Substantive change in policing will require a serious training-based input that ensures that both police and paramilitary accept that, in a democracy citizens will have grievances and their inherent right is to dialogue and question the state; more than anything else, the author argues that statutory accountability from individual officers who use brute force causing death and grievous hurt, infliction of sexual violence during crowd operations must also narrow down the sanction requirement for prosecution and create a state liability for reparation
A protest is a strange, difficult thing to be inside. It looks, from the outside, like a single body moving with a single will, but anyone who has actually stood in one knows it is nothing of the sort. It is a loose weave of grandmothers who came because a grandchild was picked up, students who saved a week of bus fare to be there, farmers who slept in trucks to be there, office workers who slipped out at lunch, drifters who came for the free water, agitators who came for the fight, and thousands of ordinary people who simply felt they could not stay silent any longer. It is tired feet and dry throats and long stretches of nothing happening. It is chants that rise and fall for no clear reason. It is bad information travelling faster than good information. It is fear and hope tangled together in the same body. What a protest is not, almost ever, is a coordinated attack on the state. And yet the standard institutional response of the Indian state, across governments and across decades, has been to treat it as exactly that.
Also, a celebration is in order. A celebration of the grit and courage the youth of this country has shown in the face of government apathy and violence by the police, all of it leading to resignation of Dharmendra Pradhan, a powerful union minister in the NDA III government.
While the main anchor entity of the movement Cockroach Janata Party has declared that government will withdraw FIRs against the protestors, a question on the responsibility for the violence against the protestors remains i.e., why were the peaceful protestors dealt with such violence to the extent of pellet guns, by the Delhi Police—a police department that works under the control of the Central Government (Home Ministry). That again begs a more structural question. How did India end up with a civilian police that is capable of unleashing such violence on peaceful protestors and since such violence is not new, how should the system be reformed?
This is the starting point from which any serious reform conversation has to begin, because the failure of Indian crowd control is not primarily a failure of individual officers behaving badly. It is a failure of an institution that has never fully accepted that a citizen with a grievance is not the same thing as an enemy of the state.
Where does the problem lie?
The Police Act of 1861 was written in the wake of 1857 with the explicit purpose of controlling a hostile native population, and its basic architecture, a paramilitary force accountable upward to the political executive rather than outward to the public, granted enormous statutory discretion and almost no statutory duty, remains largely intact. The Supreme Court’s Prakash Singh judgment of 2006 tried to open a window wherein it fixed tenures for senior officers, separation of investigation from law and order, independent complaints authorities. Every state has technically complied and almost every state has hollowed out that compliance in practice. Complaints authorities exist on paper with no staff, no budget, and no independence; officers are transferred within months of politically inconvenient decisions; the political control that Prakash Singh tried to loosen has, in many states, actually tightened. Reform in India has developed a peculiar shape wherein the letter of change manifests without the substance of it.
But locating the entire problem in colonial history is too easy, because it lets the last seventy-eight years off the hook. Post-independence India has added its own pathologies to the colonial skeleton. The most important of these is that the police have become a resource that ruling parties at the state level deploy against political opposition, and the crowd-handling function has become the sharpest edge of that deployment. A sub-inspector on the ground has almost unlimited discretion to declare an assembly unlawful under what is now Section 148 of the BNSS, to order a baton charge, to fire tear gas. There are no binding directions, no codified escalation ladder, and no procedural check that must be cleared before force is used. The following is what Section 148 (1) of the BNSS states:
Any Executive Magistrate or officer in charge of a police station or, in the absence of such officer in charge, any police officer, not below the rank of a sub-inspector, may command any unlawful assembly, or any assembly of five or more persons likely to cause a disturbance of the public peace, to disperse; and it shall thereupon be the duty of the members of such assembly to disperse accordingly.
In this very consciously worded provision of the ironically named Bharatiya Nagari Suraksha Sanhita, the brakes on arbitrary decision making are what the law starts with when it says “Any Executive Magistrate” a rank that can be seen as higher than a sub-inspector and its equal rank “an officer in-charge of a police station.” These brakes are immediately loosened as soon as the provision says “in the absence of such officer in-charge, any officer not below the rank of sub-inspector” which is essentially ‘any sub-inspector may command an unlawful assembly to disperse’. Section 148(2) states that if the unlawful assembly does not break or disperse, force may be used.
This vacuum of doctrine is filled, on the ground, by the officer’s own instincts, which are shaped by his training, and his training has taught him that a crowd is a mob and a mob is dispersed. The verbs matter. Everything downstream of that verb, the choice of equipment, the posture of the constable, the tone of the first announcement over the loudhailer, flows from an idea that treats the citizen as the problem.
Reform, then, has to begin with rewriting the verb, and only afterward can it move to the machinery. A new statutory framework, whether through fresh state Police Acts or a national code drafted with genuine input from civil liberties lawyers, protest organisers, retired officers, and public health experts, would need to establish something Indian policing has never really had i.e., an affirmative duty on the police to facilitate the exercise of Article 19 rights, with the burden of proving imminent danger sitting on the state rather than on the citizen. This is a bigger conceptual shift than I am making it out to be. It converts a protest from a privilege the police grant to a right the police protect, and it changes what a good day at work looks like for the officer on the ground. Under the current architecture, a successfully dispersed protest is a professional success. However, if facilitation becomes the way, a protest that ended peacefully with people going home hoarse and tired is the professional success, and the constable who talked a knot of angry young men out of throwing stones is the officer who is promoted.
It is here that international experience begins to matter, but it matters with far more nuance than mere adoption of any other country’s practices.
Is there a better way to handle crowds?
Swedish dialogue policing, developed after the Gothenburg riots of 2001, is genuinely interesting. A uniformed dialogue officers whose only job is to talk, weeks of pre-protest negotiation with organisers, jointly walked routes, agreed visible signals for escalation. But Sweden is a country of ten million people with a small, homogeneous police service, protests that usually have identifiable organisers, and a political culture that treats the state as broadly trustworthy. Almost none of that describes India. An Indian protest often has no single leadership to negotiate with. And that itself has many legitimate reasons. Shaheen Bagh’s power came precisely from being leaderless, and the farmers’ protest at Delhi’s borders had a rotating cast of union heads who did not always agree. A dialogue officer culture cannot be imported wholesale into a force where the sub-inspector’s authority over the constable rests substantially on his willingness to be feared, or into a political environment where the ruling party may actively want the protest broken. What can be borrowed is the underlying idea that communication is a professional skill and that the officer who de-escalates is doing the harder, more valuable job. The Indian adaptation would have to invent new institutional forms. Perhaps designated Protest Liaison Cells at the district level, staffed by officers who volunteer for this work and are trained in the sociology of Indian protests specifically, the role of caste and community in shaping crowd dynamics, the way rumours travel through WhatsApp during a sit-in, the different psychology of a farmer’s rally and a student’s march.
How do we ensure Accountability and Transparency?
Training has to move in parallel with theories and ideas, because an institution behaves the way it has been trained to behave. The average Indian constable receives around nine months of basic training, of which crowd control occupies perhaps a handful of days delivered largely as physical drill: forming a cordon, swinging a lathi, executing a charge. What is ironic is that the Delhi Police had sent its force to train on crowd control to the Rapid Action Force’s academy in Meerut. Therefore, a serious redesign would build in weeks rather than days on crowd psychology, non-violent communication, and scenario simulation, and it would bring in outside trainers whom the police service currently views with suspicion i.e., lawyers, doctors, protest organisers themselves, journalists who have been on the receiving end of a lathi charge. This means that not only is there a need to impart more democratic training to the police officers, there is also a need to first have them unlearn the anti-democratic training they undergo within the current system. Empathy is a trainable skill, and India has trained it out of the force with unusual efficiency.
Equipment is the piece where brakes are needed the most, and it is where the current system does its most visible damage. The tear gas shells commonly used in Indian crowd control operations are frequently past expiry, often fired directly at protesters rather than in high arcs, and their canisters have killed by head impact alone. The pattern is documented from Tuticorin in 2018 through Delhi in 2020 and beyond. Pellet guns, deployed extensively in Kashmir, have blinded thousands, including children, and belong to a category of weapon that no civilian police service anywhere in the democratic world uses at that scale. Reform here is partly a matter of procurement rules, independent medical review of every less-lethal munition, withdrawal of anything with a documented pattern of causing serious injury, and partly a matter of what the equipment signals. When a police unit arrives at a peaceful sit-in in full riot gear with automatic weapons visible, it has communicated something before a word is spoken, and the communication is that this is not a conversation.
Body cameras are worth thinking about carefully, because they have become the fashionable answer everywhere and their record is genuinely mixed. American police departments adopted them at scale after Ferguson in 2014, and the studies since have shown that they reduce complaints and reduce force only when the surrounding accountability system actually processes the footage. Otherwise, they simply become one more piece of evidence that the police control. An Indian body camera regime would work only if the footage was held by an authority outside the police, if turning the camera off during an incident created a legal presumption against the officer, and if selective deletion carried personal consequences. Without those scaffolds, a body camera is theatre.
Section 197 of the CrPC, now carried into the BNSS, requires government sanction to prosecute a public servant for acts committed in the course of duty, and that sanction is almost never granted for crowd-control violence. The Hashimpura convictions came thirty-one years after the killings, and only because the High Court refused to let the case die. The blinding, firings, and custodial deaths that fill the decades between then and now have produced almost no criminal consequences for the officers involved. Reform requires narrowing the sanction requirement sharply for offences involving death, grievous hurt, or sexual violence during crowd operations; creating a statutory presumption of state liability with immediate compensation from a dedicated fund; and, most importantly, building Police Complaints Authorities with the staff, the budget, and the independence to actually investigate. None of this needs a foreign model to justify it. It is a purely Indian problem with a purely Indian solution i.e., the state must be willing to let its own officers be held to account, and until it is, every other reform is decoration. And the deployment of paramilitary forces like the CRPF for what is essentially civil policing has to be scaled back, because their training is for insurgency, not for citizens.
Be that as it may, some protests turn violent and they will need force to restore peace. When such protests do happen, it is natural for the police to need split-second decision-making powers. It is here that a transparent system should be put in place that protects both the citizens against undue force and also ensures that the police are able to restore peace. What could those systems be?
A detailed, publicly accessible post incident report, with statements taken from the personnel as soon as their shift ends, by an independent protest cell would be desirable. This cell could consist of independently appointed panel of citizens, by the people’s representatives at the local level i.e., the Sarpanches, Corporators, Mandal Parishad and Zilla Parishad Members via detailed reasoning for each selection.
For added protection, each cell should be mandatorily adopt a consensus approach while making decisions. This could ensure that accountability is introduced. It is here again that Bodycams could be of assistance to show whether disproportionate force is being used or not. The police should be able to show to the protest cell all their attempts to assuage the crowd or calm it before they resorted to force. Individual complaints against any officer using disproportionate force need to be taken. For example, there have been many instances in the recent Delhi protests against NEET, where videos were shot when the police used batons to thrash protesters as they were walking without any disturbance. Such excessive and unnecessary use of force needs to be held accountable for any meaningful reform.
Conclusion
Reform of this depth is slow, and its friends inside the system are few. But the alternative is what India has now i.e., a police service whose default response to democratic expression is force, and a citizenry that has learned to expect the lathi as the state’s first sentence in any conversation about grievance. That expectation, more than any single incident, is what corrodes the trust on which a democracy runs. A protest, at its heart, is a citizen still willing to talk to the state. The reform of Indian policing, in the end, is the question of whether the state is willing to talk back.
(The author is part of the legal research team of the organisation)
Related:
Permission to Protest? Why the right to dissent needs no licence
Defectors & Democracy: A critique of the Tenth Schedule of the Indian Constitution
July 2026: A dissent that was brutally put down, a crackdown after the protest ended