Iran's crackdown: the arithmetic of fear, and the deaths it leaves behind

Tuesday 27 January 2026

When a state decides that a protest is not a political fact to be negotiated with but a security incident to be extinguished, the language of governance changes. It becomes a language of curfews and cordons, of arrests without paperwork, of communications cut at the switch. In Iran, since protests that began on 28 December 2025 spread beyond the Grand Bazaar and into streets across the country, that change in language has been accompanied by something grimmer: a rising body count that no single institution can reliably tally, and a pattern of killings described by witnesses and rights organisations as both lethal and indiscriminate. 

The first task, if one wishes to describe the scale of killing, is to acknowledge the intentional fog around the numbers. Iran’s authorities have imposed sweeping restrictions on communications, including a near-total internet shutdown beginning on 8 January 2026, alongside limits on phone connectivity and pressure on families. These measures do not merely inconvenience journalists. They reshape the evidence base, slowing the circulation of names, photographs, hospital records and burial details that make a death count more than a rumour. In that sense the blackout is not a side effect of the crackdown; it is part of its method. 

Even within that fog, the broad contours are clear. Iran’s State media has now issued an official death toll. On 21 January 2026 statements carried by State television, attributed to the Interior Ministry and the Foundation of Martyrs and Veterans Affairs, said 3,117 people were killed. The same reporting said 2,427 of the dead were “civilians and security forces”, without clarifying how many were protesters, bystanders, police or members of Iran’s paramilitary and Revolutionary Guard structures, nor explaining the remainder of the figure. 

That official number is, by design, a single blunt statistic. Outside Iran, activist networks and human rights organisations have offered higher and more granular estimates, albeit with the caveat that verification is impaired by the State’s own restrictions. A prominent example is the US-based Human Rights Activists News Agency (HRANA), which Reuters reports as having verified 4,519 deaths at the time of its 21 January dispatch, including 4,251 protesters, 197 security personnel, 35 people under 18 and 38 bystanders. HRANA also reportedly had thousands of additional deaths “under review”. 

This divergence between the official count and activist counts is not unusual in violent crackdowns, but it matters for how the killings are understood. A high but vague official figure can serve two functions at once. It can be offered externally as proof that the State is being transparent, while internally it can be used to blur categories: to merge protester, passer-by and security officer into a single mass of “victims of unrest”, and to shift attention from who fired the bullets to the fact that “chaos” occurred. The competing counts, by contrast, are attempts to restore categories that a security State prefers to dissolve.

Numbers alone, however, do not convey the nature of the killings. The more telling point is how people are said to have been killed, and who is said to have been targeted.

Witness testimony reported by Reuters describes security forces arriving and firing “randomly” towards demonstrators, with civilians nearby caught in the gunfire. In those accounts, the dead are not only those chanting or throwing stones but also those watching from pavements, sheltering in shops or trying to flee the commotion. That, in practical terms, is what “indiscriminate” looks like: lethal force applied to a space, rather than to an identified threat, with predictable consequences for anyone in the vicinity. 

Human Rights Watch has described the killings since 8 January as a “coordinated” escalation, with evidence reviewed suggesting that many were shot in the head and torso. It also reports that protesters and bystanders were killed across multiple provinces, and that severe communications restrictions have concealed the full scale of what occurred. The significance of such descriptions is not clinical detail but intent. Fire directed at the centre of the body is, in plain terms, fire intended to kill, not to disperse. 

Amnesty International, in parallel, has described security forces using rifles and shotguns, including metal pellets, alongside tear gas and beatings, and has called the use of firearms unlawful. In the Reuters reporting, Amnesty is quoted as documenting forces firing from streets and rooftops and repeatedly aiming at unarmed individuals’ heads and torsos. This is not simply the story of “crowd control gone wrong”. It is the allegation of a policy choice: a preference for lethal deterrence over containment, with little discrimination between an active protester and a person close enough to be seen, suspected or simply unfortunate. 

The inclusion of bystanders and minors in multiple tallies and reports adds another dimension to the same conclusion. When deaths extend beyond those actively participating, the State’s claim to necessity becomes harder to sustain. A government can argue, however unpersuasively, that it faced an armed insurrection. It cannot easily argue that an unarmed teenager observing events or a passer-by trying to get home was a combatant. Witness accounts of such deaths, and HRANA’s categorisation of bystanders and people under 18, make the concept of indiscriminate killing more than a rhetorical flourish. 

There is also the question of narrative control after the killing has occurred. Reuters reports families describing pressure to attribute deaths to “terrorists and rioters”. This is a familiar manoeuvre in authoritarian crisis management: to separate the State from the act of killing by assigning moral agency to an undefined enemy. It is also a means of pre-empting mourning as politics. Funerals and vigils are, in Iran as elsewhere, moments when grief can become collective and collective feeling can become mobilisation. A State that fears crowds will often fear funerals most. 

What, then, does “scale” mean in the Iranian case, beyond a contest of numbers?

It means, first, that even the State’s own account concedes thousands of deaths in a matter of weeks, which is itself an admission of extraordinary violence. 

It means, secondly, that independent and semi-independent counts, produced under blackout conditions, converge upon an assessment that the deaths are not in the dozens or the low hundreds but in the high hundreds to several thousands, with further cases unverified or still emerging. 

It means, thirdly, that a significant portion of the killing is alleged to have been carried out through tactics that predictably widen the circle of victims: firing into crowds, firing in public streets where non-protesters are present, and using elevated positions that turn a square or a boulevard into a killing ground rather than a site of policing. 

And it means, finally, that the crackdown is paired with information control that delays the reckoning. A death toll in such circumstances is not a static figure. It is a number that tends to rise in steps, as communications return, as families speak, as graves are photographed, as hospitals leak data, as names move from whisper to publication.

In the immediate term, Iran’s rulers appear to be wagering that intensity will substitute for legitimacy: that the memory of gunfire, arrest and disappearance will discourage the next march, and that international outrage will peak and pass. Human Rights Watch and Amnesty International are, predictably, calling for international action and accountability mechanisms. Whether such mechanisms will bite is an open question, but the record being assembled now matters regardless. A state can survive a protest by killing it. It survives longer by persuading the world, and its own people, that the killing was accidental, necessary or someone else’s fault. The emerging evidence, and the pattern described as indiscriminate, cuts against that defence.