Language Accents and Attitudes

Appel and Muysken proposed that “If there is a strong relation between language and identity, this relation should find its expression in the attitudes of individuals towards these languages and their users”. (1987). This is also true for accents which are key in signaling someone as different or as part of an out-group. Language attitudes are not linguistic but social. People automatically and instinctively assign characteristics to speakers of a particular language or accent based on their sterotypes and beliefs about members of that community (Bonvillian, 2003). The level of Solidarity is typically highest in languages geographically or culturally closest to one’s own; the level of Competence in a language is associated with that people’s reputation for hard-work and good education; the level of prestige or Power in language can typically be equated with the amount of riches and technology that the country has. It is natural that, for example, Swedes, have more positive attitudes to and stereotypes of those nations and peoples closest to themselves (say Norway and Denmark) – both geographically as well as culturally. Additionally, Nesdale & Rooney pointed out that language attitude research has shown that the most powerful accents in a community receive high marks in status and competence while lesser known languages and minority accents receive higher marks in solidarity and integrity (1996).

Contact

Contact and experience are obvious factors that affect our stereotypes and attitudes. Just as one breaks the cycle of prejudice through exposure and knowledge, one can “liberate people from such prejudiced discourses” by exposing them to linguistic diversity (Kubota, 2001). Much exposure today happens naturally. Chambers (2002) cited the power of mobility as a powerful linguistic force today and it is undeniable that face-to-face interactions, personal contact, and broad international experiences are happening between more and more citizens of the world. Not only are our knowledge and understanding increasing but consequently our attitudes are also altering and changing course. Markham noted that in regards to his research on accents “Experience with or awareness of foreign and native accents must also play a role” (1997, p.100). Markham also pointed out that both specific familiarity with a particular non-native accent as well as broad experience with non-native speakers may help in identifying accent.

Caveat

In a study by Boyd (2004), the relationship between accents, attitudes and stereotypes is questioned. She pointed out that in order to have an attitude towards a speaker based on their acccent and one’s own stereotype of that culture, one must first be able to correctly identify that speaker’s first language/cultural identity. While her results showed clearly that while native Swedish speakers could quickly and correctly judge the degree of accent of a second language Swedish speaker, they were only able to correctly identify two accents (and therein cultural identities/first languages) of the participants. Both of these languages, Finnish and German, are languages with a long historical connection to Sweden. She also proposed that English and French accents in Swedish would be correctly identified by Swedish speakers but that accents from languages other than these four would be difficult to identify (even for trained phoneticians). While Boyd’s study focused only native speakers’ ability to identify accents, it is reasonable to assume the same principal applies to non-native and second-language speakers.

Responses

Affective and cognitive reactions may occur when one’s emotions and reactions are connected to a particular person or situation (Cargile and Giles, 1997). Both the Austrian study by Dalton-Puffer, et al. and the Dutch study by Broeders and Gussenhoven mentioned earlier displayed similar effects – a preference for that which one knows and feels comfortable with. Dalton-Puffer also cites a Japanese study (Chiba, et al, 1995) in which students rated the accents they recognized most easily (those that were most familiar), most positively (though not their indigenous variety). These conflicting results coincide with the social identity theory that people prefer their most salient in-group but are not immune to contact factors. The amount of experience and exposure are powerful factors regarding attitudes. A study done by Byrnes, Kiger, and Manning in 1997 showed that, among other things, the amount of experience a teacher had working with minority language children positively affected their language attitudes.

Attitude Research

The term “attitude” and the study of attitudes are both borrowings from social psychology.

Edwards wrote, “Because language is one of the traditionally important social markers, it is not surprising that the study of attitudes has a central position in the social psychology of language” (1999). People have reacted to and evaluated different accents, dialects and languages since the beginning of mankind’s verbal history. Historically, two theoretical approaches are discussed regarding the study of language attitudes. The first one is the behaviorist view, in which attitudes must be studied by observing the responses to certain languages in actual interactions. Conversely, the mentalist view says that attitudes are an internal, mental state, which can explain certain forms of behavior. The mentalist view has been the one most widely followed and employed in language attitude research as it is most conductive to surveys and interviews.

One of the first groundbreaking attitude studies was made by Lambert in 1960 presenting the “matched guise” technique, i.e. the same speaker presented as “native” in one or more varieties. Since then many dialectal and bilingual studies have followed in a similar strain.

While the “matched guise” technique is still often used, it is often removed due to lack of authenticity and poor feasibility on the part of the speaker.

Indeed language is such a powerful social force that listeners, even small children, perceive and interpret linguistic and paralinguistic variation in messages as indicators of both personal and social characteristics (Cargile & Giles, 1997). Magen (1998) cited research (Flege & Hammond, 1982; Flege, 1984) that showed that listeners could detect a foreign accent after exposure to a sample of speech as short as 30 ms. However, there is no research that states how long it takes for a listener to recognize or identify a foreign accent. In most language attitude studies listeners are asked to rank or evaluate speakers on qualities relating to prestige, power, or social attractiveness by asking questions like “how friendly is this person” or “how educated does this person sound”? The results of language-attitude studies are based on the premise that languages (or linguistic varieties) are objectively comparable and intrinsically and inherently equal and the differences in subjective evaluation of speech fragments are not caused by differences in aesthetics or logic but rather by the differences in the social position of the listeners and the imposed norms their culture has bred within them. Piske, et al, (2001) stated in their review that it appears the accuracy of a non-native speakers pronunciation is dependent on the L1. They also stated that studies have shown that read speech was judged to be more strongly foreign-accented than spontaneous speech samples. Furthermore, Markham wrote that “If listeners are told that all speakers are non-native, then it seems likely that listeners will tend towards hearing accent. If they are told that native speakers are present in the sample, then some listeners may err on the side of caution and be inclined to give all speakers slightly better scores” (1997).

There are some questions as to the reliability of participants. scoring but Magen (1998) cited sources that stated untrained listeners perform reliably when judging foreign accents although experienced listeners can sometimes be more so. There are additional factors that may affect a listener’s response. Non-linguistic factors such as context, topic, and relevance of the text may alter a listener’s scoring. Degree of emotionality and humor may also factor in. Previous studies have also considered rate of speaking and musical ability among other things as possible and probable factors affecting attribute scoring.

Ground for this study

Since the start of language attitude studies in the 1960s, many new techniques, language groups, and factors have been explored. Edwards (1999) claimed that much evidence has been secured regarding the reactions of native speakers of English to various dialects and varieties of English. However there have been only a few studies made on the perception of non-native speakers of different varieties in English (Dalton-Puffer, et al., 1997). It is apparent that even fewer of these studies have been done comparing non-native English varieties with only non-native speakers as listeners and judges. This study aims to fill that gap.

Earlier language attitude studies have shown that there is often a difference in response to speakers of different varieties based on a listener’s stereotypes of the speakers and the social situation of the speakers and the listeners. The perceived Power of a speaker is normally associated with language varieties that have a lot of social power; the perceived Competence of a speaker is linked with that people’s work ethic and assumed education; the perceived Solidarity is connected with those geographically and culturally closest to one’s self. Similarly this leads to the first hypothesis:

Hypothesis 1: There is a difference in attitude to the English accents of speakers of different languages in regard to Power, Competence, and Solidarity.

Hypothesis 2: A listener’s time abroad affects their attitudes towards the English accents of speakers of different languages in regard to Power, Competence, and Solidarity.

Hypothesis 3: A listener’s time abroad affects their ability to identify a speaker as coming from a particular country.

Note: Time abroad. is defined as one month or more spent in a foreign country in contrast with less than one month spent in a foreign country.

Method

These hypotheses have been tested quantitatively on a group of Swedish female high school students using recordings of non-native female speakers together with Likert-based attitude scales and an identification task. The “matched-guise” technique was not used in this study due to lack of feasibility and authenticity. The dependent variable (DV) for the study was the students’ responses on the attitude scales. Attitude scales are indisputably the most popular way to measure attitudes and these particular scales used in this study have been widely used in accent attitude studies before and have therefore been proven both reliable and valid. The independent variables (IV) were the students’ time abroad also called “contact” (contact/no contact) and the speakers of the recordings (9 different nonnative English accents). The experiment was a 2 (contact/no contact) x 9 (speaker ethnicity: nine different European accents) factorial design where both factors were fixed.

There were linguistic variables both phonetic and phonemic, as well as extralinguistic variables, that were uncontrolled and may have affected the results. Such linguistic variables resided primarily within the speakers used and included, for example, the number of mistakes and hesitations. Extralinguistic variables may have included those regarding the speakers or the participants (social, regional, political, religion), group (size, immediacy, influence), and situation (time of day, location). Two additional external factors that may have influenced the results was my own native English speaking before the recordings and during the experiment as well as the speakers’ earlier models, i.e. what variety of English they have been most exposed to during their education. Further factors that were not taken into account are the speakers’ and listeners’ knowledge and familiarity of English as well as their time spent in English-speaking countries.

Material

The material for this experiment consisted of 9 recordings of the same text read by 8 European exchange students and 1 Swede (all female) studying at Växjö University during the spring term 2004. Their participation was voluntary. The text used is entitled “The Rainbow Passage” and, in its entirety, includes all the sounds of English. In this experiment, however, only the first two paragraphs were used. These recordings were not altered or controlled in any way in regard to time, rate of speaking, or pitch in order to produce the most natural recording possible. The speaking time for the first paragraph ranged between 30 and 46 seconds for each speaker while the time for the second paragraph ranged between 25 and 38 seconds. Also, each speaker was given the text to review for only a few minutes before recording with the possibility to ask questions about pronunciation or meaning. The result was a somewhat spontaneous reading by each speaker.

Speakers

These nine speakers were located, interviewed, and recorded within their residence halls. All 9 speakers were women between the ages of 21 and 31 currently studying at Växjö University. Excluding the Swedish speaker, they had been in Sweden for a time period of between 2 months and 1 year. All nine speakers were studying a subject other than English and had not studied English since high school. Questions were also asked that involved if, how, when, and where they had spent time in English-speaking or other foreign countries. Of the 9 speakers, 3 of them (Spain, Poland, Germany) had spent 2 months in an English-speaking country (vacation/working) while the remaining 6 had spent less than a month in an English-speaking country. Also, the women from France, Poland, and Germany had spent one month or more in a non-English-speaking foreign country. All nine women classified their English as Intermediate or Advanced and agreed that they had a typical foreign accent in English for a person from their country. The recordings were placed in the following order: Germany, Portugal, France, Poland, Sweden, Spain, Austria, Italy, the Netherlands. No control was made for age of learning, years of formal instruction, learning aptitude, order influence, or for the fact that inclusion or exclusion of one voice or accent may affect judgements of those remaining.

Participants

Participants for this study were female students in their first or second year within the International Baccalaureate Program at Katedralskolan, a high school in Växjö, Sweden in the spring of 2004. Participation was voluntary. In order to qualify for the study, the student must have been a Swedish citizen by birth or have moved to Sweden by the age of six. A total of 20 qualifying students listened to recordings and filled in the questionaire. Of these 20 students, 10 reported that they had had more than one month of continuous contact with a foreign country (e.g. vacation, language studies). The remaining 10 did not report having more than one month of continous contact. These two groups were designated “Contact” and “Non-contact” respectively. Furthermore, it is worthwhile to note that the “contact/no contact” factor is quantitative not qualitative. What kind of contact and international experience the listeners had, for example how many different contacts and what level of contact are not measured or controlled. There was also no differentiation made for whether the contact they had was with an English-speaking or non-English speaking country.

Procedure

Once the recordings and the questionnaire were complete, I contacted the head of the International Baccalaureate Program and decided on which classes would be most appropriate. Two English classes were chosen and we met accordingly. I was given the first part of the class to conduct the study, approximately 25 minutes. I was introduced to the class as a native English speaker living in Sweden. I then explained that my study was about perception of a speaker based on their voice, for example a stranger on the radio or on the telephone. They were not told that the speakers were all non-native speakers of English.

I reviewed and explained how to fill in the attitude scales and answered questions regarding the meaning of the qualitites. I then played each of the recordings (first paragraph of the text) in the order mentioned earlier. I paused after each recording and waited for all the students to finish marking the scales. In an effort to avoid fatigue, I encouraged them repeatedly to keep paying attention. As an oral and mental “palate cleanser”, I also told them a knock-knock joke. I chose knock-knock jokes since they require an active response from the class. After completing the first nine attitude scales, the students were then given instructions for the accent identification section. They were to listen to each speaker (second paragraph of the text) and then place the number of the speaker in front of their best guess for where that speaker was from. There was no space for “I don’t know” and each country was only to be used once. Following these instructions, I played up the recordings with only a brief pause in which to say the number of the next speaker and to remind them they were required to write an answer. After this section was completed I asked the participants to complete the page of demographic questions on the back. As they completed this page, I collected the packets and answered any further questions that they had.

Summary of results

The results revealed that the speakers were indeed perceived and interpreted differently by the listeners. The nine speakers scored at varying levels within the categories Power, Solidarity and Competence. The speakers with Dutch, Austrian and Polish accent fared well in regards to the three qualities Power, Solidarity, and Competence. Portugal, Italy, France and Germany were rated generally low while Spain and Sweden placed somewhere in the middle. Time spent abroad, as defined and used in this study, also had an effect on the listeners. attitudes and responses as well as their ability to identify a particular accent. The contact group was seemingly negatively impacted producing lower average score responses as well as a lower rate of correct identification, even in regard to their own non-native English speaker [Swedish – speaker number 5]. There was no strong overall tendency to evaluate Swedish highly though the Swedish speaker’s ethnicity was rather identifiable by the listeners.

While there exists a strong link between language and ethnicity, it is not a one-to-one relationship. I believe both the listeners and speakers in this study may have responded, had they been asked, differently in regards to their “grouping” and as to their strength as a member of the ethnic group in which I have categorized them. Loyalty to and identity with one’s in-group might occur at any level – local, municipal, provincial, regional, national or international. Most often, the strongest identity is found at the lower levels and then successively weakens as the in-group broadens.

These constructions are not only based on our personal history but also on functional interaction. In many ways it is not until we meet someone from an out-group that we are able to identify ourselves as part of an in-group. In particular the speakers in this study, as exchange students, may be reevaluating their “group” membership and redefining their identity based on new functional interaction. Their personal contact and experience, especially situation-based, may have drastically changed over the days, weeks, and months prior to the recordings as their identity may have shifted from a local level to a national level (e.g. a change from “I’m from Barcelona” to “I’m from Spain” except when meeting other Spaniards).

These differences may have been caused by their contact with foreigners or their time abroad. In this study the non-contact group seems to identify more strongly with other European non-native English speakers while the contact group does not. While contact and time abroad may be a decisive factor in this difference there exists perhaps some difference “at home” as well. Some of the students may have more contact with foreigners (for example, immigrants) without ever leaving Sweden.

Furthermore, the listeners’ education may have an impact on their identity. The International Baccalaureate Program is used throughout the world and embraces multiculturalism. It is offered in several languages but most prominently in English. The Swedish students who choose this program which is given in English would most likely have a particular predilection for English and internationalism and there by also an identity that lays at a level that is more national or international than local or regional. Surely these particular factors must impact one’s personal identity, one’s group identity, and therein a speaker’s accent and a listener’s reaction.

As I was recording the nine speakers, clearly they could be thought to be producing better English then than had they been unaware of the proceedings. In authentic social situations speakers should be more prone to communicate, to simply make themselves understood. It is only when evaluation comes into the picture that people, even native speakers, make an effort to produce correct English.

I believe that this choice can also be affected by the company kept. Anecdotally, I can state that many people feel awkward speaking a second or foreign language such as English with a native speaker. In a mixed ethnic group where no native speakers are present (and there is no other shared language), participants feel that English is a natural choice and that they are on a more even playing field. In a world where the number of non-native speakers of English continues to expand prolifically beyond the number of native speakers, it seems reasonable that more and more people will choose to simply communicate.

It is my personal experience that Swedes are particularly partial to near-native English as opposed to ethnic (read Swedish)-accented English. Many Swedes laugh or are embarrassed to hear Swedish journalists or politicians speaking accented English while surely the French population would be in an uproar if Chirac began speaking fluent British English.

While it is plausible for a non-native speaker to reach native-like language skills, it is neither always possible nor desirable.

 

Instruction: Don’t copyBritta Larson Bergstedt’s summary above, first of all, because she limits it to the results of her own experiment without setting connection with the first six paragraphs of the text. As there are quite a few references to interpretations of indicators of both personal and social accent characteristics, you will have to select those that may be considered valid for Bergstedt’s study. Be sure to describe the basic hypotheses and the stages of the experiment.

 

To repeat:

1. A summary begins with an introductory sentence that states the article's title and author.

2. A summary must contain the main thesis or standpoint of the text, restated in your own words. (To do this, first find the thesis statement in the original text.)

3. A summary cannot exceed 1/3 of the original.

5. A summary should contain all the major points of the original text, and should ignore most of the fine details, examples, illustrations or explanations.

4. The backbone of any summary is formed by crucial details (key names, dates, events, words and numbers). A summary must never rely on vague generalities.

5. A summary must contain only the facts and ideas of the original text. Do not insert any of your own opinions, interpretations, deductions or comments into a summary.

Important points to remember:

Make sure to include the author and title of the work.

Don't copy the article. Don't put your own opinions, ideas, or interpretations into the summary.

Be concise: a summary should not be equal in length to the original text. Instead, paraphrase. If you must use the words of the author, cite them. If you quote directly from the original text, use quotation marks. (Minimise how often you do this)

Write in the present tense.

The purpose of writing a summary is to accurately represent what the author wanted to say, not to provide a critique.

Edit what you write. Check your English grammar, spelling and punctuation mistakes.