The first case of the COVID-19 pandemic in Mumbai, India was reported on 11th March, 2020. Thirteen days later, a nationwide lockdown was announced – bringing India to a grinding halt. 

Except that is not what actually happened. 

Those who could afford it shielded themselves within their homes, rations packed to the rafters and N-95 masks stockpiled. For the over 600 million internal migrants in India–those whose homes are in villages but who work in informal labor markets in the city–the lockdown announcement triggered a mass exodus. Droves of people fled the cities they worked in to return to their rural communities, largely on foot. With their wages coming to an abrupt standstill, they left deeply fearful of what lay ahead.

Much has been written about the lack of statistics regarding this exodus. Many lives were lost to hunger, fatigue, heatstroke and, of course, disease. Yet “there are no numbers,” Santosh Kumar Gangwar, then Indian Minister for Labour and Employment, stated the same year when asked to enumerate the tragedy’s scale at a national level. Migrant workers have already long been considered “fringe figures” within the Indian urban social network. With the rupture caused by the pandemic, their existences have only been further invisibilized. 

The initial guidance provided by India’s central government was to ensure that migrants did not leave the cities. However, given the sheer volume of panicked people desperate to rush back home, this guidance was impossible to actually implement. When the stay-where-you-are orders failed, the center tried creating quarantine camps at state borders.This, too, did not prove successful. Attempts to build a database of the departing migrants were also abandoned halfway. 

The pandemic was already seen as an arithmetic problem: a problem of numbers where a solution could purportedly be reached by just pinning down the right formula. This notion was only compounded upon by the use of terms such as “rate of infection” and “doubling time” in the media, which made the actual lack of data and data collection efforts regarding migrant workers result in a particular kind of disenfranchisement. 

Despite the magnitude of the exodus, India’s national mood was to dismiss the migrants’ long march as simply an aberration. Since the event was caused by the deep distrust that migrants displayed in the state’s ability to provide them with safety nets, any acknowledgment of the tragedy’s nuances would misalign with the government’s narrative of complete control over the crisis.   

A Vocabulary of Ghostliness 

In retrospect, the lack of numbers eventually became an object of interrogation. 

A particular trope came into play within the media discourse surrounding the migrant exodus: a vocabulary of ghostliness. Words used to describe the state of the migrants essentialized their identities to solely their forced absence from the labor market. News reports in publications like the BBC and God Save the Points, spoke of “ghost workers” and “ghost towns.” In a Telegraph India essay written shortly after the first lockdown, academic Manas Ray describes the migrant workers trekking to their native villages as “ghost mutineers stalking the country in search of a home.”

“These lives are, of course, not entitled to the city's culture and taste, to its intellect and leisure; these are gross lives,” Ray writes further. The word “gross,” a mathematical term for excess, is specifically used here to capture the unnumbered migrants’ lives. 

“What seems like a relatively stable social order is constantly being modified, added, subtracted, maintained, and cleaned by the invisible labor force mostly made of migrants,” Ray continues.

While terming the migrants as ghosts evokes a certain poignancy, it also dehumanizes and homogenizes a diverse, marginalized group of people. Although the tragic scale of the exodus could not accurately be enumerated at the time, it is now possible to retrospectively analyze Indian media archives and give an approximate number to the verbiage that was in play.

As an intervention into this archive of absence, I formulated a dataset containing newspaper (e-paper) stories that appeared when I ran a Google Search with the following phrases as keywords: 

1. Migrant Haunting Mumbai 

2. Migrant Ghost Mumbai

3. Covid Haunting Mumbai 

4. Covid Ghost Mumbai

I delimited the database both spatially and temporally. The city of Mumbai became a stand-in for the urban, chosen for being the country’s financial capital. Temporally, I limited the selected articles to those published between 15th March, 2020 and 10th August, 2021. I downloaded the text from these news articles from relevant pages of search results as raw TXT data and eliminated the duplicate results, making sure that each webpage was represented only once in the TXT data file. This data was subsequently input into a Word document where, using the “Find” feature, I located the words “haunt” and “ghost,” highlighting the sections they appeared in. I further transferred these sections to columns to see the frequency of the words and the contexts they were phrased within. Finally, I color-coded repeated phrases, numbering each occurrence. 

My goal through this exercise was to locate patterns within this particular media discourse which evoked a metaphoric vocabulary of ghostliness.

The data I analyzed for these patterns encompassed roughly 106,000 words in total, including headlines, by-lines, articles, conjunctions, and prepositions over the four keyword searches. It is important for me to say that by no means did I conduct a perfect academic study which incorporated all the work that has been produced relating to the migrant exodus. The formulation of the data set was restricted by resources, paywalls, and availability of time so it is meant to be indicative rather than declarative. Therefore, this is not a quantitative analysis, but a qualitative exploration of the use of a specific vocabulary and its implications for understanding a certain media archive. 

Why is it necessary to think about the vocabulary used to describe this, or any, tragedy? 

First, without numbers, we have no other way to understand the scale of the lives lost and destroyed. Secondly, understanding language allows us to understand who is permitted to be forgotten or remembered, and who media discourse renders invisible. The absence of numbers of lives can then be understood by investigating who is made a ghost–who is seen to haunt rather than live as a full human being–and how.

When we cannot account, we must articulate. 

There is a long tradition in the social sciences of using the vocabulary of ghostliness and hauntings to explain societal relations. In a 1919 essay titled The Uncanny, Sigmund Freud describes how any change in the way society functions bring with it a sense of deep unsettlement. Karl Marx takes this even further at the beginning of the Communist Manifesto, where he terms communism itself as a specter haunting Europe, invoking ghosts to signify societal churn. 

More recent scholarship in anthropology has built on tradition, hypothesizing how societies often tell ghost stories as a way of integrating uncomfortable memories into the cultural fabric. 

In scenarios with no actual historical record or archive, hauntings and ghosts become a means to combat “institutional forgetfulness.” With the COVID-19 pandemic and migrant crisis in India, we can see deliberate institutional forgetfulness in action. Here, the vocabulary of ghostliness becomes a tool to grasp public sentiment. 

Even three years removed from the worst of the pandemic, which disproportionately ravaged the Global South, understanding its impact on human lives is to grapple with ambiguity–intellectual, pragmatic, and experiential. It is to be faced with something that is not quite historical, not quite normal, and not quite visible. It is to engage with a ghost. 

In the depths of the night, a lonely soul weeps,

Tangled in shadows, where despair seeps.

A heart, heavy with the weight of solitude's sting,

A melody of sorrow, a dirge I sing.

(Inspired by the song Gloomy Sunday composed by Hungarian pianist and composer Rezső Seress and published in 1933.)

Accounting, Articulating

Within my data set, the word “haunt” in various conjugations (haunted, haunting, et cetera) occurred 29 times. The term was used most often to describe images of the migrant exodus and how the city folk were haunted by the visuals of it. To ascribe a numeric value: out of the 29 references, 11 referred either to “haunting images” or “haunting visuals.”

As anthropologists Benjamin Smith and Richard Vokes write in their 2008 article “Haunting Images,” the photograph and the ghost “are never far apart.” The two can be interchangeable in their function, “standing in for relationships that cannot or can no longer be performed directly,” and share the similarity of embodying present absences. They further activate an “emotive force through their representation of absent objects, kin and places.”

Images from the pandemic are rife with this emotive force as they represent moments of death and tangible devastation, evoking significant grief, and by extension of the vocabulary of haunting, horror. Through images, citizens of the city are forced to reckon with the structural collapse of urban labor networks.

In my study, a second pattern emerged: the use of the word “haunting” to describe memory and recollection. There were four references to being “haunted by memories.” Comparing it to the previous pattern, where photographs produced ghosts, memory here is where the lost “normal life, or the remembrance of normality,” resides. During the pandemic, the phrase “new normal” was commonplace. In such an unprecedented time, recent memories felt historical, and indeed haunting given the sense of loss they invoked.

The word “ghost” itself appeared in my study 28 times. 

21 of these occurrences concerned a place, with 11 referring to “ghost towns,” nine to “ghost villages,” and one to the ghostly nature of abandoned roads. In media discourse during COVID19, the term ghost town was clearly used to describe the emptied urban centers, while ghost villages referred to the rural settings where the population had previously been sparse due to internal migration. During the pandemic, these became the sites of return for the working class who were seeking safety and familiarity.

In five instances across the data set, “ghost” was an epithet transferred to the laborers themselves leaving the cityscape. Coupled with migrants already being othered and alienated, this deployment of the language of haunting only served to further exacerbate their marginalization and cement their erasure. 

A 2022 report from the World Health Organisation suggested that India’s real COVID toll may never be known. According to the report, more than 4.7 million people – a nearly ten times higher statistic than estimates by Indian officials – might have died from COVID-19 infection between 1st January, 2020 and 31st December, 2021. It is not a stretch to postulate that the missing numbers from India’s state statistics might be deaths that occurred in villages or at the homes of those who could not afford medical treatment. Data paucity within India is not a new phenomenon, and it is well-documented that the ones left out are often from marginalized communities. 

A poem written by Indian filmmaker Kireet Khurana during the lockdown turns attention to the migrant crisis with the following stanza: 

“Hum to pravasi hain, kya is desh ke vaasi hain? 

Agar nahi hain insaan to maar do abhi,  

de do farmaan” 

(We are migrants, are we (not) residents of this country? 

If we are not human, kill us now, 

Give the command) 

The stanza juxtaposes “pravasi” (migrant/traveler) with “desh ke vasi” (residents of the country). The value of this wordplay comes from the etymology of the terms and their meanings. The root word for both pravasi and vasi is the same–“vas” meaning abode. Therefore, a vasi is one who is of the abode, so its negative suffix pra(vasi) implies one who is separate or othered from their place of abode.

However, the term desh ke vasi (residents of the country) often signifies being a citizen. Citizenship and residency are therefore interchangeable in this context. The poem questions the disenfranchisement of migrants by declaring “if we are not human, kill us now,” criticizing the political leadership's unwillingness to provide migrant laborers with humane means of returning to their native communities.

In his celebrated essay collection Politics of the Governed, historian Partha Chatterjee categorizes individuals afflicted by infrastructural disenfranchisement as occupying a fringe space. In this fringe or margin, they reside within the city but cannot rely on it for social 

safeguards. Thus, they are rendered beyond the comfort of being a vasi. 

This only became more explicit in India through the COVID-19 pandemic. Despite numerous assurances by the government that migrant workers would be safe within the cities, a precedent of haplessness and lost livelihoods led to large masses attempting to leave cities. For most migrant workers, the uncertainty of a treacherous journey back home was preferable to relying on the state for sustenance. The distrust created by constant erasure simply could not be erased by politicians’ promises and press broadcasts. 

Specters and those who witness 

David Torri, an anthropologist of shamanism, describes the ghost as first and foremost a story: it “needs listeners more than it needs witnesses.” 

As researchers charged institutionally with the creation of knowledge, the onus is upon us to bear witness to the lacunae within archives and acknowledge our failures in listening to those who fall through the chasms of documentation. 

India’s COVID-19 migrant exodus was a humanitarian crisis born out of rightful mistrust held by laborer populations towards urban administration. The ghosts resulting from this exodus, and further exacerbated through media discourse, are not new, but have always existed – the pandemic simply made visible the cracks within India’s neoliberal urban apparatus.  

Indian cities have continued to grapple with their failure to integrate migrant laborers into their social and cultural fabric in the three years since the pandemic. Despite the significant cost to human life, there has been no socio-political change aimed at remedying the gap between those seen as citizens of the city, and those essentialized as mere bodies for labor.

“I felt betrayed twice: by society, because no one around me lent a hand – my landlord kicked me out – and by the state,” a construction worker from Kanpur, Ram Yadav, said in a 2022 documentary made by The Guardian. At the time of the lockdown, he vowed never to return to the city he’d left. A few months later, however, he had no choice but to head back to Delhi. 

By November 2020, large sections of migrant workers, much like Yadav, had returned to the cities they had left. There was no newfound love for the urban–just desperation in the face of limited job opportunities within rural communities. The disenfranchisement they continue to face is deeply institutionalized. Within most archives their experiences are secondary. 

The fact that there are no numbers is potent; the state does not account for the working class body, neither in life nor death. In life, they have no stability or voice in the functioning of the very urban centers that rely on their migrant labor; in death, they are merely erased.  

This erasure reaffirms migrant workers as Chatterjee’s term of fringe figures, or outsiders to the city’s social and cultural fabric. Devoid of agency, the migrant becomes the object of urban anxieties, rather than a subject experiencing them. The city is thus simultaneously run by migrants yet haunted by their absence, with the urban populace haunted in particular, albeit at a comfortable distance, by migrants’ trauma. In other words, the laborer is subject to the whims of the megacity and those who administer it. They become the “other,” pitied by middle-class citizenry, yet still not seen by them as human or equal. As Jacques Derrida puts it in his book Specters of Marx (1994), disjunctures in society, like pandemics, make apparent the anxieties of a place, and the “ghosts” that emerge here are testimonies to alienated labor. 

By reconciling these specters through scholarship, at the least, we can move forward towards marking the absences within existing records. It is an attempt to integrate significant institutional failure into cultural memory.

The production of knowledge is never perfect, but the use of alternative vocabularies as interventions allows us to pinpoint deliberate erasures. Fully understanding the effect of a crisis, of course, does not encompass just metrics, even if imprecise, for its impact. Yet, it is an honest start.∎