4. Discourse Analysis
Texte intégral
1Having discussed what I consider to constitute evidence of ontological insecurity, and the securitisation of ontology (through overcoding), the purpose of this chapter is to provide evidence to this effect. In the first section I introduce the corpus data and pinpoint areas for closer examination, as well as noting to what extent the discourse has changed. In the second section I focus on the signs of indeterminacy and the ontological insecurity this produces. In the following sections I explore this further, paying particular attention to how this indeterminacy is overcoded through discourse, and how this seeks to produce a sense of ontological security.
2The attention in the second half of the chapter is thus on what the discourse does: how it creates and curates Bangladesh and how this produces, instantiates and maintains the border/difference between India and Bangladesh. In creating Bangladesh as a dystopic spectre, Indian ontological security is advanced. As I discussed in chapter one, and will return to in the next chapter, the process of nation-building in India has hinged on the need to make Partition final: to naturalise the notion of nationhood and to realise the nation as distinct and imbued with ontological status. The history of Indian nationalism and nation-building has been centred on the creation of such a presence, precisely by curating difference. But because this difference evaporates at the margins, it is asserted but ungrounded. As Bangladesh’s first Principal Secretary commented in 1985: “[we need to] establish our separateness, since we are basically the same people with the same history and culture. Because we desire a separate political homeland, negative sustenance becomes absolutely necessary” (quoted in Hassan, 1987: 180). This is as true for all three ‘children of partition’. Thus, in the following pages, I look at how the discourse establishes this separateness, generates negative sustenance and, in doing so, denies and obliterates the ambivalence that challenges it.
4.1 Overview
3Tables 1 and 2 are collocate lists of Bangladesh and from Bangladesh respectively. They denote the top 35 nouns, adjectives and verbs, with sports references and place names removed. Tables 3 to 5 are the diachronic breakdown of collocates by year, again according to nouns, adjectives and verbs, for the word Bangladesh. The MI3 score, as discussed in chapter three, is a measure of association; the higher the number, the more strongly associated the collocation is. Collocate frequency and corpus frequency are present in tables 1 and 2 and show how many times the collocate appears proximate to the node (Bangladesh or from Bangladesh) and how many times it appears overall in the corpus. This column will not be engaged with in the analysis but is included to provide a sense of frequency. This is omitted in the diachronic tables due to space constraints. These tables show how Bangladesh and Bangladeshis are predicated in the discourse. I have used these tables as an entry point into the discourse, using them to identify patterns that I then analyse in more depth.
Table 1. ‘Bangladesh’ collocation lists

Table 2. ‘from Bangladesh’ collocation lists

Table 3. Bangladesh collocation lists by year, Nouns only

Table 4. Bangladesh collocation lists by year, adjectives only

Table 5. Bangladesh collocation lists by year, verbs only

4The first thing to note is that, as expected, there is a prevalence for discourses that overcode Bangladesh and Bangladeshis. There is a consistent negative semantic preference undermined only by signs of diplomatic success (historic agreement – the Land Boundary agreement discussed below). Words such as good and brilliant are used in reference only to sport.1 Some words might be said to be semantically neutral and some words are purely functional (particularly adjectives and verbs) as would be expected. Otherwise, the majority of the collocates imply or refer to a negative representation. At a very broad level then, we can say there appears to be minimal semantic competition.
5Indeed, this negative semantic preference can be divided into at least two discourse prosodies. A discourse prosody is the semantic character that particular words keep. For instance, in Baker’s (2014) study of portrayals of homosexuality, he found that a common collocation of gay was allegations, implying that homosexuality was something shameful. However, he also noted that other collocates like slurs, confessed and smears worked in the same way. Discourse prosody enables one to systematically analyse collocation patterns and draw links between numerous collocates (Baker, 2014: 111). Using table 2, and exploring concordances, I identified two main discourse prosodies: Bangladeshis come to India escaping persecution, and Bangladeshi migrants are pernicious. Both of these represent Bangladesh negatively, either as a place unsafe for minorities, or as a place and a people mired in crime, violence and illegality.
6The first discourse prosody is evident in the collocates highlighted green in table 2. Many of the words in this prosody emerge from and are prevalent because of the discourse surrounding the 2019 Citizenship Amendment Act (CAA), that granted non-Muslims (defined as Jains, Parsis, Hindus, Christians, Buddhists and Sikhs) from Bangladesh, Pakistan and Afghanistan the ability to claim Indian citizenship. Indeed, the prosody is far less prevalent in the 2011, 2013 and 2017 sub-corpora, though it does exist (see below). It should be noted that there are no Jains, Sikhs or Parsis in Bangladesh, and so these words are collocates only because of the CAA discourse. I will return to this in more depth below, but at this point it is important to note that the CAA was successfully justified by claims that these three countries persecute their religious minorities who then come to India as refugees. The reflection is therefore that Bangladesh is either unwilling or unable to provide a safe and just environment, and that conversely, the Indian state and nation is the opposite. This discourse prosody therefore confirms the notion of overcoding, as discussed at the end of chapter three.
7The second discourse prosody, which is prevalent throughout all sub-corpora, is that of Bangladeshis as pernicious. The words highlighted red show them to be framed as illegal infiltrators who sneak into India, often smuggling fake/counterfeit currency/notes, and who indulge in criminal activities (burglar) for which they are sometimes arrested. This group is presented as an influx that is unabated and poses a demographic challenge that has swamped or hurt the Indian nation, in particular Assam, which was the premise of the Assam Accord. This is reinforced if we look only at collocates that immediately modify Bangladeshi (i.e. L1R0). The top two collocates were illegal and suspected. Repeating the search for L0R1, the top collocates were immigrant and migrant, followed by settlers and nationals. We can therefore see that Bangladeshi is primed to be modified by illegal and suspected and to be followed by (im)migrant. We can take this further by exploring it as a collocation network. A collocation network is a visualisation of chains of collocates (see Brezina, McEnery and Wattam, 2015; Baker, 2016). Following collocations, and the collocations of those collocates, I found four formulaic phrases with the word Bangladeshi:
Illegal > Bangladeshi > (im)migrant > held/arrested/caught/nabbed
Illegal > Bangladeshi > (im)migrants > are > flooding
Suspected > Bangladeshi > immigrants
Suspected > Bangladeshi > nationals
8In exploring collocation networks, we can understand how patterns of language use emerge in the discourse and how these may be said to become formulaic. Primings prime further primings and language use, while creative, is also visibly frequently patterned and primed through regular use. This is incidentally very close to post-structural ideas of intertextuality and performativity, and of discourse as a structuring regime. This collocation network also allows us to explore collocational competition, i.e. the collocates that compete to occur with any given node. Illegal and suspected compete to modify Bangladeshi. In this case, the collocates share a negative semantics and a discourse prosody for doing something they should not. If the competition were greater, and included more diverse options (e.g. intelligent, wealthy), this would evince a more varied representation of Bangladeshis in the discourse. Since it is not, we can say that Bangladeshis are represented consistently, with little variation.
9Importantly, three of these collocation networks and the lexical primings therein provide support for the presence of ontological insecurity. The second one coheres with a wider semantic preference for border movement already identified, and suggests that Bangladeshis are flooding into India. As I shall argue further below, this is indicative of two things. Firstly, it denies ambiguity by constructing the identity of the Bangladeshi as unproblematic (illegal is not modified as alleged, suspected, likely, etc.) even though this is usually far from the case. Evidence of this is seen in very low conviction rates. The Foreigner’s Act places the burden of proof on the accused. Data on how many are arrested are subsequently put before a Foreigner’s Tribunal, and how many of those are found to be non-Indian, are not available. In Assam, the conviction rate of the Foreigner’s Tribunals has been around 4% (Rix, 2016). The NRC figure for illegal migration in Assam was slightly higher at 5.75%, though this included a significant number of indigenous people. It is likely that figures for the rest of the country would be much lower given Assam’s proximity to Bangladesh. Nonetheless, this widespread ambiguity is omitted and overcoded. Secondly, and relatedly, it begins to construct the porosity of the border as an alibi; this sense of flooding is created precisely because there is a degree of indeterminacy between Indians and Bangladeshis. The border is not as porous as it is made out to be, and many of those claimed to be Bangladeshi are in fact not.
10This is pre-empted in the third and fourth formulaic phrases: suspected Bangladeshi nationals/immigrants. I suggested in chapter three that one of the indicators of indeterminacy that I expected was the frequent use of modifiers such as suspected. Evidence that suspected is in fact one of the words most primed with Bangladeshi is therefore strong evidence of this indeterminacy. Having given a brief overview of the corpus and the discourses therein, I shall now discuss what changes have occurred, before delving into an in-depth analysis, beginning with the notion of this idea of the ‘suspected Bangladeshi national’.
4.1.1 Has the Discourse Changed?
11One of the reasons for building a corpus that stretched over several years was to understand what elements of the discourse have changed or remained constant over time. My argument that the indeterminacy of the difference between India and Bangladesh is a source of ontological insecurity would be strengthened if I could show that the concern with Bangladesh, Bangladeshis and Bangladeshi immigration was not simply a temporary phenomenon, but a recurring theme in Indian politics. This is quite clear historically, looking at the origins of the Bongal Kheda and Assam Movement following Partition and rising after Bangladeshi independence in Kolkata and elsewhere, particularly after the fall of the Mujib government. Throughout the mid-1980s and 1990s this spread across India and culminated in a number of deportation drives. Although I am unable to go back this far in the discourse, I still seek to understand if the discourse has changed over the study period.
12To understand how the anxiety over Bangladesh transcends party-political positions I chose to construct a corpus that dealt with both Congress- and BJP-led eras. Importantly, I would expect to see some changes in focus over the period, because, as I argue, indeterminacy haunts different nationalisms in different ways. I therefore expected changes in terms of the relative prevalence of framings, but minimal change in the range of framings evident. The tropes and anxieties I identify have featured throughout the study years. This will be evident in the discussion of the concordance examples.
13I already noted that there appears to have been an increase in the concern over Bangladesh as majoritarian and Islamic and over ‘refugees’ who have left Bangladesh for India because of this. This has been the main change over the study years and coincides quite clearly with BJP rule and particularly with their passing of the CAA. However, the concordance lists actually overstate this increase because of the preoccupation with the CAA in the media. In fact, the discourse of Bangladesh as Islamically majoritarian and a source of refugees is evident in all study years, and, as I shall show, the CAA is evidence of this in its use of this discourse as legitimation.
14Besides this, two discourses that featured relatively frequently in the 2011 and 2013 sub-corpora but not in 2017 or 2019 concerned the Land Boundary Agreement and the sharing of the Teesta River. The former was an agreement that Manmohan Singh tried to finalise that sought to settle the issue of enclaves. The border between India and Bangladesh had around two hundred enclaves, where pieces of territory belonging to one country were surrounded by the other. The BJP were highly opposed to the issue, which they framed as ceding territory to Bangladesh, and were concerned with how many acres India would lose. Incidentally, the BJP then passed the agreement in their first year in office. The sharing of the Teesta River that runs through India into Bangladesh has been a diplomatic issue for decades, following the construction of a barrage that limited water entering Bangladesh. In 2011 Manmohan Singh negotiated a water-sharing agreement, but it faced opposition from the Chief Minister of West Bengal, and was shelved by the incoming BJP government. Neither issue is immediately relevant here.
15Additionally, the concordance lists in tables 3 to 5 would suggest terrorism has become an issue between 2011 and 2019, with both terror and Jamaat-ul-Mujahideen featuring on the 2019 list in table 3. This over-represents the data however. Partly, this is because of the way MI3 works in biasing exclusivity, and so because Jamaat-ul-Mujahideen is an exclusive collocate of Bangladesh it features in the list. Importantly, references to terror do exist in collocate lists for all other years, but do not make the top 35. The word blog in the 2017 corpus refers to attacks on secular bloggers by terrorists in Bangladesh that year. There is therefore evidence that references to terrorism and Bangladesh increased in 2019, but as I expected, this discourse did already exist prior to that.
16Lastly, discourse concerning Rohingyas was more frequent in 2017. This is because of the mass movement of Rohingyas to Bangladesh following violence in 2017. References did occur in 2013 and 2019, though less frequently.
4.2 The Obsession with Documentation and ‘Suspected Bangladeshis’
17In this section, I focus on demonstrating both indeterminacy and the anxiety this causes. One way to show this is in the challenge of identifying Bangladeshis and Indians. As noted, the prevalence of the collocate ‘suspected’ concerns precisely this indeterminacy, in that people are suspected of being Bangladeshi, but this is rarely proven. It is nearly impossible to prove conclusively the non-Indianness of those suspected of being Bangladeshi.2 This causes anxiety over how to prove nationality in a context where essentialised markers of difference do not exist, where many (poor) Indians are under-documented, where bureaucratic inefficacy renders the verification of documents problematic, and where there are very few documents that can legally prove citizenship (Roy, 2016: 348; Jayal, 2013: 72). A focus on documentation betokens this difficulty and highlights the need to identify both Bangladeshis and Indians and thus to reinsert determinacy. If determinacy were straightforward, documentation would not be so important.
18This focus on documentation is not immediately apparent in the collocate lists, though it is intimated in the adjective undocumented in table 2. To explore this further, I treated the words Aadhaar,3 voter ID, pan card, passport, ration card and ID as a single collocate of Bangladeshi, and manually calculated the MI3 score, which was 20.2. I did this because, in as far they concern documentation and proof of Indianness, they work as synonyms.4 It should be noted that this score would put it in the top ten noun collocates of Bangladeshi. However, this under-represents the prevalence of the discourse surrounding suspected Bangladeshis and documentation, because the collocate list is built with an L5R5 window. I read and manually coded 100 randomly selected articles with the words Bangladeshi immigrant or Bangladeshi migrant; 71 discussed the challenge of identification and proof of citizenship. Below are three examples:
…the intelligence bureau had conducted a vast survey in 2003 when thousands of people with doubtful credentials were traced in Varanasi and surrounding districts. It was suspected that they were Bangladeshi immigrants who had started dwelling here to earn livelihood by working as rag pickers, rickshaw pullers and labourers. However, when the exercise to detect their parental roots with the help of state police like West Bengal failed to yield any positive result, the issue was put on the back burner… the local intelligence unit had also made a similar exercise when it had succeeded in tracing over 800 suspected Bangladeshi immigrants. But, again, the efforts to trace their roots did not yield any fruitful result. (TOI, 2011)
"We are Indians and no one has the right to label us Bangladeshi. Such a claim will render us jobless. Each one of us has documents such as voter identity card and Aadhaar card. Why would anyone suspect us to be immigrants in our own motherland?" said 22-year-old Sabina. (HT, 2017)
Police in Tripura are in a fix after 18 Muslim men of suspected nationality allegedly sneaked in from Bangladesh armed with Aadhaar cards, saying they were from Murshidabad in West Bengal… Police are trying to establish if the Aadhaar cards are genuine and how the 18 reached Tripura from West Bengal via Bangladesh. Sepahijala superintendent of police Sudipta Das said they were yet to get permission to interrogate the youths who are in judicial custody till November 6. "We are trying to find their identity through UIDAI and whether or not they are Bangladeshi nationals," Das said. (HT, 2013)
19In these examples, the indeterminacy of those ‘suspected’ is evident. The ubiquity of the discourse about the inability to identify Bangladeshis is evidence that this inability is inducive of ontological anxiety over indeterminacy. In these examples, what is clear is the indeterminacy of national belonging and citizenship. There is no clarity beyond documentation as to the citizenship of ‘suspected Bangladeshis’ and ‘dubious’ Indians. The pervasive focus on documentation also signals the provisional and unstable nature of national belonging and the presence of a national self. For many people in India, the only thing that marks them out as Indian or not, is their ability to show documentation. This may sound obvious, but in many other contexts, nationhood and citizenship are more naturalised. The fact that nationality is often forcibly reduced to documentation undermines the notion of a nation that is cohered by a tangible Indianness.
20In India there is often a distinct anxiety surrounding the need for documentation, particularly in places like Assam, West Bengal, Tripura or Odisha, where people have regularly been required to prove their nationality. But this is also the case throughout India, particularly with that large section of the population who are outside the nation-proper, living in what Chatterjee (2004) calls ‘political society’. For these people, documentation becomes a top priority. In my own experience in Assam and Tripura for instance, people are fully conversant in the language of bureaucracy and citizenship. In Assam one is regularly told what admissible documents they have, what category A and B documents they possess, whether their ancestors can be found in the legacy data, etc. For members of a community that is one of the least developed, with one of the lowest literacy rates in India, the degree of fluency in the documentary regimes of the state is remarkable, and signifies the essentiality of such a regime to their life (relatedly, see e.g. Gupta, 2012). Documentation here becomes the only marker of nationhood (and the only thing separating people from citizenship and statelessness). This is precisely what is reflected in the corpus by this focus on documentation or the lack thereof. The anxiety of these people vis-à-vis documentation is a function of the state and nation’s anxiety over their uncodability.
21This inability to prove a person is Bangladeshi is the reason India has such difficulty deporting them, as Bangladesh understandably requires proof of nationality. As a result, the police not only struggle to prove their suspicions, they also have no way of dealing with those they do deem Bangladeshi. Consequently, police forces prefer displays of physical force over arrest, trial and deportation, a dynamic common to policing of the Indian poor more generally (Roy, 2010; Chatterjee, 2004). Often those accused are simply stripped of any Indian documentation they do have and are declared infiltrators (Chhotray, 2018). This then compounds the issue by destroying what little documentary evidence they did possess. Indeterminacy and lack of proof means that Bangladeshis are almost always ‘suspected’ rather than ‘proven’. One of the key ways Bangladeshis are ‘suspected’ is their use of Bengali, often spoken with a ‘Bangladeshi’ accent.
Police said that these migrants have crossed the border from Dhaka and have entered India while their advocates maintained that they were Bengalis and from Odissa [an Indian state] and were caught only because they spoke Bangla. (TOI, 2013)
The Bangladeshis settled in Delhi could speak good Hindi and you can't distinguish them from their accent. (TOI, 2011)
The… [victims] told police that the robbers were speaking in an Assamese/Bangladeshi a dialect [sic]. There is a huge population of such natives in Auragabad. (TOI, 2017)
22In the absence of evidence, statements such as this become implicit legitimisations of allegations of foreignness. Van Leeuwen (2007: 93) argues that studying legitimation involves studying ‘residues’ – “a ‘residue’ of elements that could not be said to add anything to the description of what actually went on”. Here, in lieu of documentary proof of non-citizenship, other markers are used in an attempt to restore determinacy. Their use of Bengali, or of languages (Oriya, Assamese) that sound similar, is a common means of identifying ‘Bangladeshis’. In states adjacent to India, often a Bangladeshi accent is remarked upon. Needless to say, none of these are effective ways of distinguishing nationality, and many of those who stay in India for some time learn to imitate a West Bengali accent, or become fluent in Hindi.
23Bangladeshis therefore take on a spectral quality. Their common indeterminacy with eastern Indians means that they both blend in and are able to take on an outsized appearance. Precisely because they blend in with other Indian communities, who then become branded as ‘suspected Bangladeshis’, the Bangladeshi migrant becomes hypervisible in India. The combination of hypervisibility and invisibility compounds the spectral nature of Bangladeshi immigrants, who are consequently thought to be ‘flooding in’. In Assam, for instance, there are between 9 and 13 million people who might be considered ‘suspected Bangladeshis’.5 The fact that the NRC determined only 1.9 million to be foreign, and that this is likely to reduce to around 100,000 following appeals, is evidence of this indeterminacy (Karmakar, 2020).
24Indeed, the purpose of the NRC is precisely to uncover (but really to produce) the determinacy of the nation that is obfuscated in the borderland. Everyone in Assam has had to prove that their family has been in India since 1971. As expected, evident in the discourse is the extreme difficulty encountered in implementing the NRC, and the anxiety that seems to surround issues of determining nationality. There were over 300 articles in the corpus that dealt with the difficulties of the NRC, particularly the ability to prove citizenship and illegality. Note the following examples:
A retired army officer was asked to prove his Indian nationality as Assam police branded him a Bangladeshi… A foreigners tribunal served a notice to Mohammed Azmal Hoque, who retired as a junior commissioned officer (JCO) last year after serving the army for 30 years. The Guwahati-based Hoque maintains that his family is indigenous Assamese and his father's name is mentioned on the voters list of 1966. His mother's name was listed in the 1951 national register of citizens. "I have no doubt that I will get justice. But it pains me when my daughter questions me if this is how the country treats those who serve it for so many years," he said on Sunday. Hoque enlisted in 1986... The couple's son is studying in the Rashtriya Indian Military College in Dehradun and daughter at Army Public School in Guwahati. "This incident has saddened me a lot. Even after 30 years of service to the nation, we are asked to prove our identity. This is unnecessary harassment," he said. The tribunal had served notices to government officials before and a talked-about instance is Assam police constable Abu Taher Ahmed, who was accused of being an illegal immigrant. (HT, 2017)
Earlier this month, former Congress MLA from Goalpara Shadeed Mazumdar and his wife were served notices by the NRC office to re-verify their documents. Another former MLA Siddique Ahmed also got a similar notice. (HT, 2017)
25If military officers, constables and elected politicians, those thought perhaps to be unquestionably Indian, can be at the heart of the nation one minute and alien the next, there is a codability crisis. In cases like this, it is clear it is not merely about moral panic; it is the anxiety over the possibility of annulment. In Assam, the nation comes up against its own indeterminacy. Over the decades there have been numerous attempts to verify and exclude, and even the current NRC is not necessarily final, with the government suggesting it will redo it in Assam as part of a nationwide NRC.
26The suspected nature of Bangladeshis has at least two dimensions. On the one hand, it gets to the heart of the codability crisis: the total inability to prove national belonging. On the other hand, it suggests that certain populations are suspect: that something about them seems peripheral to the nation, neither inside nor out. Bengali Muslims and poor eastern Indians are assumed by many to be Bangladeshi precisely because they are seen to have the attributes of Bangladeshiness. There is an ambiguousness that pervades the notion of nation in India. Legally, nationhood is encapsulated in citizenship, but in most nationalisms (notably Hindutva, but also Gandhian and even Nehruvian nationalism), national belonging inheres in some combination of cultural, spiritual or religious markers (Oommen, 1997; Seth, 1992; Varshney, 1993; Krishna, 1994). Of course, the incompatibility of these visions, and the subsequent ambiguity regarding the extent of the nation has led to a great degree of anxiety. Ideally, the nation and state should be coextensive, with the nation being as clear cut as the territory it supposedly inhabits. The fact that the nation is not clear cut, and that, at its many edges, the nation is either everywhere or nowhere, undermines national selfhood. If ontological status – a sense of self – relies on difference within equivalence, the inability to specify this difference, and operationalise it (what defines the nation in the first place?) threatens to annul the positivity of an Indianness. This second form is incidentally very similar to Bauman’s (1990) framing of national indeterminacy and the anxiety this generates.
27Nonetheless, indeterminacy is perhaps more acceptable in those considered to be marginal to the nation. Those who are poor, rural, low-caste, and who inhabit slums or bordering states are peripheral to the nation-proper (e.g. Jalais, 2013). Indeterminacy is particularly disconcerting when it concerns the middle-class nationalist. The following article reveals not only the spectral nature of Bangladeshis and the indeterminacy of citizenship and belonging; it also betokens an anxiety over the inability to distinguish Indian from Bangladeshi:
Concerned after an illegal Bangladeshi immigrant was found with a Unique Identity card (Aadhaar card), all police stations in the city have been instructed to check housing societies to flush out such illegal residents who may have fraudulently obtained identity documents. The discovery of the Aadhaar card has sent alarm bells ringing in the security establishments as such documents can also enable illegal migrants to get an NoC [permit] from the police in buying or renting an accommodation in registered housing societies without raising suspicion. Sources in the special branch (SB II) of the Mumbai police, which arrested the illegal Bangladeshi from a south Mumbai housing society, told HT that senior inspectors have been asked to meet office bearers in housing societies and to run a check on the lease and licence agreements. Instruction has been given to real estate agents and property brokers to alert the police whenever they find any customer's identity document suspicious. Naval Bajaj, additional commissioner of police, SB II, said the police have also written to the departments concerned in the state and central governments to incorporate stricter conditions while issuing crucial personal identification documents such as the Aadhaar and PAN cards. (HT, 2013)
28In this article there are number of the indicators identified in chapter three. Not only are there indicators of a concern with indeterminacy, there are also indicators of the anxiety this causes. In this case indeterminacy literally sent alarm bells ringing. What is particularly interesting here is the concern that Bangladeshis are hiding in plain sight, in the midst of the Indian nation. The housing society is the home of the Indian citizen proper, the middle-class Indian of civil society (Chatterjee, 2004), and South Mumbai is one of the wealthiest places in India.6 Incidentally, there is a clear pattern in the discourse concerning Bangladeshis and residence in formal dwellings. If the Bangladeshi can blend in here without raising suspicion then there really is a codability crisis. Very clear in this example is the spectral nature of Bangladeshis, who haunt the Indian nation. The discovery of a Bangladeshi who had been passing off as a ‘respectable’ Indian citizen for years under the nose of fully paid-up members of the nation set alarm bells ringing. This is because the presence and indeterminacy of Bangladeshis here threatens to annul the positivity and ontological status of Indianness and of an Indian self, antithetical to and distinct from Bangladesh. If even the centre of the national self is compromised and contingent, a borderland in its own right, then the incoherence of nation is not just something seen on the territorial edge, but something that defines the nation. It is a case of the ‘irreducible margin in the centre’. If the nation is in fact defined by its indeterminacy, then it becomes clear that it cannot have an essential presence, that it is doomed and defined by its frantic attempts to disguise its indeterminacy. This troubled relationship with indeterminacy is what underpins nationalism; it is marked by its hauntedness. In the inability to determine nationality is the annulment of value.
29In a cartoon in the TOI (Adhwaryu, 2018),7 the codability crisis of Indians and Bangladeshis, and India and Bangladesh, is made explicit. The cartoon details the exodus of Indian migrants from Gujarat, who left following violence by local communities. The migrants meet on the road, with their bundles of clothing atop their heads. One of them asks “Bangladeshi!?” and the other responds “No… Desi!!”.8 The cartoon implies a number of things. Firstly, their dialogue suggests that Indian and Bangladeshi migrants are indeterminate. Secondly, it suggests that they are essentially the same; literally walking through the country with their possessions on their backs. Both have left their homes for work, and both have been the victims of violence and exclusion. The cartoon therefore challenges the difference between Indians and Bangladeshis, supposed citizen and suspected foreigner, but also, implicitly, India and Bangladesh. Bangladesh, a country of violence and poverty where its people are forced to leave, is rendered indistinguishable from India, where the migrant is also a victim of violence and poverty, consigned to a life of itinerancy. The cartoon thus raises the question of the relevance of nationhood in a context where the supposed national and suspected Bangladeshi are functionally the same.
30In the following sections I explore the ways that Bangladesh and Bangladeshis are represented in the discourse. These representations work to construct Bangladesh and those seen to be Bangladeshis as phantasmagorical. This works to overcode the indeterminacy that seemingly pervades nationhood in South Asia. Importantly, in constructing Bangladesh as dystopic, indeterminacy becomes even more dangerous, as it raises the prospect that either the dystopia is exaggerated or that India too is dystopic (if indeed we can still speak of an ‘India’ after its elision with Bangladesh). In both scenarios, the national self risks annulment. In this way, Bangladeshis in India do cause moral panic, but this panic is especially effective because of those instances of indeterminacy. The nature of Bangladeshis as all-pervasive yet spectral is what enables this panic to be effective, and what makes it ontological.
4.3 Bangladesh as Undevelopable
31Developmentalism has been central to Indian politics and nationalism since independence, and, indeed, is intimately tied to the postcolonial project around which nationalisms cohere. Securing development has particular significance in the context of postcolonialism, as it is understood as a way of reasserting an agency, subjectivity and pride that colonial knowledge systems have long denied. The picture of Bangladesh as undeveloped (or even undevelopable) feeds into the idea of Bangladesh as a postcolonial dystopia, and figures Bangladesh and Bangladeshis as phantasmagoric. Bangladesh as poor and underdeveloped thus reoccurs throughout the corpus:
…the migration is essentially driven by abject poverty and hunger prevailing in the neighbouring country. (HT, 2013)
The endless street battles and riots that characterised its politics for decades, the steady stream of migrants into India and elsewhere, and the threat of Islamicist influence have been a reflection of the country's poverty. (HT, 2013)
The Bangladeshi youth have always been very dynamic despite their poverty, unemployment and the country's several religious issues. (HT, 2013)
32In cases such as these, Bangladesh is presented as poor, politically unstable and violently Islamic, all themes explored further below. Of particular interest are those instances where Bangladeshi and Indian development are contrasted. Numerous articles that deal with India’s poor ranking in development indices compare India’s score with Bangladesh’s often-higher score as a way to reinforce India’s failures. The fact that India measures its own development failures against Bangladesh implies that Bangladesh is commonly perceived as less developed than India. The choice of Bangladesh as a juxtaposed comparison is intentionally done so as to maximise effect and reinforce Indian failure. Observe the following:
Even Bangladesh, no economic superpower, claims it has reduced the maternal mortality rate from 320 per 100,000 live births in 2001 to 194 per 100,000 in 2010. In India… [it] continues to be as high as 230 deaths per 100,000 live births… Bangladesh has put India to shame. (TOI, 2011)
Even neighbouring Bangladesh… could put India to shame with the corresponding figures [for female school enrolment]. (TOI, 2011)
Even Bangladesh has performed better than India in many aspects of human resources development. (HT, 2013)
Even Bangladesh had banned free sale of the corrosive substance and enacted a law treating acid attack. (TOI, 2013)
India's expenditure [on mental health] is an abysmal 0.06%, less even than Bangladesh (0.44%). (TOI, 2017)
India continues to be one of the poor performers ranking at 154, much below China, Sri Lanka and even Bangladesh. (TOI, 2017)
33Alternatively, Bangladesh is grouped together with states that are stereotyped as undeveloped, particularly those in sub-Saharan Africa. There is a long record in Indian and colonial political discourse of looking down upon Africa (Gandhi’s racism is well known for instance [Krishna, 2015]) and racism continues along these lines today (e.g. Gupta, 1991). While this is of course a product of colonial thought, and while there is also a history of Indo-African solidarity, the stereotype persists, as it does throughout the world. It is therefore no coincidence that sub-Saharan countries are grouped with Bangladesh to emphasise Indian failure:
Over one in five women in India aged 20-24 had given birth before the age of 18. In countries like Bangladesh, Chad and Niger, around a third of all women aged 20-24 are married by the age of 15. (TOI, 2011)
We can take solace from the fact that a worldwide scorecard on child marriage shows that we're better off than Bangladesh, Mali and Burkina Faso. (TOI, 2011)
India is doing worse than Bangladesh and sub-Saharan Africa in terms of malnutrition. (TOI, 2013)
Global Risks Atlas 2011 on Friday described India as the 16th riskiest country to invest in for the security hazards it poses and rather embarrassingly clubs it with Niger, Bangladesh and Mali. (TOI, 2011)
34Alternatively, Bangladesh is grouped with states that are commonly understood as ‘failed’ or ‘rogue’:
An International Food Policy Research Institute (IFPRI) ranked India below countries like North Korea, Bangladesh and Iraq in its annual report on global hunger. (TOI, 2017)
35Indeed, even though Pakistan is often negatively contrasted with Indian development, there are signs that Bangladesh is perceived as even less developed. There is a common perception in South Asia that Bangladesh is the least developed in the region. The verb ‘stunned’ gives away the cognitive dissonance produced by the claims that Bangladesh is more developed than others:
Pakistanis are also stunned that many reports now rank Bangladesh… ahead of it in several economic metrics. (TOI, 2017)
36Van Dijk (2018: 240) looks at how implication often signals presuppositions: statements which “represent the knowledge shared by speaker and recipients… and which is relevant for the production and interpretation of a… discourse”. Put simply, presuppositions are those pieces of knowledge that are assumed to be true and known, and which make elements of discourse interpretable. The notion that Bangladesh is undeveloped is presupposed in the juxtaposition of Indian failure and Bangladeshi success. Saying that India ranked below Switzerland in development indices amounts to a truism that does not require highlighting, and that does not destabilise the idea of India as an agential postcolonial state (and hence India is never contrasted with Switzerland in the corpus). The idea that India ranks below Bangladesh, however, is surprising, because it jars with the implied and presupposed understanding of Bangladesh as poor and India as advanced. The juxtaposition therefore challenges the validity of the difference between India and Bangladesh, and heralds the possibility of elision and the ‘death of the self’. Here, ontological insecurity is generated for rhetorical ends.
37If the developed/undevelopable divide is an instantiation of the necessary border between the self and the not-self, the suggestion that this divide is fictitious is a threat to the notion of self, for if there is no difference on a scale of equivalence, the nation’s ontological status is questioned. The indeterminacy of nationhood finds its defence in those discourses that both picture the border as porous and overcode the Indian poor as Bangladeshi. This then works as an alibi, both for Indian failure and for indeterminacy, as shown below. It is as if Bangladesh’s underdevelopment is infectious:
A slum has been festering in a coastal regulatory zone in Dahisar despite the best efforts of locals, activists and even corporators to remove it. The 8,000 shanties of Ganpat Patil Nagar on Link Road, Dahisar, apart from being an eyesore, are causing immense environmental damage… “illegal hutments kept proliferating in the area till the BMC [Bombay Municipal Corporation] razed the structures in 2003. But the activity started again right afterwards," said Abhishek Ghosalkar, a corporator. “The main occupants of the slum are migrants from Bangladesh and Nepal”. (TOI, 2013)
38Six months later, the TOI reports on the same slum:
this pocket of 8,000 shanties on mangrove land was a threat to the environment and a hotbed for water and power thefts. Dahisar residents had also said that the slum comprising migrants mainly from Bangladesh, served as a vote-bank. (TOI, 2013)
39There are several points worth highlighting. Firstly, the use of words such as festering, eyesore and hotbed work to emphasise the abhorrence of the settlement and the people who live there, who, like disease or bacteria, fester (like a sore) and proliferate. As though it was cancerous, they had tried to remove it, but it came back again. Secondly, the claim that the occupants (in contrast to the residents of the area) were mainly Bangladeshi is asserted but no evidence is provided, implying that it is easily believable and requires no elaboration. This is despite the fact that the Dahisar slum is one of India’s largest, and that it is highly unlikely that the majority of occupants are non-Indian. It is also noteworthy how the quote in the original article lists both Bangladeshis and Nepalis as the main occupants, but that this is reduced simply to Bangladeshis in the second article.
40The association between Bangladeshis and poverty, and the indeterminacy of Indians and Bangladeshis, also works through presupposition. Note the following:
The demand for detection and deportation of suspected illegal immigrants in Karnataka… is not new. The BJP's… Arvind Limbavali has for long claimed that there are many illegal Bangladeshi immigrants in the city. A slum where migrants from West Bengal reside, was to be demolished, but was eventually stayed by the Karnataka high court. (HT, 2019)
41Here, an article about Bangladeshi ‘infiltration’ ends with a sentence describing a slum where migrants from West Bengal reside. No link is made between this sentence and the rest of the article. Without knowledge of the presupposition that Bangladeshis ‘pretend’ to be West Bengali, and that by association poor Bengali migrants are themselves ‘suspect’, and without the association between Bangladesh(is) and poverty, this final sentence makes no sense. Here, once again, the figure of the Bangladeshi migrant provides an alibi for Indian poverty by casting doubt on the Indianness of the supposed residents. In another article, the link between Bangladesh and slums is developed further:
A settlement in Bhayander, named as ‘Bangladesh zopadpatti’ [slum] on government documents, has recently come to light. Interestingly, the slum has been carrying the name for over 35 years and is occupied mostly by people from the interiors of Maharashtra. Civic officials and residents of the slum, however, have not found anything objectionable in the use of the word ‘Bangladesh’ while referring to the locality. Two other ‘Bangladesh’ exist in Utan and Chowk villages of Bhayander. Here too, the occupants are from rural Maharashtra… “No one knows how these places came to be referred to as Bangladesh. No one has ever raised any objection.” Residents say that no Bangladesh national lives in any of these slums. A number of illegal Bangladeshis, however, have been fished out of various hutments in Mira-Bhayander. Civic officials said that the term Bangladesh was being used even before the municipality came into existence. (TOI, 2013)
42Mehta (2006) notes how the localities of another Mumbai slum took on the meanings of South Asian geopolitics. The square in the Muslim area became known as the ‘Parliament of Pakistan’ and the drain that separated Muslim from Hindu houses was commonly known as the ‘India-Pakistan border’, with yet another being the ‘Line of Control’. Similarly here, we have very poor areas becoming known locally (and even officially) as little Bangladeshes despite the absence of anyone from Bangladesh. It is no coincidence that these areas are some of the poorest and least ‘developed’ in India, and one can assume that, just as areas become named after Pakistan because of the religion of their residents, so these areas became named after Bangladesh because of the destitution of those who live there. This is prime evidence that Bangladesh and Bangladeshis have become overcoded in India, and it also subliminally creates a contrast space where Indian difference and selfhood become possible.
43Ultimately, the association of poverty, uncleanliness and unsightliness with Bangladesh feeds into an anxiety to secure the India of development and progress:
Two ambitious apartment schemes of the Lucknow Development Authority - Dhenumati and Kalptaru apartment - would be executed by construction giants Larsen & Toubro and Nagarjuna... The apartment tower, which would rise to a height of approximately 150 feet would look over a park magnificently carved out of red sand stones... For any apartment complex, security happens to be a major concern. Dhenumati apartment here takes the cake. It comes just adjacent to the Bahukhandi Vidhayak Niwas (MLA residence) and opposite Butler palace. The only sore for the scheme would be a huge slum opposite to the site. The slum spreads over an area of 10 acres and essentially has Bangladeshi immigrants …Each unit would have an area in the range of 96.18 sq mts to 149.62 sq mts and the sale pricing in the range of Rs 21.50 lakh to Rs 32.50 lakhs. (TOI, 2011)
44The security they speak of is not just a material one (though given Bangladesh’s association with crime this is clearly implied), it is also a conceptual one. It is about securing India as a progressive, agential and developing nation. The division between the rights-bearing citizen who resides in formalised dwellings and the Bangladeshi ‘infiltrator’ who festers in proliferating slums is evident. The use of the word sore here coheres with the general semantic preference for Bangladeshis as diseased (or more aptly as being a disease). It is the spread of this disease that is the security concern.
45Keeping the Bangladeshi out becomes crucial to keeping the Indian nation on the path of progress and development. It is not about any one of these slums or instances and their factuality as ‘Bangladeshi’. Rather they are emblematic of the threat of Bangladesh and Bangladeshis in India. By coding these areas as Bangladeshi, the problem of poverty in these areas is externalised, as a foreign problem to be removed rather than a key element of the Indian nation. Moreover, and more importantly, it provides a contrast space wherein India gains a presence in contrast to the poverty of Bangladesh. This contrast however, in part relies upon the viability of the order that separates India from Bangladesh. If this order were shown to be fictitious, if the contrast space were to collapse, there would be nothing to separate aspirant India from diseased Bangladesh. The implosion of the contrast space is the implosion of the self. This implosion was threatened in accounts of indeterminacy presented above, and the example of the housing society is particularly apposite here.
46Of course, I am not suggesting that poverty is always or even ordinarily branded as Bangladeshi. It is not. But these instances where it is serve to indicate the perception of Bangladesh itself as festering, where the slum becomes a synecdoche for the state and nation. In this context, the subsequent indeterminacy of the border and the citizen present a threat. If the presence of Bangladeshis threatens to undermine development in India, the collapse of the border threatens the difference between exceptional India and benighted Bangladesh. But precisely because India’s presence is a function of its difference, the threat to this presence by indeterminacy is a threat to the viability of a national self. The coding of Bangladeshis as always poor thus both signals a concern over the border and works to resurrect the border conceptually.
4.4 Bangladesh as Anarchic and Communal
47The idea that Bangladesh is underdeveloped feeds into an image of weak statehood, where the state is unwilling or unable to fulfil its historic role. Again, this discourse is important to the extent that it creates presence through difference for ‘India’. Perhaps the clearest example of the Bangladeshi state being represented as weak or anarchic is the discourse that Bangladesh is violent, communal and non-democratic. This was hinted at in table 2 where from Bangladesh was primed to co-occur with words like Hindu(s), religious minority and persecute, implying a discourse prosody of people fleeing persecution in Bangladesh. Indeed, there has been a steady rise in references to Bangladeshi minorities and their persecution. This is unsurprising given the BJPs ‘advocacy’ of Hindus outside India and their understanding of India as the rightful home of all Hindus. As noted, they passed the CAA in 2019 on this basis. The collocate list of from Bangladesh for the 2019 sub-corpus is dominated by references to the CAA. The same list from the 2017 sub-corpus has reference to religious minorities, Hindu(s) and persecuted but far less frequently. By 2013 none of the top 35 collocates include reference to religion or persecution, though in 2011 Hindu and refugee reappear. There is therefore a rise in discourse about minorities in the corpus, but the discourse itself is not new. Note the following:
Given that the persecution of religious minorities is rampant in these countries… (HT, 2019)
So why was this amendment necessary for the minorities of the three Muslim-majority republics of the subcontinent? Because they face extinction-level persecution. (HT, 2019)
…the formation of Pakistan and Bangladesh as Islamic nations led to persecution of Hindus and other minorities in the two countries, which necessitated the legislation. (TOI, 2019)
48Here, discussions of the CAA use a pre-existing discourse of religious violence to justify the granting of citizenship to non-Muslims (those that are considered Hindu under the terms of Hindutva). The presupposition is that the discourse of Bangladesh as a centre for Islamic fundamentalism and violence (often with the complicity of the state) is well known in India. This presupposed discourse then works to legitimise the CAA. Incidentally, the act has nothing to do with a concern for refugees, and everything to do with the BJP’s electoral politics in east and north-east India. This is evident in the fact that persecution is not mentioned in the wording of the act. The assumption is that it is so obvious that non-Muslims are persecuted in these countries that it does not need to be explicitly mentioned.
49Ever since at least the partition of 1905 there has been a sense of ascending political Islam and anti-Hindu sentiment in East Bengal. Indeed, incidentally, the spread of Islam in Bengal was largely due to Brahminical oppression of Buddhists and low-caste Hindus for whom the egalitarian ideas of Islam, particularly Sufism, offered an alternative (Uddin, 2006; Eaton, 1993; Roy, 1983). The partition of 1905 was supported by many East Bengali Muslims and ‘Hindu’ dalits because it ostensibly signalled a change from Hindu domination (Ludden, 2012). This support was viewed by many high-caste Hindus as a betrayal of Bengali identity and a capitulation to British policies of divide and rule (Sengupta, 2012; Chatterji, 1994). Since then, instances of communal violence and support for the 1947 partition have compounded the sense that East Bengal is a ‘hotbed’ of Islamic fundamentalism and majoritarianism, which has become woven into the narrative of Partition and nationalist history.
50In the colonial period, Bengal witnessed some of the worst instances of communal violence, particularly the ‘Great Calcutta Killings’ of 1946. News of these riots caused reprisals as far away as Punjab. Further reprisals in Noakhali and Khulna (present-day Bangladesh) also sparked violence elsewhere, with Gandhi himself spending several months in East Bengal trying to pacify the situation. Ultimately these instances informed India’s partition (Sengupta, 2012). Since independence, Bengali and Islamic majoritarianism have been key dynamics both in East Pakistan and Bangladesh, with isolated but not infrequent violence towards religious minorities (which has also been the case in India, importantly).
51This is of course a partial reading. As Ghosh (2015) argues, the violence of colonial Bengal was not necessarily ‘about’ religion, and was tied up with economic issues which were simplified in the press as communal.9 Moreover, as Bandyopadhyay (2004: 198-205) notes, the mobilisation of communal sentiment in 1940s East Bengal was due in large part to the Hindu Mahasabha10 seeking to instil a sense of Hindu consciousness among low-castes. Most of the Hindus in Bengal, particularly in the east, were (and remain) low-caste or dalit; they had far better relations with Muslim communities and did not identify as Hindu until the 1940s. Indeed, in 1906, Namasudra leaders11 petitioned the colonial government supporting the 1905 partition, saying: “Namasudras and Mohamedans are the predominating communities of Eastern Bengal, and the latter unlike the Hindus possess a good deal of sympathy for the Namasudras” (quoted in Ludden, 2012: 508). Furthermore, the fact that there was minimal violence following Partition in 1947 and that most non-Muslims remained in East Pakistan after Partition (in stark contrast to the west) point to the extent to which the idea of East Bengal as communal is partial.
52Moreover, the struggle for Bangladeshi independence was, on many levels, a struggle for secularism and democracy, and a continuation of the long-established notion of Bengaliness as a unifier. As one commentator wrote of Bangladeshi independence, the idea of Muslim nationalism, which had been “formulated in the middle-class living rooms of Uttar Pradesh, was buried in the Bengali countryside” (Ali, in Fazal, 2015: 82). The victory of independence (1971) and Bangladeshi nationalism represented a victory of Bengali identity over and above Islamic identity (van Schendel, 2009), and the fact that today 12% of Bangladeshis are non-Muslim is indicative that, though far from perfect, the country is not the communal dystopia it is made out to be. This story has sadly been sidelined, and, as a consequence, the image of Bangladesh as majoritarian, rife with violence against minorities, is commonplace in India, to the exclusion of a more complex view.
53Today, many point to a decreasing proportion of Hindus in Bangladesh as evidence of majoritarian violence. While not denying the challenges faced by minorities in Bangladesh, it is important to note that in fact the number of non-Muslims in Bangladesh has increased over the last decade, from 10% to 12%. Moreover, the categorisation of Hindu migrants in India as ‘refugees’ often misrepresents the myriad reasons minorities have moved to India.12 Another regular feature in the corpus was the common reference to Bangladesh, Pakistan and Afghanistan as Islamic Republics. Bangladesh is not an Islamic republic, and while it has Islam as the state religion it remains a secular country. Nonetheless, the grouping of Bangladesh with Pakistan and Afghanistan suggests similarities between both the treatment of minorities and the political stability in these countries. These similarities are hard to argue. While Bangladesh does have ‘radical elements’ (as does India), they do not operate at the level they have done in Pakistan or Afghanistan, and religious minorities have not faced the same degree of persecution. Moreover, while much of Afghanistan and parts of Pakistan have been or are controlled by Islamist groups, this has never been the case in Bangladesh, where state power has never meaningfully been challenged (except in the Chittagong Hill Tracts where a secular insurgency occurred).
54My point is not to suggest that minorities in Bangladesh do not face insecurity and discrimination, but rather to reinsert nuance, complexity and similitude between India and Bangladesh. When Bangladesh can no longer figure as a communal dystopia, and when Indian claims to rule of law and minority protection can be shown to be similarly partial, the ability to differentiate on a scale of equivalence becomes infinitely more challenging. Indeterminacy of sorts once again raises its head in the challenge of coding indeterminate difference. Instead, however, this indeterminacy is banished as Bangladesh becomes overcoded.
55The focus therefore is on what this discourse does. Overcoding Bangladesh as a place rife with Islamic violence and terrorism is useful to the extent that it helps India understand itself as politically stable and secular, or to enable Hindutva claims that ‘Hindu civilisation’ is more virtuous, and that Islam is necessarily perverting of Indian potential. The discourse of religious violence thus works to bolster ontological security for both secular and Hindutvavadi nationalism. For secular nationalists, the idea of Bangladesh as violently majoritarian contrasts with and makes possible their construction of India as secular, democratic and founded on the principal of inclusivity. Consider the following:
The riots that have broken out against the harsh but unsurprising sentences being passed down by Bangladesh's war crimes tribunal are a reminder of the continuing struggle between the country's Islamicist elements and its more secular Bengali instincts. (HT, 2013)
The endless street battles and riots that characterised its politics for decades, the steady stream of migrants into India and elsewhere, and the threat of Islamicist influence have been a reflection of the country's poverty. This poverty is no longer deemed inevitable… The present trial, whose credibility has already been undermined by incidents that included the disappearance of a defence witness at the court's gates, underlines how fragile Bangladesh's progress continues to be and how much more proactive India needs to be. (HT, 2013)
Haider's murder has enraged secular Bangladesh and split the nation into two. (HT, 2013)
56In these examples, Bangladesh is presented as a country split into two, with an Islamist half and a secular half: as if the murder of Haider (an atheist blogger) was supported by half of Bangladesh. There also seems to be a tension between ‘Islamists’ and secular Bengali instincts as if being Islamist (or even Muslim) and Bengali is a contradiction in terms. This implicitly feeds into the Hindutvavadi idea that the subcontinent is not naturally Muslim; there is an Indian core (i.e. Hindu civilisation), an instinct that has been perverted by the imposition of Islam. In this context, India needs to be more proactive in supporting fragile Bangladesh. India here is positioned as a benevolent but superior neighbour, whose role and right is to intervene in Bangladeshi politics. This feeds into the reason for the original souring of Indo-Bangladeshi relations; a backlash against Indian attempts at hegemony. Also of note here, is the construction of Bangladesh as anarchic and violent; the focus on riots and street-battles converges with an image of the state that is either unwilling or unable to provide order and security, captured by Islamist influence. As then Indian prime minister Manmohan Singh commented, “at least 25 percent of the population of Bangladesh swear by the Jamiat-ul-Islami and they are very anti-Indian, and they are in the clutches, many times, of the ISI” (Hazarika, 2011).
57Similarly, for Hindutvavadi nationalists, the idea of Bangladesh as Islamic and dangerous feeds into their mythology of Islam as destructive of virtuous Hindu culture and antithetical to a strong and virtuous political and social order. An HT opinion piece makes this clear:
The CAA is an affirmation of our diverse and inclusive Indian civilisation, which dates back nearly 7,000 years. Our existence as a modern nation-state, anchored in a constitutional republic, represents only 1% of our history. The inclusive character of this civilisational heritage is reflected by the fact that Malabar Jews, Syrian Christians, Parsis of erstwhile Persia, or my parents, fleeing the violence of Partition, have all found a safe home in our land. It draws inspiration from, and is a continuation of, this civilisational value system. Given that the persecution of religious minorities is rampant in these countries, the CAA seeks to secure those individuals in India who have fled from Pakistan and elsewhere due to the harsh conditions they lived through on account of their religion. (HT, 2019)
58This idea of the Indian state as the continuation of a 7,000-year-old civilisation is firmly rooted in Hindutva ideology, and the ‘history’ of Hindu civilisation created by Savarkar (1923). Like Savarkar, the writer characterises Hindu civilisation as inclusive and generous, bound together by a “civilisational value system”. The writer juxtaposes the virtuosity of ‘Indian’-as-Hindu civilisation with the persecution, intolerance and violence of Pakistan and Bangladesh.13 These countries are not party to this civilisation, despite sharing the same land (Akhand Bharat), ancestry, languages and, presumably, 99% of their history. This contrast works to place Bangladesh and Pakistan as inferior and dangerous, but importantly it does so in a way that elevates and exceptionalises India. Bangladesh and Pakistan are presented almost as a ‘heart of darkness’ characterised by savagery and violence. In both secular nationalism and Hindutva, Islam is seen as a symptom and a cause of Bangladesh’s failure to move from ex-colony to post-colony. The contrast provides an ontological space for understanding India. The overcoding of India and Bangladesh as wholly different reinserts determinacy into a space constantly undermined by the ambivalence of the uncodable.
59Similarly, secular India defines itself largely in reference to the non-secularism of its neighbours, particularly Pakistan and Bangladesh. The imbrication of Islam with politics in these countries is seen to necessarily end up perverting good governance and rule of law. Pakistan and Bangladesh therefore become the examples of what happens when a state strays from the path of secularism; they are the reminder of what could happen in India.
4.5 The Infiltrator
60The construction of Bangladesh as a ‘hotbed’ of radical Islam is linked to the prevalent notion of Bangladeshis as infiltrators, as noted in table 2. The rhetoric of infiltration first emerged in relation to Pakistan, and the supposed movement of Pakistani-backed insurgents into Indian Kashmir. In Assam, where Bongal (meaning foreigner, later Bengali) was the common term between 1940 and 1960, infiltration became increasingly popular from the 1980s onwards, perhaps as a strategic move to frame the issue more effectively, nationally. In Assam, the term is used for people of all religions. By the early 1990s, the BJP popularised the term infiltration in reference to Muslim Bangladeshi immigrants across India, and it has now become mainstream parlance, to refer both to Bangladeshis in general and to Bangladeshi Muslims. For Hindutva, the term is inherently tied to the religion of the migrant. Modi made this quite clear in 2014:
Two types of people have come from Bangladesh – the refugees who have been thrown out in the name of religion and the infiltrators… In any country of the world if there are Indians in whose blood the colour of India runs, if they are ethnic Indians, whatever be the colour of their passport. Should not they come to India and be greeted with open arms? …Those who are thrown out of Bangladesh, those who observe Durgastami and speak Bengali, they are all our Mother India’s children. (Indian Express, 2014).
61There is an explicit conflation of Indian and Hindu here, to the extent that only those who observe Durgastami (Bengal’s largest Hindu festival) are those in whose blood the colour of India (presumably saffron) runs. Muslims here are absurdly denied a blood relation, or even an ethnic relation, to India and Indians. The difference is black and white, no indeterminacy is permitted. The Muslim Bangladeshi is the latest antagonist in the ‘history’ of Muslim invasion. They are infiltrators to the extent that they present a demographic challenge to the Indian-as-Hindu nation.
62Indeterminacy is key here however; Hindutva is all about creating determinacy. The creation and curation of Hinduism as a monolithic entity with an ontic reality (a presence), and the creation and curation of Muslims as outsiders, is fundamentally a means of producing determinacy and overcoding the inherent indeterminacy that undermines such a view. In fact, Hinduism is not cohered by sanatana dharma, and even the key mythologies of Hinduism (e.g. the Ramayana) are understood in very different ways across India. In many places, Hindu practices have a distinct Islamic/Sufi influence, just as many strands of Islam in India incorporate practices more familiar to Hinduism. This curation of Hinduism and Hindu was seen very clearly in the previous section, in the way dalit and low-caste communities were gradually and often forcibly interpellated as ‘Hindus’, and as the opposite of Muslims.
63The point in this tangent is that Hindutva itself is deeply involved in the production and maintenance of determinacy: of overcoding these complex and disparate dynamics to produce determinacy based on a division of Hindu and Muslim. Coding poor east Indian Muslims as Bangladeshis therefore fits into their quest for determinacy. Ultimately, however, there is still a great anxiety over this. The Indian nation is not a Hindu nation; much to the BJP’s chagrin, the NRC is the latest proof of this. The indeterminacy of the Bangladeshi is disturbing because it reveals the indeterminacy of the Indian, particularly of the Indian Muslim. Hindutva constructs the Muslim as inherently non-Indian, as alien. The presence of Indian Muslims and the fact that many chose to remain in India (thus rejecting the two-nation theory) is uncomfortable for Hindutvavadi politics. This of course feeds into the ambiguity surrounding the concepts of citizenship, nationhood and belonging, and the fact that in Hindutva terms, while Muslims may be outside to the nation, they are nevertheless inside of it as far as citizenship is concerned. The disjuncture between the imaginary of nation and the institution of nationality is once again a form of indeterminacy. The idea of Bangladeshi infiltrators and of a porous border works to restore Hindutva ontological security by enabling the construction of poor Muslims as foreign.
64However, secularism in India also often has an implicit closeness to Hinduism (Udayakumar, 2005). Even the purest secularist, Nehru, could not ultimately escape from a reliance on defining Indianness in terms clearly couched in ‘Hindu culture’ (Varshney, 1993). Muslims still form the minority that needs to be accorded protection, or modified as ‘Indian Muslims’ (you never hear the term Indian Hindus because it is like saying Indian Indians) (Pandey, 1999). As Asad (1993: 257) notes, “to speak of cultural majorities and minorities is… to make the implicit claim that members of some cultures truly belong to a particular politically defined place, but those of others (minority cultures) do not”. Muslims have never been un-tenuously part of the Indian ‘we’ and have always needed to prove their Indianness; that is to say that they have always been marked by indeterminacy.
65Moreover, the term infiltration also captures the anxiety that surrounds (perceived) Bangladeshi immigration, and the indeterminacy of those ‘suspected’. Its dual semantics of harm and secrecy set up the image of the Bangladeshi as a threat, in part because of their activity, but also in part because of their ability to hide in plain sight. Consider the following:
To keep a check on the rising crime graph in the city, police have decided to set up vigil and start the verification drive of Bangladeshi migrants. The Local Intelligence Unit (LIU) has prepared a list of 19 ‘sensitive’ areas that happen to be a safe haven for Bangladeshis… Officials in LIU told TOI that Bangladeshis usually infiltrate Assam to get their ID cards, passports and voter cards made. From there, they migrate to semi-urban areas in the country. They live in slum areas near residential colonies and work as rag-pickers. They usually strike during the night, and after committing the crime, flee to other states… Bangladeshis hide in the guise of Assamese. (TOI, 22 April 2017)
66This article clearly draws on numerous stereotypes common to representations of Bangladeshis in the corpus. Here, Bangladeshis are not only presented as a ‘threat’, but also held responsible for an increase in crime in the city (Lucknow), and for poverty. In this regard the ‘suspected Bangladeshi’ serves an ontological function by allowing the externalisation of crime and poverty as foreign problems. However, just as interesting is the claim that “Bangladeshis hide in the guise of the Assamese” (which concurs with the statements on language above). Together these claims signal the ambiguity that pervades the border/order and as such induces a sense of cartographic anxiety about the ontology (being-ness) of the Indian nation and self.
67Implicit here is the surface indiscernibility of Bangladeshis and Indians: the tacit acknowledgement of ambiguity. The indiscernibility is only superficial; the Bangladeshi cannot possibly share in the values and culture of the Indian citizen, just as they are excluded from the ‘7,000-year-old Indian civilisation’. Terms like guise and infiltrate reveal the anxiety of indeterminacy, but they also reassert certainty. Thus, while a Bangladeshi may ‘don the guise’ of the Indian, may ‘infiltrate’ and thereby gain the appurtenances of Indianness, such acts fail to change the fundamental difference of Indianness and Bangladeshiness that supposedly lies beneath. The tell-tale signs of the Bangladeshi give away the falsity of this surface likeness; the poverty and lawlessness of the subject cannot help but confirm their Bangladeshiness. Legal standard of proof becomes irrelevant – their association with poverty, lawlessness and their ‘dubious’ accent/appearance is proof beyond reasonable doubt. The common use of the word infiltrate may bespeak an ontological anxiety, but it also serves an ontological purpose. It suggests that those who inhabit the liminal, or who come to India and live as Indians, cannot be understood as Indian.
4.6 Agency
68The tropes and representations I have analysed so far have all suggested, in one way or another, that the Bangladeshi state is weak. Terrorism, communal violence, human trafficking, economic migration, poverty and illegality all bespeak a state unable to fulfil its proper role as the guarantor of development and progress (Krishna, 1999). The Bangladeshi state is therefore non-agential. I argue that this perception is also evident in the way Bangladesh is narrated. Proportionally, India is succeeded by a verb 27% more often that Bangladesh, with the latter having almost half the range of verbs than India. Moreover, the use of verbs that are most associated with agency (as opposed to passivity) – do, will, can, go, make – are even less proportionate, with these being used 40% as frequently for Bangladesh as for India. Indeed, Bangladesh very rarely collocates with verbs, and when it does, it is often in coordination with other countries, notably India. This suggests that Bangladesh is not often presented as a country that acts or possesses agency, and that when it does act, it does not do so unilaterally. This is in sharp contrast to India which very often collocates with verbs. Additionally, India is more than twice as likely to be the subject (67%) compared to Bangladesh (26%), which instead is strongly primed to be used as an object.14 Indeed, even when Bangladesh is the subject, and thus supposedly the actor, it is sometimes represented more as a receiver/reactor than a protagonist. For example:
Bangladesh grapples with a huge influx of refugees;
New Delhi 's support would help Bangladesh resolve all of their issues.
69Alternatively, it is modified or made conditional. For example:
when Bangladesh buys its first nuclear reactor;
if Bangladesh takes effective actions against the militants;
Though Bangladesh plans to launch its geostationary communication satellite.
70The repeated denial of a subject position to Bangladesh works to construct a sense that Bangladesh is non-agential and passive, in contrast to the agential and active India (see Partington’s [2015] analysis of the ‘The Arab World’). Hoey (2005: 13) argues that words are primed not just to collocate with other words (as in ‘suspected Bangladeshi’) but also “to occur in (or avoid) certain grammatical positions, and to occur in (or avoid) certain grammatical functions”. From the analysis above it is clear that Bangladesh is primed to avoid both grammatical positions which confer subjecthood or agency and to occur in positions that objectify and disavow the potential for agency. Thus, even at the grammatical level, we can see the instantiation of a line between the agential India and the passive Bangladesh. This line, as I have argued, is crucial to maintaining the notion of the sovereign, present self.
4.7 The Standard of Proof – A Summary
71To conclude this chapter, it is apposite to return to what I suggested would provide evidence of ontological insecurity, and to assess to what extent my analysis of the discourse has provided it. In chapter two I discussed how to differentiate the anxiety of ontological insecurity from moral panic used for instrumental political ends. I suggested that, while the two are not mutually exclusive, signs of ontological insecurity would be reinforced if I could show that the discourses were not temporally bound to elections or government terms. I have shown throughout this chapter that the anxiety around Bangladeshis and their indeterminacy has continued largely unchanged across the study period. I also suggested that such concerns should not be found solely or mainly in quotes of politicians. As my concordance and other examples have shown, this is not the case. Thirdly, I suggested that my ability to show the ontological nature of anxiety would depend on the ability to link it to indeterminacy. At the end of chapter three I discussed what would constitute evidence, at a discursive and linguistic level, of indeterminacy, and of anxiety over this.
72Indicators of indeterminacy were identified as issues of identification and misidentification (seen in the common reference to identity documents or lack thereof), the capacity for mistakes in the identification of nationality in the NRC, and the prevalent use of modifiers to cast doubt on the nationality of people. All three of these indicators were evidenced throughout. There was a strong prevalence and a discourse prosody for identity documents that I showed was linked to issues of identification and misidentification. I noted the frequency of articles concerned over the NRC and the difficulty it was having in identifying citizens and foreigners, and I found that Bangladeshi was closely primed with suspected.
73Indicators that this indeterminacy caused anxiety included the sheer prevalence of discourse concerning the challenge of identification, which would bespeak an obsession with indeterminacy. This prevalence has been clearly shown. I also suggested that a semantic preference for shock or alarm upon finding that Bangladeshis had been passing off as Indians would reinforce my argument for ontological insecurity over indeterminacy. Evidence to this effect was found and shown in numerous concordance examples.
74Lastly, I suggested that evidence of ontological insecurity caused by indeterminacy might be seen in the way India and Bangladesh are overcoded. The second half of this chapter has dealt primarily with this overcoding. This was also evident in the way nationality was often asserted as if proven. I particularly suggested that overcoding could be said to occur if there was minimal semantic variation and discursive competition, and that this could be said to create a contrast space for Indian ipseity if Bangladesh was portrayed in a consistently negative light, particularly in comparison to India. This has very much been the case, with Bangladesh figuring without variation, as poor, non-agential, communal and anarchic.
Notes de bas de page
1 By looking at concordance lines it is possible to understand in what ways particular words are used. Just because war collocates with Bangladesh does not necessarily mean that we can infer the context in which the two words are used. Far from suggesting a tense relationship, concordance analysis shows that war actually refers to the Bangladesh Liberation War, in which India played a key role. Looking at concordance lines (i.e. individual collocations in context) thus constitutes a plausibility probe. All subsequent collocation analysis follows analysis of concordance lines.
2 Information on how many of those arrested are then prosecuted, and how many of those are found definitively to be Bangladeshi, is not readily available. The deportation rate (~580/year) could be an indicative measure, given that Bangladesh requires the highest standard of proof (Hindustan Times, 2017). Bangladeshis constitute 75% of deportations. To put this into some context, the EU, which has one third of the Indian population, deported 158,000 people in 2018 (European Commission, 2019). The point therefore is that the deportation rate is very low, suggesting that adequate proof of foreignness is rare.
3 Biometric cards.
4 Legally, none of these are proof of citizenship, but they are often used nonetheless as a symbol of citizenship.
5 This is extrapolated from census data but is very difficult to calculate because the census distinguishes religion and language. Most of the Hindu Bengalis speak Bengali in Assam, but many of the Muslim Bengalis speak Assamese. In addition, some of the Muslims in Assam are not Bengali, while the Barak Valley has always been Bengali-speaking.
6 Housing societies in India can be highly political, with many people de facto refused residency if they are Muslim, lower-caste or otherwise deemed ‘unrespectable’. See e.g. Banerjee et al. (2015).
7 The cartoon cannot be shown here due to Swiss copyright law.
8 Desi means indigenous, or in this context, Indian.
9 Indeed, “As ever, the role of the popular press was significant in generating and punctuating narratives and time-lines. The ‘Great Calcutta Killing’, almost exactly a year before the Partition, is remembered for its violence and political significance, to be sure, but certainly also for this very memorable moniker and mnemonic, which was generated years after the event(s) by a Calcutta newspaper” (Ghosh, 2015: 270).
10 A Hindu chauvinist organisation.
11 A dalit/avarna community.
12 My own work with Hindu and Buddhist minorities who have left Bangladesh suggests that numerous reasons motivated the choice, only one of which was insecurity.
13 Indeed, Hindutva often presents ‘Hindu culture’ as a truer manifestation of secularism in its claimed focus on humanitarianism, peace and ‘assimilation’. As Modi tweeted, after the passing of the CAA: "This bill is in line with India's centuries old ethos of assimilation and belief in humanitarian values" (HT, 2019). Hindutva nationalists also commonly refer to others as ‘pseudo-secularists’.
14 This was calculated by randomly down-sampling to 100 instances of both terms.

Le texte seul est utilisable sous licence Creative Commons - Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International - CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.
The Development of International Refugee Protection through the Practice of the UN Security Council
Christiane Ahlborn
2010
The SWIFT Affair
Swiss Banking Secrecy and the Fight against Terrorist Financing
Johannes Köppel
2011
The Evolving Patterns of Lebanese Politics in Post-Syria Lebanon
The Perceptions of Hizballah among Members of the Free Patriotic Movement
Fouad Ilias
2010
La justice internationale à l'épreuve du terrorisme
Défis, enjeux et perspectives concernant la Commission d'enquête internationale indépendante (UNIIIC) et le Tribunal spécial pour le Liban
Sébastien Moretti
2009
Aut Dedere, aut Judicare: The Extradite or Prosecute Clause in International Law
Claire Mitchell
2009