When Police Kill Civilians: How Media Frames 'Encounter' vs Murder
TL;DR
When a Delhi Police head constable shot dead a 21-year-old delivery agent in Dwarka on April 26, 2026, some headlines called it a "firing incident" while others named it murder. This linguistic gap isn't accidental. Research shows that the words media chooses when reporting police violence directly shape whether the public demands accountability or shrugs and scrolls past.
A Birthday Party Ends in a Bullet
Around 2 AM on April 26, 2026, a group of young men stood on a road in Jaffarpur Kalan, Dwarka, waiting for a ride home after a birthday celebration. Head Constable Neeraj, a Special Cell officer with the Delhi Police, walked toward them, argued, pulled out his service weapon, and shot Pandav Kumar in the chest. Pandav, 21, was declared dead on arrival at the hospital. His friend Krishna took a bullet to the abdomen and remains critical.
The facts were blunt. An off-duty police officer, possibly intoxicated, killed an unarmed civilian over a noise complaint. A murder, by any honest reading.
But look at how the story was told.
The Headline Test
Here is a sampling of how major outlets framed the same event within hours of it happening:
| Outlet | Headline |
|---|---|
| India TV | "Delhi Police constable opens fire at labourers in Dwarka; One dead" |
| Deccan Herald | "Delhi cop opens fire at labourers after altercation in Dwarka, 1 dead" |
| NewsX | "Delhi Shocker: Head Constable Opens Fire After Birthday Party Argument" |
| DNA India | "Delhi Police head constable arrested for shooting dead food delivery agent" |
| Business Upturn | "Tragedy in Dwarka: Labourer killed, another injured in late-night firing" |
Notice the patterns. "Opens fire" makes the shooting sound almost procedural. "Altercation" implies mutual aggression. "Tragedy" frames a deliberate killing as something that simply happened, like a natural disaster. "Late-night firing" strips the sentence of its subject entirely. Who fired? The headline doesn't say.
Only DNA India's headline comes close to naming the act for what it was: a police officer shooting dead a civilian.
This isn't nitpicking. It is the central issue.
The Science of Obfuscatory Language
In 2022, researchers Bocar Ba, Patrick Bayer, Aurelie Ouss, and Jonathan Moreno-Medina published a landmark study in the Quarterly Journal of Economics titled "Officer-Involved: The Media Language of Police Killings." They analyzed thousands of American television news broadcasts covering police killings between 2013 and 2019 and found something striking.
Media outlets systematically use what they call "obfuscatory language" when reporting police killings, far more than when reporting civilian killings. These structures include:
- Passive voice: "A man was shot" instead of "An officer shot a man"
- No-agent constructions: "A man was killed" with no mention of who did the killing
- Nominalizations: "An officer-involved shooting occurred" instead of "An officer shot someone"
- Intransitive verbs: "A man died" instead of "An officer killed a man"
The researchers didn't stop at documenting patterns. They ran controlled experiments. Participants who read obfuscatory descriptions of the same event were measurably less likely to hold the police officer morally responsible. They were less likely to demand penalties or accountability.
The effect was strongest in cases involving unarmed victims, precisely the situations where accountability should matter most.
Journalist Radley Balko coined a term for this pattern: the "exonerative tense." It sounds academic, but it describes something visceral. Language that exonerates the perpetrator before any investigation has even begun.
And here is the kicker from the study: the obfuscation doesn't originate with journalists' editorial choices or audience demand. It traces back to the original narratives crafted by police departments. Reporters parrot police press releases, and the language of those releases is engineered to deflect.
India's Unique Euphemism: The "Encounter"
The United States has "officer-involved shooting." India has the "encounter."
The word "encounter" in Indian policing doesn't mean what it means in English. It doesn't refer to a chance meeting or a confrontation. It is a euphemism for extrajudicial killing, so deeply embedded in Indian media vocabulary that most readers don't even pause at it.
In February 2026, UN human rights experts issued a warning about India's policing practices, specifically calling out this terminology: "Even the use of terms such as 'encounters' and 'half-encounters' masks a pervasive and dehumanising practice that risks normalising unlawful violence and eroding public trust in law enforcement."
The typical Indian encounter follows a script so familiar that it has become a template. A "criminal" was "evading arrest." They "opened fire at police." Officers "retaliated in self-defense." The suspect "succumbed to injuries." Almost every police version across every state parrots this exact sequence. The narrative is so standardized that the Pulitzer Center documented it as a tool of institutional cover-up.
The numbers tell a grim story. Between 2017 and 2023, 183 people were killed in encounters in Uttar Pradesh alone. Over 5,000 were injured during the same period, a disproportionate number of them shot in the left knee. This isn't coincidence. It is policy dressed up as policing.
In Assam, the Gauhati High Court was informed in 2023 that 161 "police action" incidents occurred in just 13 months, leaving 51 dead and 139 injured. In many of these cases, FIRs were registered against the victims, not against the officers.
In May 2025, the Supreme Court intervened directly, ordering the Assam State Human Rights Commission to investigate all alleged fake encounters in the state. "The use of excessive or unlawful force by public authorities against a victim cannot be legitimised," the bench ruled, setting aside the Gauhati High Court's earlier dismissal of a PIL seeking accountability. The Court directed the AHRC to issue public notices in national and vernacular newspapers, inviting victims' families to come forward with evidence.
The political backdrop makes this worse, not better. Assam Chief Minister Himanta Biswa Sarma publicly encouraged police to "shoot criminals in the leg," claiming it was legally permissible. UP Chief Minister Yogi Adityanath promised he would "help criminals meet Yamraj." When chief ministers openly endorse extrajudicial violence, the "encounter" becomes not just a euphemism but a performance. The language of the press release follows the signal from the top.
The Accountability Desert
India's data on police accountability is not just disappointing. It is almost comical in its bleakness.
According to NHRC data submitted to Parliament in 2026, between 140 and 176 deaths in police custody are reported every year. Over five years (2021-2026), exactly one case of disciplinary action was recorded against a police officer for a custodial death.
One.
The Status of Policing in India Report 2025, published by Common Cause, describes a pattern that has remained essentially unchanged for decades: deaths are recorded, cases are registered, magisterial inquiries are conducted by local magistrates who depend on the same police force they are supposed to investigate, and the file gathers dust.
India still hasn't ratified the UN Convention Against Torture (UNCAT). It lacks a standalone domestic law criminalizing torture as a distinct offense. The Supreme Court's D.K. Basu guidelines, issued in 1997, laid out 11 safeguards for police conduct during arrests, from visible name tags to mandatory medical examinations every 48 hours to informing a detainee's family within 12 hours. Nearly three decades later, as the Supreme Court itself has lamented, courts are still "forced to restate" these principles because police routinely ignore them.
The Jayaraj-Bennix Verdict: A Rare Exception
In April 2026, India witnessed something extraordinary. A court sentenced nine police officers to death for the 2020 custodial murders of P. Jayaraj and his son J. Bennix in Sathankulam, Tamil Nadu.
The facts of the case are harrowing. Jayaraj was arrested for allegedly violating COVID-19 trading hour restrictions. His son Bennix went to the police station to check on his father and was detained as well. Between 7:45 PM and 3 AM, both men were beaten and subjected to sexual violence. Bennix, 31, sustained 13 external injuries and died on June 22, 2020. Jayaraj, 58, had 17 injuries and died the next day.
The police initially claimed the injuries resulted from a "scuffle while resisting arrest." It took a head constable named Revathi, who witnessed the torture firsthand, to break the wall of silence.
Judge G. Muthukumaran of the First Additional District and Sessions Court in Madurai called it a "rarest of rare" case, noting the officers acted with "vengeance" and "deliberate intent to instill fear."
Amnesty International called the verdict "a rare moment of accountability," but cautioned that the death penalty is "a deflection from the deeper reforms urgently required to ensure police oversight and accountability."
Between 2018 and 2023, zero convictions were recorded in custodial death cases across India. The Jayaraj-Bennix verdict is the exception that exposes the rule.
And notice how language played a role even in this case. For six years, the official version was that Jayaraj and Bennix sustained injuries "while resisting arrest." That phrase appeared in initial FIRs and early news reports. It was only after the CBI took over, examined 105 witnesses, and the eyewitness testimony of head constable Revathi became public that the narrative shifted from "resistance" to "torture." The language of the initial police report shaped six years of public perception before the truth caught up.
How Language Becomes Policy
The connection between media language and policy outcomes isn't abstract. It is mechanical.
When a police shooting is described as a "firing incident," the public processes it differently than when it is called a murder. When a custodial death is reported as "the accused died during interrogation," the passive construction literally removes the killer from the sentence. The reader's brain, processing the information quickly, registers an event without an agent. Something happened. Nobody did it.
This matters because public opinion drives political will, and political will drives reform. If every encounter killing is framed as a gunfight between brave officers and dangerous criminals, why would any government invest political capital in police accountability? If every custodial death is reported as a medical event, why would voters demand an anti-torture law?
The NBER study's most uncomfortable finding was that the source of obfuscatory language in media isn't editorial bias or audience preference. It is police public relations. Reporters receive a press release written in the exonerative tense, and they reproduce it. The police control the first draft of history, and in breaking news, the first draft is often the only one that matters.
In India, this dynamic is amplified by the relationship between police and press. Crime beat reporters depend on police sources for their daily stories. Challenging the official narrative risks losing access. The incentive structure pushes toward stenography, not scrutiny.
The Dwarka Shooting Through This Lens
Return to the Dwarka case. DCP Kushal Pal Singh described the incident as follows: the accused "walked toward" the group and "began arguing," then "took out his weapon and shot the victim in the upper body."
This is, by Indian police briefing standards, unusually direct. The officer is named as the agent. The act is described in active voice. There is no "encounter" framing, no self-defense claim.
Why? Because the killer was off-duty. He wasn't on an operation. There was no official narrative to protect. The police department itself treated the case as a straightforward murder, arresting the accused and registering a murder FIR.
This contrast is revealing. When a police officer kills outside the line of duty, the framing is clear and direct. When they kill on duty, during an "encounter" or in custody, the language shifts. The passive voice descends. The euphemisms multiply. The officer vanishes from the sentence.
The honesty with which Dwarka was reported is, paradoxically, the strongest evidence of how dishonestly "encounters" are reported.
What Readers Can Do
Media literacy isn't just about spotting fake news. It is about reading the grammar of real news and understanding what it hides.
Here is a practical checklist for reading any report about police violence:
1. Check the voice. Is the sentence in active voice ("Officer X shot Y") or passive voice ("Y was shot")? Passive voice in police reporting almost always obscures responsibility.
2. Look for the agent. Every killing has a killer. If the headline or lede doesn't name who did the killing, ask why.
3. Watch for euphemisms. "Encounter," "police action," "firing incident," "succumbed to injuries" are all terms that soften the reality. Substitute the plain-language version and see how the story reads differently.
4. Note the source. Is the article based primarily on a police press release? Has any independent witness, lawyer, or family member been quoted? Single-source reporting from police is stenography, not journalism.
5. Compare across outlets. The same event reported by five outlets can read like five different events. The variation tells you where editorial choices are being made.
6. Follow the follow-up. The initial report of a police killing is almost always based on the police version. The truth, if it emerges, comes later. Read the follow-up stories, not just the breaking news.
The Reforms That Would Actually Matter
India's police accountability gap is not a mystery waiting to be solved. The solutions have been identified, debated, and shelved repeatedly. A few that would make a structural difference:
Ratify UNCAT. India is one of the few democracies that hasn't ratified the UN Convention Against Torture. A standalone anti-torture law would criminalize what is currently prosecuted under general assault provisions, if at all.
Mandatory body cameras. Several states have piloted body-worn cameras for police. Making them mandatory during arrests and encounters, with footage preserved as evidence, would break the monopoly police have over the narrative.
Independent investigation bodies. Magisterial inquiries conducted by officials who work alongside police are structurally compromised. An independent body, modeled on the UK's Independent Office for Police Conduct, would provide genuine oversight.
Enforce D.K. Basu. The Supreme Court issued its guidelines in 1997. Enforcing them with real consequences, not just contempt proceedings, would save lives. CCTV in every police station, as the Supreme Court has ordered, would be a start.
Language guidelines for newsrooms. News organizations could adopt internal style guides that prohibit passive voice in reporting police killings and require naming the officer as the agent in the first sentence. The Associated Press and Reuters have moved in this direction. Indian newsrooms have not.
Separate the crime beat from the police beat. The fundamental conflict of interest in Indian journalism is that reporters who cover crime depend entirely on police sources. Creating dedicated accountability desks that cover police conduct separately from daily crime reporting would break this dependency. Some digital-first newsrooms in India are beginning to experiment with this model, but legacy outlets remain locked in the access-for-coverage exchange that produces stenographic reporting.
The Bottom Line
Words are not neutral carriers of information. When a police killing is called a "tragedy" instead of a murder, when a custodial death is reported as a medical event, when an extrajudicial execution is labeled an "encounter," the language does political work. It deflects responsibility, suppresses outrage, and shields institutions from accountability.
Pandav Kumar, 21, was shot in the chest by a Delhi Police head constable named Neeraj on the night of April 26, 2026. He died. That sentence names the victim, the perpetrator, and the act. It is the simplest form of accountability: telling the truth in plain language.
Every reader deserves that clarity. Every victim deserves that honesty.
Sources: India TV News, Deccan Herald, OHCHR, NBER Working Paper 30209, Quarterly Journal of Economics, FACTLY, Common Cause SPIR 2025, LiveLaw, Amnesty International, News9Live
