhibit far more negative evaluations towards their own variety than native speakers would. In this situation, possible remedies are suggested.
Key words: language attitudes, Learner Englishes, L2 acquisition of linguistic perceptions, method mix

1   Introduction
Language perceptions or language attitudes are crucial in the study of language because speakers’ social evaluations have been shown to have an impact on how language is put to use in the community and how it changes over time. In his seminal work of the community of Martha’s Vineyard, William Labov demonstrated that the ongoing centralization of /ay/ and /aw/ diphthongs was correlated not only with various socio-demographic characteristics of the community members but more importantly, with their positive or negative evaluations of the local community and its traditional values (Labov 1972: 1–42). People’s attitudes furthermore play a crucial role in the formation of linguistic stereotypes, which are subsequently manifested as distinctive social meanings in the community. To give one example, native speakers of English have been shown to harbour negative feelings toward some urban English accents associated with the working class such as, for example, Birmingham English, labelling them as significantly less friendly, interesting or cool (Clark & Schleef 2010: 311). Such linguistic stereotyping is believed to play a key role in the acquisition of language evaluations (imposed norm hypothesis, Ladegaard 1998: 253). For this reason, sociolinguists have joined their efforts with those of specialists in related fields, notably applied linguistics and social psychology, in an attempt to ascertain how native English speakers evaluate different linguistic varieties and, more importantly, how these evaluations are acquired by a community of native speakers. This strand of research has been extremely prolific, having produced a plethora of studies over the past four decades (e.g. Giles 1970, Trudgill & Giles 1978, Zahn & Hopper 1985, Giles & Coupland 1991, Lippi-Green 1997). Still relatively unexplored, non-native Englishes, Learner Englishes in particular, present the researcher with an excellent opportunity to explore learners’ attitudes toward native and non-native English and the channels through which these attitudes may be transmitted in a social environment where English is not the main language of the dominant population group. Together, these issues contribute to the building of second-language acquisition and sociolinguistic theory.
That said, the current study sets out to explore attitudes towards native and non-native English by German students at the University of Mannheim (Germany). We focus on English spoken in Germany because previous studies examining this learner variety provide very detailed descriptions of the amount and types of exposure to English by secondary-school and university students (Erling 2007, Grau 2009). The other strand of research has investigated the history of contact with English as well as its major domains of use in the country (Hilgendorf 2007). Moreover, studies examining attitudes towards English in Germany are largely descriptive in nature (Erling 2007, Hilgendorf 2007), and there are to date no studies using experimental data to explore this issue.
More specifically, we focus on the differences in the perceptions of native and non-native English by German learners. The native varieties chosen for the elicitation of attitudes include standard British English (BrE) and mainstream American English (AmE). The non-native varieties targeted in the study are Indian English (IndE), a second-language variety, and German English (GerE), a foreign-language variety.[1] In this study, we chose to target IndE because it is a variety that is relatively well-known in Germany through job-related experiences and portrayal in the media. GerE was a logical option since it is a form of English with which English learners in Germany have an inherent affiliation.
The present paper is organized as follows. Firstly, an overview of existing research in Learner Englishes and learners’ perceptions towards different forms of English is provided. Having described attitudes as an object of study, we present the research questions and hypotheses, while discussing the major methodological approach employed here and commenting on its benefits. In the next step, the results of the three case studies will be presented, i.e. a survey, a verbal guise test and sociolinguistic interviews. The paper is rounded off with the discussion of the major findings and their relevance for the field of sociolinguistics and second-language acquisition.

2   Learner English and Learners' Attitudes towards Native and
     Non-Native English: State of the Art
The field of World Englishes has witnessed a steady increase in the studies exploring learner varieties of English over the past few years (Szmrecsanyi & Kortmann 2011, Davydova 2012, Edwards 2014, Kasztalska 2014). The pertinent interest in these forms of language has been mostly motivated by the need for contrastive comparisons from a typological perspective, as in Szmrecsanyi & Kortmann (2011), or comparisons with indigenized English, as in Davydova (2012) and Edwards (2014). Advances in corpus linguistics have produced a plethora of corpus-based studies exploring various types of English. As a result of that development, studies focusing on learners’ speech have gained momentum (see, for instance, Mukherjee & Hundt 2011).
In contrast, studies looking at learners’ perceptions of different forms of English are still relatively few (Ahn 2014: 196). For instance, Ladegaard (1998) and Clark & Schleef (2010) explore the perception of various native English accents by non-native speakers, whereas McKenzie (2008a, 2008b) looks at the perceptions of standard and non-standard forms of native English as well as Japanese English by Japanese learners. Several studies look at the differences in the perceptions of BrE and AmE by Chinese learners (Ng & He 2004, Hu 2005, He & Ng 2013).
Previous studies report highly positive attitudes toward both AmE and BrE in East Asia. McKenzie (2008b) paints a detailed portrait of learners’ attitudes towards English in Japan, showing that Japanese learners highly favour UK English and US English in terms of status but express “a high degree of solidarity with heavily-accented Japanese speech” (McKenzie 2008b: 75). Studies looking at learners’ attitudes towards English in the European context maintain that standard BrE is a highly prestigious variety, especially when placed against other standard and non-standard varieties of English (Dalton-Puffer, Kaltenboeck & Smit 1997, Ladegaard 1998, Clark & Schleef 2010). More specifically, in his study of English learners in Denmark, Ladegaard (1998) shows that standard BrE receives highly favourable ratings on the dimension relating to status and competence as well as aesthetic characteristics, but is actually downgraded with respect to social attractiveness. In contrast, AmE received fairly high ratings for social attractiveness compared to BrE. Ladegaard concludes that his findings indicate “obvious similarities with social stereotypes found in Anglophone contexts” (Ladegaard 1998: 259). In their study of native and Polish-born adolescents from London and Edinburgh, Clark & Schleef (2010) replicate the finding that standard BrE is the most highly evaluated variety with respect to status that nevertheless receives low ratings on the solidarity dimension. This is, again, consistent with what we know about evaluations of English varieties in native contexts. Overall then, the foregoing researched has yielded some highly interesting findings showing how learners evaluate various native forms of English. Relatively little is, however, known about how learners perceptually differentiate between native and non-native forms of English. More specifically, there are to date no studies investigating how European learners perceive the differences between native and non-native varieties of English.


3   Language Attitudes as an Object of Study

Language attitudes are psychological constructs (Garrett 2010: 20) that cannot be observed directly as one can, for instance, observe a water molecule under a microscope. They are essentially perceptions of and judgments about language or its various aspects (phonological features or accents, morphosyntactic structures, and lexicon). As such, attitudes are mental constructs.
People’s attitudes toward language have an evaluative component (e.g. This kind of language is appropriate for the use in this specific context) and an affective component (e.g. This variety sounds so cool / funny / sophisticated). The former includes people’s cognitive judgments about the amount of status and social prestige that a given linguistic variety enjoys in the community. The latter is related to empathy and is indicative of the level of social attractiveness and solidarity assigned to specific languages. Some perceptions of language can furthermore be understood as evaluative-affective judgments. These often pertain to people’s evaluations of their own selves as well as the forms of language that they think they speak and identify with.
Finally, existing research suggests that attitudes can be experienced both consciously and subconsciously by the speakers. This observation has direct repercussions for the methods through which attitudes can be, and indeed have been, studied. Conscious attitudes can be assessed directly either with the help of a questionnaire or an interview. In such cases, the researcher poses straightforward questions, asking people what they think of a variety X. Unconscious attitudes can be arguably accessed through evoking tacit associations with linguistic forms and varieties. This is achieved through the technique called verbal guise test (VGT).

4 Research Questions and Hypotheses
The present study seeks to discover if there are any differences in the way German speakers of English perceive native and non-native varieties of English including their own. This very general issue yields a set of more specific questions and hypotheses.
RQ 1:   Do German speakers perceive the differences between mainstream native-speaker varieties / accents in a fashion similar to that attested for native speakers of English?
Previous research (Ball 1983, Huygens & Vaughan 1983, Stewart et al. 1985), which tested native speakers’ attitudes towards native-speaker accents, has shown that RP speakers were rated highly on the scales covering social status and prestige, whereas AmE speakers were highly rated for social attractiveness.
H1a: German speakers of English will rate the RP accent / standard BrE highly on the scale of social status, prestige, and competence.
H1b:  In contrast, AmE can be expected to receive higher ratings for social attractiveness and solidarity in comparison to BrE.
RQ 2: Can German speakers identify native and non-native accents by region?
H2:    They can be expected to identify BrE and AmE accents quite well. They can also be expected to be able to localize their own accent. It is furthermore hypothesized that they will be able to classify South Asian / IndE accent given its presence in the media.
RQ 3:   Will their perception of and attitudes toward non-native Englishes other than GerE be different from their perception of native varieties / accents?
Previous research (Tan & Castelli 2013) has demonstrated that native speakers exhibit much higher positive attitudes towards South and Southeast Asian English accents than speakers from South and Southeast Asia who tend to negatively evaluate their own accents.
H3:    Similarly to native English speakers, non-native German speakers of English can be expected to exhibit a generally positive attitude towards South Asian English, particularly with respect to social attractiveness and solidarity but less so with respect to status, prestige, and competence.
The expected findings can be of relevance to the advertising of India-related products to the German consumer that employs the use of IndE voice effects.
RQ 4:   Are speakers’ evaluations of their own variety / accent, GerE, different from their evaluation of other non-native and native varieties?
H4:    German speakers of English are hypothesized to evaluate their own variety / accent positively on the dimension called ‘identity’.
Whether or not they will positively evaluate their own variety / accent on all other dimensions including status, competence, and solidarity is determined in the course of this study as there is no foregoing research regarding this issue.

5   Methods
In order to address the issues outlined in the foregoing section, we relied on both direct and indirect approaches to measuring attitudes (Garret 2010). More specifically, two quantitative methods were employed in this study:
(i) a survey eliciting conscious attitudes towards a given variety and
(ii) a VGT tapping into covert perceptions of speakers towards native and non-native forms of language.
Additionally, attitudes towards varieties of English in question were obtained through qualitative statements about different forms of English. These statements were elicited through sociolinguistic interviews. The method mix used was implemented as an attempt to obtain converging evidence stemming from different cognitive domains (i.e. declarative, conscious knowledge vs. non-declarative, unconscious knowledge, Percy 2012: 70). Such a procedure, in turn, allows for sound generalizations about learners’ overt and covert perceptions of native and non-native forms of English.
Secondly, on a methodological note, the employment of a method mix is crucial as it allows the researcher to look at data from complementary, yet related perspectives (see also Garrett 2010: 201). In a related line of thinking, Labov (1972a) states that sound sociolinguistic knowledge can only be obtained “by convergence of several kinds of data with complementary sources of error” (Labov 1972a: 97, cited in Schilling 2013: 66).
The following sections depict the main steps of analysis carried out in each case study. The data used in this project were obtained from students aged between 18 and 25 enrolled in Bachelor and Master Programmes in Language & Business Administration and Language & Media Communication at the University of Mannheim (Germany). Approximately one third of the entire sample population was studying EFL so as to become teachers. All these students were brought up in Germany, having come into first contact with English through a classroom context. All of them had German as their mother tongue. Five speakers included in the survey were brought up bilingually with German and Turkish.[2] The data for the survey were obtained from 94 respondents (72% of them being female and 28% male); the VGT data were elicited from 65 respondents (69% of them being female and 31% male); the qualitative data were obtained through sociolinguistic interviews from 24 speakers (70% of them being female and 30% male). By the time the data were collected, all the subjects had been exposed to the English language in an academic setting for at least ten years. Furthermore, an overwhelming majority of these students (77% in total) reported having visited an English-speaking country at least three times. Their English language proficiency can be described as upper-intermediate / advanced (B2 / C1 CEFR levels[3], respectively).

6   The Study

6.1 Case Study 1
6.1.1 The Survey
The survey elicited learners’ conscious attitudes towards BrE, AmE, IndE, and GerE on the dimensions of (i) status / prestige (evaluative judgments), (ii) social attractiveness / solidarity (affective judgments), and (iii) identity (evaluative-affective judgments). Each dimension was represented by two statements about each variety (e.g. I think X English is a high-status variety). The respondents had to assess each variety on each dimension by placing a cross somewhere on a 6-point Likert scale, as illustrated in Figure 1:

Quickly read the following statements about British English and decide to what extent you agree with each statement.

1. I think British English is a high-status variety.
 1                   2                   3                   4                  5                 6
    I strongly disagree                                                                        I strongly agree

2. I think British English is prestigious.
 1                   2                   3                   4                  5                 6
    I strongly disagree                                                                        I strongly agree

3. British English is socially attractive.
 1                   2                   3                   4                  5                 6
    I strongly disagree                                                                        I strongly agree

4. I use British English to express my solidarity with others.
 1                   2                   3                   4                  5                 6
    I strongly disagree                                                                        I strongly agree

5. British English is a form of English that I speak.
 1                   2                   3                   4                  5                 6
    I strongly disagree                                                                        I strongly agree

6. British English is a form of English that I strongly identify myself with.
 1                   2                   3                   4                  5                 6
    I strongly disagree                                                                        I strongly agree
Figure 1:  Instructions to participants and the Likert scale employed in the survey
The collected data were then fed into the SPSS spreadsheet, checked for outliers, tested for normality (Wilk Shapiro test) and homogeneity of variance (Mauchly’s test of sphericity). After that, it was subjected to repeated measures analyses of variance (ANOVA), which compared judgment means for each statement across the four varieties studied here.

6.1.2 Results: Emerging Trends and Patterns
First and foremost, both native-speaker varieties received higher ratings on all three dimensions studied here, and these differences are statistically significant at p = .000, as reported in Table 1:
Statement
Mean Scores
F-value
D.F.
P-value

BrE
AmE
GerE
IndE



Dimension: status / prestige
1. I think X is a high-status variety
4.50
3.62
2.97
2.21
70,101
2.8, 254.8
.000
2. I think X is prestigious
4.36
3.39
2.78
2.00
89.126
2.9, 267.9
.000
Dimension: solidarity / social attractiveness
3. X is socially attractive
4.00
4.39
2.87
2.06
66.598
2.9, 268.2
.000
4. I use X to express my solidarity with others
2.30
3.58
2.56
1.29
46.075
2.7, 252.0
.000
Dimension: identity
5. X is an English that I speak
2.78
4.25
3.22
1.14
58.545
2.2, 202.2
.000
6. X is an English that I strongly identify with
2.55
3.75
2.37
1.47
45.820
1.9, 181.2
.000
Table 1:  Repeated measures ANOVAs of survey data comparing mean evaluations of two native and two non-native varieties by German students (N 94)

A closer examination of evaluation means further reveals some noteworthy patterns. BrE receives higher scores on the dimension called status / prestige in comparison to AmE; the post-hoc comparisons (with Bonferroni adjustment) show that these differences are significant (mean difference 0.88, CI 95% 0.45 to 1.30, p = 0.000 for statement 1 above, and mean difference 0.97, CI 95% 0.59 to 1.33, p = 0.000 for statement 2). This result substantiates Hypothesis 1a, stating that BrE is a variety that enjoys more status and social prestige.
In contrast, AmE receives higher ratings on the solidarity / social attractiveness dimension when compared to its British counterpart; the post-hoc comparisons reveal that these differences in means are significant at least with respect to statement 4 above (mean difference .038, CI 95% 0.16 to 0.93, p = 0.361 for statement 3, and mean difference 1.28, CI 95% 0.70 to 1.85, p = 0.000 for statement 4). This finding adds weight to Hypothesis 1b, suggesting that AmE appears to be more socially attractive when compared to BrE.
Perhaps one of the most remarkable findings stemming from the survey is the higher ratings received for AmE on the dimension labelled identity. Results unequivocally show that German students believe that the form of English that they speak and strongly identify with is AmE. The post-hoc comparisons show that the mean differences (i.e. AmE vs. another variety) are significant (p < 0.0005) for both statements. Thus, Hypothesis 4 could not be substantiated, as GerE does not seem to be the form of English that German students aspire to speak or strongly identify with. The low prestige of GerE is attributable to the negative associations evoked through puristic public discourse, advising against using typically German expressions in English and strongly encouraging following the native-speaker norms instead (e.g. Littger 2014 and Meyerhöfer 2014).
Finally, IndE is a variety that receives the lowest ratings on all dimensions and the mean differences (i.e. IndE vs. another variety) are significant (p < 0.0005) for all six statements. Hypothesis 3 could not be empirically substantiated. Overall, IndE appears to have little prestige and social attractiveness for German learners. Interestingly, these results are consonant with those reported in Bernaisch for Sri Lanka where IndE “generally comes last in the attitudinal ranking” (Bernaisch (2012: 286). This similarity in attitudes is even more striking, given that the German learners tested in this study judged a geographically and socially remote variety of English, whereas Sri Lankans have a much more immediate exposure to IndE.
In sum, the survey data yield four main conclusions:
(i) BrE is judged as having significantly more status,
(ii) AmE is more socially attractive,
(iii) GerE is not a form of English that German speakers aspire to speak or strongly identify with, and
(iv) IndE has the least prestige and social attractiveness in Germany.
However, one might wonder if these claims will still stand once we attempt to tap into speakers' unconscious beliefs about specific forms of English. These unconscious beliefs are particularly important to tap into because they are free from the social desirability bias, which is sometimes contained in straightforward answers (Garrett 2010: 44–45). In order to tackle this issue, a VGT study was carried out.

6.2 Case Study 2
6.2.1 Verbal Guise Test (VGT)
Developed in the 1960s by a team of social psychologists from Canada, the VGT approach is a fairly subtle technique that explores people’s attitudes to language without placing overt emphasis on the target of the investigation, i.e. English varieties or accents. In this case study, students had to listen to recordings of a short text that was read out loud by native speakers of standard BrE, mainstream AmE, educated GerE, and educated IndE. The recordings were obtained from the website of the International Dialects of English Archive, IDEA (http://www.dialects-archive.com/; accessed on June 1, 2015). The text was produced at roughly the same speech rate by each speaker. The speakers were matched for demographic characteristics (male, middle-aged, and middle-class). We also controlled, as far as possible, for the speakers’ voice qualities: The speakers chosen for the VGT experiment were tenors with a light timbre.
Perceptions of four accents (standard British, mainstream American, German, and Indian) were elicited indirectly through a semantic differential scale tapping into different attitudinal dimensions (Clark & Schleef 2010: 310, Garrett 2010: 66). As  underlying mental concepts can be represented by different semantic features in different communities, we needed to establish the traits that would be meaningful to the judges in the study. Students from the same academic community were asked to describe the four guises in their own words. We also asked the pre-study judges to provide a list of adjectives that closely described the German character in order to tap into the concept of German identity. We then selected the most frequently occurring lexical items, carefully checking them against the list of items tested in the previous studies.
In groups and in friendship pairs, the judges were then exposed to each guise once and were instructed to rate each speaker in terms of twelve features (adjectival pairs) on the scale consisting of 100 dashes. In contrast to the more traditional 5-, 6-, 7- or 9- point Likert scales, this technique arguably allows for a more differentiated analysis of such complex phenomena as covert attitudes (Clark & Shleef 2010: 305).  The VGT task is illustrated in Figure 2:
Figure 2:  Instructions to participants and the semantic differential scale employed in the VGT
The collected data were entered into an SPSS spreadsheet, where they were checked for outliers. The data were then subjected to the principal component analysis (PCA), a data reduction technique which helped to identify the correlational relationships between individual features and, thus, to establish the major components, the so-called super-variables, in the data. The dimensions of attitudes were elicited through a combination of two statistical procedures: (i) a scree test and (ii) an items-on-factor technique (McCroskey & Young 2006: 381). As illustrated in Table 2, the PCA revealed the presence of four components accounting for 71.27% of the variance, which were labelled status and competence, social attractiveness, identity, and superiority. Within these main domains, individual features cluster together in statistically significant ways; although given the relatively small dataset, these dimensions should be understood in terms of “patterns and tendencies rather than absolutes” (also Clark and Schleef 2010: 308).

Feature
Component 1
Status and competence
Component 2
Social attractiveness
Component 3
Identity
Component 4
Superiority
1. reliable
.818
-.147
-.112
.012
2. trustworthy
.812
-.069
-.074
.185
3. competent
.806
-.165
-.166
.262
4. educated
.778
.059
.200
.176
5. disciplined
.735
.258
-.030
-.165
6. intelligent
.642
-.048
.360
.325
7. posh
.045
-.082
-.011
.938
8. relaxed
.118
.826
-.079
-.048
9. friendly
.405
.714
-.072
.078
10. cool
.101
.591
-.494
-.051
11. serious
.246
-.552
.527
.132
12. unemotional
.-112
-.115
.914
-.070
Eigenvalue
4.4
2.4
1.05
0.78
% of variance explained
36.31
19.63
8.81
6.52

   Table 2:  Results of Principal Component Analysis 
                      (rotated component matrix, varimax rotation)
Table 2 reveals that the features reliable, trustworthy, competent, educated, disciplined, and intelligent strongly load on status and competence, whereas the feature posh shows a very close association with superiority. In contrast, the features relaxed, friendly, and cool strongly load on social attractiveness. This is consistent with previous findings involving evaluations of native and non-native speakers of English. Finally, the features serious and unemotional pattern together, giving rise to the dimension labelled identity.
In the next step, we tested for normality (Wilk-Shapiro test) and homogeneity of variance (Mauchly’s test of sphericity) and compared the mean evaluations of each feature across four accents (twelve calculations in total) by carrying out repeated measures ANOVA. The Bonferroni post-hoc tests were further employed in order to determine which specific contrasts the spotted differences were attributable to.

6.2.2 Results: Emerging Trends and Patterns
Generally, as shown in Table 3, native speakers got much more favourable ratings for all the features that were taken to represent the dimension status and competence and social attractiveness. The repeated measures ANOVA showed a main effect for all these features (nine in total). BrE furthermore received the highest rating in terms of superiority. Taken to represent the dimension identity, the features serious / funny, emotional / unemotional show diverse patterns. Overall, Hypothesis 4 could not be substantiated.

Feature
Mean scores
F-value
D.F.
P-value
100
0
BrE
AmE
GerE
IndE



Dimension: status and competence
reliable
unreliable
73.40
64.25
54.35
44.94
23.193
2.7, 169.8
.000
trustworthy
untrustworthy
70.97
66.23
55.22
43.86
22.407
2.7, 175.4
.000
competent
incompetent
77.82
63.45
39.71
35.34
71.410
2.5, 163.5
.000
educated
uneducated
78.69
61.03
51.69
41.68
41.957
2.6, 166.8
.000
disciplined
undisciplined
70.88
57.45
60.09
51.16
8.282
2.7, 170.5
.000
intelligent
unintelligent
71.48
64.51
50.87
45.89
30.158
2.7, 171.2
.000
Dimension: social attractiveness
relaxed
tense
60.12
68.14
30.60
47.43
26.821
2.7, 172.8
.000
friendly
unfriendly
70.60
72.08
48.97
51.89
18.719
2.7, 176.2
.000
cool
not cool
51.40
66.83
27.77
33.63
38.39
2.6, 169.1
.000
Dimension: identity
serious
funny
65.52
44.23
74.05
61.70
17.802
2.5, 159.5
.000
unemotional
emotional
61.45
44.58
72.05
79.56
27.018
2.4, 149.4
.000
Dimension: superiority
posh
common
41.05
33.09
33.48
28.51
3.225
2.7, 175.6
.028

   Table 3:  Repeated measures ANOVAs of verbal guise data comparing mean evaluations of two native and two non-native accents by German students (N 65)

Results reported in Table 3 furthermore point to clearly discernible patterns:
(i)   BrE receives higher ratings on the dimension status and competence and superiority when compared to AmE,
(ii) AmE receives higher ratings on the dimension social attractiveness when compared to BrE,
(iii) GerE receives higher ratings on the dimension status and competence when compared to IndE, and finally
(iv) IndE receives slightly higher ratings on the dimension social attractiveness when compared to GerE.
With respect to the statistical significance of these data, we report results of the Bonferroni post-hoc tests, focusing on the four trends outlined above.
As shown in Table 4, the BrE guise was rated as more reliable (mean difference 9.14, CI 95% 1.01 to 17.26, p = 0.019) in comparison to all the other guises. It is also perceived as somewhat more trustworthy although this difference does not come up as significant in the post-hoc comparisons (mean difference 4.73, CI 95% 4.43 to 13.90, p = 0.985). The BrE speaker is furthermore perceived as more competent (mean difference 14.36, CI 95% 6.63 to 22.10, p = 0.000) and more educated (mean difference 17.66, CI 95% 9.76 to 25.56, p = 0.000), more disciplined (mean difference 13.42, CI 95% 4.09 to 22.75, p = .001) and more intelligent (mean difference 6.96, CI 95% 0.31 to 13.62, p = 0.035) in comparison to AmE. Overall, Hypothesis 1a is substantiated by the results of the post-hoc analyses of our data.

Feature
Mean Difference
Confidence interval 95%
Std. Error
P-value
Dimension: Status and competence
Lower bound
Upper bound


BrE as more
reliable
9.14
1.01
17.26
2.98
.019
trustworthy
4.73
4.43
13.90
3.36
.985
competent
14.36
6.63
22.10
2.84
.000
educated
17.66
9.76
25.56
2.90
.000
disciplined
13.42
4.09
22.75
3.43
.001
intelligent
6.96
0.31
13.62
2.44
.035
Table 4: The Bonferroni post-hoc tests comparing evaluations of the British accent and the          American accent on the dimension status and competence by German students (N 65)

BrE was furthermore higher assessed on the dimension superiority in comparison to AmE, although, as reported in Table 5, this difference is not statistically significant.

Feature
Mean Difference
Confidence interval 95%
Std. Error
P-value
Dimension: Superiority
Lower bound
Upper bound


BrE as more
posh
4.9
5.00
14.9
3.66
1.00

Table 5:  The Bonferroni post-hoc tests comparing evaluations of the British accent and the    American accent on the dimension superiority by German students (N 65)

Importantly, the foregoing results are very much in line with those reported in Bernaisch (2012) and Ladegaard (1998). The former is a survey tapping into attitudes of Sri Lankan English speakers and contrasting four English varieties, i.e. BrE, AmE, IndE, and Sri Lankan English. It reveals that BrE is most highly evaluated in terms of status and competence as it receives the highest ratings for features such as prestige, educatedness, smartness, flawlessness, and sophistication in comparison to AmE (and the other two varieties). Ladegaard (1998) looks at tacit evaluations of standard and non-standard English accents by Danish learners, showing that RP “received the most favourable evaluation on all dimensions relating to status and competence” (Ladegaard 1998: 258).
Preliminary comparisons also show that the AmE speaker receives higher ratings on the dimension social attractiveness when compared to BrE, as shown in Table 3. Table 6 reveals further details demonstrating that American accent is perceived as cooler when compared to the standard British accent (mean difference 15.43, CI 95% 4.92 to 25.94, p = 0.001). It is also perceived as more relaxed and somewhat friendlier than the BrE accent although these differences are not significant in the post-hoc comparisons (mean difference 8.01, CI 95% 2.56 to 18.59, p = 0.259 for relaxed and mean difference 1.48, CI 95% 7.40 to 10.35, p = 1.00 for friendly). Although the data show a clearly discernible trend (AmE is evaluated as more socially attractive than BrE), Hypothesis 1b could be substantiated only partially by this dataset.

Feature
Mean Difference
Confidence interval 95%
Std. Error
P-value
Dimension: Social attractiveness
Lower bound
Upper bound


AmE as more
relaxed
8.01
2.56
18.59
3.89
.259
friendly
1.48
7.40
10.35
3.26
1.00
cool
15.43
4.92
25.94
3.86
.001

Table 6:  The Bonferroni post-hoc tests comparing evaluations of the British accent and the American accent on the dimension social attractiveness by German students (N 65)
Interestingly, standard British accent is downgraded on the dimension social attractiveness in comparison to other varieties including AmE in Ladegaard’s (1998) study. In Bernaisch’s (2012) study, BrE is evaluated as less friendly than AmE and Sri Lankan English.
On the other hand, the German speaker received slightly better ratings for the adjectival pairs representing status and competence when placed against the Indian guise. The German guise was consistently rated as somewhat more reliable, trustworthy, and competent than the Indian guise. This difference is significant for the feature trustworthy, as indicated by the post-hoc comparisons (mean difference 11.36, CI 95% 2.33 to 20.36, p = 0.006). The differences are, however, not significant for the feature competent (mean difference 4.36, CI 95% 3.6 to 12.3, p = 0.84) and for the feature reliable (mean difference 9.41, CI 95% 1.22 to 20, p = 0.113). Furthermore, the German speaker was rated as more educated, and this difference is weakly significant (mean difference 10.01, CI 95% 0.47 to 20.50, p = 0.069). The German speaker is also rated as more disciplined and intelligent than the South Asian speaker, although the differences are not significant (mean difference 8.938, CI 95% 1.49 to 19.36, p = 0.136 for disciplined and mean difference 4.9, CI 95% 3.4 to 13.39, p = 0.66 for intelligent). A summary of evaluations of non-native speaker accents along the dimension status and competence is given in Table 7.[4]

Feature
Mean Difference
Confidence interval 95%
Std. Error
P-value
Dimension: Status and competence
Lower bound
Upper bound


GerE as more
reliable
9.41
1.22
20.05
3.90
.113
trustworthy
11.36
2.33
20.36
3.31
.006
competent
4.36
3.6
12.3
2.92
.844
educated
10.01
0.47
20.50
3.85
.069
disciplined
8.94
1.49
19.36
3.82
.136
intelligent
4.9
3.4
13.39
3.08
.664
Table 7:  The Bonferroni post-hoc tests comparing evaluations of the German accent and             the Indian accent on the dimension status and competence by German students (N 65)

The foregoing analyses thus reveal a very suggestive pattern, with the German guise being placed slightly above the Indian guise on the dimension status and competence. Overall though, both German and Indian speakers are perceived as having less status and competence when placed against native English speakers.
A highly interesting pattern gleaned through the comparisons of tacit perceptions is that the Indian speaker received somewhat better means for social attractiveness in comparison to the German speaker. As reported in Table 8, the post-hoc comparisons show that the Indian speaker was rated as significantly more relaxed (mean difference 16.8, CI 95%, 5.4 to 28.2, p = 0.001). The Indian speaker was also rated as a bit friendlier and a bit cooler than the German speaker, although these differences are not significant (mean difference 2.9, CI 95% 8.8 to 14.6, p = 1.000 for friendly and mean difference 5.9, CI 95% 2.9 to 14.7, p = 0.45 for cool). Thus, the VGT data have yielded at least partial evidence in favour of Hypothesis 3, stating that South Asian English should receive more positive ratings for social attractiveness. This claim, however, only applies to non-native speakers’ comparisons.

Feature
Mean Difference
Confidence interval 95%
Std. Error
P-value
Dimension: Social attractiveness
Lower bound
Upper bound


IndE as more
relaxed
16.8
5.4
28.2
4.18
.001
friendly
2.9
8.8
14.6
4.30
1.00
cool
5.9
2.9
14.7
3.25
.455
Table 8:  The Bonferroni post-hoc tests comparing evaluations of the German accent and the Indian accent on the dimension social attractiveness by German students (N 65)

What is intriguing is that these findings actually parallel those reported in Bernaisch (2012), who has been able to show that IndE “obtains higher ratings for humbleness and friendliness” (Bernaisch 2012: 286), two qualities associated with the expression of solidarity and, by this token, highly socially attractive.
That said, the foregoing comparisons of non-native guises yield two hypotheses that need to be subjected to further empirical testing through the VGT method. Non-native speakers can be expected to rate their own accents more positively in terms of status and competence in comparison to other non-native accents. In contrast, they can also be expected to give the other non-native variety more positive affective ratings, as an attempt to express their solidarity with other non-native speakers and forms of language that they speak. We will address this issue in the discussion of sociolinguistic data.

6.2.3 Variety Recognition Task
After hearing each guise, respondents were also asked to identify the region where they thought the speaker came from. The task was an open-ended, rather than a forced-choice, question and the informants were instructed to be as specific as possible. Each respondent then tried to localize each speaker from the verbal guise. Evaluating this assignment was no straightforward task. The speaker taken to represent the standard variety of BrE was born in Surrey, a county in the southeast of England and spent the 25 years of his adult life in the city of Brighton, Sussex. The speaker taken to represent mainstream AmE was born in Virginia, spent three years in Washington D.C., 14 years in Philadelphia, and 10 years in New York. The speaker from the German guise was raised in Münster but had lived in Hamburg for more than 10 years by the time of the interview. The speaker from the Indian guise was raised in Mumbai but had lived in the United States for four years by the time of the recording.
While evaluating respondents’ answers to the open-end question, we adopted a scale of 0-3, which reflects the level of correctness in terms of region identification. Table 9 further illustrates the methods used in this study to quantify correct identification.

Guise example
Response
Level of identification
Points awarded
British guise
America, Brazil
Incorrect country
0

Britain
Correct but unspecific
1

England
Correct country
2

Southeast of England, London
Correct region
3
American guise
England
Incorrect country
0

US
Correct but unspecific
1

US, East Coast
Correct region
2

US, New York
Correct local area
3
German guise
Brazil, India
Incorrect country
0

Non-native speaker
Correct but unspecific
1

Northern Europe
Correct region
2

German
Correct country
3
Indian guise
Africa
Incorrect country
0

Non-native speaker
Correct but unspecific
1

Asia
Correct region
2

India
Correct country
3
Table 9:  Measuring the correct identification of guises
Table 10 shows the identification values for all four guises. Overall, the task yielded 260 responses (65 responses for each guise). The recognition rates for all varieties are generally very high. Thus, the overwhelming majority correctly identified the British guise as 'British' by indicating that the speaker’s place of origin was either Britain (44.62%) or England (32.31%). Only 10.77% of respondents (7 respondents in total) were able to correctly identify the speaker as coming from southeast England.


BrE
AmE
GE
IE
Value
count
%
count
%
count
%
count
%
0
8
12.30
4
6.15
1
1.54
4
6.15
1
29
44.62
51
78.46
1
1.54
1
1.54
2
21
32.31
9
13.85
3
4.62
11
16.92
3
7
10.77
1
1.54
60
92.31
48
73.85
No response
-

-

-

1
1.54

Table 10:  Identification values for four guises among 65 respondents participating in the VTG study

As for the American guise, 78.46% of respondents correctly believed that the speaker came from the US and 13.85% could furthermore localize the accent to the East Coast area. The non-native guises did not present German learners of English with any specific difficulties as most of the informants correctly identified the country of origin for both German and Indian speakers. Overall, Hypothesis 2 could be substantiated.
In sum, the speakers correctly identified the variety of a particular guise, and this type of evidence suggests that overall, the learners associated a specific guise with a specific country, i.e. Great Britain / England, the US, Germany, and India, while providing evaluative and affective judgments of the guise.

6.3 Case Study 3
6.3.1 Sociolinguistic Data
The main goal of this case study is to provide qualitative substantiation for the quantitative results, drawing on two complementary approaches to the study of language attitudes. We also attempt to find explanations for the patterns detected in the survey and experimental data. The clear-cut advantage of this approach is that the explanations that the researcher provides while working with sociolinguistic data are deeply rooted in the personal experiences of the community that speaks the variety under study. In other words, it is always context-related.
The qualitative data were collected with the help of so-called sociolinguistic interviews, a data-collection technique that aims at eliciting in-depth discussions of specific questions grouped into modules. These sociolinguistic interviews are part of the Mannheim Corpus of German English (MaCGE).[5] For this study, all the questions clustered around attitudes towards English. The interview schedule included modules on the role of English in the world, English language awareness, experiences with English, uses of English, students’ affiliations with different Englishes, and their own linguistic identity. The speakers were asked to comment on the role of and attitudes toward English in Germany, and on the differences in their perception of BrE and AmE. They were finally asked if they thought non-native speaker varieties such as IndE or GerE were perceived differently than native Englishes. All 24 speakers, recruited for this case study, were interviewed in pairs by two female student assistants, each interview lasting approximately one hour. The recorded data yield 12 hours of spontaneous speech material.
The following conclusions could be drawn from the examination of quantitative data:
(i) native Englishes have far better evaluations on all dimensions than non-native Englishes,
(ii) BrE is revealed as a highly prestigious variety in terms of its social status,
(iii) AmE is socially attractive to German learners of English,
(iv) IndE often evokes stigmatized connotations, and
(v) GerE is a variety about which German learners apparently have ambivalent feelings, as revealed by neither very high nor particularly low mean scores.
Relying on the work with sociolinguistic interviews, we explain in what follows what motivates the patterns obtained through quantitative data.
In their discussions, speakers refer to BrE as a sophisticated, educated, intelligent, posh, and a highly prestigious form of English. Some speakers believe this is largely due to the English values and traditions, and their literature. This is illustrated in (1) through (6):
(1) MaCGE/GE002: But British English is considered to be this sophisticated (,) better in inverted commas than the American English or any other types or varieties of English (.)
(2) MaCGE/GE001: I think it’s also linked to literature (,) we think that British English is more literal language (,) and we often connect it with kind of literature like Victorian literature (,) very eloquent literature and for American we think […] American English developed so late we don’t have a culture around it (.)
(3) MaCGE/GE004: But I still believe that British English has its prestige (,) it’s a very high prestige (,) MaCGE/GE003: it’s sophisticated and posh (,) MaCGE/GE: 004: Yeah, definitely (,) definitely (!)
(4) MaCGE/GE020: British English is more official (,) it sounds like they are saying something important (,) […]
(5)            MaCGE/GE018: […] the British English I learned is the Oxford English (,) which is really like (,) well the perfect English (,) sort of […]
(6)            MaCGE/GE008: England as a society also has quite good standing in Germany with its values and traditions (.)
AmE is in turn described as cool, laid-back, and easy-going, partly because this is how informants experienced the Americans in face-to-face communication, and partly because, according to the informants, this is the image of the US and American English projected through the media. Excerpts from discussions of AmE are presented in (7) through (12).
(7) MaCGE/GE103: Americans (,) usually easy-going (,) like, ‘Yeah, we can do that!’ and like always really nice (,) it’s like their way of treating people (.)
(8) MaCGE/GE006: What it stands for the country behind (.) America is still perceived maybe as the new world with countless (,) numerous opportunities (,) it’s got everything and so it’s the cool country (,) the cooler language (.)
(9) MaCGE/GE016: When I was a little girl I thought like, ‘Wow! You have to go there. All the celebrities live there. And you are gonna be famous when you’re there and whatever.’
(10) MaCGE/GE015: I think also that young people always think that America’s cooler than Britain is (.) And that’s because (,) I don’t know (,) how it’s portrayed across media and everything (.) Because of all the celebrities (,) because of all (,) if you ask people they probably think that they are all from America (.)
(11) MaCGE/GE019: American English is more socially attractive (,) I would say (.) MaCGE/GE020: I agree and British English is when you want to impress somebody (.)
(12) MaCGE/GE023: And then of course the American accent is more seen like the cool thing because the cool movies (,) cool music (,) and everything comes from the US mostly (.) So I think it’s true that it’s the cool English […]
Interestingly, while discussing IndE, informants almost universally reported that they perceived it as a funny form of English. The reason why it is perceived as funny is the portrayal of IndE in the global media, as testimonies in (13) through (15) reveal.
(13) MaCGE/GE006: Indian English sounds funny (,) Interviewer: Why (?) MaCGE/GE005: Because it’s funny (laughs) MaCGE/GE006: The Germans do not perceive it as one of the English languages (,) it’s either American or British (.)
(14) MaCGE/GE008: Especially concerning Indian English (,) it is always being made fun of (.) Interviewer: Why (?) MaCGE/GE008: I am not sure about that (,) because it sounds funny to most of the people and there are examples in pop culture (,) for example this guy from Big Bang Theory […] they are always presented as funny figures (,) they are not taken seriously (.)
(15) MaCGE/GE016: I mean we make a lot of jokes about like (,) what is his name (,) Kajel Lana (,) has this role being Indian person (,) raji (?) or I don’t know (,) so he speaks funny and then when we hear it somewhere else (,) we kind of reflect and associate […] with funny things so (…)
The above examples demonstrate that the fairly negative evaluations of IndE by the German learners may indeed be a direct result of how this form of English is depicted in the global media. That said, some learners seem to realize that these representations are mere stereotypes and are more careful in their judgments, as in (16) and (17):
(16) MaCGE/GE015: I don’t know but I would say Indian English is funny (,) basically (.) I don’t think we would discriminate them (,) MaCGE/GE016: No (,) no (,) no (!) MaCGE/GE015: that they speak Indian English (,) we would just maybe smile a little more (.) and it’s just really different from anything else (.)
(17) MaCGE/GE018: it’s again by watching the Big Bang Theory or the Simpsons in English with Indian people which is (,) I mean it’s stereotypically and everything but their English is just (,) it’s a funny way of pronouncing words (,) this is how it is perceived or it’s how I perceive it (.) […] there are so many representations of stereotypes of who these are (...)
On a more general level however, the interview data show that non-native Englishes, including their own variety, are generally perceived as less correct Englishes when compared to the target varieties of English.
(18) MaCGE/GE019: Indian English is just another kind of accent (,) it’s just like German English (,) it’s nothing to aim at (,) so I don’t know that makes a difference probably (,) yeah (.) It’s not the ideal (,) I mean […]
The learners lament the fact that GerE is nothing to emulate, and believe that their acquisition of English should be guided by the norms laid out by native Englishes. At the same time, some informants believe that their English is more correct, i.e. closer to mainstream English than, for instance, IndE or other forms of non-native English spoken in Asia and Africa, while at the same time acknowledging that IndE is nice and there are speakers of IndE who speak very fluent, standard-like English.
(19) MaCGE/GE004: Yes (,) I also realise that Indian English has some funny timberings and I also laugh (chuckles) […] but nevertheless they speak fluently and this is the most important thing I think when you speak English and that’s also one of the things that I admire when non-native speakers speak in English (.)
(20) MaCGE/GE007: I guess it’s because we try to put a plane somewhere else and say, ‘Okay, our English is much better than the Indian English but Indian English is as good as ours.’ MaCGE/GE008: […] We consider ourselves better in assimilating to the English standards than Indians (.) MaCGE/GE007: That’s the same principle ‘we and the other’ again (...)
These qualitative evaluations echo the VGT results which placed the German speaker above the Indian speaker in terms of status and competence, but made the Indian speaker look socially a bit more attractive when contrasted with the German guise (Table 3). Overall, while providing qualitative judgments of non-native Englishes in sociolinguistic interviews, German learners seem, both consciously and unconsciously, to navigate between the two push-pull forces, their desire to be portrayed as speaking the correct, standard-like English on the one hand (they are learners, after all, and would like to learn what, in their view, is the right kind of English) and their wish to express solidarity with other non-native speakers.

7   Discussion and Conclusions
The overarching finding of this study is that German learners generally harbor far more positive attitudes towards native Englishes than towards non-native English varieties. This finding has two implications. Firstly, it provides empirical substantiation for the claim that GerE is a variety with exornomative or external-target orientation (Schneider 2003 and 2007 for further details). While constructing their linguistic identities, learners try to align themselves with the linguistic norms made available by the native-speaker communities, notably Great Britain and the United States, and promoted through the language learning industry. As these forms of language function as role models in educational contexts, learners’ attitudes toward these forms of English are steered and reinforced by teachers. Secondly, the finding is also indicative of the existence of “an inferiority complex” exhibited by non-native speakers over their own varieties of English (Tan & Castelli 2013). This phenomenon is by no means new and has in fact been commented on in the previous literature. In their study of attitudes to Singapore English, Tan & Castelli (2013) report that
the judgments of respondents from South-East and East Asia are often more negative than those of English speakers of Inner Circle varieties. (Tan & Castelli 2013: 177)
An important question at this point is what can be done to help learners of the English language to overcome their inferiority complex in order to fully appreciate the diversity of the English language and, more importantly, their own form of English? How can we help them transcend this very limited view that non-native Englishes are mere surrogates, “less correct” linguistic varieties and embrace a more sustainable outlook on the English-speaking world?
The most straightforward answer is to do so by helping them to learn more about different Englishes as they are encountered all over the world. It is at this point that our role as professionals who are committed to the study of language comes to the fore. It is our task as linguists and also as university teachers to increase students’ linguistic awareness of diverse forms of English by designing and teaching courses, organising workshops on World Englishes, organising fieldwork trips and excursions to countries where English is spoken as a second language, introducing the topic of World Englishes as part of introduction-to-language courses (Hazen 2014: 383–384) and, even more importantly, by devising teaching materials for students with minimal or no prior knowledge of linguistics. Over the past few years, attempts have been made to fill this gap on the market of English textbooks. Siemund et al.’s (2012) The Amazing World of Englishes. A Practical Introduction, for example, is an exercise-based introduction into the field of World Englishes. It focuses not only on the theory of World Englishes but actually provides a plethora of activities including reading, listening, and viewing comprehension, fostering learners’ metalinguistic skills. The textbook has been designed in this way to meet with the expectations of “the generation of on-line, video-oriented readers” (Proshina 2013: 4). Seargeant’s (2012) Exploring world Englishes: Language in a Global Context is another notable example. It is a textbook providing a comprehensive overview of the theoretical research on World Englishes and can be used as a starting point for students’ individual projects tackling English-language diversity.
The main quantitative and qualitative findings of this study resonate well with those reported for native speakers of English (Ball 1983, Stewart et al. 1985) and also for English learners (Ladegaard 1998, Clark & Schleef 2010). The evidence obtained through the method mix reveals that German speakers perceive standard BrE as a high-status variety and mainstream AmE as a socially attractive one. Given these results, the crucial question is how those distinctive evaluations of the two mainstream varieties are acquired by German learners. More importantly, how do social perceptions of different Englishes come about in an environment that does not feature English as the dominant language of communication?
To begin with, qualitative data suggest that learners’ ability to reflect on different English varieties is grounded in their personal experiences with respective cultures and in media-transmitted stereotypes rather than in their metalinguistic knowledge of the varieties in question. This observation is consistent with the Garrett’s (2010: 22) position that attitudes are learned through personal experience and the respective social environment, including the media.
Moreover, BrE is taught as the first and often the only target variety at German schools and this is apparently how speakers learn to perceive this variety as more intelligent, competent, and educated than other native Englishes. Knowledge about the United States and AmE is reportedly acquired through exposure to the media and stays abroad. On the one hand, informants claim that they perceive the Americans they have met as easy-going, nice, and friendly. On the other hand, many informants believe that they think of AmE as a cool variety simply because this is how it is portrayed in the media. In other words, positive affective evaluations of AmE are acquired in more informal settings which might explain why this variety is perceived as more socially attractive when compared to standard BrE. The implication of this finding is that the social context, i.e. formal vs. informal settings, might indeed govern the acquisition of speakers’ evaluations of distinctive forms of English or any other language. By this token, a variety acquired in highly formal settings such as classroom instruction is more likely to evoke the feelings of respect, if not reverence, thereby securing its position as a high-status, prestigious variety. In contrast, varieties experienced through more informal settings such as face-to-face interactions are far more likely to establish themselves as more socially attractive forms of English in the mind of a second-language learner. Two major caveats need to be spelt out here. Firstly, the outlined generalization applies to native-speaker contrasts only (for instance, BrE against AmE; AmE against Australian English, and so on). Secondly, this generalization should not be taken for granted but rather be subjected to further empirical testing as it highlights the role of the social context in which a given variety is acquired as an independent variable with a potential impact on learners’ evaluations of different linguistic varieties. If accumulated systematically, resulting evidence becomes highly relevant for construction of the sociolinguistic theory of second language acquisition. For example, in this study, AmE was downgraded for status, prestige and competence as compared to BrE. The question, however, is, whether the pattern will still hold once we consider a group of learners who are exposed to AmE as the main and only target variety in a classroom setting from the first days of schooling?
Other factors which arguably affect the acquisition of varietal evaluations and are, thus, worth investigating include age and the amount of exposure to a given variety in a given context. Given what we already know about the acquisition of a second language, it is not unreasonable to suggest that early and prolonged exposure to a specific form of English is likely to result in more permanent and also more coherent attitudes towards a given variety.
On a methodological note, the study has shown that language attitudes call for an integrated programme (Garrett 2010: 201) because only a multiple-method approach allows for a comprehensive analysis of speakers’ conscious and unconscious evaluations of linguistic varieties. The term comprehensive analysis entails working with data that will hopefully yield reinforcing evidence from which sound generalisations about attitudes can be drawn. When implemented properly, different methods can complement each other, painting a richer picture of language attitudes in a specific community, and will slowly but surely foster our understanding of the non-native forms of English.

References
Ahn, Hyejeong (2014) Teachers’ attitudes towards Korean English in South Korea. In: World Englishes 33(2): 195–222.
Ball, Peter (1983) Stereotypes of Anglo-Saxon and non-Anglo-Saxon accents: some exploratory Australian studies with the matched-guise technique. In: Language Sciences 5: 163–184.
Bernaisch, Tobias (2012) Attitudes towards English in Sri Lanka. In: World Englishes 31(3): 279–291.
Clark, Lynn, and Schleef, Erik (2010) The acquisition of sociolinguistic evaluation among Polish-born adolescents learning English: evidence from perception. In: Language Awareness 19(4): 299–322.
Dalton-Puffer, Christiane, Kaltenboeck, Gunther, and Smit, Ute (1997) Learner Attitudes and L2 pronunciation in Austria. In: World Englishes 16: 115–128.
Davydova, Julia (2012) Englishes in the Outer and Expanding Circles: A comparative study. In: World Englishes 31(3): 366–385.
Erling, Elisabeth J. (2007) Local identities, global connections: affinities to English among students at the Freie Universität Berlin. In: World Englishes 26(2): 111–130.
Grau, Maike (2009) Worlds apart? English in German youth cultures and in educational settings. In: World Englishes 28(2): 160–174.
Hilgendorf, Suzanne (2007) English in Germany: contact, spread and attitudes. In: World Englishes 26(2): 131–148.
Edwards, Alison (2014) The progressive aspect in the Netherlands and in the ESL/EFL continuum. In: World Englishes 33(2): 173–194.
Garrett, Peter (2010) Attitudes to language. Cambridge: Cambridge University Press.
Giles, Howard (1970) Evaluative reactions to accents. In: Educational Review 22: 211–227.
Giles, Howard, and Coupland, Nikolas (1991) Language: context and consequences. London: Taylor and Frances.
Hazen, Kirk (2014) An introduction to language. Malden, Oxford, Chichester: Wiley Blackwell.
He, Anping, and Ng, Sik Hung (2013) Language attitudes in China toward English. The social meanings of language, dialect and accent. International perspectives on speech styles, ed. by Howard Giles, and Bernadette Watson, 125–141. New York, Washington D.C./Baltimore, Bern, Frankfurt, Berlin, Brussels, Vienna, Oxford: Peter Lang.
Hu, Xiaoqiong (2005) China English, at home and in the world. In: English Today 21: 27–38.
Huygens, Ingrid, and Vaughan, Graham (1983) Language attitudes, ethnicity, and social class in New Zealand. In: Journal of Multilingual and Multicultural Development 27: 413–429.
Kasztalska, Aleksandra (2014) English in contemporary Poland. In: World Englishes 33(2): 242–262.
Labov, William (1972) Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press.
Lippi-Green, Rosina (1997) English with an accent. London: Routledge.
Littger, Peter. 2014. Zehn typisch deutsche Englisch-Patzer. Karriere Spiegel, 13.11.2014. http://www.spiegel.de/karriere/ausland/wenn-deutsche-englisch-sprechen-10-typische-denglisch-patzer-a-1001528.html, accessed: January 17, 2015. Video contribution.
McKenzie, Robert. 2008a. The role of variety recognition in Japanese university students’ attitudes towards English speech varieties. In: Journal of Multilingual and Multicultural Development 29(2): 139–153.
McKenzie, Robert. 2008b. Social factors and non-native attitudes towards varieties of spoken English: A Japanese case study. In: International Journal of Applied Linguistics 18(1): 3–88.
McCroskey, James C., and Young, Thomas J. (2006) The use and abuse of factor analysis in communication research. In: Human Communication Research 5(4): 375–382.
Meyerhöfer, Franca Levina. 2014. “Denglisch für Better Knowers”: So absurd klingen deutsche Sprichwörter auf Englisch. Huffington Post, 25.06.2014. http://www.huffingtonpost.de/2014/06/25/denglisch-deutsche-sprichwoerter-englisch_n_5528208.html, accessed: January 18, 2015.
Mukherjee, Joybrato, and Hundt, Marianne (eds.) (2011) Exploring second-language varieties of English and Learner Englishes – Bridging a paradigm gap. Amsterdam and Philadelphia: John Benjamins.
Ng, Sik Hung, and He, Anping (2004) Code-switching in trigenerational family conversations among Chinese immigrants in New Zealand. In: Journal of Language and Social Psychology 23: 28–48.
Percy, Larry (2012) The role of emotion in processing advertising. Advertising theory, ed. by Shelly Rodgers, and Esther Thorson, 69–84. New York / London: Routledge.
Preston, Denis (2003) Language with an attitude. The handbook of language variation and change, ed. by Jack K. Chambers, Peter Trudgill, and Natalie Schilling-Estes, 40–66. Oxford: Wiley-Blackwell.
Proshina, Zoya G. (2013) Book Review. Exploring world Englishes: Language in a global context. Philip Seargeant. London and New York: Routledge, 2012, xiii + 218 pp. The amazing world of Englishes: A practical introduction. Peter Siemund, Julia Davydova, and Georg Maier. Berlin and Boston: De Gruyter Mouton, 2012, ii + 283 pp. In: World Englishes, 19 AUG, 2013. DOI: 10.1111/weng.12048.
Schilling, Nathalie (2013) Sociolinguistic fieldwork. [Key Topics in Sociolinguistics] Cambridge: Cambridge University Press.
Schneider, Edgar W. (2007) Postcolonial English. Varieties around the world. Cambridge: Cambridge University Press.
Schneider, Edgar W. (2003) The dynamics of New Englishes: From identity construction to dialect birth. In: Language 79 (2): 233–281.
Seargeant, Philip (2012) Exploring world Englishes: Language in a global context. London and New York: Routledge.
Siemund, Peter, Davydova, Julia, and Maier, Georg. 2012. The amazing world of Englishes. A practical introduction. Berlin/Boston: Mouton de Gruyter.
Szmrecsanyi, Benedikt, and Kortmann, Bernd. 2011. Typological profiling: learner Englishes versus indigenized L2 varieties of English. Exploring second-language varieties of English and Learner Englishes: Bridging a paradigm gap, ed. by Joybrato Mukherjee, and Marianne Hundt, 167–187. Amsterdam: John Benjamins.
Stewart, Mark A., Ryan, Ellen Bouchard, and Giles, Howard (1985) Accent and social class effects on status and solidarity evaluations. In: Personality and Social Psychology Bulletin 11, 98–105.
Tan, Ying-Ying, and Castelli, Christina (2013) Intelligibility and attitudes: How American English and Singapore English are perceived around the world. In: English World-Wide 34 (2): 177–201.
Trudgill, Peter, and Giles, Howard (1978) Sociolinguistics and linguistic value judgement: correctness, adequacy and aesthetics. Functional studies in language and literature, ed. by Frank F. Coppieters, and Didier L. Goyvaerts, 167–190. Gent: Story-Scientia.
Zahn, Christopher J., and Hopper, Robert (1985) Measuring language attitudes: The speech evaluation instrument. In: Journal of Language and Social Psychology 4(2): 113–123.


Author:
Dr. Julia Davydova
Postdoctoral Researcher
WOVEN funding scheme, University of Mannheim
L13, 9; R. 206
68131 Mannheim





[1] Second-language varieties are varieties spoken in countries where English was introduced during the colonial era and has been widely used in the government, administration, and education, developing, as a result, a plethora of unique but widely accepted features of its own. Furthermore, English plays an important role in informal interactions and everyday encounters in such countries. Foreign-language varieties or Learner Englishes are in contrast spoken in countries where English is primarily regarded as an important lingua franca for international communication and is taught as one of the subjects at schools. In these countries, the knowledge of English is decisive in academic contexts and for communication with foreigners.
[2] Similar to monolinguals, these speakers learned English as a foreign language at school and are thus included in the study.
[3] CEFR stands for ‘Common European Framework of Reference for Languages’, see also http://-en.wikipedia.org/wiki/Common_European_Framework_of_Reference_for_Languages, accessed January 20, 2015.
[4] One of the commentators noticed that these ratings of the German speaker broadly conform to the cultural stereotypes attached to German people and associated with the German culture. Given that the judges were generally aware of the speakers’ origins (see Variety recognition task), it is not unreasonable to assume that the L1 German raters may have expressed their natural cultural self-perceptions through these ratings.
[5] The corpus is compiled as part of the DFG-funded project DA 1678/1-1 “Determinants of sociolinguistic variation in the ESL/EFL English: Evidence from two academic communities”.