Identity or Prestige : The Chameleon Effect on EFL Pronunciation

Although there is a current tendency to deal with English as an international language, EFL pronunciation seems to be motivated by conceptions on identity and prestige. In this respect, the following study investigates the effect of identity and/or prestige on EFL speakers' English speech (or usage) in two different settings. Male and female participants from Kuwait University (63 in total), were given debate topics to engage in. Both their meetings and actual debates were recorded. The participants also filled a short questionnaire. The recordings were analyzed using Praat, while the questionnaire results were obtained using SPSS. The results show that the EFL speakers use different accents and various realizations of English consonants as a reflection of what is known as the Chameleon Effect. This study aims at helping teachers understand the difference between the speakers’ language knowledge and their actual performance.


Introduction
An accent, in simple terms, is the articulation and shape of a language specifically connected to a certain speaker.The Oxford dictionary defines an accent as being "A distinctive way of pronouncing a language, especially one associated with a particular country, area, or social class."(https://en.oxforddictionaries.com/definition/accent).While the phonology of English has focused on accentual differences to identify social backgrounds, the idea became more complicated with speakers' knowledge of the importance of the implications of an accent.
Everyone has an accent; this is a fact that many people might not realize (Gluszek & Dovidio, 2010).Regardless the source of these accents, accents are considered an important identity marker.Kinzler et al. (2009) defines accents as being a social-indicator of several variables such as ethnicity, nationality and even social status.While categorizing people according to accent is natural, as the minds processes the available and salient linguistic information to form social assumptions, this process of stereotyping has drawn people to change their accent when needed, whether they are aware of it or not.
The theory behind the prestige of an accent suggests that if a person speaks in a standard accent, which is usually the accent of the dominant group in society, this speaker would be given a higher social status than those who speak in a less popular accent.While the standard accent produces a positive reaction, any other non-standard accent might produce a different reaction and attitude (Anderson et al., 2007).To avoid certain negative reactions, some individuals hide their own accents, trying to imply their belonging to a certain social category.will be exhibited; hence making ground for a brief conclusion of outcomes.

Research questions:
This research investigates the affect people have on our choice of accents.If people affect our choice of accent, can this theory be utilized as a strong device in EFL language classes to teach students languages and correct pronunciation?In order to find answers to these questions, the researchers investigated the occurrence of the Chameleon Effect reflected in the speech of EFL Kuwaiti speakers.The questions that need to be addressed were,  Do Kuwaiti speakers of English change their accent in certain contexts?
a. What accentual features are affected?
b. What contexts ignite these changes?
c. What factors affect the speakers' accent?
d. Are these changes intentional?
e. What is the Chameleon Effect on the recipients?

The Chameleon Effect
When the saying goes "monkey see, monkey do", it refers to conscious mimicry.This imitation is usually goal-directed.People seem to like those who imitate them (Neufeld & Chakrabati, 2016).Several studies have shown that mimicry of posture, body language and linguistic features helps boost pro-social behavior toward the imitator.Studies on mimicry are not current in psychology however the impact of mimicry on judgment and people's attitude is current.Many researchers have found that people mimic the verbal and nonverbal expression of others intentionally and unintentionally.Giles and Powesland (1975) discovered that humans mimic the accents of their peers.Verbal characteristics, such as speech tone (Webb, 1972) and speech rhythms (Cappela & Panalp, 1981) are mimicked unconsciously in certain social contexts.In addition, numerous studies investigated the contagious effect of laughter (Bush, Barr, McHugo, & Lanzetta, 1989;Provine 1992) and plenty of experiments have determined that using canned laughter leads an audience, in return, to laugh longer and more frequently (Cialdini, 2001).Posture and behavior have also been mimicked in social interactions, (Bernieri, 1988;Lafrance, 1982).For example, when associates have been ordered to smoke in a bar toilet, it was observed that participants imitate the smoking conduct of the associates (Harakeh, Engels, Van Baaren, & Scholte, 2007).Quigley and Collins (1999) investigated the act and amount of alcohol consumption.They found that alcohol amount and type of drink was induced by their peers in social interaction.
Although people mimic verbal and nonverbal expressions and conduct of their peers in various social contexts, studies additionally found that mimicry is given with a highly positive evaluation of the mimicker.Chartrand and Bargh (1999) engaged individuals in an assignment with an associate who was told to either mimic the behavior of the participants or to show impartial, nondescript mannerisms.Their experiment showed that the participants who were mimicked by the associates reflected more liking of the associate compared to individuals who were not mimicked.The mimicked participants felt that the mimicker was smoother and extra harmonious.This result is consistent with the previous work of Maurer and Tindall (1983) who investigated the mimicry of hand and leg position.They found that customers who have been mimicked by an agent felt more empathy towards the agent than the non-mimicked customers.
Not only was mimicry connected to social interaction, but also digital mimicry has been investigated.Interacting with a personified artificial confederate in interactive digital reality that mimicked people's behavior was found sufficient in influencing the evaluation of the confederate.In an experiment conducted by Bailenson and Yee (2005), a virtual confederate would verbally supply a persuasive argument to a player who interacted with the confederate.The virtual agent mimicked the player's head movements at a four seconds postponement half the time.After the contact, the player had to indicate his/her agreement with the message introduced by the agent and gave his/her opinion about it.It was clear that the mimicking digital confederate was more persuasive and gained better ratings than the non-mimicking agents.Kulesza et al. in (2014) and(2015) found that verbal mimicry is a very powerful tool.The effect of verbal mimicry has reflected positive attitude as well.A good example of that would be an experiment conducted by Van Baaren, Holland, Steenaert and Van Knippenberg (2003) who investigated the effect of verbal mimicry on the recipient.They found, in the first experiment, that when the waitress repeated the order with exact literal verbal behavior as the customers, she received higher tips than when she did not.In another experiment by Van Baaren, Holland, Steenaert and Van Knippenberg (2004), the mimicry was of posture of people evaluating an advertisement.A similar experiment was conducted, and similar results reflecting participant satisfaction were found.The experimenter mimicked the posture of the participants' (or did not mimic) during the advertisement evaluation.
After the evaluation was over, the experimenter dropped 6 pens.Where the participant was mimicked, the participant picked up the pens (100%), and only 33% of the non-mimicked participants helped pick up the pens.
The idea behind the Chameleon Effect was set by Chartand and Bargh in 1999.they first demonstrated the saying "monkey see, monkey do" to refer to intentional mimicry of goal-directed activities that these primates do.However, they noticed from conducting several experimental studies, that sometimes mimicry is unintentional.They then invented the term "Chameleon Effect" as it represents unintentional mimicry.The term "Chameleon Effect" was taken from Woody Allen's play "Zelig", who plays the role of a human chameleon-changing its color to match its current surroundings.This act is also defined in literature as unintentional mirroring.They then defined the Chameleon Effect clearly as being "the nonconscious mimicry of the postures, mannerism, facial expressions and other behaviors of one's current social environment."(Chartand & Bargh, 1999: 893).This mimicry is usually aimed at getting acceptance from the group or society a person is involved with.

Identity and Prestige
Since the focus in this study is on the relationship between a particular dialect, General American English, and linguistic or social prestige, it is necessary to explore some of the studies that have looked at this relationship.
Nordquist (2017) defines linguistic prestige as "the degree of esteem and social value attached by members of a speech community to certain languages, dialects, or features of a language variety."In his famous study on the accents of the people of New York, Labov identified two types of prestige, covert and overt, where the first is an aspiration to access a more dominant group and the latter is showing loyalty to a subculture or a secret limited membership group (Eckert, 2011).In studying the relationship between English and prestige in the Pakistani youth's language in social media, Sadiq (2015) finds a shift from Urdu to English among both the educated and uneducated users, explaining how English came to symbolize power, prestige, and status.
Academically, an accent is extremely powerful so much so that it might be a barrier towards learning.According to Al-Saidat (2009), language learners will be hindered if they have negative attitudes towards the variety of English used in the classroom.In his study, 56.8% out of 426 Jordanian college students do not see fluency and attaining native-like pronunciation as a priority or even a necessity despite their beliefs in the importance of the English language especially for instrumental goals such as a degree or a job opportunity.Although the attitudes towards English are positive, the Jordanian students did not find it more preferable than their mother tongue, Arabic.Language attitudes and preferences of a variety over another depend on experiences as well as emotional liking (McKenzie, 2010).Carrie (2017) indicates that both standard and non-standard varieties of English can have positive associations for native English speakers, since standard varieties are prestigious and non-standard varieties are socially attractive (solidarity).For example, she found that Received Pronunciation is a variety of prestige while General American is a variety of solidarity.She, also, speculated that the EFL learners participating in her study might benefit from both varieties of English, PR and GA, for both instrumental as well as integrative objectives respectively.
EFL learners in Zarrinabadi & Khodarahmi's (2017) study expressed clear preference for native-like accents and found them prestigious and linguistically knowledgeable.On the other hand, strong accents are stereotyped negatively as uneducated and illiterate.It is considered a privilege to speak like a native.Moreover, Sa'd and Modirkhamene (2015) explored attitudes of Iranian language learners towards accented and non-accented English.
They listed a number of reasons for this preference: intelligibility, beauty, resemblance to native speakers, English global status, originality, fluency, interest, sign of mastery, and prestige.On the other hand, 40 participants in the study preferred the mother tongue accent for also several reasons as follows: linguistic and ethnic identity, importance of communication over accent, and lack of importance.Interestingly, the researchers encouraged language teachers to improve their target language accent for credibility; at the same time, they suggested balancing it out by valuing L1 due to its social and cultural associations (2015).
In terms of the gender variable, women tend to be more sensitive to accent and tend to conform to standard language variety.In a study conducted by Zhang & Ding (2011) on Chinese speakers, it was found that male non-native speakers (47%) did not mind their Chinese accent yet the female participants (46%.7) did not like their Chinese accent.According to Kirkova-Naskova (2010), acquiring L2 pronunciation is not an easy process due to the interaction between a number of elements including L1 and L2 sound systems, language experiences, aptitude, and the learning environment.Accented English alone does not complicate intelligibility and communication like a combination of factors such as grammar, pronunciation, and prosodic errors.
Language is as much a part of identity and self-representation as culture, ethnicity, and religion.The multiple and diverse linguistic variables allow speakers to choose what expresses their identities and enables them to reach certain goals.Therefore, choosing to speak one language over another or one dialect or accent over another has deeper social, political, or personal implications.In some cases, the choice of accent is involuntary.
To identify a person's social affiliation, such as, nationality, ethnic background, regional membership and even age (Kinzler et al, 2009), the human mind tends to receive the language and process it efficiently to obtain salient and available information (Allport, 1954).Rakie, Stenffens & Mummendy (2011) found in a study they conducted, on an Italian looking man who speaks German fluently, that a language is a more important indicator of social variables than appearance.Another study conducted by Adank, Stewart, Connell & Wood (2013) found that regional accents are stronger implications of the affiliation to a regional community than real ethnic affiliation.In their study, they found that a speaker of the standard local accent, which is spoken by most individuals in the community, is perceived as more attractive socially and more competent than speakers of regional or foreign accents.
Stereotyping is the assumption that people with specific linguistic features belong to a certain identity.Kinzler, Dupoux & Spelke (2007) identified what they called "own accent-bias", which is the favoring of a person who speaks their accent, and hence are perceived positively.They found that infants as young as five months old had native accent preference.This idea enhanced the prestige of certain languages/accents over others.
It is well known that accents are identity cues.A person is given an identity as soon as he/she is born.When we use language, we do so as individuals with social histories.Our histories are defined in part by our affiliation to a collection of social groups into which we are born.These groups usually involve gender, social class, religion and ethnicity.For example, we are born as male or female and into a distinctive income level that defines us as poor, middle class or well-to-do.Likewise, we may be born as Christians, Jews, Muslims or with some other religious affiliation, and thus presume a specific identity ascribed to us by our particular religious connotation.Even the geographical region in which we, or even our grandparents, are born provides us with a certain group membership and upon our birth we undertake specific identities such as, southern, northern, and so on.Within national boundaries, we are defined by association in regional groups, and we take on identities; for example, in the Kuwaiti social context, people are defined by being from Ahal Sharq (Kuwaitis who use to live in an area called Sharq) or Ahal Jibla (Kuwaitis who used to live in an area called Jibla).
We hold similar expectations about what others are likely to do and not do as members of their particular groups.
The linguistic resources we use to communicate, and our interpretations of those used by others, are shaped by these mutually held perceptions.In short, who we are, who we think others are, and who others think we are, mediate in important ways our individual uses and evaluations of our linguistic actions in any communicative encounter.We chameleonize, unknowingly, by code switching, changing intonation and even sound production to adhere to our identity "requirements".In this study, we will explore the Chameleon Effect in light of accent change and consonants realization in an EFL context by trying to reach an answer to a set of questions.

Methodology
To answer the questions above, a methodological structure was designed.The study was conducted for a full academic year in Kuwait University, in two language classes.The participants were 37 female speakers and 28 male speakers, ages around 19-22.The students were given three topics to debate: 1.The use of smart phones is an unhealthy phenomenon.2. Crash diets are never successful.3. Early marriage is a disaster.The students were asked to discuss the topics in groups, and then participate in the in-class debates.The students were asked to record their group discussion out of class when they prepare for the debate.The teacher informed the students that in-class debates were recoded.Where the students were asked to record their group work, they used their smart phones.
The teacher, on the other hand, used an Olympus VN-7200 digital recorder.The groups were divided according to gender into groups of 4-5, and within these groups, they were divided systematically to have mixed GPAs in one group.A questionnaire, in the end of each term, was distributed asking the students about the importance of accents, and its reflection of intelligence.Some questions about debating were added to camouflage the focus of the questionnaire.Students of private schools were excluded later as it appeared that private schools graduates had a very different criteria.
The data obtained from the recordings was divided into three sets; namely, accent change (including vowel change), consonant [p] change, consonant [v].The recordings data were entered in SPSS, where every occurrence of the targeted sound was marked either as consonant or vowel change, and given a point per realization.When the consonant and vowel did not change in another context, this was also marked as one point per realization.Hence, the percentage of change was obtained.The questionnaire consisted of 18 questions, some of which required more detail, thus the students were provided space to elaborate.Independent factors of age, gender and GPA were also considered, to obtain significant results obtained from the correlation between dependent and independent factors.The results from the questionnaire were also entered into SPSS for frequency, mean, standard deviation and significance calculations.

Results
The results might not be so surprising at first.When listening to the out-of-class group discussions, male students used a lot of Kuwaiti, and would code-switch when writing their points.Their accent was majorly very heavy Kuwaiti-English with consonant changes.For example, [p/b], [v/f], rhotic [r] to an Arabic trill.
During the out-of-class discussions, the mean for pronouncing [p] as such in the male groups was 36.71.The use of [b] for [p] was excessive in their own groups, yet in class this realization changed and in the same words, such as in the words "program", "picture" and "place".The correct realization of [v] was a high m=65.61.When compared to the in-class discussion significance was found in the realization of [p] at f ≤0.01 (where significance was calculated at f=0.05).The mean realization of [p] raised to 43, while for [v] it only raised to 66.86.

Figure 1. The realization of [p] and [v] in two contexts
As can be seen from Figure 1, very little difference occurred in the realization of [v].
From the recordings, it was found that only two male students spoke in an American like accent out of class.Yet, in class, most male participant toned down their heavy Kuwait-English accent, displaying two more American-like accent speakers.As for the female participants, a serious attempt at speaking in English most of the time was highly evident.A lot of stuttering and reluctance in the girls' speech reflected the amount of effort they put into speaking in English most of the time.The accent fluctuated, reflecting the forceful attempt to sound American, dark apical [ɫ], hyper correction of English consonants also was clear with words like "viper", "possiply".
The As can be seen inFigure 2, the difference in [p] was more than it in [v].
In class, the female groups were under the pressure of winning a debate, their strong attempt to speak in an American accent seemed less, especially in the realization of dark apical [ɫ].It was difficult to recognize the number of students who spoke in an American accent out of class, 6 students seemed to have a good command of the accent, and similarly, in class, 6 female participants showed good command of the American-like accent.It appears from the data collected that female speakers believe in the prestige of the native like accent, in their groups and in class, they do their best in depicting what they know.In groups, however, they put a little more effort to sound like the group leader, who speaks very good English as the group leaders where chosen for their good language abilities, which increased hyper-correction by the other members.The Chameleon Effect in the speech of the female participants was shown in the increased change during one out-of-class session.It was clear from the word "program" for example, that one speaker changed its realization three times during a one-and-a-half-hour meeting from /brɔgram/ to / prɔgraem/, and then finally settling at /pɚɔgraem/.The majority of participants seemed to believe that the group leader knew best, and their speech would usually reflect hers.The students with less confidence, who had low GPA, would be those mostly under the influence of mirroring.Male speakers, on the other hand, have shown the Chameleon Effect significantly, in accent, intonation and even production of consonants regardless of their English fluency level.They mainly spoke in a Kuwaiti-English accent out of class, which changed somewhat into a toned-down Kuwaiti-English, and sometimes American-like accent.
In order to find the reason behind these changes in realizations of sounds and accents, a short questionnaire with 18 questions, 11 of which on accents, was given to the participants (administered by the authors).The first question enquired about the source of accent gained by the participants.The majority of the participants stated that they gained it from the media.A one-way ANOVA by gender in correlation with the first question reflected significance of f≤0.01.A closer look at the mean by gender indicated that male participants mostly replied "school" to this question, where m=1.93; whereas for the females m=2.27 (school=1, media=2).In terms of GPA, it appears that most A students stated that they learned their accent from the media, as opposed to the D students who mainly believe their accent is acquired from school.Statistically, the higher the GPA, the more the speaker feels that his/her language was learned from the media.
In the second question, the students were asked about their accent.Almost half the participants stated their accent was a Kuwaiti-English one, while the other half believed it was American.When splitting the data by gender, it appears that 53.6% of the male participants stated they speak in a Kuwaiti-English accent, as opposed to 43.2% of the female speakers.35.7% of the male students believed they speak in an American accent, and 54.1% of the female students stated they speak in an American accent.In a one-way ANOVA by gender in question 2, high significance was found at f≤0.01.This significance reflects that many male students believe they speak in a Kuwaiti-English accent (m=2.32), while most females believed they speak in an American accent (m=1.89).As seen in Table 1, the choices of accent seem to go around American and Kuwaiti English.
It appears from the analysis of the GPA and accent, that 70% of the A students believe their accent is American as opposed to about the same percentage of D students who stated they speak in a Kuwaiti English accent.As with the first question, it is clear that the higher the GPA, the more the speaker believes his/her accent is American.
When it comes to the Chameleon Effect, it was necessary to find whether the students were aware of the changes that occurred in and out of class.Hence, question three asked the participants if they found that their accent changed according to listener?44.6% of the participants replied sometimes, while 30.8% stated they never change their accent.When analyzing this question in terms of gender, it appears that the female participants believe they change their accent depending on the listener slightly more than the male speakers.This difference was not found to be significant.In addition, the question requested clarification.Many participants stated that they might change their accent with non-Kuwaitis, like when ordering at a restaurant or when speaking to the housekeepers at home.Only two male speakers stated that they change their accent when sitting in a group because speaking in an American accent is "not cool" and "stigmatized" among Kuwaiti youngsters.This finding proves that most speakers are touched by the Chameleon Effect, hence their accent changed in their groups and in class unconsciously.As evident in Table 2, the students with a GPA around 1.00 mostly believe that accents reflect intelligence.
It appears that 43.1% of the participants rarely feel that speaking in a native accent is prestigious across gender and GPA.However, 50.8% of the participants believe that a Kuwaiti with a native like accent is smart.When analyzing this question by students' GPA, it was found that the lower the GPA, the more the participant thinks that the native accent reflects intelligence.Many male students wrote in the explanation box that their friends don't find it "cool" or "smart" if they speak in an American accent.One participant wrote "if I speak like an American, I'd look like a show off not smart".Another participant said that it feels "acquired" when a Kuwaiti speaks like an American.This attitude was only reflected in the male groups, the females reflected pride in the American accent, even though some of them did not speak in an American accent but clearly believe they do.
The next three questions in the questionnaire aimed at investigating the importance of a native-like accent to the speakers, which would identify a reason for "chameleonizing" speech.The questions investigate the importance of acquiring a native accent "would you like to learn a native accent?","do you think it's important to speak in a native accent?', and "do you think you should be taught a native accent in school?".Although in the first two questions, the replies concentrated on rarely and sometimes, where the results in the first question were 21% sometimes and 22% rarely, and 26% and 18% respectively in the second question.A detailed investigation of this data shows discrepancy in terms of gender and GPA.
Figure 3.The question on the importance of language by gender Gender differences in the importance of native accents were eveident from the statistics.While 42% of the male participants rarely feel that they need to learn a native accent, and only 3.6% always would like to learn a native accent; 32.4% of the female speakers sometimes feel that learning an accent is desirable and 21.6% always have that desire.Nonetheless, most male and female participants stated that native accents should be taught in schools (sometimes by 51.4% and 42.9% successively).
When it comes to GPA, statistics show that the students with lower GPA are those who would like to learn a native accent (66.7%).The students with GPA of B (with GPA around 3.00) and C (with GPA around 2.00), mostly felt that they rarely would like to learn an accent (42.1% and 46.2%).Most students from different GPAs believe that it is sometimes important to learn a native accent, conversely, the students with the highest GPA were the least to believe that teaching an accent in schools is important.
The final two questions aimed to investigate the effect of chameleonizing on the recipients.Due to the importance of the question, the researcher included the language instructor (an American) in this part.The participants were asked to think of who they preferred most from their group and who they preferred least, and write the reasons in detail for their preference.The instructor chose two group leaders as her favorite, her reasons being their language ability.The students she chose had very good American-like accents.The male students, on the other hand, preferred the students who where "like the rest of us", humorous, and did not show off.One student named his preferred participant stating, "he spoke in Arabic, not like some who insisted to discuss the points in English (he) knows English but doesn't need to twist his tongue to be intelligent".Group leaders in the male group were the least preferred by most male participants because they used English in out-of-class group discussions most.However, female students mostly reflected on the students they liked least.The majority found that the students who did not participate, and "sat watching only" were considered not friendly and a burden.The female group leaders anonymously believed that the girls they preferred were smart and "intellectual", choosing the girls who tried speaking in an American accent.One female active student with an American-like accent stated that "the group leader is the best, very friendly, and we had a lot in common".Her group leader spoke in a similar accent.Some female students, who did not participate much in the out-of-class discussion, reported that the group leader was their least preferred as s/he "showed off" with her/his "American".

Discussion
This study provided a view and an indication of the types of factors that influence Kuwaiti speakers into unintentionally changing their accents when speaking English.A process we refer to in the present study as the 'Chameleon Effect' drawing from Chartand and Bargh (1999).One interesting aspect that stands out from the findings is the sociolinguistic factor of gender differences.Our results replicate previous findings on speech accommodation and speech variation, which argue that, unlike male speakers, female speakers have a higher tendency to make an effort to modify their speech based on social factors, such as who their hearer is and the type of interactional setting they are in (see Trudgill, 1986;Willemyns et al., 1997;Namy et al., 2002).Moreover, the study found that chameleonizing did affect the recipient, as generally the similarity of accents was a reason for being acceptable as a member of a group.
The data in this study reveal that female speakers have a high regard for the (standard) American English accent evidenced by the efforts exerted into moving away from a heavily Kuwaiti accented English and more towards sounding as American as they can possibly manage.Despite difficulties in pronunciation due to the obvious lack of equivalence of numerous English sounds in the Arabic phonetic system, such as [p] and [v] (among other sounds, see Taqi et al., 2015 and2018), the female participants in the present study seem willing to 'stutter' and 'hesitate' while trying to ensure that their realizations of these sounds (including the American dark apical [ɫ]) are almost native-speaker-like.Many of the female speakers identify themselves with and emulate the speech of their group leaders, who have been purposefully chosen to be among the most proficient in class.These students reported their liking of the "American" speaking peers.The group leaders in turn identify themselves with American native speakers, and also preferred the students who sounded like them.Whether linguistically competent or not, most female speakers taking part in this study, undeniably regard American English as a prestige accent.In fact, the less competent female students who struggle with low GPAs seem to consider their more fluent female classmates as intelligent simply because they can speak English with an American accent (see Table 1), yet at the same time, they reflected their dislike of the students who spoke in an American accent considering it a show off.And how better to counter an argument about language attitudes and identity than to bring forth perspectives from the other end of the continuum: speakers who linguistically identify more with their ethnic background as reflected by previous research too (see Hall, 1996, Eckert & Rickford, 2001;Eckert, 2005;Bucholtz & Hall, 2005;Rindal, 2010).The male participants regularly followed the theory of Chameleon Effect by ensuring that even the preferred member of their group, like them, spoke English with aKuwaiti accent rather than an American accent.
Altogether, the results support the idea that the male speakers in this study are far less likely to pronounce correct realizations of [v] and [p].Yet this aspect is less prevalent with the former sound ([v]) both inside and outside the classroom, while it is more prevalent with the latter sound ([p]) mainly outside the classroom, where students are generally more at ease (see Chart 1).In fact, the male speakers were far less concerned with correct realizations of [p] and [v] that, unlike female speakers; they demonstrated almost no instances of hypercorrection.The findings in the present study suggest that hypercorrection is a more prominent feature in female participants' speech.Moreover, according to comments made by male speakers when answering the questionnaire, using an American accent when speaking English among the male population in the present study is not seen favorably and even shied away from as much as possible.This, yet again, reinforces views from previous research, which point to male speakers favoring the use of non-standard regional accents when speaking English (see Gordon, 2005;Coates & Spittal, 2015).These comments are further emphasized when the male speakers' responses from the questionnaire tended to undermine the importance of acquiring correct English language pronunciation.This brings us to the second part of our discussion, namely the perceptions of male and female speakers over the importance of learning to speak English with a native speaker accent.
Regardless of the methods with which Kuwaiti speakers learn correct American accent pronunciation (although it has to be pointed out these methods are mostly, either schooling and/or daily exposure to American English through the media), only those who are highly motivated to learn will do so, and quite successfully.Another interesting perspective observed in our findings, which is again consistent with other findings across the gender variable, is that many female participants seem to believe that learning to speak English with an American accent is important (see Table 1).However, students with high GPAs from the female population are less keen on acknowledging this importance.Likewise, male participants prefer not to place much priority on being taught native speaker English pronunciation (see Figure 3).In general however, Kuwaiti EFL learners (regardless to gender) are much more likely to acquire an American accent from the media than they are from school; given that English high school teachers in Kuwait do not usually have native English speaker competency.
In accordance with our results, it is mostly the high achieving female speakers who link the factor of regular exposure to American movies and soap operas with successfully 'picking up' an American accent (see Carvalho, 2004;Rindal, 2010).A phenomenon, which seems to be increasingly prevalent among college students in Kuwait and hence worth further investigation in future research.The other participants in the present study, namely the male and female speakers with low GPAs seem to think that correct pronunciation when learning English can only be taught in school.This is an unlikely prospect since the vast majority of the students who enter public colleges in Kuwait tend to have a very basic command of both English pronunciation and syntactic competence.Our discussion points to the conclusion that not all the participants in the present study value speaking English with an American accent as an asset, but those who do, are highly motivated learners and are therefore usually self-taught via frequent exposure to American media.

Conclusion
The following study investigated the occurrence of the Chameleon Effect, and its implications and causes in a Kuwaiti context.65 young Kuwaiti male and female participants recorded their meetings and debates in and out of the classroom.The participants were also given a questionnaire to comment on the importance of native accents, and the implications of accents in their points of view.The data analyzed reflected a clear change of accents according to context, which was a reflection of the Chameleon Effect.The majority of the male and female participants, as revealed in the questionnaire, seemed to prefer the students that spoke "like them".In the male group, most participants reverted to the use of a Kuwait-English accent to assert their identity and belonging to the group.The female participants, on the other hand, attempted to speak in an American-like accent in order to gain prestige in their group.In either way, most participants believed that the ability to speak a native accent might be a reflection of intelligence.
As English teachers and learners, our views on the importance of pronunciation significantly determine how we should approach teaching and learning pronunciation.Although it is generally recommended to approach English as an International Language, avoiding the quest for a native-like accent, many students in this study reflected their preference to be taught a native accent.While not stressing the issue of accent learning, in the classroom, teachers could work on prosodic instruction to teach general English (either American or British) and/or they could apply more focused instruction of individual consonants and vowels while students listen and speak out loud carefully to be able to hear their own realizations of sounds.While outside the classroom, listening to English songs can be the best accent teacher because songs have rhythm and tone that would allow learners to spontaneously tune in.Also, let us not underestimate the effect that the media (especially, American movies and soap operas) has on regular viewers.
It must be said that while teaching and learning correct and appropriate standard pronunciation is recommended, it should not supersede intelligibility.Successful communication is everyone's potential goal and unlike the idea of accented speech, it is not influenced by sociolinguistic factors such as identity or peer pressure.Interestingly, the students in the present study intentionally spoke English with a Kuwaiti accent believing in doing so that it is an identity preservation strategy especially when interacting with likeminded peers and striving to be accepted by the group, which can at times be overwhelming for a learner.
When engaging in group discussion in class, teachers could encourage learners to decide for themselves whether they would like to speak English with an American accent or a Kuwaiti one so long as they speak comfortably, confidently and intelligibly.Affecting an accent when possible (for not all learners are able to do so) is mostly an individual's choice and tends to occur in connection with various sociolinguistic factors such as identity, ethnic background, social class, gender and so on.Educators have to become more aware that the Chameleon Effect in specific settings (such as the one in the present study) is the student's way of wanting to make a public statement.
In fact, students may view performing a certain accent like a linguistic accessory akin to choosing one's dress code to express their identity to the world outside.

Figure 2 .
Figure 2. The realization of [p] and [v] by female in and out of class never don't know

Table 1 .
Accents spoken according to participants

Table 2 .
The percentage of question 3 by GPA