Structural Changes in Bengali–English Bilingual Verbs through the Exploration of Bengali Films

Abstract

This paper investigates structural changes in the use of Bengali–English bilingual verbs through the exploration of Bengali film scripts from three decades: the 1970s, 1990s and post-2010. Previous research has shown that the increase in use of bilingual verbs, especially involving embedded language lexical verbs with matrix language helping verbs possibly results from increased bilingual proficiency. Over the past years, the use of English, including code-switching between Indian languages and English has increased dramatically in Indian society. Given this development, this paper explores film data to ascertain the extent to which the use of bilingual verbs, closely connected to code-switching and bilingual speech, has also changed diachronically in Bengali–English speech and if these changes occur from increased levels of bilingual proficiency. In connection with structural change, this paper also discusses the sociolinguistic factors that may be related to bilingual verb use. Results show a massive increase in bilingual verbs in the films post-2010, especially the ones involving English lexical verbs. Increased use of code-switching may have played a role in these changes and sociolinguistic factors related to the use of bilingual verbs seem to be less relevant now as compared to the earlier decades.

1. Introduction

As with many languages around the world, bilingual verbs are a common occurrence in several South Asian languages. These constructions have been documented in languages, such as Hindi, Punjabi, Tamil and Bengali [1,2,3,4,5,6,7]. In this paper I investigate the use of Bengali–English bilingual verbs in terms of the changes that these constructions have undergone diachronically. Bilingual verbs in Bengali–English speech consist of English lexical nouns or verbs that occur alongside Bengali helping verbs such as do and be. Examples (1) and (2) below illustrate two bilingual verb constructions. In Example (1), the English lexical verb, renovate is combined with the ‘do’ verb koreche from Bengali, which acts as a helping verb to integrate the non-native verb in the language. Example (2) is that of a bilingual verb where an English noun, discussion occurs alongside the helping verb. These constructions are also referred to as bilingual complex verbs because they have a complex structure.

1.	Ora	park-ta	renovate	kor-ech-e
	3PL	park-CL	renovate	do-PFV-3P P
	‘They have renovated the park.’

2.	amra	tɔkhon	discussion	kor-chi-lam
	1PL	then	discussion	do-PROG-PST.1P
	‘We were discussing then…’

Bilingual complex verbs (BCVs) have been the subject of a copious amount of literature not only because they occur widely in several language pairs across the world (see Backus [8]; Muysken [9,10]; Wichmann and Wohlgemuth [11]; Wohlgemuth [12]), but also because they can shed light on the bilingual mind and its ability to juggle two or more grammars. This paper therefore investigates changes in bilingual verbs to get a better understanding of the connection between BCV use and bilingual proficiency, including code-switching specifically in context of the Bengali–English speech community in West Bengal, India.

Some scholars, such as Backus [8] have also argued that bilingual complex verbs undergo a process of grammaticalization of the helping verbs, because they initially have an N + ‘do’ form, where the lexical element from the embedded language (EL) is a noun, but gradually turn into V + ‘do’ constructions, where the lexical element is a verb. It has also been hypothesized that balanced and imbalanced bilinguals differ in their use of BCVs, because imbalanced bilinguals (speakers for whom the native language is more dominant) have more N + ‘do’ structures, while balanced bilinguals (speakers for whom both languages are on the same footing) tend to use V + ‘do’ structures more. These differences constitute changes in bilingual speech, because in many matrix languages1 especially Indian languages, a native verb cannot occur before the helping verb ‘do’ [5]. Therefore, given that BCVs occur most frequently in bilingual speech, with the increase of bilingualism, specifically code-switching within a community as well as bilingual proficiency, we expect not just an increased use of BCVs in general in the bilingual Bengali–English speech community of West Bengal, India but also specifically the occurrence of more English verbs than nouns in those BCVs.

In the case of Indian society, the use of English and code-switching with English has intensified greatly over the past decades, including in the Indian media and films, due to factors such as the advent of technology in the last couple of decades, the spread of a global economy, and the economic liberalization policies of the Indian government post-1990. Although English has been a part of India’s linguistic scenario since the eighteenth century, it is only in recent years, with the advent of a global economy, technological innovations and influences from the Western media that the use of English in India has witnessed massive growth (e.g., [13,14]). On the other hand, regional Indian languages such as Bengali and Hindi (among many others) also exist and are widely used. This situation has led to stable bilingualism among a section of the Indian population, for whom the first language (or the native language) is an Indian language or languages and the second language is English, by virtue of being educated in English instruction schools. The state of stable bilingualism is also reflected in the Indian media and in films, which are produced in great numbers every year, thereby constituting an abundant source of data that could be used to observe trends in the speech patterns of a community.

With this in view, this paper examines the use of BCVs in Bengali films from three decades, the 1970s, the 1990s and post-2010 to determine possible changes in their structures over the years. The specific research questions driving this paper are: (1) Is there an increase in BCV use in contemporary films as opposed to films from the 1970s and 1990s? (2) Are there differences in the distribution of N + ‘do’ and V + ‘do’ constructions in the films of the three decades? (3) Does the switch to more V + ‘do’ constructions indicate higher bilingual proficiency? (4) What sociolinguistic factors relate to the use of BCVs in the films?

The organization of the paper is as follows. In the next section, I present a brief background of Bengali and the history of English in India both in the past and in contemporary times. In Section 3, the methodology of this paper is presented including a discussion of the films that were used for this study as well as a rationale for using the media as a source of data. Section 4 discusses complex verb structures of monolingual Bengali and then provides a comparison between monolingual complex verbs and their bilingual counterparts. Section 5 first presents a quantitative and a qualitative analysis of BCV use in the films of the three decades, including a discussion of the findings. Section 6 presents the conclusion of the paper.

2. Background of Bengali and English Use in India

Bengali is a prominent eastern Indo-Aryan language spoken in India and Bangladesh2. It is the official language of the state of West Bengal in India and is used as the main language of communication within the state. Hindi, which is the most widely spoken indigenous language in India is followed by Bengali as the second most widely spoken language [15] (pp. 235–236). Bengali has come in contact with several European languages over the past centuries, including Portuguese, French, Dutch and English, and of all these languages that have come in contact with Bengali, however, it is with English that contact has been the most intensive. Contact between Bengali and English dates back more than three hundred years. English was first introduced to India’s linguistic environment by the traders of the East India Company following a charter of Queen Elizabeth in 1600 allowing trade with India. It was the third European language that came to India after Dutch and Portuguese [16]. The language arrived in Bengal when the East India Company set up its base in the city of Calcutta in 1698 [17]. The East India Company gained full political control of Bengal in 1784 following which the city of Calcutta (now the capital of West Bengal and its urban center) became the center of trade and commerce and the capital of the British administration [18].

English was formally introduced as a medium of instruction in India’s higher education system in 1835, a policy that continues to this day. As a result, English education saw a major expansion in the country. Since India’s independence from British rule, English was adopted as the official language of the country alongside Hindi. After the end of British rule, English has overcome its primary role as a tool of colonization and is now used in the domains of education, law, government, science, and technology, media and in inter-personal communication in urban areas. As a result of English being widely used in the education system (both before and after independence), it has become a second language for most educated people across urban India. This has led to widespread bi/multilingualism between English and Indian languages. However, people who are not literate or who have not been schooled in English, especially those living in rural areas, have little to almost no knowledge of it. Some sources say that only about five to 10 percent of the Indian population actually speaks English [19] (p. 791), [20]. Additionally, code-switching between English and Indian languages is a regular occurrence in the day-to-day speech of English educated bilinguals in India [19].

Even though the use of English in India’s higher education system has lasted for nearly two hundred years, it has experienced unprecedented growth in the country in the past two decades following the spread of technology (satellite television, mobile phones and the internet), globalization and economic liberalization policies introduced by the India government in the 1990s. English has always been a prestige language in India but in recent years, the need to speak English fluently has become a prerequisite for any white-collar job. Consequently, the need to maintain a ‘pure’ or unmixed use of Bengali or for that matter other Indian languages has become less important. Therefore most urban, English educated speakers, nowadays, either speak English when their jobs demand it or use code-switching between Bengali and English as a mode of speech in contexts where English is not required. Given this situation, the increased use of bilingual complex verbs is certainly not unexpected.

The next section presents the methodology that was used to examine the research questions of this paper.

3. Methodology

For the purpose of this paper, I analyze a total of 6 Bengali films from three decades, the 1970s, the 1990s and post-2010. I examine two films from each decade. I chose to examine films from three different decades in order to get a diachronic perspective of bilingual verb use and of possible changes that may have occurred.

Like the film industry based in Mumbai, popularly known as Bollywood, there are several regional language based film industries in India, which produce numerous films every year and enjoy a large amount of viewership. Among these, is the Bengali film industry based in the city of Kolkata, sometimes referred to as Tollywood. The films I analyze for this paper were all produced in this industry within the past forty years and include Chadmabeshi (1971) [21], Basanta Bilap (1973) [22], Agantuk (1992) [23], Unishe April (1994) [24], Maach Mishti & More (2013) [25], and Hawa Badal (2013) [26]. The films were chosen based on similarities in their settings and subjects. Firstly, the stories of these films are all based in the urban areas of West Bengal and depict narratives of urban characters. Secondly, the characters of the films typically belong to the same socio-economic status groups, specifically the middle/upper-middle class. All of the films deal with issues of inter-personal relationships within the family and between lovers and touch upon secondary themes such as material and intellectual aspirations in life, memories of the past and the search for one’s identity. The films also all belong to the genre of comedy-dramas and have comparable lengths in that they have a run time of around two hours.

A study by Si [13] reported a diachronic investigation of code-switching in Bollywood films from three decades, the 1980s, 1990s and 2000s. This study found a massive increase in the overall use of English in the films of this period, specifically observable in the speech of young speakers. It was further reported that the code-switching patterns in these films became increasingly complex over the years involving more alternations between chunks of Hindi and English as compared to insertions of English words in the structure of Hindi sentences. Therefore, the study serves as a good illustration of how media data can be used to ascertain broad trends occurring in the speech patterns of a community. The present study has a similar purpose, which is to determine changes in the use of bilingual verbs over time and makes use of a similar methodology.

Coding and Data Analysis

In the films that were analyzed for the paper, all the occurrences of complex verbs with the X + ‘do’ form were coded. These included Bengali complex verbs such as bikri kɔra ‘sale do’, which means to sell. I use the form N(Beng) + ‘do’ to represent these verbs. The other type of complex verb that was coded included bilingual complex verbs such as, renovate kɔra ‘to renovate’, which is a combination of an English verb and a ‘do’ verb from Bengali, represented with the form V(Eng)+‘do’. The example discussion kɔra ‘to do discussion’, which includes an English noun and a ‘do’ verb from Bengali, represented with the form N(Eng)+’do’ is the third type of complex verb that was coded. The fourth type of complex verb was record kɔra ‘to record’, which includes an English lexical element that could be analyzed as both a noun and a verb (specifically in the context of Indian English) with a ‘do’ verb from Bengali, represented with the form N/V(Eng)+‘do’. In standard varieties of English, there is a distinction between récord, where the stress is on the first syllable, when it is used as a noun and recórd, where the stress is on the second syllable when used as a verb. However, such a distinction is usually not found in varieties of Indian English, where stress is usually on a heavy syllable and in some cases even arbitrarily placed depending on the L1 of the speaker and other sociolinguistic factors e.g., [27]. Therefore, a word like record usually always carries a stress on the second syllable, irrespective of whether it is being used as a noun or as a verb. The following forms of complex verbs were coded and analyzed: N(Beng) + kɔra; V(Eng) + kɔra; N(Eng) + kɔra; N/V(Eng) + kɔra.

Following the coding of the complex verbs, the data was analyzed quantitatively by calculating the frequencies of occurrence of these verbs for the films of each decade to ascertain the overall distribution of complex verbs in the data. The frequencies of occurrence of complex verbs in each decade were then compared with those of the other decades to find out the changes that have taken place in the use of bilingual verbs over the years.

4. Complex Verb Constructions in Bengali

In this section I first provide a discussion of the nature of complex verb structures in Bengali and then focus on the structures as they occur in monolingual Bengali and in bilingual Bengali speech.

Complex verb structures have been shown to occur frequently in Indo-Aryan languages [28,29,30,31,32] such as, Urdu, Hindi, Bengali, Marathi, Gujarati, among others languages in the world. These structures contain two or more than two elements such as nouns, verbs or adjectives, which together predicate as a single unit; that is, “their arguments map onto a monoclausal syntactic structure” [30] (p. 2). Their structure can be of the ‘Noun + Verb’ (N + ‘do’) or ‘Verb + Verb’ (V + V) form, making them complex structurally but not necessarily so in terms of semantics, as I discuss below. The term ‘complex verb’ is often used to refer to the N + V and V + V structures together because of their complex structures but separate terms for the individual N + V and V + V structures are also used. Independently, the N + V structures have been called conjunct verbs [33] and V + V structures are called compound verbs [32,34,35].

In an N + V structure, the noun is followed by a light verb, such as ‘do’ or ‘be’, which bears inflections and turns the N + V structure into a single predicate. This is a productive process in many Indo-Aryan languages where nouns are verbalized with the help of the ‘do’ or ‘be’ verbs. The V + V structure consists of two or more verbal elements, where the first verb or the main verb in the construction provides the main semantic content to the construction. This verb is followed by a light verb that bears inflections, affects the aktionsart of the joint predication and adds, instead of its inherent lexical meaning, more subtle semantic information such as benefaction, suddenness, volitionality, forcefulness, etc. (e.g., [29,30,36]).

Table 1. Verbal system of modern Bengali.

Verb Types	Sub Types	Components of Verbs	Examples of Each Type
			Example	Gloss	Translation
Simple Verbs		1 verb	bɔla	speak	‘to speak’
Monolingual Complex Verbs	Conjunct verbs	N + kɔra/hɔwa N + kɔra/hɔwa + (V)	bikri kɔra jiggesh kore newa	sale do question do take	‘to sell’ ‘to ask for oneself’
	Compound verbs	V + V	khe newa	eat take	‘to eat for oneself’
Bilingual Complex Verbs	Conjunct verbs	N (Eng) + kɔra/hɔwa N (Eng) + kɔra/hɔwa + (V)	drawing kɔra manage kore jawa	drawing do manage do go	‘to draw’ ‘to manage continuously’
	Compound verbs	V (Eng) + kɔra/hɔwa V (Eng) + kɔra/hɔwa + (V)	deny kɔra decide kore phæla	deny do decide do throw	‘to deny’ ‘to decide (completely)’

below illustrates these constructions. However, in this paper, I do not deal with the V + V structure and instead focus only on the N + ‘do’ structure because it is the most frequently used light verb in Bengali–English CS. To do away with the confusion of using several different terms to refer to these different verbal constructions, I use the term complex verb as a cover term to refer to the X + do structures examined in this paper.

The verbal system of modern Bengali (both monolingual and bilingual speech) consists of different types of verbs, including simple and complex ones. The complex verbs can be categorized into two main types, monolingual verbs and bilingual verbs. The monolingual verbs consist of N + do + (V) and V + V structures where all the components are from Bengali. The bilingual verbs also consist of N + do + (V) and V + do structures, but crucially in these verbs the main element that provides the semantic content to the construction is from English. Table 1 summarizes and provides examples of these different structures.

Modern Bengali also has single-word verbs, which I call simple verbs (following Thompson [32]) such as khæla ‘play’, or shona ‘hear’ just like in English, an example of which is provided below in (3). These verbs consist of a single word denoting an action or a state. Apart from simple verbs, are the complex verb structures, which are discussed in more detail below.

3.	she	gach	theke	por-ech-e	(simple verb)
	3SG	tree	from	fall-PFV-3P
	‘He/She has fallen from the tree.’ (Thompson 2010)

4.1. Monolingual Complex Verbs

• (N(Beng) + kᴐra/hᴐwa ‘do’/’be’)

The N + ‘do’/’be’ verbal structures are two-part and three-part verbs that occur in addition to simple verbs in modern Bengali. In these structures, nouns (or adjectives) occur alongside helping verbs such as kᴐra ‘do’ or hɔwa ‘be’. The nouns provide the semantic content to the complex verb and the helping verbs turn the constituent N + V into a verbal predicate. The helping verb also bears inflection. An example is cheshta kᴐra ‘to try’, formed by the noun cheshta ‘trial’ or ‘trying’ and the verb kᴐra ‘do’. When a complex verb has hɔwa ‘be’ after the noun, it is a passive construction, an example of which is nihɔto hɔwa ‘killed be’ or ‘was killed’. In Bengali, other helping verbs such as take, cut, give, get, etc. also occur e.g., [32]. However, in this paper, I focus only on the structures with the actional helping verb ‘do’.

Modern Bengali shows many occurrences of these N + V structures in addition to simple verbs. It must also be noted that Bengali does not have any V + kᴐra ‘do’ structures. For example, kine kᴐra ‘buy do’ (intended meaning ‘to buy’) would be an impossible structure because kine is a verb in Bengali and does not need a helping verb to express the action of buying.

Some complex verbs also have a third element, occurring after the noun and the helping verb, forming an N + ‘do’ + V structure. This third element is a verb that alters the aktionsart of the complex verb and in some instances provides additional subtlety of meaning. This verb has been referred to as a vector verb [34,35], or it has been treated as part of the class of light verbs [30], including verbs such as kɔra ‘do’ and hɔwa ‘be’. Examples of two-part and three-part complex verbs from modern Bengali are shown in (4a) and (4b).

4.a.

robi

lok-er

ɔnek

upokar

kor-ech-e

(N+‘do’)

Robi

people-GEN

much

help

do-PFV-3P

‘Robi has helped people a lot.’

b.

bæpar-ta

na

jene

jiggesh

kor-e

bosh-ech-i

(N+‘do’+V)

thing-DEF

NEG

knowing

question

do-PFV.PTCP

sit-PFV-1P

‘Without knowing, (I) suddenly (and unintentionally) asked him/her about the thing.’

The verbs in each of the examples (4a) and (4b) function as a single predicational element including the noun upokar ‘help’ or jiggesh ‘question’ providing the main semantic content of the construction. The kɔra ‘do’ verb, functioning as a verbalizer, carries tense, aspect, and person inflections as shown in (4a). The third verb or vector verb in example (4b) alters the aktionsart of the construction and adds subtle semantic content. For example in this sentence, it indicates that the action of asking was done suddenly and unintentionally. Unlike in example (4a), where kɔra ‘do’ carries inflections, in (4b) it is in perfective participle form and the vector verb carries inflections of tense, person and aspect. It must be noted that not every noun adjacent to a verb forms a complex verb. In regular object-verb constructions, the nouns are only arguments of the verbs (4c), and together do not form a complex verb construction (for relevant distinctions see [33] pp. 347–350).

c.	uni	amar	theke	shahajjo	ni-ech-e-n	(O V)
	3SG.HON	1SG.GEN	from	help	take-PFV-3P-HON
	‘He/She has taken help from me.’

4.2. Bilingual Complex Verbs

• (N(Eng) + kᴐra/hᴐwa ‘do’/’be’; V(Eng) + kᴐra/hᴐwa ‘do’/’be’;
• N/V(Eng) + kᴐra/hᴐwa ‘do’/’be’)

Another type of complex verb in modern Bengali is the bilingual complex verb. These bilingual verbs of Bengali combine English nominal or verbal lexical elements in the bare form with helping verbs from Bengali. These verbs in modern Bengali consist of either two-part or three-part structures and the first lexical element is always from English, combining with helping verbs such as kɔra ‘do’ or hɔwa ‘be’ from Bengali. Sometimes a third verb from Bengali may also occur after the helping verbs ‘do’ and ‘be’ (see [37]) for relevant discussion on the third verb in bilingual complex verbs). The examples below show the range of lexical elements from English that can occur in these structures. Example (5a) illustrates English nouns, (5b) English verbal nouns, (5c) English phrasal verbs and (5d) an English verb. All of these nominal and lexical elements may occur beside the kɔra ‘do’ verb from Bengali. I use the term verbal noun to refer to elements such as shopping and skipping, and phrasal verbs to refer to elements such as build up, work out, hang up, as used by Romaine [3].

5a.

shei

moment-e

operation

kor-l-o

(N(Eng)+’do’)

that

moment-LOC

operation

do-PST-3P

‘In that moment, (he) did the operation (surgery).’

b.

tui

o-r

shɔŋge

shopping

kor-b-i?

(Verbal noun(Eng)+’do’)

2SG

3SG-GEN

with

shopping

do-FUT-2P

‘You will be shopping with her?’

c.

cinema-ta

je

bhabe

build up

kor-echil-o

(Phrasal V(Eng)+’do’)

cinema-DEF

COMP

way

build up

do-PST.PFV-3P

‘The way they had built up the movie…….’

d.

ora

bol-ech-e

application-ta

renew

kor-b-e

(V(Eng)+’do’)

3PL

say-PFV-3P

application-DEF

renew

do-FUT-3P

‘They said (they) would renew the application.’

The Bengali elements that occur in the bilingual verbs are the helping verb kɔra ‘do’, in active structures, and hɔwa ‘be’, in stative structures, in line with Romaine’s [3] observation of Punjabi–English bilingual speech. Each complex verb together expresses one single event, the English nominal or verbal element, which is in the bare form, provides lexical meaning and the helping verbs bear tense, person, and aspect inflections.

As we would expect, bilingual complex verbs occur much more frequently in bilingual Bengali–English speech, than they do in the speech of monolingual Bengali speakers or speakers who do not have a lot of exposure to English. Some bilingual verbs correspond to N/V(Eng) + ‘do’ structures, in which the N/V are lexical items in English that cannot be clearly identified as either a noun or a verb. For example, shift kɔra ‘to shift’ where shift can be analyzed either as a noun or a verb. I have included both verbal nouns and phrasal verbs in this category. Pillai [6] points out that a verb such as give can also occur as a helping verb in bilingual complex verbs. However, if give occurs as a helping verb in a bilingual complex verb, it is not clear if the construction is a complex verb or a regular object verb construction. Therefore, for the purpose of this paper, I restrict the discussion and analysis only to the X + do constructions; that is, constructions involving a noun and/or verb along with the helping verb ‘do’.

4.3. Literature Review on Bilingual Complex Verbs

In this section, I review some previous studies on bilingual verb structures.

Apart from Bengali, bilingual complex verbs also occur frequently in CS between other Indian languages and English. Romaine [3] reports the occurrence of these verbs in bilingual Punjabi–English speech. In her data, the complex verbs consist of English nouns, verbs, verbal nouns (such as lobbying) or phrasal verbs (such as pick up) alongside Punjabi helping verbs such as karna ‘do’ and hona ‘be’. Romaine (using the term operators for these verbs) shows that ‘do’ and ‘be’ distinguish the stativity of the structures, in that ‘do’ occurs in actional constructions and ‘be’ in stative ones. The operators modify the English nominal or verbal elements and bear inflections. Romaine uses the term compound verb to refer to these structures and says that there is a tendency for English verbs to occur more frequently than English nouns in such structures.

Annamalai [4], in his analysis on Tamil–English bilingual speech, distinguishes between structures such as try pannu ‘try’ (V + ‘do’) and business pannu ‘do business’ (N + ‘do’). He argues that bilingual Tamil–English V + ‘do’ structures are different from bilingual N + ‘do’ structures because accusative case can be optionally added to business, but not to try, and so he argues that business is a noun and not a verb. In addition, dummy verb (Annamalai uses this term for helping verbs) ‘do’ is added after a main verb as a carrier of inflection. In a later study Annamalai [5] furthers a hypothesis about balanced bilinguals and imbalanced bilinguals (people who are stronger in Tamil than in English) in their use of bilingual N + ‘do’ and V + ‘do’ structures. He says that groups of bilingual speakers can be distinguished on the basis of the types of mixed compounds they use. This is because imbalanced bilinguals conform to the N + ‘do’ structures of native Tamil by using structures such as ‘reservation do’ whereas balanced bilinguals would tend to use structures such as ‘reserve do’, which are changes in bilingual speech, because in monolingual Tamil a V + ‘do’ construction cannot occur.

Bilingual N + ‘do’ structures have been distinguished from bilingual V + ‘do’ structures in many language pairs across the world [9]. Muysken [9] argues that the insertion of foreign or native nouns alongside helping verbs and the use of foreign verbs alongside helping verbs in the recipient language constitute different processes. In the noun insertion strategy, there is a process of insertion of the foreign noun in the position of a native noun. However, the occurrence of a foreign lexical verb beside the helping verb from a matrix language (usually considered to be the dominant language of the bilingual speaker) is a process of adjunction of the foreign verb to a matrix language verb ‘do’. He argues that this process of verb adjunction is an example of alternational code-switching between the lexical verb from the embedded language (usually the non-dominant language of a bilingual) and the helping verb from the matrix language. The foreign verbs in the complex verb structures are not borrowings, because they are not phonologically adapted to the recipient language. They also happen to be quite productive in the recipient language [9] (pp. 193–206).

It has also been postulated that some bilingual complex verbs emerge as a result of the grammaticalization of the helping verbs. That is, these constructions initially start as N + ‘do’ constructions but later go on to become V + ‘do’ constructions [9] (p. 213), [5] (p. 51). In his investigation of three generations of Turkish immigrants in the Netherlands, Backus [8] found that more proficient bilinguals of Turkish and Dutch had more V + ‘do’ constructions than less proficient bilinguals or mostly monolingual speakers, who had more N + ‘do’ constructions (similar to Annamalai’s [5] observation of Tamil–English bilinguals). The first generation of Turkish immigrants that Backus studied consisted of people who were dominant in Turkish and knew minimal Dutch. These people immigrated to the Netherlands as adults. The second generation of Turkish immigrants were born in Turkey and immigrated to the Netherlands at a very young age. The third generation of immigrants consisted of people who were born in the Netherlands to Turkish immigrant parents. The second and third generation speakers were bilingual in Turkish and Dutch. When these people spoke Turkish, Backus found that the third generation speakers had mostly V + ‘do’ constructions in their speech. The root verbs were from Dutch and the ‘do’ verbs were from Turkish. Backus found that the first generation of Turkish speakers used mostly N + ‘do’ and the second generation of speakers had a mix of both N + ‘do’ and V + ‘do’ constructions in their Turkish. This finding lends support to the hypothesis that the V + ‘do’ constructions derive from their nominal counterparts.

There have also been opposing claims about the status of the foreign verbal element in the complex verb structures. Scholars such as Moravcsik [38] have argued that if a language borrows a verbal element from another language, then it will be borrowed in the form of a nominal element. According to Moravcsik, verbs cannot be borrowed in the form of verbs, and a borrowed verbal element will always behave like a nominal element in the recipient language, since they require a helping verb to appear beside them, which then verbalizes the construction.3 More recently, Wichmann and Wohlgemuth [11] and Wohlgemuth [12] have argued that verbs are often borrowed as non-verbs in recipient languages, and not necessarily as nouns as Moravscik [38] has claimed. They argue that the foreign elements that occur in these bilingual structures can either be nouns or verbs (that is, they are identical in form to the nouns and verbs of the foreign language) but the recipient language does not specify their part of speech [11] (pp. 2–5), [12] (pp. 104–111).

The literature discussed above shows two different strands in the analysis of bilingual verbs. While Muysken [9], Romaine [3], Backus [8] and Annamalai [5] distinguish between N + ‘do’ and V + ‘do’ constructions structurally, for others, these two are similar structures. Based on the work of Muysken, Romaine, Backus and Annamalai, there seems to be a structural distinction between an N + ‘do’ construction and a V + ‘do’ construction in bilingual speech. In line with these observations for other language pairs, Bengali–English bilingual speech shows a higher frequency of occurrence of English verbs than English nouns (as also shown by Annamalai [5] in the case of balanced and imbalanced bilinguals of Tamil and English).

Below I try to analyze the bilingual verbs in the Bengali films as a domain involving structural changes. I partially draw from previous analyses of bilingual verbs in my explanations of the bilingual verbs of Bengali.

5. Analysis of BCV Use in the Bengali Films

This section provides a quantitative account of the occurrence of complex verbs in the films from each decade, with a specific focus on the occurrence of bilingual verbs. I first discuss the occurrence of complex verbs in the films of the 1970s, then the films of the 1990s and then in those of post-2010. Following that I compare the occurrence of complex verbs in the films across the three decades and discuss how the occurrence of bilingual verbs has not only increased considerably over the years but also undergone structural changes.

5.1. Distribution of Complex Verbs in the Films of the Three Decades

5.1.1. Distribution of Complex Verbs in the Films of 1970s

In the films of the 1970s we find that the occurrence of bilingual complex verbs was much less frequent than the occurrence of monolingual Bengali complex verbs.

Table 2. Frequency of occurrence of complex verbs in the films of 1970s.

Films of 1970s	Complex verbs N = 236
	N(Beng) + ‘do’ %	N(Eng) + ‘do’ %	N/V(Eng) + ‘do’ %	V(Eng) + ‘do’ %
	85.16	1.69	9.74	3.38

below provides the frequency of occurrence of complex verbs in the films of this period.

As Table 2 above shows, about 85% of the complex verbs in the films of the 1970s were monolingual Bengali in nature, that is they had the N(Beng) + ‘do’ form. The remaining 15% of complex verbs were bilingual verbs or mixed verbs. Among these verbs, roughly 10% had the N/V(Eng) + ‘do’ form, which means that the lexical element from English, occurring before the ‘do’ verb of Bengali, could either be analyzed as a noun or a verb as shown in example (6) below.

6.	aapni	ei	shɔb	tribes der ke	study korechen
	2SG.HON	DEM	all	tribes PL ACC	study do-PFV.3P-HON
	‘You have studied all of these tribes/You have done studies on all of these tribes.’

In this example, the word study cannot be analyzed as only a noun or a verb, since the word study could be used both as a noun, such as in the sentence, I did the study only recently or as a verb in the sentence, I studied this only recently. Therefore, examples such as these have been categorized as N/V(Eng) + ‘do’. In addition, only about 3% of the bilingual verbs in the films from this decade involved an English verb and about 2% of them involved an English noun.

5.1.2. Distribution of Complex Verbs in the Films of 1990s

The frequency of occurrence of complex verbs in the films of the 1990s is provided in the table below.

From

Table 3. Frequency of occurrence of complex verbs in the films of 1990s.

Films of 1990s	Complex verbs N = 214
	N(Beng) + ‘do’ %	N(Eng) + ‘do’ %	N/V(Eng) + ‘do’ %	V(Eng) + ‘do’ %
	71.02	7.94	8.87	12.14

we see that the percentage of occurrence of bilingual verbs not only increased in the films of the 1990s but also doubled in comparison to those of the 1970s. That is, almost 30% of the complex verbs in the films of the 1990s were of the bilingual type and about 70% were of the monolingual Bengali type. Of the bilingual verbs that occurred in the films of the 1990s, about 9% had the N/V(Eng) + ‘do’ form and about 8% involved an English noun, while about 12% involved an English verb. Therefore, those bilingual verbs, which involved an English verb, had a slightly higher percentage of about 4% than those involving an English noun. Even though the difference between N(Eng) + ‘do’ and V(Eng) + ‘do’ structures in the films of the 1990s is not very much and therefore perhaps not noteworthy, in the films post-2010, the difference increases dramatically.

5.1.3. Distribution of Complex Verbs in the Films post-2010

The frequency of occurrence of complex verbs in the films post-2010 is provided in

Table 4. Frequency of occurrence of complex verbs in the films of 2010s

Films of 2010s	Complex verbs N = 297
	N(Beng) + ‘do’ %	N(Eng) + ‘do’ %	N/V(Eng) + ‘do’ %	(Eng) + ‘do’ %
	34.34	14.47	18.87	32.32

. Table 4 shows that in the films of the 2010s, only about 34% of the complex verbs are of the monolingual Bengali type, while almost 66% are of the bilingual type. This is a drastic increase from the films of the 1990s where only about 30% of the complex verbs were bilingual. This increase is even more in comparison to the films of the 1970s where only about 15% of the complex verbs were of the bilingual type. Furthermore, the frequency of the V(Eng) + ‘do’ structure has also increased in the films of this decade and is at about 32%, which is almost double than that of the frequency of the N(Eng) + ‘do’ structure. The N/V(Eng) + ‘do’ structure has a frequency rate of about 19%. The frequency of the V(Eng) + ‘do’ structure in the post-2010 films is also double the frequency of the same structure in the films of the 1990s. This means that not only has the share of bilingual verbs increased in general in recent decades, but the share of this particular structure has also increased as compared to other bilingual verbal structures involving English nouns.

Figure 1 summarizes the frequency of occurrence of complex verbs in the films of the three decades.

Figure 1 shows that the share of bilingual verbs has increased greatly in the recent decades and that almost 65% of the complex verbs in Bengali are now bilingual in nature. Similarly, the increase in the frequency of the V(Eng) + ‘do’ structure in recent years is also noteworthy because in the 1970s, only about 3% of the complex verbs were of this type, while in the 1990s, only about 12% were of this type. However in the films post-2010, more than 30% of the complex verbs have the V(Eng) + ‘do’ structure. The rate of increase of other bilingual verbs involving the N(Eng) + ‘do’ structure and the N/V(Eng) + ‘do’ structure is also important to point out since in the films of the 2010s, these structures also doubled in comparison to the 1990s. This massive increase in the use of bilingual verbs leads us to examine the factors that may have caused this increase and what this means for the Bengali verbal system. However, before I delve into that issue, I will briefly examine the data in qualitative terms.

5.2. Qualitative Examination of Bilingual Complex Verb Use in the Films of the Three Decades

In this section I focus more closely on the types of complex verbs used in the films from the three periods. As we saw from Figure 1 in the previous section, almost 85% of the complex verbs in the 1970s films were of the monolingual Bengali type. This percentage included verbs, which occur regularly in everyday speech such as shuru, which means ‘start’ and bæbohar, which means ‘use’ as well as less frequently occurring verbs such as jorimana ‘fine’ and apshosh ‘regret’. Most bilingual verbs in the films of this decade were reserved for English words that either do not have Bengali equivalents or words for which the Bengali equivalents are rarely used, such as phone, post, money order, telegram and application.

In comparison, the films of the 1990s had a higher share of bilingual verbs (almost 30%) and a closer examination of those films reveal that in several cases the use of bilingual verbs are context specific. This means that specific contexts triggered a higher use of bilingual verbs as compared to other contexts where bilingual verbs were not as commonly used. For example, the central character in the 1992 film, Agantuk (The Stranger) [23] is an anthropologist who has extensively researched tribal cultures around the world. Therefore, when this character describes his travels, studies and research of tribal cultures and customs, he uses several English verbs, such as sponsor, prove, verify, introduce, study, etc. Topics such as, education, professions, jobs often trigger the use of English words since higher education in India is conducted almost entirely in English. Therefore, the use of English words in this context is not entirely surprising and in fact is even expected. We also find that the use of bilingual verbs in the films of this decade is limited to topics such as education, medicine, photography and law, whereas in the case of discussions on food, household chores, family values or descriptions of people, bilingual verbs do not figure that often.

The use of bilingual verbs in the films of recent years is however surprising. In these films, we find that bilingual verbs are not only used in instances where they are expected because of the context or topic (such as discussions connected with education or professions) but also in cases where they are not. For example, in the 2013 film, Hawa Badal, bilingual verbs such as employ kɔra, ‘to employ’, publicize kɔra, ‘to publicize’ and promote kɔra, ‘to promote’ are used in the setting of a multinational company. This is expected because English is the language of the global economy and using English verbs in such a context is normal. However, we also find that bilingual verbs are used in the context of a family setting, which are also in the case of frequently occurring actions such as use, start, try, stop, etc. Even for the same complex verb, when the films of the 1970s use a monolingual Bengali verb, the films post-2010 use its bilingual counterpart, such as use, start, try, stop and deny. Also as expected, most bilingual verbs occur in the speech of young, English educated speakers as compared to middle aged or older speakers, since these young speakers are connected through technology to a globalized world, which requires them to speak English on a daily basis.

The rise in the use of bilingual verbs in the films post-2010 is reflective of the deep inroads that English has made in Indian society, especially in the speech of young urban dwellers. While in the previous decades the use of bilingual verbs were much more context specific and topic specific, such as education and employment, nowadays they are used even within informal interpersonal communications such as, within the family.

5.3. Discussion of the Results

This section discusses the results of the quantitative and qualitative examination of bilingual verb use in the films of the three decades as was laid out in the previous sections. What we observe from the previous sections is the fact that the share of bilingual verbs has increased drastically in recent years as compared to the films of the 1990s and 1970s where only as much as 30% of the complex verbs were bilingual. First, and perhaps most obviously, since English use has in general increased drastically in India in recent years, particularly due to the spread of globalization, technology and Western influences through the media, the increase in the use of bilingual verbs is expected. As a consequence, we also expect that the media and films, which draw from the real world, will reflect a similar phenomenon, which is what the films examined in this paper show. As mentioned before, in a closely related study by Si [13] on the use of code-switching in Hindi films, we find a very similar trend, where Hindi films of the 2000s show a major increase in code-switching including more complex code-switching as compared to Hindi films of the 1980s.

Secondly, an increase of bilingual verb use in the films of the recent decades indicate a restructuring of the Bengali verbal system where the majority of complex verbs are bilingual in nature, and where English lexical elements (nominal and verbal) are allowed to freely occur alongside the helping verbs of Bengali.

Based on the studies by Muysken [9], Backus [8], Annamalai [4,5] and Romaine [3], one could also argue for another dimension to the morphosyntactic change in the X + do verbal structures, comparing the recent films to those of the 1970s. This change lies in the occurrence of the V(Eng) + ‘do’ structure; in monolingual Bengali only Bengali nouns (and adjectives) can grammatically occur in complex verbs before the ‘do’ verb, while in bilingual Bengali different English elements including nouns, verbs, verbal nouns and phrasal verbs are allowed as part of complex verbs. I argue that the English verbs can now combine with Bengali helping verbs as a result of a structural reanalysis in modern Bengali speech. This is because Bengali furnishes the X + do structure, but the native X + do construction only licenses a noun in the place of X, but in contact with English, the co-occurrence of English verbs with Bengali ‘do’ in V + ‘do’ is not only licensed but also far exceeds the occurrence of English nouns in N + ‘do’ complex verbs (as evidenced in the post-2010 films).

I would like to propose a hypothesis about why more English verbs than nouns are used in the bilingual X + do structures in recent times. Intuitively (based on my judgments as a native speaker of bilingual Bengali–English), a distinctive, yet subtle interpretation seems plausible for Bengali–English V + ‘do’ complex verbs, in examples such as advertise kɔra ‘to advertise’, as opposed to N + ‘do’ complex verbs such as advertisement kɔra ‘doing advertising’. In the former case, the action of advertising is being done, whereas the latter seems like a more indirect, periphrastic way of expressing the action of advertising. Semantically, the latter structure indicates that one is doing an action, which happens to be advertising. Extending this interpretation to examples with other verbs or nouns that occur in the films, it seems that in Bengali–English bilingual verb structures, some English nouns are incapable of contributing the same semantic information to the complex verb as English verbs are able to. Three examples are provided below which attempt to illustrate this subtle difference. In (7a), the monolingual Bengali complex verb is ɔsshikar kɔra, ‘denial do’ which means denying in English. In bilingual Bengali speech however, the same complex verb has a V+’do’ structure, deny kɔra ‘to deny’ (7b) because it seems to be a more straightforward way of expressing the act of denying. In comparison, the bilingual N + ‘do’ denial kɔra might not be the preferred choice (7c), because intuitively, it seems like a more complex, periphrastic way of expressing the act of denying something. In (7c), the meaning of the complex verb corresponds to ‘doing an act of denial’, which sounds too awkward for use.

7a.

tomake

bhalo

bashi

sheta

ɔsshikar

korchi na

(N(Beng)+’do’)

2SG-ACC

good

love-PRS.1P

that

denial

do-PRS-1P-NEG

‘(I) am not denying that I love you.’

b.

tui

to

baar baar

deny

kor-e

jacchish

(V(Eng)+’do’)

2SG

FOC

constantly

deny

do-PFV.PTCP

do-PRS.PROG-3P

‘You are constantly denying it.’

c.

chatro-ta

accusation-ta-ke

denial

kor-ech-e

(N(Eng)+’do’)

Student-DEF

accusation-DEF-ACC

deny

do-PFV-3P

‘The student has denied the accusation.’

Other examples of English nouns and verbs that have these subtle meaning differences include permit kɔra/permission kɔra; react kɔra/reaction kɔra; decide kɔra/decision kɔra; object kɔra /objection kɔra among several others.

This brings us to one of the possible reasons why we find more English verbs in the X + do structures of the films of recent times. Firstly, given that more bilingual verbs have the V + ‘do’ structure, than the N + ‘do’ structure, there may be something special about verbs that nouns do not appear to contribute in this specific case. Secondly, because speakers have the option of choosing between two grammatical structures that have subtle meaning differences, a monolingual and a bilingual one, it may be the case that they prefer structures that appear to be more semantically direct than others. As I discussed regarding (7), some bilingual combinations involving English nouns do not have the same referential meaning as combinations involving English verbs (see Chatterjee [40] for more discussion on the distinction between V + ‘do’ and N + ‘do’ structures in Bengali–English speech).

This observation also connects with what Annamalai [5] had proposed about balanced bilinguals using more English verbs than nouns in Tamil–English bilingual speech. Because balanced bilinguals have more lexical access to different English constructions, they may use those that are more semantically applicable, while the monolingual Bengali speakers having little knowledge of English use the Bengali nominal counterparts of English verbs. Since English use in the 1970s was not as pervasive and extensive as it is today, it is possible that speakers were not equally proficient in both Bengali and English and therefore resorted to more Bengali verbs. The rise in use of bilingual verbs, especially the rise of the V(Eng) + ‘do’ structures in recent years in Bengali–English speech is comparable to what Backus [8] observed for Turkish-Dutch bilingual speakers. In his study on three generations of Turkish-Dutch bilinguals, he observed how the use of V + ‘do’ structures increased as bilinguals became equally proficient in both languages. Another study by Balam [41] on Spanish–English CS relates to the observations presented here. In his cross-generational analysis of 1750 Spanish–English BCVs produced by 56 native bi-/multilinguals from Northern Belize, Balam found a drastic increase in the use of BCVs among the younger generations who are more proficient in English and Kriol. The striking cross-generational change in the use of BCVs was attested not only in the frequency of BCV use but also in the syntactic verb types and pronominal forms in BCVs.

If societal bilingualism increases and speakers start to code-switch often then it is not surprising that the use of bilingual verbs will not be limited to specific contexts and topics. Since bilingual verbs are most commonly observable in code-switched speech, the use of bilingual verbs in such contexts is normal. This is what we observe for Bengali–English bilingual speech in recent years. Since code-switching has become an extremely common phenomenon in the speech of urban Indians, the simultaneous increase in the use of bilingual verbs is also a common feature. The films of post-2010 illustrate this very fact. As a consequence of this, the sociolinguistic factors that governed the occurrence of bilingual verbs in the earlier decades (such as age, context, topic, etc.) become less relevant in recent years. Together, the present study, Backus [8] and Balam [41] strongly suggest that an increase in societal bilingualism (including societally-sanctioned code-switching) and bilingual proficiency among younger generations results in a corresponding increase in the use of BCVs, which may be accompanied by the emergence of innovative patterns and/or forms, such as the use of the V(Eng) + ‘do’ structures (for examples of innovations in Spanish/English CS, see Balam [41,42]). Given the discussion so far, one could argue that since all the films examined in this study were based in urban areas with educated middle and upper-middle class characters, the English proficiency of the characters has remained unchanged and that only code-switching has increased in recent years due to which the use of V + ‘do’ structures has increased. This could be a possible reason behind the V + ‘do’ structures always outnumbering the N + ‘do’ structures even in the films of the 1970s. However, two things need to be pointed out in that regard. First, due to the smaller number of tokens of bilingual verbs in the 1970s and 1990s films, this difference between V + ‘do’ and N + ‘do’ structures is probably not statistically significant. Second, given the nature of the data examined in this paper, it is difficult to tease apart the extent to which increased bilingual proficiency or increased societal CS is responsible for the higher use of V(Eng) + ‘do’ structures in the films post-2010. It is likely that because the movies were deliberately selected to center around middle/upper-class speakers, it is their proficiency in English that is responsible for the greater number of V + ‘do’ structures in all decades. Further studies could test this issue by following particular English word pairs over the decades to ascertain if nouns tend to change to verbs over time, that is if there are more N(Eng)+’do’ structures such as revision kɔra ‘to do revision’ in the 1970s, but more V(Eng) + ‘do’ structures such as revise kɔra ‘to revise’ in the 2010s. If this is the case, then the higher number of V(Eng) + ‘do’ structures in recent times may be due to higher proficiency in English. If not, it might simply be a case of an increased English lexicon that is available for incorporation into Bengali–English bilingual speech, or in other words, a case of diachronic change in the mixed code itself.

6. Conclusion

This paper has examined the use of Bengali–English bilingual verbs through the exploration of Bengali films from three decades. Two films each from three decades, the 1970s, the 1990s and the 2010s were examined for their use of complex verbs in order to gain a diachronic perspective on possible changes or developments in these structures. The results show that the use of bilingual verbs have greatly increased in the films post-2010 in comparison to the films of the earlier decades, where only as much as 30% of the complex verbs were of the bilingual type. Not only has the share of bilingual verbs increased in recent years, it is a specific type of bilingual verb, the V(Eng) + ‘do’ structure, involving an English verb with a helping verb ‘do’ from Bengali, which has the highest rate of occurrence of other types of bilingual verbs involving English nouns. Previous studies have linked the use of more bilingual V + ‘do’ structures to higher bilingual proficiency among speakers, which is partly relevant also for Bengali–English bilingual verbs. Since English use has increased greatly in recent years in India due to the spread of globalization, economic expansion and technological innovations, speakers have become more proficient in English and have started code-switching extensively between Bengali and English. Given this, the use of more bilingual verbs in Bengali–English speech in recent years is not unexpected. Therefore, while in the films of the earlier decades bilingual verbs were mainly used for verbs which either did not have Bengali equivalents or used in specific contexts and topics (such as education or employment), in recent years, bilingual verbs are used in a range of topics and not necessarily limited to specific contexts. The sociolinguistic factors that governed the occurrence of bilingual verbs in earlier decades have become less relevant in recent years.

The use of more V + ‘do’ structures in bilingual speech as compared to N + ‘do’ structures may also be linked to the hypothesis I proposed based on semantic differences. It seems, intuitively, that bilingual V + ‘do’ structures are more straightforward semantically than bilingual N + ‘do’ structures, which amount to a more complex way of expressing an action. Therefore, bilingual speakers who have access to both Bengali and English lexical items may prefer to use structures that are more direct and ones they are more familiar with. Since urban bilingual speakers are often required to speak fluent English in order to be competitive in the employment sector, several Bengali words have fallen into disuse in recent years. The subtle semantic differences between bilingual V + ‘do’ and N + ‘do’ structures, including the easy access that bilingual speakers have to English vocabulary items may together be responsible for the use of more bilingual verbs and specifically more bilingual V + ‘do’ structures.

There are several research questions that could be investigated further in terms of bilingual verb use in Bengali–English speech. For example, the bilingual V + ‘do’ and N + ‘do’ structures could be analyzed from the perspective of a theoretical framework. Secondly, the hypothesis I presented in terms of the differential use of V + ‘do’ and N + ‘do’ structures could be investigated empirically in a more grounded way, the results of which would provide us with insights on many previously unanswered questions. However, in spite of these outstanding issues, this paper has not only furthered our understanding of bilingual verb use in the context of Bengali–English speech, it has also illustrated that media data can be used successfully in understanding the trends in the speech patterns of a community.