“You Belong to Gutters,Not Facebook or Twitter”: Recovering Dalit Histories From the Shadows of Social Media

Concealment of Context

The first narrative pertains to a common feeling among the Dalits we interviewed and observed as being “unseen”; a sense of insignificance brought about by being overlooked by social media platform architectures. Dalits felt that they were unable to construct historical self-knowledge due to an inherent lack of caste-based context among the algorithms that curate and control social media content. Dalits regularly considered how platforms privileged dominant castes, sharing examples whereby hashtags promoting Dalit culture and Dalit history (e.g., #DalitHistory #DalitHistoryMonth) were either removed or failed to appear in trending topics on social media platforms. In contrast, hashtags associated with dominant castes, such as the celebration of popular Hindu festivals like Diwali (#diwali) and lord Ganesh festival (#ganpatibappa), were visible and highly engaged with. Dalits felt that many of these festivals had historical connotations that were inherently “discriminatory against lower castes,” (Suman), but this historical context was largely overlooked. While Dalit history was acknowledged through posts that commemorated important figures in the community (i.e., B.R. Ambedkar on his birth anniversary), such recognition was transient, with many feelings that “discrimination was practiced for the rest of the year” (Pawan), meaning that attempts to celebrate Dalit history remained “unseen.” This is the essence of “context concealment.”

Hashtags used by Dalits, such as #AmbedkarJayanti (#AmbedkarBirthday), were regularly hijacked, being taken over by trolls and those with “Hindutva” ideology; deep-rooted ideologies that promote distorted accounts of Ambedkar to offend Dalits. Our participant, Suman, felt crowded out in social media: “[dominant caste members] quickly take over any discussion and shift the narrative, and then you are again invisible.” Caste discrimination occurred frequently, but subtly, whereby trolls employed indirect or milder terms to avoid detection by platforms, pacifying caste-based discrimination as “cultural differences.” This led Ashok to comment, “Caste is untouchable . . . it is too harsh or blunt, but they do not like to call it caste . . . and will never accept that caste-based discrimination is the real evil.” Caste, and individuals experiencing caste-based marginalization, were “unseen.”

Dalits were aware that the complex nature of the caste system, characterized by numerous subcastes, hierarchies, and regional/community-specific practices and norms, posed challenges for algorithms to identify caste-based abuse. In many ways, the concealment of their context was somewhat expected. Yet many recounted examples where Dalit views had been “mislabeled” or “mischaracterized” by social media platforms. Naveen recounted how the reporting of the disruption of a Dalit’s marriage procession was incorrectly tagged as “fashion” on Twitter (Figure 1). Similarly, another post on Facebook (Figure 2), designed to educate dominant castes on language that may be offensive to lower castes, was flagged as “inappropriate” by social media algorithms, not only limiting its reach but conveying a painful message to Dalits that resistance was futile, making them invisible.

OPEN IN VIEWER

Figure 1.

Atrocities on Dalits Mislabeled by Social Media Platform.

OPEN IN VIEWER

Figure 2.

Dalit Voice Flagged as “Inappropriate” by Social Media Algorithms.

Trolls regularly evaded algorithmic detection, further intensifying the exclusionary potential of social media spaces for Dalits. Abuse failed to be identified and acted upon by social media platforms through utilizing “code” words, such as Dolit or D@lit instead of Dalit, abbreviations such as “R people” to convey Dalit who avail reservation (affirmative actions) or BRA (an undergarment) to shorten Bhim Rao Ambedkar’s name, or disguised language such as “Hinglish” (a hybrid language of Hindi and English commonly used by Hindi-speaking individuals in northern India). “Hinglish” is evolving rapidly, numerous Dalits recounted, as more and more Hindi native speakers learn and practice English. They felt that such language use, facilitated through social media, showed dominance over those who do not speak English and was a strategic way to mask marginalization. For instance, “Bhimte”—a slur word used against followers of Bhim Rao Ambedkar—that has risen in popularity in recent years and is often associated with a call to violence against Dalits. Calling someone a “Bhimte” belittles Ambedkar’s legacy and mocks the pride Dalits take in Ambedkar’s contributions for human rights, reducing a movement for equality to a slur. ⁸ Figure 3 provides a post shared with us using the phrase “Bhimte,” designed to deride Dalits of their educational abilities and to insinuate that they are incapable of success without reservation policies. Such posts were neither labeled as being problematic nor taken down by social media platforms, further fueling feelings of being “unseen.”

OPEN IN VIEWER

Figure 3.

“Bhimte” Used as Abuse on Social Media.

Intersectional dimensions further complicated the ability of algorithms to detect abuse. Dalit women, such as Radhika and Maya, shared how they faced discrimination online based on both caste and gender, as well as discrimination within the Dalit community itself, as Dalit women lie at the bottom of the caste hierarchy and patriarchy. Many Dalit women confided in us their experiences of sexualized caste-based abuse, such as Figure 4; a post employing the term “Bhangi”—Balmiki, derogatorily known as bhangi, is a Dalit caste group that traditionally was restricted to manual scavenging jobs—a pejorative term used to demean lower castes and pertaining to harmful stereotypes that Dalits do not pay taxes. The post also refers to Dalit women’s bodies, invoking the painful history of the “breast tax” imposed on Dalit women until the early 1900s if they wished to cover the upper part of their bodies. The post also includes messages from the sender, such as “yeh dekh bhangi” (see this, manual scavenger) and “hijde” (transgender). These phrases are deeply embedded in India’s caste norms and heteronormative social orders and are attempts to reduce Dalits’ identity to a stigmatized caste-ascribed role. The term “hijde,” when weaponized in caste-based trolling, reflects intersectional oppression where transphobia becomes a tool to police lower caste self-expression, where gender nonconformity becomes a site of humiliation. In allowing such posts to prevail, algorithms invalidated the pain of Dalit users, further exacerbating the feeling of being “unseen.”

OPEN IN VIEWER

Figure 4.

“Bhangi” Used as a Slur on Social Media.

Context Recovery

To counter these feelings, Dalits acknowledged the limitations with platform design and began circumventing social media entirely to cultivate their own digital archives. “Context recovery” involved bypassing mainstream platforms’ exclusionary algorithms through initiatives such as RTI (roundtableindia.co.in; founded in 2011), Velivada (velivada.com; founded in 2017), Equality Labs (equalitylabs.org; founded in 2015), BAWS (baws.in; founded in 2022), The Mooknayak (themooknayak.com; founded in 2022), and The Ambedkarian Chronicle (theambedkarianchronicle.in; founded in 2024). Here, Dalits felt “free” to document scholarly critique, oral histories, protests, and cultural practices, preserving Dalit experiences outside of hostile digital environments. As Manoj noted, “on these sites, our words do not get deleted or hidden, and [we] can write without being judged.” These caste-aware content repositories housed rich Dalit narratives—especially those that documented trauma or challenged dominant historiography—ensuring long-term access to Dalit historical and current struggles. These platforms have also been influential, with thousands of followers on social media, and have been able to impact policy decisions, such as in 2023, Equality Labs’ report on “Caste in the United States” was used by the California State Senate to pass a bill banning Caste-based discrimination (Outlook India, 2024), and Velivada and RTI websites provided space to hundreds of Dalits to contribute by writing articles. Anand compared these websites to libraries, “these are our digital libraries, built by us, for us.” While limited funding and technical expertise restricted scalability, Dalits remained committed to nurturing tangible resources that recovered—and protected—their historical context.

In sum, the narrative of unseen captures the pervasive sense among Dalit users of being rendered invisible by social media platforms. This invisibility was not accidental but rooted in the structural design of platforms that actively concealed historical context. Dalit voices were deprioritized in content visibility, often bypassed by algorithms, and neglected in moderation processes, reflecting how platform structures reinforce dominant-caste norms. Context recovery involved Dalits circumventing mainstream platforms to construct their own digital archives.

Concealment of Context

The second narrative demonstrates how Dalits felt “unheard” in social media spaces; a feeling of isolation where their concerns were disregarded and dismissed. Dalits felt that they were unable to rectify errors in the construction of their history online, or seek remedy, due to lack of contextual insight within the social media platform policies and processes. This further led to a concealing of the Dalit context. Raising concerns about inappropriate content with platforms was felt to be unnecessarily complex, and many spoke of feeling “exhausted” (Meena) and “defeated” (Ashok) by the constant battle they faced with social media platforms, and particularly their content moderation policies. As Kalpana expressed, “I utilize the block and report options on social media, but I am unable to find an option to report posts related to caste abuse or caste harassment.” Even if complaints did reach content moderators, most Dalits felt that a lack of cultural understanding of caste-based abuse prevented any form of recourse. Many Dalits recounted unsatisfactory responses they had received from content moderators, claiming that reported posts or profiles “did not violate platform policies” despite Dalits explaining why the content caused offense. As Pawan shared,

I reported that fake profile to Twitter [X] a couple of times, highlighting the tweets [messages on X] that were simply copy-paste of my tweets with altered meanings and opposite to the message that I was conveying . . . [but] Twitter replied that such messages and profiles do not violate their policies. I could not do anything much at that time.

For Dalits, policies to protect against hate speech were “empty words; these platforms’ promises to protect us do not match the reality” (Rajeev). Praveen felt like, “it is like they are just pretending to care,” and for Radhika, “their consistent silence [on caste abuse on social media] feels like an endorsement of casteism.” Those who had studied social media anti-discrimination policies felt that they downplayed the significance of caste discrimination due to an inherent lack of context, with major social media platforms in India having no mention of caste in their anti-discrimination policies at all.

Undeterred, Dalits continue to be visible online, and in doing so actively recover their historical context. However, they found that content moderation processes were swift to remove their attempts to engage in caste-based resistance. Many users recalled where their accounts had been disabled or blocked. In one instance, Praveen was banned on Facebook after dominant castes objected en masse to his posts. His post that attracted backlash was about highlighting discriminatory quotes from religious scriptures, leading Praveen to lose access to his Facebook account. Some even felt that the rise in atrocities against Dalits was directly linked to their activism online. When Dalits attempted to reappropriate terms used against them, such as “Chamar”—a stigmatized caste group among Dalits in northern India who traditionally were involved in skinning dead animals and tanning leather—they were met with inconsistency. As Meena recounted, “sometimes those posts were kept hidden or marked as sensitive and sometimes not,” and Savita reminded us, “context matters.” Savita’s remark that “context matters” referred to the inconsistent ways platforms read and classify caste-related content, where identical phrases could be treated differently depending on accompanying text, images, or user histories.

In the absence of support from platforms, Dalits took matters into their own hands and attempted to educate other social media users on the intricacies of Dalit history and to provide more context as to why certain content was offensive. Pawan recounted how they “tried to give [trolls] the reasons, argue and try to have a sensible disagreement that why calling someone ‘bhangi’ (manual scavenger) is not cool.” This left them feeling “tired” and “demoralized”; “I do not think it makes much difference for them . . . they continue to behave in the same fashion” (Sunil). Many Dalits commented on the fear of physical abuse and violence prompted such interactions demonstrating the risk associated with context recovery: “Once you have been attacked by the army of dedicated trolls, they will make sure your life is miserable” (Radhika). Here, Dalits recalled feeling “scared” and “stressed” for their and their family’s safety at the hands of “abuse” and “death threats.” What made these instances even more distressing was that sometimes the perpetrators were high-profile individuals or groups who had a well-established social media presence, a substantial number of followers, and even verified accounts. For example, Jyoti recounted a time when Hindutva actors publicly encouraged followers to demand the lynching of Dalits and Muslims. In 2020, a video went viral in which Hindutva followers were seen inciting violence not only against Muslims but also against Dalits during the Delhi riots (Lalwani, 2020). Such abuse allowed high-profile and influential social figures to “shape the narrative” of Dalit history online, resulting in Dalit counter-narratives being “unheard” in social media platforms.

Platform policies intersect with the legal protections for Dalits, such as the “Scheduled Castes and the Scheduled Tribes (Prevention of Atrocities) Act 1989” (also known as the SC/ST Act). Yet Dalits rarely registered complaints for their belief that a lack of context on caste-based abuse prevailed, and conviction rates were low. Radhika commented, “our confidence goes down in the abilities of social media and legal systems to handle our issues.” This led many Dalits to believe that the legal system was “eyewash” and “just talk with no action” (Focus Group 2), fueling feelings of being “unheard.” As Jyoti recalled, “I reported to the platform about the abusive posts, but I was blocked for some time . . . I wanted to take legal action . . . but then I realized it is not worth it when . . . courts are biased … they will not take it seriously and not understand.”

Context Recovery

An important part of context recovery involved Dalits forming digital counterpublics to create safe spaces for education, dialogue, and resistance. Platforms like Dalit Human Rights Defenders Network (dhrdnet.org) and International Dalit Solidarity Network (idsn.org) foster community-driven spaces where users share experiences and advocate for caste-sensitive policies, and ultimately, to be heard. As Maya explained, “these networks give us the courage to speak without fear of being silenced.” IDSN, in particular, operates internationally to coordinate advocacy, raise awareness, and influence UN and government bodies on Dalit rights issues, and has given “chance to Dalits to participate in policy development at international level” (Maya). Additionally, Dalit-Bahujan scholars and activists, who possess some form of social capital, utilize digital platforms for network-building and epistemic assertion. For example, the offline and online debates surrounding Arundhati Rao’s introduction to Ambedkar’s Annihilation of Caste illustrate that digital engagement also reflects negotiations, contestations, and, at times, conservative or reactionary turns (Mane, 2015). During the debates, Dalits argued that Roy misrepresented Ambedkar’s views on Adivasis, compared Ambedkar to Gandhi (both ideological rivals), and that the original book, Annihilation of Caste, first published in 1936, did not need an introduction, as it dilutes the book’s importance. Through these, Dalits construct “Ambedkarite public culture,” which is a mental and physical space that cultivates the autonomous Dalit assertive self and emancipatory world view (Guru & Sarukkai, 2012, p. 8284, cited from Gaikwad, 2021). Furthermore, Ambedkarite public culture, long expressed through commemorations, songs, and symbols around Ambedkar, now flourishes on social media, where digital art and livestreamed events globalize its reach. As Prakash described, “sharing Ambedkar’s words online feels like lighting a lamp that others can see, no matter where they are.” Social media platforms amplify Dalit counterpublics, enabling everyday assertion of dignity and rights while transforming local practices into transnational solidarities that challenge caste hegemony.

In sum, Dalits expressed feeling unheard on social media, where their concerns were routinely dismissed. Many felt they lacked access to effective redress mechanisms, particularly when trying to challenge dominant narratives or to recover accurate historical context. This silence was not just interpersonal but structural, shaped by platform policies that lacked contextual sensitivity to caste. It is for this reason that Dalits sought contextual recovery via the formation of counter-publics and safe spaces.

Concealment of Context

Our final narrative of “unspoken” captures how Dalits engaged in self-censorship; their voices, thoughts, and feelings were repressed and not expressed aloud within social media platforms. Dalits often gave into the concealment of their historical context due to confusion and fear. For instance, expressing their identities, using social media in their mother tongue, or sharing personal pictures were activities fraught with danger. One disabled Dalit bravely shared their caste and disability struggles on social media as an act of inclusion. A barrage of hurtful comments ensued, with their disability mocked and the term “langda” (slang for crippled; someone with leg-related problems unable to walk) employed. This term is highly offensive to Dalit communities that face higher rates of disability (Pal, 2010) and are seen as less than human. It also carries a distinct caste-based meaning in India, where bodily disability is historically and symbolically intertwined with caste status and social devaluation, untouchability and impurity, and “ability” is intertwined with high castes.

As the Dalits we spoke to recounted, examples such as there were not merely attacks on bodily difference but codified caste insults, deeply entangled with how Dalit bodies are rendered abject and unworthy in dominant public imaginaries (Pal, 2010). Unlike generic abuse, these insults also draw on cultural memory rooted in religious scriptures, myths, and everyday discrimination, where caste “pollution” designates social inferiority and is rationalized through notions of karmic retribution, which refers to the belief that Dalits’ lower social status or perceived impurity is the result of past actions or sins, framing caste oppression as a moral or cosmic inevitability rather than a socially constructed injustice. When this language enters social media through trolling, it is not merely a form of Dalit hate speech; it is the re-inscription of caste ideology through digital tools. Witnessing such hate led Dalits, “avoiding unnecessary troubles” (Rajeev) by changing their surnames to fictitious upper caste names, concealing their profile pictures, or even creating false personas, contributing to true Dalit identities—and caste-based context—becoming “unspoken.”

Most Dalits were painfully aware of the digital divides they experienced, such as limited (high-speed) Internet access, computer/laptop access and lower digital literacy levels. Such features meant that even if they wanted to “speak” online, they often remained “unspoken” due to very practical challenges. Most Dalits involved in our study lived in rural areas and primarily accessed social media via mobile phones, without access to video and image creation software. We learned from Manoj and Kalpana that Government and NGO-funded digital literacy programs in India, such as the Digital India initiative, often failed to reach rural Dalit communities. This left Dalits reliant on their own personal resources to educate one another and mobilize. Many Dalits were simply unaware of how to handle abusive comments beyond reporting abusers. As Kalpana commented, “I am quick to block and report those who misbehave and create problems . . . what else is there to do?”

It was, for many, not surprising that Dalit communities were often depicted through “photos of visits to Dalit slums” (Rajeev) or the glorification of the broom as a symbol of dedication to work. In social media narratives, worshipping the broom is presented as an extension of the general Indian principle of respecting one’s tools, as they are the means by which a livelihood is earned. However, this narrative conceals the historical caste oppression in which broom work was assigned to Dalits, recasting structural oppression as cultural pride. Most Dalits were impacted by the narrative of “if you are not successful, it implies you are not working hard enough” (Prakash); narratives that suggested that Dalits were not successful due to their lack of effort, as opposed to structural inequality. Such images—virally shared in social media—and narratives were a constant reminder of longstanding, historical caste divisions in wealth and education; images that Dalits were keen to rally against, yet were rarely successful. Dalit perspectives were often labeled as “unverified” when they did speak out, with comments questioning the “divisive” nature of perspectives and the “validity of posts” in ways not applied to dominant-caste users (Suresh). Even when Dalits shared their life experiences, those appeared exaggerated and agenda-driven to others. Dalit views were regularly ridiculed and undermined (Vikas; see also Majumdar, 2022), further concealing their context. One Dalit shared how they had, on a few occasions, attempted to “tag” the United Nations on Twitter with posts highlighting caste-based discrimination, but they received opposition, with comments suggesting that the participant was “an agent of external forces, trying to ruin the reputation of India” (Praveen). Further, Praveen recounted how TikTok became known as “Shudra of the Internet”—signifying its association with lower caste users, who were believed to be low-status and culturally inferior within the online hierarchy—in India, because lower castes were using it more to amplify their voices online.

It is for these reasons that many Dalits lost faith in social media platforms or, at worst, completely abandoned social media to “detox the mind” (Maya), further fueling Dalit “unspokenness.” Dalits even recounted experiences of being arrested for protesting online (The Hindu, 2022); “the chances that you get into trouble remain high if you try to challenge religious scriptures that promote caste . . . that is a no-go area” (Praveen). When Dalits expressed their support for reservation policies, for instance, they were ridiculed. Abusive posts included comments such as, “reservation is the reason that talented engineers and doctors are moving out of India” (Anand), and “you belong to gutters, not Facebook or Twitter” (Anita). For example, in his posts, a cricketer turned politician, Gautam Gambhir, mocked only those who use the caste-based reservation and received hundreds of reshares (Figures 5 and 6). This episode not only reflected a persistent anti-reservation sentiment in India’s upper-caste public discourse but also revealed how digital publics were curated, shaped, and rendered meaningful in alignment with the values of the dominant castes, leaving Dalits unspoken. Indeed, caste-privilege, demonstrated by actors such as Gambhir, became more than just personal opinion; it entered mélange history as a meme-able, shareable, “truth,” shaping the digital historical record in a way that marginalizes the very groups affirmative action is meant to support.

OPEN IN VIEWER

Figure 5.

Mocking Affirmative Action Policies on Social Media.

OPEN IN VIEWER

Figure 6.

Ridiculing the Constitutional Provisions That Support Dalits.

Context Recovery

Finally, to protect and recover their context from the shadows of social media, Dalits engaged in digital activism using hashtags such as #DalitLivesMatter and #SmashBrahminicalPatriarchy and ran campaigns to foster historical self-knowledge. For instance, the #DalitWomenFight campaign empowered Dalit feminists to address caste-gender intersections, building solidarities that challenged mainstream digital discourses. While many Dalit hashtags were canceled, these particular hashtags became successful because of timely responses to caste-based atrocities, the strategic use of prominent figures to amplify marginalized voices, and the creation of compelling visual content that resonated with a broad audience. For instance, #DalitLivesMatter, first used around 2014, gained substantial momentum in 2020 following the #BlackLivesMatter movement and the 2020 Hathras case in India, where a Dalit teenage girl was gang-raped and murdered. Similarly, #SmashBrahminicalPatriarchy gained attention after Twitter CEO Jack Dorsey was photographed, along with Dalit women, holding the slogan during his 2018 visit to India, a moment that strategically leveraged celebrity visibility to amplify the campaign’s reach. Even if these hashtags faced opposition, Dalits felt they enjoyed sharing stories of resistance and identifying new alliances across the Internet. Pawan reflected on it as, “posting with these hashtags made me feel like part of something bigger.” These hashtags enabled Dalits to work against feelings of being unspoken, often considering and if and how to associate themselves with other castes.

In sum, the narrative of unspoken highlights how Dalits often engaged in self-censorship on social media, silencing their own voices out of fear, uncertainty, and emotional fatigue. Faced with the persistent misrecognition of caste-based abuse and the absence of a caste-sensitive context among users and platform operators, many Dalits withdrew from social media. As a result, their thoughts and experiences were concealed. Attempts to recover the Dalit context through digital activism provided a temporary respite, although no panacea.