Professional Documents
Culture Documents
ABSTRACT: In the context of pronunciation teaching, the relevance of a standard native accent as
teaching model and norm needs to be reconsidered in view of the learners' attainment as well as their
wishes and expectations. In order to test attitudes to native and non-native varieties of English in Austria, a
language attitude study was undertaken with 132 university students of English. In response to the
modified matched-guise test, the subjects evaluated three native accents RP (Received Pronunciation),
near-RP and GA (General American) and two Austrian non-native accents of English. The results confirm
the low status the non-native accents have amongst their users and the overall preference for the three
native accents. Generally, the respondents rate the accent best with which they have become familiar at
school and/or during stays in English-speaking countries. Reflecting historical and geopolitical circum-
stances, the majority of the subjects support RP as their favourite model of pronunciation. Furthermore,
the study sheds light on the importance of personal exposure to English in its native environments. While
the evaluations of the students with EFL experience reflect rather rigid stereotypes, those students who
have spent some time in English-speaking countries reveal more individualized, situation-linked attitudes.
1. INTRODUCTION1
The teaching of pronunciation plays a crucial role in second language learning and has, in
fact, become an integral part of almost any language course. Good pronunciation is
indeed indispensable for adequate communication in a foreign language and is, moreover,
to a large extent responsible for one's first impression of a learner's L2 competence. The
importance of pronunciation teaching is even more strikingly obvious in the context of
language teaching at university level, which involves the training of future language
teachers. While teaching methods have considerably evolved over the years with the focus
of interest having shifted somewhat more to suprasegmentals (Dalton and Seidlhofer,
1994), pronunciation teaching is still an area which is experienced as problematic by both
teachers and learners.
Like most other European countries, Austria belongs to the `expanding circle' (Kachru,
1992: 356) of the English-speaking world. Over the past decades the teaching of English
has indeed expanded in Austria. Part of it has been a general trend towards modern (living)
languages, but English has also gained ground at the cost of other modern languages,
notably French. Today, the teaching of English as a foreign language is firmly anchored in
the Austrian school system with all pupils starting to learn English at primary level. At the
vast majority of secondary schools English is the main (and sometimes the only) foreign
language taught. On entering university, students will typically have had ten years of
formal instruction in English.
For a university course in English pronunciation it is the assessment of students'
spoken performance that represents an inherently problematic issue, especially with
regard to the type of accent to be attained. The implicit aim of most pronunciation
courses at Austrian universities is for students to achieve a native-like accent (i.e.,
# Blackwell Publishers Ltd. 1997, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA.
1/5/97 - 16:49 96/89 WENG 89_11 SueH
Theoretical considerations
`Attitude' is a construct which has played a central role in social psychology for the
better part of this century. Given that language is a social practice of prime importance it is
only natural that the study of language attitudes has proliferated. Without entering into a
review of the various controversies regarding the exact nature of (language) attitudes, we
would like to mention a few aspects which feed directly into the research reported in this
paper. We regard attitudes not as monolithic entities but as composites of a number of
cognitive, emotive and/or conative factors (McGuire, 1969; Smit, 1994). They are mental
constructs acquired through experience, predisposing a person to certain feelings and
reactions in response to certain situations, persons or objects. There have been a number of
attempts to single out the exact share of the three components (beliefs, feelings and
intentions) in the constitution of an attitude, but these have not been very successful. While
this may be a productive difficulty on the theoretical level, it poses several problems in the
development of a research methodology. The way out of the conundrum has usually been
to adopt an operational definition of attitude in terms of observable responses as, for
example, `[a]n attitude is an evaluative reaction a judgement regarding one's liking or
2. METHODOLOGY
The present research question with its emphasis on subjects' reactions to target language
accents suggested the use of a modified matched guise methodology, the `verbal guise'
method (Cooper, 1975: 5; Teufel, 1995: 75). Instead of one speaker assuming different
guises, several speakers are used on the stimulus tape. In second language contexts it is
practically impossible to find speakers who are equally convincing in several guises. This
means of course that variables like voice quality can be controlled only minimally.
The speakers who were finally chosen for the stimulus tape were all university-educated
females who were between 30 and 40 years old. They had been screened for obvious
differences in voice quality and reading rate. The two non-native (Austrian) speakers were
rated to have weak but recognizable Austrian accents. Speakers 35 spoke educated
standard native varieties. The reason behind the fact that there is only one American but
two British voices lies in the problematic status of what counts or does not count as a
standard pronunciation of British English (Shibles, 1995). Speaker 3 thus represents the
neutral radio announcer type of accent (RP) which is present on many tapes accompanying
# Blackwell Publishers Ltd. 1997
1/5/97 - 16:49 96/89 WENG 89_11 SueH
conventional course books. Speaker 5, on the other hand, represents the type of standard
accent (far more frequent among the native speaker community) which has a slight local
colouring (near-RP).
The speakers were asked to read a short text on the topic of bilingualism, which was
emotionally neutral and which also tied in with the university setting of the study. Both,
readers and listeners, were given matching instructions saying that the taping was done in
the interest of finding speakers for the publishing of an audio-book on child language
development. In our opinion this decision has advantages on two levels. In the actual test
situation the given context makes it more plausible that the subjects should evaluate a
person on grounds of her voice alone. The main gains of this methodological step,
however, lie on a more theoretical level. Given that attitudes are a social phenomenon it
seems paradoxical to obtain evaluations of speech in a situational vacuum (Giles, 1992: 36;
Smit, 1994: 176). In reality, people react to speech in specific situations and the same voice
or speaker may well get different evaluations in different contexts. As context-free speech
evaluation is not the normal case, it is more than likely that the subjects will (consciously
or subconsciously) construct a context for themselves. Interpreting the subjects' evalua-
tions against the researchers' understanding of the situational context would be an
additional source of misinterpretation. Even though the exact workings of the situational
influence are hard to pin down, predefining the situational context gives the researcher a
certain amount of control over the conditions of the subjects' and the researcher's
interpretations of the situational context.
The subjects' evaluations of the five speakers were captured in two ways (See Appendix
for the questionnaire). A list of 12 (mostly adjectival) attributes was given and subjects
were asked to indicate to what degree this attribute applied to the speaker. The adjectives
were chosen so as to reflect both status and solidarity values (Ryan and Giles, 1982: 6). The
exact format used is a variant of the semantic differential technique. It does not contain a
`neutral' option, since attitudes cannot, by definition, be neutral. The zero option was
formulated as `Does not apply to the speaker' and placed at the rightmost end of the scale.
This has the advantage of undercutting respondents' tendency to go for the middle values
in case they cannot quite make up their minds. The subjects were asked to respond to each
voice immediately after hearing it. The evaluations captured in this way are assumed to be
mostly emotive in nature thus answering to one component of `attitude.' The list of
adjectives was followed by two summary statements `I think this person would be a good
radio presenter,' and `I would like this person as a friend.' By virtue of their being more
abstract, these two statements are taken to access a more cognitive aspect of the attitude
construct. Altogether, the elicitation methods employed are indirect rather than direct.
The test was administered as part of the opening session of several courses at the English
Department of Vienna University in October 1995. In case there were any hidden `traumas'
regarding departmental expectations about pronunciation `standards,' subjects were told
that researchers from another Austrian university had approached us to collect data for
their study.
3. DISCUSSION
The test population consisted of 132 students of English, most of whom are L1 speakers
of German and between 19 and 22 years of age. Reflecting our department's general
student population, about 65 percent of the respondents intend to become English
119
1/5/97 - 16:49 96/89 WENG 89_11 SueH
teachers, and the female respondents outnumbered their male colleagues by almost 7:1 (for
the relevant figures see Table 1).
The responses showed that British English, traditionally preferred in Austria, is still the
most popular model: more than two-thirds of the respondents attempt to learn British
English and its standard accent RP. This orientation towards British English is, of
course, also supported by the British Isles' geographical closeness to Austria. While about
30 percent of the respondents have already spent more than one month in the UK, only
17 percent have been to the USA, and a mere 4 percent to Canada, Australia or other
English-speaking countries further afield (see Table 2). This means that almost half of the
respondents has not had the chance to experience English in one of its native environ-
ments. It is interesting to note that, of those who have chosen RP as their model, even more
students, namely 55 percent, have not spent more than one month an in English-speaking
country, while of those attempting to attain an American accent only 34 percent have not
been on an extended stay abroad (see Table 3). This means, that personal experience is
much more important in choosing GA as one's model than RP, which is, of course,
another indicator for the preponderance of British English at Austrian schools.
As the respondents were asked to evaluate varieties of a foreign language, it could not be
presupposed that they would be able to identify them correctly. Misidentifications
especially undetected ones could have led to misinterpretations of the data. Conse-
quently, the respondents were asked to identify the speakers' places of origin while
listening, for a second time, to the first sentences of the speech samples. In general, the
respondents did not have problems with this task, and the hit-rate lies above 85 percent for
all the speakers, except for speaker 2. This speaker's Austrian American accent (OEAm)
was so convincingly American to most of the students that only 17 percent recognized her
Stays abroad
Preferred accent None GB USA GB&USA Total
as Austrian. Even among respondents whose preferred accent is GA, the majority of over
70 percent thought they had listened to a native American (see Tables 4a and 4b).
On average, the responses given by the 132 respondents show very clearly how the five
speakers were evaluated as regards their abilities as readers of audio books. The speaker of
RP (speaker 3) was rated most positively throughout. While its position as first native
accent to be rated might have enhanced the positive reactions to it, the overall preference
for it cannot be discarded as a study-inherent error since it reflects the fact that, of the
accents included in this test, it comes closest to the standard British accent on ELT tapes
used in Austrian classrooms. Speaker 3 was followed in rank by the two other native
speakers of GA and near-RP (speakers 4 and 5), and the OEAm speaker (speaker 2). The
most negative evaluations were given to speaker 1, the speaker of Austrian British English
(OEBr). While not all of these numerical differences have tested as statistically significant
(see Table 5), the ratings for all the attributes and the two statements illustrate the same
order of preference RP first, OEBr last, and the other three accents inbetween. Due to
this consistency, the evaluations can be interpreted as reliable enough to be acceptable as
representing the test population's attitude pattern.
Figure 1 illustrates this order of preference for intelligent and likeable as well as for the
two statements, `friend' and `radio presenter' (the lower the mean the more positive the
evaluation). While, on the solidarity level (`friend'), native and non-native accents are not
very clearly distinguished, the reactions to `radio announcer' reveal the clear distinctions
made between the accents as regards their social status: RP is judged much more positively
than the other two native accents and, in fourth place, the OEAm accent. By far the least
attractive accent is OEBr, i.e., the one most often heard in Austria and spoken by the
students themselves.
The consistency of evaluations is not only a feature of the complete test population but
also of the sub-groupings according to the two relevant independent variables, namely
Preferred Accent and Stays Abroad. The former divides the students into two subgroups
Correct 126 96.2 22 16.9 109 87.2 121 93.8 106 85.5
Wrong 5 3.8 108 83.1 16 12.8 8 6.2 18 14.5
Total 131 100.0 130 100.0 125 100.0 129 100.0 124 100.0
Table 5. Evaluation of the five different speakers according to variety of English used
Significances (calculated by t-test for two independent samples): numbers in ascending order stand for five
speakers (from left to right); highly significant differences (1% level of error) between 2 means are marked by
hyphens; slashes indicate means with which means on other side of hyphen have tested significant; means in
brackets stand for significant differences (5% level of error).
3.5
2.5 EBr
EAm
2
RP
1.5 GA
Near-RP
1
0.5
0
Intelligent Radio presenter Friend Likeable
Figure 1. General evaluations of five voices (for intelligent, `radio presenter,' `friend,' likeable)
according to the model of pronunciation they have chosen, and, so we postulated, they
prefer personally either British English or American English. For ease of reference, these
two subgroups have been labelled `Brits' and `Amis' respectively. The second independent
variable is based on the information the respondents offered on where and how often they
have spent time in English-speaking countries. It thus divides the population according to
degree of personal exposure to English in its native environments into four subgroups no
visits of longer than one month (None), longer stays in Great Britain (GB), longer stays in
the USA (USA), and longer stays in Great Britain and the United States (GB&USA).
Similarly to the general evaluations described above, the response patterns of the
subgroups to the various attributes either do not reveal statistically significant differences
or they are consistent with the ratings of the two statements. We will therefore concentrate
on the two statements in the following description of the relevant attitude patterns.
Figures 2 and 3, which illustrate the evaluations of the two statements according to
Preferred Accent (see Tables 6a and 6b), give evidence of the fact that the minority group
3.5
3
2.5 Brits
2 Amis
1.5
1
0.5
0
EBr EAm RP GA Near-RP
Figure 2. Evaluations of five voices for `radio presenter' according to Preferred Accent
3
2.5 Brits
2 Amis
1.5
1
0.5
0
EBr EAm RP GA Near-RP
Figure 3. Evaluations of five voices for `friend' according to Preferred Accent
`Amis' is in general more tolerant in their ratings. They are more positive to the non-
native accents, and also to the other native accent which is not their personal favourite. In
particular, the more localizable British English is equally acceptable to Amis and Brits.
From a statistical point of view, however, the only significant differences in evaluation
concern the native American accent. Here the lower means for Amis can be interpreted as
genuinely representing more positive attitudes.
3.5
2.5
None
2 GB
1.5 USA
GB&USA
1
0.5
0
EBr EAm RP GA Near-RP
Figure 4. Evaluations of five voices for `radio presenter' according to Stays Abroad
2.5
2 None
GB
1.5
USA
1 GB&USA
0.5
0
EBr EAm RP GA Near-RP
The graphs in Figures 4 and 5, which present the evaluation of the two statements
according to Stays Abroad (see Tables 7a and 7b), illustrate interesting differences between
None, i.e., the respondents with very limited exposure to English in its native environments,
and the other three subgroups, GB, USA and GB&USA.2 The group None evaluated the
five speakers in the same way on both social status and solidarity levels, thus revealing a
rather rigid stereotype. The RP speaker was placed first, followed by the other two native
and the OEAm speakers, and, in fifth place, the Austrian British speaker. On the other
hand, the other subgroups differentiated between the speakers as radio presenters or as
possible friends. For instance, the three groups rated the RP speaker very positively as a
radio presenter, but only the GB group did the same on the solidarity level. The GA speaker
was rated equally positively as a friend by all the groups, but only the USA and GB&USA
groups evaluated her positively as a radio announcer. These differences show that the
groups generally supported their respective model of pronunciation.
With regard to the respective outgroup model, however, the groups differed in their
evaluations according to social status and solidarity levels. The GB group rated the GA
speaker positively as a potential friend and more negatively in the professional setting,
while the USA group rated the RP speaker positively in the professional setting and more
negatively in the group solidarity one. These evaluations clearly reflect the outgroup
stereotypes of, on the one hand, the friendly and easily approachable American speakers of
lower-status GA and, on the other, the distanced British speakers of higher-status RP. The
generally attested distinction made in Britain between RP and more localized accents the
latter achieving worse ratings in social status settings but better ones in solidarity situations
(e.g., Honey, 1989: 5859) is also reflected in the responses given by the respondents who
have been to the UK. While they rated the RP speaker more positively as possible radio
presenter than the near-RP speaker, the reverse is true when asked to rate the speakers as
potential friends.
To sum up this discussion, the test population displayed a consistent attitude pattern:
the native speakers were preferred to the non-natives. More specifically, the RP speaker
was rated best, the OEBr speaker last and the other three inbetween. This means that the
OEAm speaker was ranked similarly to the two native speakers; a result that has to be seen
in correlation with the high percentage of the respondents misidentifying her as a native
American. That the general attitudes towards the three native speakers are not represen-
tative of all respondents became clear when comparing the responses of the two subgroups
Amis and Brits. While the latter, numerically much larger, subgroup rated the RP speaker
first and the GA speaker only slightly better than the Austrian British speaker, the Amis
rated the GA speaker more positively than the speakers of British accents. But, in contrast
# Blackwell Publishers Ltd. 1997
1/5/97 - 16:49 96/89 WENG 89_11 SueH
to the Brits, they were in general more tolerant towards other accents maybe thereby
revealing a (stereo)typical American attitude. Similarly differentiated attitudes were
attested when comparing the results of the subgroups None, GB, USA and GB&USA.
While the None group voiced the stereotypes perpetuated in Austrian classrooms RP as
most prestigious, followed by other native and finally non-native accents in all circum-
stances the other groups differentiated their preferences according to, first, the dimen-
sions of social status and solidarity and, second, in- and out group status of the respective
accents. In other words, the respondents who have gained personal experience in English-
speaking countries revealed more individualized, situation-specific attitudes than the
rather rigid stereotypes of the EFL learners who have not had this kind of exposure.
4. CONCLUSION
The results of the language attitude test reported in this article show clearly that the null-
hypothesis that the subjects do not have different attitudes towards different accents of
English is not tenable. Advanced Austrian EFL learners display negative attitudes towards
their own non-native accent of English. This parallels the results attained by Chiba et al.
(1995) and Forde (1995). Among the native accents the respondents prefer the one with
which they are most familiar. The results furthermore show the importance of personal
contact with the target language group for the development of differentiated attitudes
rather than rigid stereotypes.
Due to traditional preferences and present models offered at schools, the preferred
accent is mostly RP. For pronunciation teaching this means that the norms put up by
English teachers in Austria are also widely accepted by their students.
Although these native accents are firmly in place as models for EFL learning and
teaching, the level of achievement amongst university students of English does not concur
with the attitude patterns obtained in the study. The greater part of the learners do not
seem to be able to attain the standard pronunciation they evaluate so positively.
Successful L2 phonology learning thus cannot be attributed exclusively to the existence
of positive attitudes towards the target accent. Clearly, there must be other factors exerting
an influence on students' level of achievement. Although there is no generally accepted
model of the factors influencing language achievement (Gardner, 1982, 1988), it can be
assumed that learners' motivation for studying English will have to be considered more
comprehensively than just focusing on their attitudes. As our main concern is learning
pronunciation, i.e., the aspect of language that most obviously expresses social identity and
group membership, further studies will also have to incorporate more deeply-seated socio-
psychological factors connected to questions of `self' and identification with the target
group.
APPENDIX
A child learning a language is learning about the world, about how it is organized and how it
works. This is very different from the adult learning a second language: he brings his world with
him and uses the language to try to express it. One of the greatest advantages of bilingualism is
that even very small children realise that the relationship between words and the objects they refer
to is not a necessary one, that the same things can have different names. It does seem that this
early exercise in abstraction does give the bilingual the mental flexibility and openness which has
frequently been reported by experimenters and psychologists.
After each voice you are asked to fill in one of the questionnaires according to your perceptions of the
speakers (start with Voice Number 1).
NOTES
1. A previous report on this project has been published in Vienna English Working Papers (Views), 4(2), 1995.
2. As too few of the respondents had spent time in other English-speaking countries, they could not be considered
in this analysis.
REFERENCES
Chiba, Reiko, Matsuura, Hiroko and Yamamoto, Asako (1995) Japanese attitudes towards English accents.
World Englishes, 14, 7786.
Cooper, Robert L. (1975) Introduction. International Journal of the Sociology of Language, 6, 59.
Dalton, Christiane and Seidlhofer, Barbara (1994) Pronunciation. Oxford: Oxford University Press.
Dalton, Christiane, Kaltenboeck, Gunther and Smit, Ute (1995) Language attitudes of L2 learners to native and
non-native varieties of English. Work in progress on pronunciation teaching at the English Department,
University of Vienna. Vienna English Working Papers (Views), 4(2), 414.
Eisenstein, M. and Verdi, G. (1985) The intelligibility of social dialects for working-class adult learners of English.
Language Learning, 35, 287298.
Flaitz, Jeffra (1988) The Ideology of English: French Perceptions of English as a World Language. Berlin: Mouton
de Gruyter.
Forde, Kevin (1995) A study of learner attitudes towards accents of English. Hong-Kong Polytechnic University
Working Papers in ELT & Applied Linguistics, 1, 5976.
Gallois, Cynthia and Callan, Victor J. (1981) Personality impressions elicited by accented English speech. Journal
of Cross-Cultural Psychology, 12, 347359.
Gardner, Robert C. (1982) Language attitudes and language learning. In Attitudes towards Language Variation.
Social and Applied Contexts. Edited by Ellen B. Ryan and Howard Giles. London: Edward Arnold. pp. 132
147.
Gardner, Robert C. (1988) Attitudes and motivation. Annual Review of Applied Linguistics, 9, 135148.
Gardner, Robert C. and Lambert, Wallace (1972) Attitudes and Motivation in Second Language Learning. Rowley,
MA: Newbury House.
Giles, Howard (1992) Current and future directions in sociolinguistics: a social psychological contribution. In
Sociolinguistics Today. Current Perspectives. Edited by Kingsley Bolton and Helen Kwok. London: Routledge.
pp. 361368.
Giles, Howard and Powesland, P. F. (1975) Speech Style and Social Evaluation. London: Academic Press.
Giles, Howard and Ryan, Ellen B. (eds.) (1982) Attitudes towards Language Variation. Social and Applied
Contexts. London: Edward Arnold.
Honey, John (1989) Does Accent Matter? The Pygmalion Factor. London: Faber & Faber.
Kachru, Braj B. (1992) Teaching World Englishes. In The Other Tongue. 2nd edition. Edited by Braj Kachru.
Urbana: University of Illinois Press. pp. 355365.
Lambert, Wallace, Hodgeson, R. C., Gardner, R. C. and Fillenbaum, S. (1960) Evaluational reactions to spoken
languages. Journal of Abnormal and Social Psychology, 60, 4451.
McGuire, William J. (1969) The nature of attitude and attitude change. In The Handbook of Social Psychology.
Vol. III (2nd edition). Edited by G. Lindzey and E. Aronson. Reading, MA: Addison-Wesley. pp. 136314.
Oller, John W., Baca, L. and Vigil, F. (1978) Attitudes and attained proficiency in ESL: a sociolinguistic study of
Mexican Americans in the Southwest. TESOL Quarterly, 11, 173183.
Oller, John W., Hudson, Alan J. and Fei Liu, Phyllis (1977) Attitudes and attained proficiency in ESL: a
sociolinguistic study of native speakers of Chinese in the United States. Language Learning, 27, 127.
Pierson, Herbert D., Fu, Gail S. and Lee, Sik-yum (1980) An analysis of the relationship between language
attitudes and English attainment of secondary students in Hong Kong. Language Learning, 30, 289316.
Ryan, Ellen B. and Giles, Howard (eds.) (1982) Attitudes towards Language Variation. Social and Applied
Contexts. London: Edward Arnold.
Shibles, Warren (1995) Received pronunciation and Realphonetik. World Englishes, 14, 357376.
Shuy, Roger and Williams, Frederick (1973) Stereotyped attitudes of selected English dialect communities. In
Language Attitudes: Current Trends and Prospects. Edited by Roger Shuy and Ralph Fasold. Washington D.C.:
Georgetown University Press. pp. 8596.
Smit, Ute (1994) Language attitudes, language planning and education: the case of English in South Africa.
Unpublished Ph.D. thesis. University of Vienna.
Teufel, Gunda (1995) Language attitudes of Anglo-Australian high-school students towards German-accented
English. Vienna English Working Papers (VIEWS), 4(2), 5973.
Tucker, G. R. and Lambert, Wallace, E. (1969) White and Negro listeners' relations to various American-English
dialects. Social Forces, 47, 463468.
Weber, Ann L. (1992) Social Psychology. New York: Harper Collins.