Changes in Code-Switching Patterns among Hindi-English Bilinguals in Northern India

This study investigates changes in code-switching (CS) patterns among Hindi-English bilinguals in Northern India. There is a dearth of studies of naturalistic Hindi-English CS, which this study attempts to address. By recording a group of three older speakers and one of three younger speakers, the study identifies generational differences in CS behaviour. Echoing Si’s (2010) results, it finds younger speakers using more English overall, preferring to alternate between fully English or fully Hindi clauses, while older speakers insert English items into Hindi clauses more often. Finally, the paper investigates how CS reflects and aids speakers’ (linguistic) identity construction in modern India.


The Hindi-English Bilingual Community in India
Anyone wanting to study changes in code-switching (CS) patterns over time has the great challenge of finding historical sources of informal conversational data.Aung Si's (2010) study of Hindi-English CS in Bollywood films across three decades (1980s, 1990s, 2000s) allows some insight into change in this medium and has inspired the present study of generational differences in real-life CS among Hindi-English bilinguals.Si describes a general trend towards a "more English-intensive form of code-switching" (Si 2010:399).The increased amount of English in commercial film CS is mainly produced by young, urban upper-middle-class characters (Si 2010).This younger generation of the post-2000 films bears close resemblance to the educated younger speakers of the current study, born in the 1980s and instructed completely through the medium of English in the 1990s and 2000s.The characters of the older films are the young adults of the 1970s and 1980s.They correspond to the older urban middle-class generation of speakers of this study, born largely in the 1950s and educated at Hindi-medium high schools and English-medium universities in the 1960s and 1970s.Kachru (2004, cited in Graddol 2010:66) suggests that now over one third of Indians, equalling some 333 million people, "use English", while during most of the twentieth century only about 5% had been thought to do so.The most dramatic increase has followed the economic liberalisation of the 1990s, as English has become ever more popular and firmly established itself, especially as the language of the modern, urban young (Sonntag 2000:140).Demand for English-medium education is continually rising in India, as more and more people strive for their share of India's growing wealth (Simpson 2007).Increases in English-medium education go hand in hand with new opportunities, especially in the information and communication technology sectors (Graddol 2010), and the Internet brings English from all over the world into an increasing number of Indians' homes, making bilingualism in English ever more common.
CS patterns and the amount of English in CS produced by Hindi-English bilinguals in India are necessarily determined by age and level of English-medium education, as well as by socio-economic status.Those competent in English can be placed on a spectrum: the lower classes are eager to acquire English as they aspire to a higher social status (Vaish 2008), while the upper/middle classes are marked out by virtue of already being highly competent in English.Thus, particularly in Northern India, "English ostensibly pits urban elites against subaltern masses" (Chand 2011:13).The six individuals recorded for this study belong to two different generations, and while their socio-economic status is comparably middle class, English has played and plays different roles in their lives.The younger generation is more likely to consume English-language media and be dependent on English as an entrance tool into the job market.Furthermore, the older speakers have not had English-medium schooling, while the younger ones have.All speakers are English-medium university educated; however, all spoke Hindi as their primary home language when growing up, and their regular production of sentences and turns entirely in Hindi clearly indicates that they do not lack any proficiency in it.1Therefore, changes in CS patterns among these groups are unlikely to be the result of changes in Hindi proficiency, but rather indicative of social, cultural, and personal identity-based shifts in the use and role of Hindi and English.

Literature Review and Aims of the Study
The existence of older Hindi-English CS studies (Verma 1976, Gumperz 1977, Kachru 1978) clearly indicates that the phenomenon has been commonly observed for many decades.However, older studies often put forward strictly domain-oriented interpretations of CS, identifying Hindi as the language of intimacy and English as the language of status (Malhotra 1980).This perspective, based largely on colonial power structures, cannot be taken for granted anymore.Instead, in modern India, bilingual mixing rather than double monolingualism "has now become a legitimate interactional practice, especially among the young, English-knowing bilingual middle class" (Bhatt 2008:179).In his analysis of Hindi usage in English-language newspapers during the 1998-2002 heyday of the Hindu-nationalist government's Hindutva2 ideology, Bhatt (2008) introduces the notion of a third space, created through innovative language mixing, as a means for non-Hindu-nationalist speakers to align their political views with their linguistic behaviour and express new hybrid identities suspended between modernity and traditionalism.
In his study of Hindi-English CS in Bollywood films, Si (2010) categorises switches into alternations and insertions; more precisely, he counts turns as 1.Hindi-only, 2. English insertions into a Hindi sentence, 3. Hindi/English alternation, 4. Hindi insertion into an English sentence, or 5. English-only.Si's (2010) central claim is that the overall amount of English increased in the more recent films he examined.While the oldest film had 77.2% Hindi-(or Punjabi-) only turns, the most recent one had 69.5% turns involving some English and even 23.2%purely English turns (Si 2010:403).More specifically, the older films frequently featured English insertions into Hindi sentences, such as in the 1982 film Masoom: "Discuss kar lĕge problem" ('We will discuss the problem') (Si 2010:394).However, speakers, especially the younger ones, in more recent films were more likely to insert Hindi items into English clauses or to alternate between fully Hindi and fully English clauses (Si 2010).
(1) Excerpt from the 2001 film Dil Chahta Hai: Samir: "kitnī khūbsūrat jagah hai".('It's such a beautiful place.')Akash: "Hmm . . .it's beautiful."Samir: "You know what, hamĕ har sāl kam se kam ek hafte ke liye goa ānā cāhiye."('…we should come to Goa for at least one week every year.')(Si 2010:404) The main expectations of the current study are thus to find a higher overall amount of English in CS among younger speakers, a prevalence of English insertions into Hindi clauses among older speakers, and a greater alternation between purely Hindi and purely English clauses among younger speakers.Furthermore, for arthouse films specifically, Si (2010) identifies a "turnover" phenomenon whereby the matrix and the embedded language swap roles over the period studied,3 so that in the newer art-house films there are more Hindi insertions into English, rather than vice versa as in the older art-house films.Si (2010) is cautious of giving any definitive explanations for this phenomenon, but suggests that two factors are at work: the large overall increase in the use of English in general and the recent emergence of new types of CS, indicating the development of an English-dominated ethnolect among young speakers.The present study will allow for some brief reflection on whether the collected CS data support this claim or not.
The present study adapts Si's (2010) methodology and, taking up Bhatt's notion of a third space, analyses CS as a means which aids Hindi-English bilinguals in modern India to actively construct their (linguistic) identities.This is undertaken in three steps.Firstly, the Results section will present quantitative data outlining the generational differences in the amount of English and Hindi used, highlighting differences in the types of CS produced.Secondly, the Discussion section offers a closer examination of how the different generations use the two languages, and the option of switching, by discussing several typical examples.Lastly, both quantitative and qualitative analyses will be drawn together to investigate how CS is employed by Hindi -English bilinguals in the construction of their linguistic identities.In its analysis of the role, function, and patterns of CS, this study will refer to three factors that have been identified by Gardner-Chloros (2009) as influencing CS behaviour.Firstly, she identifies those factors independent of the particular speakers and circumstances, such as the power and prestige associated with a particular variety in a community and linguistic "market forces".The second set of factors relates to the speakers as individuals and as members of sub-groups, including their education, competences, attitudes, social networks, and perceptions of self and other.Thirdly, Gardner-Chloros (2009) identifies factors influencing CS patterns within the conversation, namely, where it is used as a structuring device.All three factors play a role in the analysis but are found to be of varying relevance to the CS behaviour observed in this study.All strands of analysis combined help to explain not only the observed CS behaviour but also the generational shift.

Methodology
Over the period of one month, circa 13:45 hours of material were recorded, involving more than 30 different speakers.The data chosen for the present study comprise six conversations that were especially suited to its aims, as they consist of highly naturalistic code-switched speech between six different native Hindi speakers of the appropriate age and socio-economic brackets.All recordings were made in Dehradun (Uttarakhand) in July 2012.Participants were recorded in an informal setting, most often their own homes or the home of close relatives or friends, often within the same neighbourhood.Written consent to being recorded for the purposes of an undergraduate dissertation in the field of linguistics, with the potential for future use of the recordings for other research-related work that may be made available to the wider public, including potentially online, was obtained from the participants.The exact nature of the research, i.e., its focus on CS, was only disclosed in advance if speakers expressed a specific interest in knowing this; in all other cases, disclosing this detailed information ahead of recording was avoided as it was thought likely to consciously or unconsciously influence the speakers' behaviour.If, following the recording, speakers expressed an interest in knowing more about the purpose and nature of the recordings, they were debriefed accordingly.The participants were also given a short questionnaire to gather information on their linguistic backgrounds.All speakers were female and natives of Northern India.While the selection of only female speakers was not intentional, this characteristic of the participants is significant.As argued by Labov (2001:291), females tend to be at the forefront of linguistic change, being "quicker and more forceful in employing the new social symbolism" of a particular linguistic variety.Therefore, studying an all-female group of speakers might be particularly indicative of ongoing and potential future shifts in the linguistic behaviour of a speech community.
The recordings vary in length between 00:16 and 00:32 hours, resulting in 02:07 hours in total of material analysed.Three conversations feature pairs from an older group of speakers (AB, PS, and VKD), while the other three feature pairs from a younger group of speakers (M, S, and K) (see Tables 1 and 2).All speakers, except K, are related to each other, as well as to the author, albeit sometimes very distantly.The atmosphere was as relaxed as possible with a microphone present, and it was emphasised that the speakers should feel under no obligation to speak any particular language; instead, they were encouraged to speak to each other as naturally as possible while trying to ignore the presence of the author.Nevertheless, the author's identity as half-insider, being a family member and understanding Hindi, and half-outsider, coming from Europe and not speaking much Hindi herself, means that some influence over speech behaviour cannot be ruled out.For this study, the categorisation used by Si ( 2010) is adapted, counting clauses instead of turns.This is because turns could be any length, making quantification unreliable.While the categories of "English-only" and "Hindionly" are maintained, instances of alternation between them were not counted; instead, an increased amount of clauses in either language was taken as an indication that more alternations from one language to the other had taken place.Clauses with insertions of any kind from one language into the matrix of the other language are counted as "Mixed English" and "Mixed Hindi" clauses, respectively.In keeping with Si (2010), no distinction has been made between insertions and borrowings, counting all instances of English usage within a Hindi clause and vice versa as insertions.In order to investigate whether the types of insertions differ across the generations, data regarding the preferred types of insertion have been collected and analysed and are presented in the Results section.

Results
Echoing the results of Si's (2010) study, Figure 1 shows that the younger participants in the present study use more English overall and alternate between fully Hindi and English clauses.Younger speakers produce 75% Englishmatrix clauses and only 25% Hindi-matrix clauses, while the older speakers produce a much larger percentage of Hindi-matrix clauses, 68%, and only 32% English-matrix clauses. 4This almost inverse distribution of ML among younger and older speakers clearly indicates the younger group's preference for English-matrix clauses.
Figure 2 gives a more detailed breakdown of the types of clauses produced by older and younger speakers with fully English clauses in blue, fully Hindi ones in orange, Mixed English clauses in grey, and Mixed Hindi ones in yellow.As can be seen, the younger speakers show a clear preference for fully English cla uses (63.89%), while only producing 10.72% of clauses in either Mixed English or Mixed Hindi.They also produce 14.66% of clauses fully in Hindi.The older speakers, however, produce a large amount of clauses fully in Hindi (36%) and in Mixed Hindi (31%), while producing a much smaller percentage of fully English clauses than the younger speakers (26%) and only a very small amount of Mixed English clauses (7%).As Figure 2 shows, younger speakers use more clauses that are fully English than any other type o f clause, confirming the expectation that younger speakers' behaviour exhibits an overall increase in the use of English as compared to older speakers' behaviour.
Figure 2 also proves insightful in investigating Si's (2010) second central finding, explaine d in Section 1.1, regarding a shift over time away from insertional switching towards alternational switching.The older group of speakers in the present study produces Mixed Hindi clauses almost as frequently as they do Hindi ones (31% versus 36%, see Figure 2).Thus, it can be established that English insertions into Hindi are a frequent feature of the older CS pattern.The younger speakers, on the other hand, tend towards alternational switches at the clause boundary.This is indicated by the fact that purely English and Hindi clauses, as opposed to clauses containing insertions either way, together make up almost 80% of all clauses.With regard to Si's (2010) claim of a turnover between the matrix and embedded languages of the older and the newer art-house films, as referred to in Section 1.1, Figure 3 allows an interesting observation.As it gives a breakdown of individual speaker behaviour, with categories of clauses marked in the same colours as in Figure 2, it can be used to investigate particular speaker idiosyncrasies, while also providing confirmation of the overall trends already identified in Figures 1 and 2. Of course, real-life conversations cannot be classified into film genres; therefore, no direct comparison with Si's (2010) turnover claim is possible, or sensible.However, it is interesting to note that, while Mixed English (grey) is infrequent among all speakers, it is used somewhat more by the younger speakers than by the older ones.While S produces 12% of Mixed English, and M and K produce 10% and 11.592% respectively, AB only produces 4%, VKD 5.228%, and PS 9.458%.Whether this constitutes enough evidence to support Si's (2010)  however, it undoubtedly fits with the overarching observation that younger speakers use more English across the board, and do so in innovative ways.The analysis of the types of English insertions into Hindi clauses has shown that the most frequently inserted items were nouns or noun phrases (NPs).In a conversation between PS and VKD, 84% of the English insertions into Hindi clauses were nouns or NPs.This trend is confirmed in the other two conversations between older speakers, with 69% and 71% English noun-and NP-insertions into Hindi.The pattern is also repeated in the conversations among younger speakers.Though their use of English insertions into Hindi is lower overall, these are 74%, 51%, and 76% nouns or NPs, respectively.

Discussion
The current section discusses the results in a more qualitative light, examining the insertions and their qualities in the CS data in 4.1 and the differences in alternation between the generations in 4.2.

Aspects of Insertional CS
As laid out above, the English insertions into Hindi preferred by older speakers were most commonly found to be single noun-or NP-insertions.
(2) Typical illustration of older insertion-dominated speech VKD: "lekin kal ek discussion ye bhi horaha tha ki laḍkiyõ ke liye, jo hamara time tha, wo phir bhi broad-minded tha" ('But yesterday we also had this discussion, that for girls, in our time, it was actually more broadminded') In this example, two of the three English insertions into the Hindi clause are nouns ("discussion" and "time").This is consistent with both the results of Si's (2010) analysis and the apparently universal finding, confirmed in numerous CS studies, that nouns are generally the most frequently switched items (Gardner-Chloros 2009).Explanations for this preference for noun/NP-insertions include the noun's relative freedom from syntactic restrictions and the fact that it is the most easily accessible word class, even to bilinguals with only a minimal grasp of the donor language (Gardner-Chloros 2009).
Given India's colonial past, nouns and NPs describing technological, cultural, or social concepts introduced under British rule are unsurprisingly common among the switched items.In keeping with Si's (2010) methodology, the present study counts all instances of English usage within a Hindi clause as insertions.However, there is a smooth transition between borrowings and insertions, and it is quite possible that numerous of the insertions  (Bauer 2008), and Gardner-Chloros has sensibly argued that "the distinction between code-switching and loans (…) [should be] of a 'more or less' and not an absolute nature" (cited in Bauer 2008:6).A useful definition, however, is that favoured by Shana Poplack and colleagues, who argue that a borrowing is a word frequently and consistently used by different speakers; phonologically, morphologically, and syntactically integrated into the other language; and whose equivalent in the other language, though existent, is often not used (Boztepe 2003).Such borrowings have been a staple of even the most elementarily educated Indian's vocabulary for so long that virtually no conversation passes without at least a few occurrences.To give but a few examples: (3) "Engineer hai, doctor hai…" ('she is an engineer, is a doctor') (Old speaker) (4) "jaise colony ban ni wali hai" ('a colony is about to be built') (Old speaker) (5) "to usme likha rahta tha, school ka nām, bachche uniform me" ('written on it was the school's name, children in uniform') (Old speaker) These correspond well with Si's (2010:395) examples, which include botal 'bottle', ṭren 'train', miniṭ 'minute'.
The other main source of noun/NP borrowings are recent technological innovations.Such items, without Hindi equivalents, are termed "cultural borrowings" by Myers-Scotton, as opposed to "core borrowings", which describe words for which equivalents do exist in the recipient language (cited in Bauer 2008:6).Examples for such cultural borrowings are also very frequent: (6) "Ye to mujhe batā hai ki mobile sab ke pās hai" ('This I know, that everyone has a mobile (phone)') (Old speaker) (7) "uske Sheikh Sarai ke car park me dekhta koi ghāḍi aise nahin millegi" ('if you look in that Sheikh Sarai car park you won't find any such car') (Old speaker) (8) "Pankha aur air-conditioning, aur bahar jānē ki bāt hi nahin karsaktē" ('Fan and air-conditioning, and no talk even of going outside') (Old speaker) (9) "mai to callijaungī, tū mujhe mail likh denā" ('I'll go, you write me an email') (Young speaker) Insertions have been found to be used indiscriminately, namely, without discernible patterns, "triggers", or typical contexts.This supports the argument that these types of switches are unmarked and part of the established mixed medium that prevails among the speakers.However, there are idiosyncratic differences between the speakers' speech behaviour.As Figure 3 shows, the three older speakers consistently used more English insertions into Hindi, between 27% (PS) and 36% (VKD), supporting Si's (2010) findings.Nevertheless, PS also exhibits a high amount of English-only and Hindi-only clauses (34.37% and 28.95%, respectively), that is, a high amount of alternation, while AB exhibits a significant preference for Hindi-only (60%).VKD produces the most Mixed Hindi, as typical of her generation (36%).
(10) A typical turn, exhibiting both English insertions into Hindi and alternation PS: "Isliye abh har schools me ye activities start ho kardiya, because tenth is based on that only, class ten, you have to do lot of debate, declamation, recitation-" ('Therefore, now, in every school, they have started these activities…') The first part of this turn (preceding the subclause beginning with "because") contains two English nouns and one English verb in a Hindi-matrix clause, which is followed by an alternation to English.It is likely that these particular insertions by PS are governed, in part, by the English context of the topic discussed, i.e., "school", and PS's biographical background as a teacher.The occurrence of the English verb "start" in this example further serves to illustrate another interesting type of insertion that is frequently reported in CS studies, namely, that of " [v]erb in one language" + "verbal operator meaning 'make/do' in the other language" (Gardner-Chloros 2009:34, previously identified in Verma 1976:163).In the example above, in the phrase "start ho kardiya", the English verb "start" is combined with (a past tense form of) the Hindi verbal operator karnā ('do/make'), rendering the meaning 'have started'.This phenomenon is found fairly frequently in all conversations, both among young and old.Examples include: (11) "To mujhe call karna, ānē ke bād" ('So call me after you've arrived') (Young speaker) (12) "tū mere ko usme bhējnā, phir reply karnā, thīk hai?" ('You send me [an email] on that [email address], and then reply, ok?') (Young speaker) (13) "agar yahan ke lōg khudi apne ko apni help karnai ke sōcle…" ('if the people from here started to think about helping one another…') (Old speaker) Again, a similar phenomenon is observed by Si (2010), confirming the notion that V + karnā is an established and thus unmarked form of insertional CS.His example from the 2001 art-house film Monsoon Wedding takes the form of an English object insertion immediately followed by the V + karnā construction: "pahle āp Constitution change karvāiye" ('Get the Constitution changed first') (Si 2010:395-396).This construction of Object (E) + Verb (E) + karnā is also a frequent feature of the CS recorded for this study, used by both older and younger speakers.Other examples include "lifestyle intrude kardiya" and "college join nahin kiya" (the latter one illustrating that the construction is also comfortably used with negation).Both older and younger speakers draw on broadly the same fields for their English insertions, such as technology, education, business, politics, colonial imports, numbers, dates, time, and weekdays.Interestingly, "India" and "Indian" are exclusively used in the English form, whether in an English or a Hindi-matrix clause.The Hindi equivalent, bharat, is not used once.This may appear curious; however, it can be seen to support the idea that the speakers use their mixed code to emphasise their identities as modern, cosmopolitan Indians who do not want to be associated with the Hindu nationalism of the Hindutva ideology, and therefore prefer not to use the ideologically more charged, Sanskrit-derived name for their country.Though many English insertions, mainly nouns and NPs, obviously stem from India's colonial history or its more recent, globalisation-driven emergence onto the world stage, there are frequent instances of evidently random insertions.They are most often adjectives or adverbs, and more rarely gerunds and past participles.No system is discernible behind insertions, which include anything from "young", "good", "brave", "fresh", "fancy", and "flexible" to "nervous", "congested", "remote", "traditional", "hard-working", and "honest".Evidently, whichever word comes to the speakers' minds first, whether Hindi or English, is used, thus signalling that the pattern of English insertions into Hindi is not in itself different between the generations; rather, the quantity is.In the following section, this is found to differ with regard to alternations, in the employment of which qualitative differences between the generations can indeed be found.

Aspects of Alternational CS
The data have revealed that younger speakers use much more English than older speakers.This is most often realized in the form of English-only clauses, which often alternate with Hindi-only ones at the clause boundary, while insertions, of both kinds, are much less frequent.( 14) Typical example of the alternational CS preferred by younger speakers S: "Ya, right, sahi bōl rahi ho tū.Haan, he keeps forgetting the things, no?" ('Ya, right, what you are saying is right.Yes, he keeps forgetting the things, no?') In this example, an English discourse marker ("right") introduces a clause entirely in Hindi ("sahi bōl rahi ho tū"), followed by a Hindi discourse marker ("Haan") and an alternation to a clause entirely in English.An examination of the distribution of English in the speech of the younger speakers provides insight into the position English holds among this younger generation.It is obvious that English is present throughout their conversations-it features, at least to some degree, in all of the subjects covered, confirming the expectation that for the younger generation, English is by and large perfectly acceptable as the main medium of casual conversation.However, it is not a purely English medium but a mixed code that acts as the unmarked choice, as Hindi, far from being abandoned by the younger generation, is still an essential part of their discourse.There are some noticeable cases where a switch to Hindi appears to be somewhat marked as it coincides with a change in topic, thus apparently functioning to direct conversation (Auer 1998).For example, following a stretch of mostly English turns discussing plans to buy property, the topic is exhausted and the speakers search for a new one.This is signalled by S's switch into Hindi in asking: (15) S: "Aur to?Aur kya ho raha hai?" ('And so?What else is going on?') K takes up the new language for one clause: (16) K: "Tū batā" ('You tell') In this utterance, she carries on the search for a new topic, indicating her own inability or unwillingness to end it.However, immediately following this, K reverts back to English.Introduced by the discourse marker "so", K seems to have decided on a new topic herself: "what's your plan for the fitness and all?".The conversation then carries on in its previous English-dominant style.Similarly, in the same conversation, after a discussion in the English-dominant medium on mountaineering instructors, S switches into Hindi as she abruptly changes the topic: (17) S: "Aur apnē-, teri mummy, dekhigī tera-, wahan pe?" ('And your-, your mum will see your-, there?')K does not immediately follow the change of topic and asks for clarification, also in Hindi: (18) K: "Kahan pe?" ('Where?') S proceeds to explain: "Your house-", and K, satisfied with the explanation, is happy to go along with the change in topic and the conversation continues again in its previous, English-dominant style.
It is noticeable that, while English dominates the discourse, at some critical points in the conversation either English or Hindi appear to be viable choices for the speakers, supporting the claim that they comfortably use both languages even in contexts that touch upon very personal or intimate topics, thus further dismantling the outdated domain model.Following an English-dominant medium, K produces a purely Hindi sentence: (19) "Mujhe rōnā bhi ārā tha us din" ('I also cried that day') After this, English again dominates.Similarly, a short while later, K says: (20) "Gita, wo, uski tabyat thīk nahin thi" ('Gita, she, she wasn't well') These instances of switches into Hindi could be interpreted as indicating that very personal subject matters, to do with emotionality and illness, are still more easily expressed through the traditional medium of intimacy, Hindi.However, in the same conversation, S is quite comfortable addressing similar topics in English, saying for example, "You cried also, I think" and "She vomited a lot, I believe?".Similarly, the mention of the religious concept of holy men, sadhus, does not call for a switch to Hindi beyond the obvious insertion of the lexical item itself; however, when talking about the unpredictability of the weather, M jokingly switches into Hindi to indicate that the only option is to pray: "I think you should just pray hard ki Bhagwan achcha rahengē" ('… that, God, let it be nice').Furthermore, a lengthy story about a scary incident involving being stopped by policemen and having to pay bribes to be released is told by M almost entirely in English, with the notable exception of reported speech which is often rendered in Hindi.On the one hand, this indicates that Hindi need not be the medium of a conversation involving emotional themes, like fear and serious danger, thus supporting the overall trend towards a more English-intensive form of CS among younger speakers; on the other hand, it is certainly also due to M's particularly high English proficiency that she is able to choose to use only English for the majority of the story.An examination of other narratives embedded in the recorded conversations appears to support the claim that generational differences in English proficiency contribute to governing such choices, rather than what Gardner-Chloros (2009) identified as her third set of factors (conversation structuring).Not one of the older speakers chooses to switch to an almost exclusively English medium in order to embed a narrative within the conversation; rather, narratives are rendered in the familiar Mixed Hindi medium.The fact that older speakers do not use alternation to structurally set apart a narrative from the rest of the conversation indicates not only a quantitative but also a qualitative difference between older and younger CS behaviour.While no qualitative differences in the use of insertions were found between the generations in Section 4.1, there is, thus, a qualitative difference between the generations with regard to alternations.

The Hindi-English Bilingual Community Revisited
The great value of this study's rare, highly naturalistic data set lies particularly in relating what has been found and discussed above back to the bilingual situation in the community more broadly.The first of Gardner-Chloros's (2009) CS-influencing factors, linguistic market forces, is evidently highly relevant.The older speakers, though not educated at English-medium schools, clearly have English at their disposal.All three older speakers are English-medium university educated.Moreover, AB and PS have taught at English-medium schools, while VKD has worked in the United States and frequently conducts business internationally.Their competence in English should not therefore be doubted.They frequently insert crucial lexical items into their Hindi discourse; however, they choose to not use English much on its own-a language that may be seen as the prerogative of the younger, English-medium educated generation.The older speakers also perceive their own English competence as lower than that of the younger generation, as illustrated by the following exchange between VKD and PS: (21) VKD: "(…) ham Hindi medium school me paḍhā, thīk hai, science paḍhli (…), whatever, and we became somebody, lekin jab mū se Angrezi nikalti hai, na, English, mujhe lagta hai, Hay Ram, ham na-, hamne kyo nahin paḍhā hua?Thorasa English medium me paḍhliya-" ('We studied at a Hindi medium school, fine, we studied science, whatever, we became somebody, but when English comes out of our mouths, I think "Oh God, why didn't we learn, why didn't we study in English medium at least a bit"') PS: "Aur English now is a international language, so why not?" ('And…') VKD: "Haan, why not?" ('Yes, …') PS: "Why not?Because tomorrow if your child goes somewhere, he will be left out, because if he can't converse, if he can't say or, uh, anything, so then what is the-" VKD: "And converse properly, effectively…" The younger speakers were entirely educated through the medium of English and have grown up exposed to a much larger amount of English.English, to them, is not only a daily professional requirement but also plays a major role in their private lives, including their travel and entertainment choices, hobbies, and social circles, meaning they are competent and comfortable enough to choose to use English in large parts of their casual conversations.
Finally, what Gardner-Chloros ( 2009) identifies as her second set of factors, those pertaining to the individual speaker's education, competences, and attitudes, and their ways of constructing their identities through language, play a particularly significant role.In a Hindi-English CS study conducted by Vineeta Chand (2011), highly elite participants are reported to construct their identity in a way that does not value Hindi particularly highly: they do not see it as a potentially unifying force, nor do they regard Hindi as useful for them, given their English-dominant lives and jobs.They even claim to have only limited competence in Hindi, partly due to their English-medium education, but also for political reasons, namely, to distance themselves from right-wing nationalist Hindutva ideologies (Chand 2011).This behaviour has not been observed among the present speakers.Hindi is a perfectly accessible option for them, as it is present, to a greater or lesser degree, in all conversations.Furthermore, all conversations contain ample examples of purely Hindi clauses, indicating full competence and comfortableness in its use as the members of the current group of speakers all grew up speaking Hindi, by contrast with Chand's (2011) elite Delhiites, who mostly spoke English at home and actively encouraged their children to do so too.The ease with which these speakers use Hindi, however, must not be taken as an indication of a nationalist outlook.Though no explicit questions regarding ideology were asked in this study, it is clear that Hindi-English CS gives these speakers the opportunity to balance two identities-the one traditional and the other cosmopolitan (Bhatt 2008).They can retain Hindi, a language that is important to them, and embrace what it means to them, while at the same time avoiding the risk of being branded nationalistic by acknowledging the universal importance of English in modern India.This awareness is again illustrated by the exchange between PS and VKD (Excerpt 21), in which the two acknowledge the central importance of English to life in modern India.Immediately following this passage, the speakers go on to praise the high quality of English acquired by the younger generation, further supporting the claim made in this study that there has been a shift towards more English-heavy CS and more whole-English clauses in conversation.As such, the present speakers' CS behaviour should be taken as an indication of their third space modern identity construction, as they choose to use English to a significant extent without denying the importance of Hindi to their sense of self in a nation suspended between tradition and modernity.
What this paper has shown in its examination of CS among two generations of Hindi-English bilinguals in Northern India is how CS behaviour has changed from the one generation to the next, and how the speakers employ the option of switching, or rather the option of using a mixed medium, in order to align their linguistic selves with a nation in linguistic, cultural, and economic flux.In particular, the study has found an increased amount of English being used by the younger generation, which is related to the increase in English-medium education and the changed market capital that English holds in modern India, as it does globally.Furthermore, a change in the pattern of CS has been observed between the generations.While the older generation tend to insert English items into Hindi clauses, the younger generation prefer alternating at the clause boundary between fully English and fully Hindi clauses.While this shift may reasonably be due to the bigger role English plays in the lives of the younger generation, this does not alter the fact that the mixed medium of Hindi and English functions as the unmarked choice in casual conversations for all of the sampled speakers, indicating the importance that both languages, rather than one over the other, hold for the speakers, who confidently use them in novel ways to reflect their linguistic identities in diverse contexts.

Figure 2 :
Figure 2: Percentages of English, Hindi, Mixed English, and Mixed Hindi among younger an d older groups.⃰

Table 1 :
Demographic information about older speakers

Table 2 :
Demographic information about younger speakers turnover claim is debatable; 4 Percentages out of total number of clauses (N).
counted here are, in fact, stable borrowings from English into Hindi.The definition of a borrowing has been contested