Language Choice in Online Written Communication among Maldivian Professionals

Language choice is a sociolinguistic phenomenon which refers to the selection of language(s) for different purposes in various contexts. This study examines the language choice of Maldivians in online written communication in the domains of family, friendship and work. Dhivehi is the unique Indo-Aryan language spoken almost exclusively in the Maldives along with English as the unofficial second language. A total of 150 professionals from the fields of education, administration and health took part in the study with academic qualifications ranging from diploma to doctorate degree. The data was collected using a survey questionnaire and analysed quantitatively. The findings revealed that Dhivehi is the dominant language in the family domain while English is dominant in the friendship and work domains. There is also a relationship between language choice and age, and English language proficiency. The high preference for English in online written communication indicates a need for a language policy that would emphasise and increase the use of Dhivehi online as a language maintenance effort


INTRODUCTION
Language choice occurs in situations where there are two or more languages to choose from, in various contexts and for different purposes.A language that is suitable for one purpose or a certain situation may not be suitable for another.The choice of language is also influenced by broader sociolinguistic factors such as the status of English as a 'global' language (Seargeant Tagg & Ngampramuan 2012).
In this era of technology, online communication via Instant Messaging (IM) applications such as email, Viber, Facebook and WhatsApp among others are quite common and preferred by many people as they are faster and more convenient.It is obvious that similar to face-to-face communication, interlocutors in online communication have a choice of language in a bilingual or multilingual context.Thus far, only a few studies have been carried out to investigate the language choice in an online context (Androutsopoulos 2006, Gerrard & Nakamura 2004, Warschauer, El Said & Zohry 2002).Even fewer studies have been done in a bilingual context like in Maldives with a unique language -Dhivehi.
Dhivehi font is not widely available in the IM applications that are popular in Maldives.Recently, a new feature introduced in some IM applications allows users to use Dhivehi font, but it is not commonly utilised probably due to the difficulty in placing the diacritic marks for each of the letter or the alphabet.Those who do use Dhivehi online, usually use English alphabet for writing statuses and chatting.
Many Maldivians are avid users of the Internet and being bilingual the speakers have a choice on which language to use in different situations (Holmes 2004, Muhammed, Farrag, Elshamly & Abdel-Ghaffar 2011, Seargeant, Tagg & Ngampramuan 2012).Working professionals are one category of Internet users who are required to communicate online in different situations.They would be the ones who need to give importance to English as a lingua franca while at the same time maintaining their native language.The respondents of this study come from various educational backgrounds and workplaces and they have different levels of English language proficiency.This study would provide insights into how language proficiency and educational backgrounds influence their language choice.In the workplace, all online written communication should be carried out in Dhivehi as it is the official language.Therefore, there is a need to find out whether Internet has an impact on the language choice of this speech community and more importantly, how it has affected the use of Dhivehi in online communication.

LINGUISTIC BACKGROUND OF MALDIVES
Maldives is a small island nation comprising 1192 islands and is homogenous with regards to ethnicity, religion, culture and most importantly the language.Dhivehi is an Indo-Aryan language spoken exclusively by the population of 336,921 people in Maldives except for the 10,000 speakers in the island of Minicoy in the Union territory of Lakshadweep of India.Dhivehi is a language closely related to Sinhalese (spoken in Sri Lanka) although these two languages are not mutually intelligible.
There are four major dialects in Dhivehi which include Huvadhu, Mulaku, Addu and Male' which is the standard dialect.Language use in Maldives can be considered as diglossic where varieties of a language are used for different social functions in a speech community (Ferguson 1959).The standard variety Male' is used for all official purposes and in schools to teach Dhivehi language and Islamic studies.The writing system is also based on the standard variety.Dhivehi is written using Thaana script which consists of 24 letters some of which are derived from Arabic numerals (Salahuddin 1928 as cited in Mohamed 2006).Vowels are indicated with the Arabic system of diacritics which appear above or below the consonants.Similar to Arabic, Dhivehi is written from right to left.
Before gaining full independence in 1965, Maldives was dominated by the colonial powers of Portugal, the Netherlands as well as Britain .Due to the fact that Dhivehi has no use outside the country, Maldivians gave great importance to foreign language learning and as a consequence English was made the medium of instruction (Mohamed 2013).Proficiency in English is widely regarded as the means to succeed academically, by providing opportunities to go abroad for higher education and in securing profitable employment in both private and public sectors (Mohamed 2013).
It is stated in the constitution of Maldives that the official language of the country is Dhivehi.However, almost every Maldivian understands English to some extent as it has been the medium of instruction for the last 50 years and owing to the booming tourism industry.The use of English is becoming increasingly popular with the fast growth of technology especially on social networks and email communications (Mohamed 2013).Internet was first introduced to the country in 1996.Since then, Maldives has built one of the most advanced telecommunication systems in the region with 229,417 internet subscribers as of 2015.In addition, the government has implemented various policies that emphasise the use of ICT in teaching.

LANGUAGE CHOICE AND DOMAIN ANALYSIS
In defining or explaining language choice, domains are very important.Fishman (1965) defined domains as "institutional contexts or socio-ecological co-occurrences" and proposed that a particular language or varieties of language are used in different domains for different purposes (p.73).He also explains that domains are a collection of factors such as participants, location and topic.Fishman (1972) defined domain analysis as "who speaks what language to whom and when those speech communities that are characterised by widespread and relatively stable multilingualism" (p.437).
In this regard, domain analysis is also related to diglossia which is the use of two dialects or two closely related languages by the same speech community.In such a community, the low language or variety is usually used in the family domain while the high language or variety is used in more formal domains such as education.A study by Parasher (1980) in India showed that English dominated high domains like government, employment and education as well as some low domains such as friendship.Hohenthal (2003) as cited in Rahman, Heng and Abdullah (2008) found similar results in a study of the role of English in India.It was reported that different languages were used differently according to the domain.According to Dumanig, David and Shanmuganathan's (2013) study on Filipino-Malaysian families, English is the dominant language used and the choice of language is based on the acceptance of English as the dominant language over the other native languages.Sahgal's (1991) study on three communities, namely, Hindi, Bengalis and Tamilians in Delhi revealed similar results.English was found to be the main language in the friendship and institutional domains and, to an extent, in the family domain.The obvious implication that can be gauged from this is that English is "gradually becoming associated with intimacy, spontaneity, and informality, along with its use in education, administration and the mass media" (Sahgal 1991, p. 305).
At this juncture, it is important to highlight the differences between face-to-face communication and online written communication.It is believed that there are various features in face-to-face communication like non-verbal cues and gestures that are not present in online written communication.However, it can be argued that the Internet users have found a better alternative for those features in the form of emoticons, stickers and animations that are available in IM application these days.
One of the earliest studies on these differences is Warschauer's (1995) study on ESL students which revealed that they used syntactically and lexically more complex and formal language in electronic discussions than they did in face-to-face communication.However, more recent studies showed how the language used online is a combination of both written and spoken features (Crystal 2006;Tagliamonte & Denis 2008).Due to these hybrid features, the language used in online written communication was termed 'Netspeak' by Crystal (2006).The most important factor that is discussed in many studies is the response time or the speed of the replies while chatting.The speed may be slower in online communication but due to the spoken language features and the sense of urgency to reply, the reaction time is quite fast (Danet 2001,Walther 2007).

LINGUISTIC DIVERSITY ON THE INTERNET
Ever since the beginning of the global spread of the Internet, there have been debates on its effect on various issues.The most common issue is the possible impact of English on other languages.While some studies focus on the Internet leading to the dominance of English and eventually crowding out small or indigenous languages, others believe that it actually helps in revitalising such languages (Warschauer 1998(Warschauer , 2001)).Herring (2002) focused on the effects the Internet has on the global linguistic ecology.Her two focal points were the Internet accelerating the global dominance of English and other languages catching up or surpassing English on the Internet.It is an undeniable fact that English has the historical primacy on the Internet in terms of technology, that is, all the fonts, code, browsers and search engines are in English (Yates 1996 as cited in Herring 2002).
The predominance of English language websites and non-native speakers creating websites in English has led to the increase in the frequency of English used by non-native speakers on the Internet.English has a wide influence on other languages as they borrow English computing terms, use English computer-mediated communication (CMC) conventions and mix English with other languages both in email and chat (Herring 2002).However, there is a change recently, especially with most of the major languages having a number of websites of their own, owing to technological advancement.
Androutsopoulos ' (2006) study on the language used by some migrant groups in Germany showed that the migrants chose German instead of their native tongue or even English.This shows that situational factors play a role in the language choice and might also be because German is one of the major languages in the world.Similarly, Climent et al.'s (2003) study shows that Spanish is the lingua franca among different ethnic populations in Spain.However, for speakers of mutually unintelligible languages, English is found to serve as a lingua franca (Durham 2003, Wodak & Wright 2007).
In some instances, Internet users may feel that English is not semantically appropriate to express the exact ideas they want to convey and they might have to switch back to their mother tongue.Warschauer, El Said and Zohry (2002) reported that participants were not very comfortable communicating certain personal thoughts in English.

LANGUAGE CHOICE IN ONLINE WRITTEN COMMUNICATION
Only a few studies were conducted on language choice in online written communication.Warschauer, El Said and Zohry (2002) investigated the use of English and Arabic online by a group of young professionals in Egypt.Their findings indicated that the participants used mostly English in formal email but a Romanised version of Egyptian Arabic in informal email and chats.Peel's (2004) study on students from a higher educational institute in United Arab Emirates also revealed similar results.He found that although the students used mainly English for purposes related to their studies, Arabic was used when emailing or chatting with friends.The intriguing finding is that they chose to use Arabic in Roman alphabet (due to lack of Arabic keyboard or font) rather than English.Despite the use of English in some situations, the results indicated that the shift to English was not as extensive as was expected.Holmes (2004) in his study found that 2267 students in eight different countries who consistently used their native language instead of English in the Internet applications are from countries with prestigious national languages with a vast number of speakers.However, a more recent study by Muhammed, Farrag, Elshamly, and Abdel-Ghaffar (2011) showed that users find it easier and faster to type in English characters than in Arabic characters as they feel Arabic characters are not technologically friendly.
It was found that English is used online by people who share the same first language.Warschauer, El Said and Zohry's (2002) study revealed that even in informal online conversations, the participants code-switched between Arabic and English.A similar study by Yusuf, Natsir and Yusra (2016) on the linguistic and discoursal features of short messaging service indicated that English was the dominant language among the five different languages used.Language choice online may not be the same as in face-to-face conversations as shown by Seargeant, Tagg and Ngampramuan's (2012) study on a native Thai speaking community.The participants consistently code-switched to English in online situations even though they asserted that they would use Thai in real conversations.However, Seargeant, Tagg and Ngampramuan (2012) posit that rather than a matter of choosing codes, it was more about "drawing in various ways on a shared set of semiotic resources which in this case include English alongside their first language" (pp.528-529).English can also be more prevalent when the participants are more proficient in English (Alabdulqader, Alshehri, Almurshad, Alothman & Alhakbani 2014).

THEORETICAL FRAMEWORK
The theoretical framework for this study is based on Fishman's (1965) concept of domains and domain analysis.Fishman (1965) explains how language choice within multilingual groups is not simply random, rather the "'proper' usage dictates that only one of the theoretically co-available languages or varieties will be chosen by particular classes of interlocutors on particular kinds of occasions to discuss particular topics" (p.67-68).Although the topic of the conversation, place and interlocutor are important elements of the choice of language, it is the concept of 'domain' that actually determines the language choice.Thus, Fishman (1972) introduced the five domains of language choice in a speech community, namely, family, friendship, religion, education, and employment.In various studies, some of the domains such as family and work are referred to as home and office or work.Although education may also be a popular domain in online communication, some of the respondents of this study come from professions other than education (i.e.health, administration), therefore, education is categorised under work domain.Religion was excluded as it is not discussed usually as a separate topic in online conversations.Hence, the three domains of family, friendship and work are used in the study as these domains are deemed the most common in online communication.
The current study seeks to examine the language choice of Maldivians in online written communication.It aims to address two research questions: 1) Which language is more dominantly used by Maldivians in online written communication in the domains of family, friendship and work?2) What is the relationship between age, gender, level of educational attainment, English language proficiency and the language choice in the domains of family, friendship and work?

METHODS
The study was conducted in Hithadhoo island of Addu City, Maldives.It is the largest island where major schools and government offices are found.Convenience sampling was used to select the respondents due to their close proximity to the researcher.

RESPONDENTS
There were 150 professionals working in the fields of education, health and administration who took part in the study.The age of the respondents ranged from 20 to 59 years old.The majority of the respondents were in the age range of 20 to 39.All the eligible respondents, with a minimum diploma qualification, were selected from three schools, one university, one hospital and the city council office.This generation of Internet users would provide more valid and insightful data as compared to the younger or older generations.This is because the teenagers would use more English due to the exposure at school and those who are above 60 may not know English at all.

INSTRUMENT
A survey questionnaire was used to find out the respondents' online language choice.The questionnaire was adapted from Yeh, Chan and Cheng's ( 2004) study on language use in Taiwan and Warschauer, El Said and Zohry's (2002) study on language choice online in Egypt.The survey questionnaire comprises 23 items which are divided into three parts.The first part elicits the demographic profile of the respondents while the second part elicits information on the level of English proficiency and the frequency of use of the Internet for chatting and email.The last part seeks information on the language choice in the three different domains of family, friendship and work.A six-point Likert scale is used ranging from 'Never' to 'Always' where 1= Never, 2= Rarely, 3= Sometimes, 4= Frequently, 5= Always and 0= Not Applicable.The option 'Sometimes' is given to avoid a forced choice.'Not Applicable' is also used as one option as some of the situations like communicating online with parents might not be applicable to some respondents.The questionnaire was piloted and the reliability index was 0.78 (Cronbach's Alpha).It was later distributed to all the respondents over a duration of two weeks.

DATA ANALYSIS
Out of the 180 questionnaires that were distributed, 163 questionnaires were returned.Due to a few problems like missing data and missing signature on the consent form, 13 were discarded.Descriptive and inferential statistics were generated using SPSS 20.
The English language proficiency of the participants was obtained by the participants' self assessment of the four skills of language -listening, speaking, reading and writing.A five-point Likert scale was used where 1= No skills at all, 2= Unsatisfactory, 3= Satisfactory, 4= Fluent and 5= Very fluent.An aggregated score was assigned to each participant based on the marks they obtained for each skill, with a possible minimum mark of four and maximum mark of 20.The final score for each participant was again categorised into three levels: Low (1 to 10 marks), Average (11 to 15) and High (16 to 20).
Language choice of the participants was obtained through the six-point Likert scale.The respondents were required to indicate on the scale for both the Dhivehi and English languages.A total aggregated score for each situation was calculated by adding the scores for all the respondents.
Chi-square test of independence was conducted to determine the relationship between the language choice patterns in the domains of family, friendship, and work with age.In order to determine whether there is a relationship between the variables, χ 2 test of independence was carried out with α = .05as level of significance.
Cramer's V was used to assess the strength of the relationship between the two variables.For this research, the effect size for Cramer's V was measured using Rea and Parker's (1992) effect size values and descriptors.

RESULTS AND DISCUSSION
Among the 150 respondents who took part in this study, the respondents ranging from age 20 to 29 were the highest (44.7%) closely followed by the 30 to 39 age range (40%).Only 11.3% of the respondents were in the 40 to 49 age range.A minimum qualification of a Diploma was a requirement for all who took part in the study which may be the reason for the low number of respondents in the 50 to 59 age range (4%) who started working at a time when such qualifications were not given much importance.The percentage of the female respondents is far greater (84.7%) than the male respondents (15.3%) which reflects the female dominance in the fields of education and health.
The percentage of the respondents with diploma level qualification was the highest (52.7%) followed by graduates (36.7%).As for the master's level qualifications, there were only 9.3% of the respondents.The PhD qualifications were the lowest (1.3%) which is not surprising as PhD qualifications are not available in the country yet.The English language proficiency of the respondents was obtained from the respondents' self-assessment of the four skills of language -listening, speaking, reading and writing.The majority of the respondents (62%) have very high proficiency in English which may be due to the English medium of instruction that is practised in Maldives.The percentage of the respondents with average English language proficiency is 34.7% while only 3.3% have a low proficiency.

LANGUAGE CHOICE PATTERNS IN THE DOMAINS OF FAMILY, FRIENDSHIP AND WORK
Figure 1 shows the language choice in the domains of family, friendship and work.Dhivehi is the dominant language in the family domain chosen by 51.4% of the respondents while English was chosen by 48.6%.On the other hand, in the friendship domain, English is the dominant language compared to Dhivehi (55.2% and 44.8% respectively).Similarly, in the work domain English is the dominant language chosen by 61.4% respondents while Dhivehi was chosen by 38.6%, which is more marked than the family and friendship domains.Therefore, it is evident that language choice is domain specific even in online written communication.It is apparent that Dhivehi is only dominant in the family domain.The difference between the use of Dhivehi and English in this domain is much smaller compared to the differences in the friendship and work domains.This might be because some of the respondents feel more comfortable using Dhivehi to express their personal thoughts, similar to the findings of Warschauer, El Said and Zohry (2002).As the family domain would consist of members of different age groups with various levels of English language proficiency, it is not surprising that the use of Dhivehi and English is almost at the same level.There would be situations where the younger family members would need to use Dhivehi in their online written communication, as the older members might not understand them or maybe because English is not semantically appropriate to convey their exact message.In the friendship domain, there is not a vast difference between the use of English and Dhivehi compared to the work domain.People from the more rural areas may not be as proficient in English as the respondents.Hence, the respondents might have to use Dhivehi while communicating with them in the friendship domain.According to Alabdulqader et al. (2014), English was the dominant language due to the participants' proficiency in English.This might be the reason why the respondents used mainly English in the work domain.

FAMILY DOMAIN
There are eight items in the questionnaire which sought information on the language choice while communicating online with parents, siblings, spouse/fiancé and relatives.The topic of conversation was found to be an important factor influencing language choice in previous studies (Appel & Muysken 1987, Saville-Troike 1989).In communicating with parents on personal and general topics, it can be seen that Dhivehi is used more often than English in both situations (refer to Table 1).This could be due to the fact that most of the parents of the respondents may not be fluent enough to communicate in English online.Therefore, their children have to resort to using Dhivehi when chatting online.It is quite interesting to see that preference for English on personal topics was slightly higher (39.1%) than Dhivehi (36.9%) while preference for Dhivehi was higher for general topics (63.1%) than personal topics (60.9%).On the other hand, when communicating with siblings on both personal and general topics English is the preferred language.The difference could be because most of the siblings are more familiar with technology and they use the Internet to communicate with each other.In terms of the topic of conversation, there is not much difference except Dhivehi is used slightly more in communicating general topics (46.9%) rather than personal topics (46.7%).
As indicated in the results, English is used more often than Dhivehi in online communication with spouse or fiancé on both personal and general topics.There is a slight difference in the way both languages are used.Dhivehi is preferred more for expressing personal matters (47.6%) than general topics (47.3%) while the reverse is true for English.However, these differences are too negligible to be considered.
The last situation in the family domain is the language choice in communicating with relatives.In this case, Dhivehi is used more than English in communicating both personal and general topics.For personal topics, 52.4% chose Dhivehi while 47.6% chose English.For general topics, 52.9% chose Dhivehi while 47.1% chose English.Similar to the situation for parents, the high preference for Dhivehi could be due to relatives, such as aunts and uncles, from the older generation who do not speak English fluently.

FRIENDSHIP DOMAIN
There are four items in the questionnaire which sought information on the language choice while communicating online with close friends and other friends (refer to Table 2).The first situation in this domain is on language choice in communicating with close friends.The results indicated that English is used more than Dhivehi in communicating personal and general topics.For personal topics, 55.4% chose English while 44.6% chose Dhivehi.For general topics, 55.7% chose English while 44.3% chose Dhivehi.Unlike parents and relatives, most of the respondents would be proficient enough in English and more tech savvy.Therefore, English is preferred as it is also a language commonly used on the Internet.Similar to the situations above, the use of English when communicating online with other friends is considerably higher than the use of Dhivehi.For personal topics, 55.3% chose English while 44.7% chose Dhivehi.For general topics, 54.4% chose English while 45.6% chose Dhivehi.There is not much difference between choice of language when communicating with close friends and other friends.This could be due to the high proficiency in English among the same circle of friends.These results are in line with previous studies on language choice in the friendship domain where English was found as the dominant language (Parasher 1980, Sahgal 1991).

WORK DOMAIN
There are six items which sought information on the language choice while communicating online in the work domain.As the focus was on workplace communication, personal and general topics were not taken into consideration; instead, the respondents were required to provide information on different online situations that they might come across in the work domain.As indicated in the questionnaire and personally clarified by the researcher, all the responses were based on communication only among Maldivians.It can be seen that the respondents used English more than Dhivehi in all the situations (refer to Table 3).For writing formal and informal email, English was chosen by 69.7% and 66.2% of the respondents, respectively.Both percentages for English doubled the percentages for Dhivehi.The use of English when writing formal email may be due to the fact that the official language is Dhivehi and emailing is not available in the Dhivehi font.Letters and documents written in Dhivehi are attached to the email.The accompanying message with the attachment is written in English which might be the reason for the high percentage of English preference in this situation.Similarly, even in writing informal email, English is used more than Dhivehi for the same reasons.
The results also show that English is used more often than Dhivehi when communicating online with colleagues and supervisors.English was chosen by 59.3% and 58.7% of the respondents respectively.Those working in the professional environment are expected to have sufficient proficiency in English.Hence, it is not surprising that English is used more when communicating online with colleagues and supervisors.As the majority of the respondents are from the education and health sector, they are required to use English a lot even in face-to-face communication.This could be the reason for the dominant use of English in online communication as well.
English is preferred while communicating with clients and customers although the difference is not as much as in the previous situations.English was chosen by 52.8% while Dhivehi was chosen by 42.7% of the respondents.
In the last situation of the work domain, which is communicating online with students, it can be seen that English is used more often than Dhivehi.English was chosen by 60.2% while Dhivehi was chosen by 39.8% of the respondents.As the medium of instruction is English, students are not expected to speak in Dhivehi even in face-to-face communication which might be the reason for the high percentage of English use online.

RELATIONSHIP BETWEEN LANGUAGE CHOICE PATTERNS AND AGE, GENDER, LEVEL OF EDUCATIONAL ATTAINMENT AND ENGLISH LANGUAGE PROFICIENCY
Chi-square test of independence was conducted to determine the relationship between the language choice patterns in the three domains of family, friendship, and work with age, gender, level of educational attainment and English language proficiency.

RELATIONSHIP BETWEEN LANGUAGE CHOICE PATTERNS AND AGE
Table 4 shows the results from the Chi-square tests.According to the χ 2 test of independence, there is a very strong evidence of a relationship between age and language choice in the family domain, Likelihood Ratio  Table 5 shows the strength of the relationship between language choice and age.For the family domain, Cramer's V = .311indicates a moderate association between the variables.For the friendship domain and the work domain Cramer's V are .449and .425respectively, indicating a relatively strong association between the variables.The results from the analysis are in line with previous studies that found a relationship between the language choice and age (Huang 1988;Yeh et al. 2004).Although the aforementioned studies were done in an offline context, it is interesting to note that the results are similar even in an online context with written communication.There is no evidence of a relationship between gender and language choice in the family domain, χ 2 (1, N = 150) = .026,p> 0.05 (refer to Table 6).There is also no evidence of a relationship between gender and the language choice in the friendship and work domains with the Likelihood Ratio χ 2 (1, N = 150) = .556,p> 0.05 (friendship domain) and Likelihood Ratio χ 2 (1, N = 150) = .421,p> 0.05 (work domain).The results from the analysis are similar to Chan's (1994) cited in Yeh et al. (2004) and Rahman's et al. (2008) studies.It was proven in those studies as well that gender is not a contributing factor to language choice.

RELATIONSHIP BETWEEN LANGUAGE CHOICE PATTERNS AND LEVEL OF EDUCATIONAL ATTAINMENT
There is no evidence of a relationship between the level of educational attainment and the language choice in the family domain, Likelihood Ratio χ 2 (3, N = 150) = 4.048, p> 0.05 (refer to Table 7).No evidence of a relationship is found in the friendship and work domains as well with the Likelihood Ratio χ 2 (3, N = 150) = 2.041, p> 0.05 (friendship domain) and Likelihood Ratio χ 2 (3, N = 150) = 6.681, p> 0.05.There is a very strong evidence of a relationship between English language proficiency and the language choice in the family domain, Likelihood Ratio χ 2 (2, N = 150) = 14.592, p< 0.05 (refer to  Table 9 shows the strength of the relationship between language choice and English language proficiency.For the family domain, Cramer's V = .300indicates a moderate association between the variables.For the friendship domain and the work domain Cramer's V = .497and .489respectively, indicating a relatively strong association between the variables.The result on the relationship between proficiency and language choice was similar to studies conducted in face-to-face communication (Rahman et al., 2008;Yeh et al., 2004).Warschauer, El Said and Zohry's (2002) research on the use of English and Arabic in an online context also showed similar results.

CONCLUSION
The findings prove that language choice does depend on the domain of use, in line with Fishman's (1972) domain analysis theory.Dhivehi is the dominant language in the family domain while English dominates both friendship and work domains.As these three domains are commonly associated with online communication for the respondents in this study, this implies that English is prominently used in online communication in Maldives.
Even though Dhivehi is dominant in the family domain, it is mainly used to communicate with parents and relatives whereas English is used for communicating with siblings and spouse/fiancé.Previous studies have focused on the reaction time of the participants while chatting online (Danet 2001, Walther 2007).The results from the present study indicate that even with the fast reaction time during online chatting, the dominant choice of language is English, which in turn has serious implications.As these respondents represent the present generation, the use of Dhivehi in online written communication might be completely replaced by English in the near future.In addition, the Chi-square tests show an association between language choice and age and English language proficiency of the participants in all the three domains.
Therefore, it is crucial that Dhivehi is used and maintained in online communication although proficiency in English is important in this globalised world.This becomes more vital as Maldives is a homogeneous country with regards to language, ethnicity, culture and religion.In this present era, a large part of the daily life is spent online.The increased use of English online may affect the use of Dhivehi in the long run.Hence, it is imperative that Maldivians use online chatting applications as a platform to maintain their native language and regard the Dhivehi language as a valuable resource to be transmitted to the next generation.Furthermore, due to the English medium of instruction used in Maldives, there is a need for a language policy to place the learning of Dhivehi on the same pedestal as the learning of English.

SUGGESTIONS FOR FURTHER RESEARCH
As this study focused only on professionals' language choice online, it would be interesting to find out the results that will be generated if this study is carried out on the current generation of teenagers who are more technology savvy and started using Internet from a very young age.Furthermore, this research could be extended to study the language maintenance efforts of Maldivians, as this research has revealed that English is preferred compared to Dhivehi in the three investigated domains.Thank you for your cooperation

TABLE 1 .
Family domain

TABLE 3 .
Work domain

TABLE 4 .
Relationship between language choice and age

TABLE 5 .
Strength of relationship between language choice and age

TABLE 6 .
Relationship between language choice and gender

TABLE 8 .
Relationship between language choice and English language proficiency

TABLE 9 .
Strength of the relationship between language choice and English language proficiency Family domain 6.When I communicate online with my parents on personal matters, I use