Daniel sent us this one — and it's heavy. There was an attack in Belfast yesterday. A Sudanese immigrant attempted to behead a random person in the street. Bystanders subdued the attacker and are being credited with saving the victim's life. The question is: in the very unlikely circumstance where you find yourself in something like this, what are the actual recommendations for how to behave? We've all seen those videos where something horrific happens and people just walk past, and there's this gut-level revulsion. But intervening is dangerous. So what do the people who study this actually say you should do?
The first thing to say is that the Belfast bystanders did something remarkable, and we should sit with that for a second before we get clinical about it. They saw a man attempting to decapitate someone in broad daylight and they ran toward the danger. That's not normal human behavior. Most people freeze. There's decades of research on this — it's called the bystander effect, and the core finding is that the more people who are present, the less likely any individual is to intervene. Everyone assumes someone else will do it.
Diffusion of responsibility. "Surely one of these other thirty people has called the police." And then nobody has.
Bibb Latané and John Darley established this back in the late nineteen sixties after the Kitty Genovese case, though the Genovese story was later found to be significantly misreported — far fewer witnesses than the original thirty-eight figure, and several did call the police. But the underlying mechanism they identified holds up. People look to others for cues. If nobody else is reacting, we interpret the situation as less serious than it is. Then there's evaluation apprehension — fear of looking foolish if you misinterpret what's happening. And what they called "pluralistic ignorance," where a group collectively misreads an emergency because everyone is outwardly calm while inwardly panicking.
Step one is basically: don't look around to see what everyone else is doing, because they're all looking at you doing the same thing.
That's actually a really clean summary of the intervention. The research on active bystander training — and there are now formal programs for this, things like Green Dot and Hollaback's bystander intervention model, though those are mostly geared toward sexual harassment and bullying prevention — consistently emphasizes that the single most powerful thing you can do is break the diffusion of responsibility before you do anything else. Point at a specific person and give them a specific task. "You in the red jacket, call nine nine nine now." "You, get the defibrillator from the station by the entrance." Specific person, specific instruction. It shatters the bystander effect instantly.
I've heard this described as the difference between shouting "someone call an ambulance" and pointing at one person and saying "call an ambulance." The first one, everyone assumes someone else will do it. The second one, that person now has a job and social pressure to do it.
And this is where the Belfast case gets interesting. In a violent attack, you don't have the luxury of organizing a response team. The bystanders there appear to have acted on instinct — multiple people physically engaged the attacker. And the standard advice from law enforcement and security professionals on this is actually quite nuanced. The UK's National Counter Terrorism Security Office, which publishes the "Run, Hide, Tell" guidance for terror attacks, doesn't really cover citizen intervention in street attacks. But there's a broader body of recommendations from self-defense instructors, law enforcement trainers, and psychologists who study emergency response.
Let's get concrete. You're walking down the street. You see someone with a knife attacking another person. What's the actual playbook?
The consensus breaks down into a hierarchy, and I think it's important to frame it this way because people tend to imagine the binary choice — either you're a hero who tackles the guy or you're a coward who does nothing. The reality is there are about five tiers of intervention, and you escalate only as the situation demands and your capacity allows. The first tier is always: be a good witness. Call emergency services immediately. Give them the location, description of the attacker, description of the victim, nature of the injuries. Stay on the line. If you do nothing else, you've already done something that materially affects the outcome.
Because response time is everything with traumatic injuries. A person can bleed out from a severed artery in three to five minutes.
Less, with the carotid. Two to three minutes. So that phone call is not a small thing. It's the difference between the ambulance arriving while the victim can still be saved and arriving to a body. The second tier is verbal intervention. This is underrated. Shouting "Police are on the way" or "Drop the weapon" from a safe distance can disrupt an attacker's focus. Attackers are not typically in a rational state, but they are goal-directed. Introducing an unexpected variable — a loud voice, the suggestion that authorities are imminent — can break that focus long enough for the victim to get away or for others to intervene physically.
The verbal equivalent of throwing a rock into the gears.
That's it exactly. There's also something that Gavin de Becker talks about in "The Gift of Fear" — he's the security consultant who pioneered threat assessment for public figures. He writes about how predators rely on the compliance and silence of bystanders. A loud voice announcing that this is being witnessed and reported disrupts that assumption. You're telling the attacker they're no longer in a permissive environment.
De Becker's whole framework is about listening to the intuitive signals that something is wrong before it escalates to violence. Which is relevant here because the third tier — before we get to physical intervention — is probably the most accessible for most people.
The third tier is what's called indirect physical intervention. This is where you're not making contact with the attacker but you're changing the physical environment. Chairs, bottles, whatever's at hand. Creating obstacles between the attacker and the victim. Using a fire extinguisher — not just as a projectile but the spray itself is a fantastic disorienting tool. It creates a visual barrier, the noise is startling, and the chemical irritant can temporarily blind. This is something security professionals recommend because it lets you maintain distance while actively disrupting the attack.
You're basically creating chaos for the attacker. Making the environment unworkable for what they're trying to do.
And this connects to something that William Aprill, who was a former deputy sheriff and defensive tactics instructor before he passed away, used to teach. He talked about how criminals — especially violent ones — are looking for a "good victim." Someone who won't resist, someone they can process quickly. The moment the environment becomes complicated, the attack becomes less viable. They may not flee, but they have to divert attention to the new threat. That buys time.
The fourth tier is where most people's imagination goes immediately — direct physical intervention — and I suspect this is where the recommendations get really specific and non-obvious.
And the first recommendation is: do not go in alone if you can avoid it. Numbers are the single biggest force multiplier. Two average people are far more effective than one strong person because you can attack from different angles, you can grab limbs, you can create confusion. The Belfast attack is a case study in this — multiple bystanders overwhelmed the attacker through sheer numbers, not through skill. The second recommendation: if you're going to engage physically, use improvised weapons. A belt with a heavy buckle. A backpack swung by the strap. A metal water bottle. A cane or umbrella. An actual chair. The goal is to create distance between the attacker's weapon and your body.
This is the "everything is a weapon if you're creative enough" principle, but the actual guidance seems to be: find something long. Something that extends your reach.
Reach is everything against an edged weapon. If someone has a knife and you have empty hands, you are at a catastrophic disadvantage. Every self-defense instructor I've ever read or spoken to is unequivocal about this: you will get cut. The question is whether you get cut somewhere survivable while doing something decisive. There's a well-known training concept from law enforcement called the "twenty-one foot rule" — though it's more of a guideline than a rule — which says that an attacker with a knife can close a distance of twenty-one feet and stab you before you can draw and fire a holstered firearm. That's seven meters. Most people dramatically underestimate how fast a committed attacker can move.
If you're within twenty-one feet and you're not armed and trained, you're already in the danger zone.
You're in the reactionary gap, and you've lost it. Which is why the fifth tier — and this one is genuinely controversial and jurisdiction-dependent — is deadly force. In places where citizens can legally carry firearms, the calculus changes. But even there, the legal aftermath of using deadly force in a third-party defense situation is extraordinarily complicated. You're not a police officer. You don't have qualified immunity. You could save someone's life and still face criminal charges or civil liability. It's not a reason not to act, but it's something you should understand going in.
That's the thing nobody wants to say out loud: even the "right" thing can ruin your life. You can be legally in the clear and still spend years in court and tens of thousands in legal fees.
There's a saying in the self-defense law community: "You may beat the rap, but you won't beat the ride." And that's assuming you're in a jurisdiction with strong self-defense laws. In the UK, the situation is different from the US. The UK has stricter weapons laws, including restrictions on carrying anything for the purpose of self-defense. If you carry a tactical pen or a pepper spray — which is classified as a firearm under UK law, incidentally — you can face charges just for having it, regardless of whether you use it. So the improvisation angle becomes even more important there. You're not carrying a weapon; you're carrying a metal water bottle because you're hydrated.
The hydration defense. I like it.
It sounds absurd but it's legally relevant. If you grab a chair in a pub during an attack, nobody's going to ask why you had a chair. If you pull out a collapsible baton, that's a different conversation with the police afterward.
What about the psychological piece? Because we've outlined the tactical hierarchy — witness, verbal, indirect physical, direct physical, deadly force — but none of that matters if you're frozen. What does the research say about overcoming that initial paralysis?
This is where the training meets the psychology. The US Department of Homeland Security's "Active Shooter: How to Respond" materials emphasize a simple decision loop: figure out the nearest exit, commit to it, move. The commitment is the key. People freeze because they're trying to find the optimal response, and the situation is moving faster than their decision-making. The antidote is having pre-made decisions. If you've already decided, at some point in your life, "if I see an active attack, I will call nine nine nine and then look for a fire extinguisher," you've short-circuited the freezing problem. You're not deciding; you're executing.
Pre-loaded scripts.
The military calls it "mental rehearsal" or "visualization." Athletes do it too. You run through scenarios in your head so that when something resembling that scenario actually happens, your brain has a template to reach for. It's not about being paranoid — it's about having done the thinking in advance so you don't have to do it under duress.
Which is uncomfortable. Nobody wants to sit around imagining knife attacks while they're waiting for the bus. But the alternative is arriving at that moment with no prepared response at all.
The research on first responders backs this up. Police officers, paramedics, firefighters — they train repeatedly on simulated emergencies specifically so that the cognitive load of "what do I do" is reduced in the real event. The psychologist Gary Klein, who studied decision-making in high-stakes environments, calls this "recognition-primed decision making." Experts don't compare options; they recognize patterns and act on the first workable option that matches their mental model.
For a civilian with no formal training, the closest you can get is to have thought about it once. To have read the hierarchy. To have decided, "I'm the kind of person who calls nine nine nine and then looks for something to throw.
That identity piece matters more than people think. There's research on what differentiates people who intervene in emergencies from those who don't, and one of the strongest predictors is whether they see themselves as the kind of person who would intervene. It's not about physical courage — it's about self-concept. If you've decided in advance that you're an intervener, the threshold for action drops significantly.
The flip side of the bystander effect, then, isn't just breaking diffusion of responsibility in the moment — it's having already accepted responsibility as a default.
This is where I want to complicate something. We talked about the videos that make the rounds — the ones where people walk past someone being attacked and everyone watching feels disgusted. And sometimes that disgust is warranted. But we should also be careful about judging from a phone screen. What looks like indifference from a distance can be freezing, can be terror, can be someone who is cognitively overwhelmed and not processing what they're seeing. The research on acute stress response shows that about fifteen percent of people will act calmly and effectively in a crisis. Another fifteen percent will become completely incapacitated — screaming, freezing, dissociating. The remaining seventy percent will be stunned and disoriented but capable of following direction.
Which loops back to your point about giving specific instructions. If you're in the fifteen percent who can act, your first job might be to activate the seventy percent.
Point at them and tell them what to do. "You, call nine nine nine." "You, help me lift this." "You, go to the corner and flag down the police when they arrive." You're not just getting tasks done — you're pulling people out of that stunned state by giving them a role.
Let's talk about the Belfast attack specifically, because it's the prompt and because it's unusual in how it played out. What do we actually know about what happened?
The details are still emerging, but what's been reported is that a man — described as a Sudanese immigrant — attacked a random person on the street and attempted to behead them. Multiple bystanders physically intervened and subdued the attacker until police arrived. The victim survived. The attacker is in custody. That's the broad outline as of this morning.
The "attempted beheading" detail is significant because it tells you something about the attacker's intent. This wasn't a robbery that escalated. This was a targeted act of extreme violence against a random person. The motivation hasn't been officially established, but the method suggests something ideological or psychotic rather than instrumental.
Which matters for the intervention calculus. Someone who is attempting to behead a stranger is not going to be deterred by a stern verbal warning. They're in a state where they've already crossed a massive psychological threshold. The Belfast bystanders understood this, whether consciously or not, and went straight to coordinated physical intervention. And it worked. The victim is alive.
There's something worth noting here about the political dimension that's going to swirl around this. A Sudanese immigrant attacking a random person on a Belfast street — this is going to become part of a larger conversation about immigration, integration, and public safety, and it probably should. But for our purposes, the question Daniel's asking isn't about policy. It's about what you do in the moment when policy has already failed and violence is happening in front of you.
The attack has happened. The question is: you're there, what do you do? And I think the answer has to start with an honest acknowledgment that there is no guaranteed safe option. The only thing you can control is whether you've prepared yourself to be useful. Everything after that is circumstances and luck.
Let's go deeper on the physical intervention piece, because I think this is where people have the most misconceptions. Hollywood has taught us that a hero disarms a knife-wielding attacker with some kind of choreographed move. What's the reality?
The reality is that disarming someone with a knife is extraordinarily difficult and dangerous even for trained people. Rory Miller, who wrote "Meditations on Violence" and spent years working in corrections and security, is blunt about this: if someone is committed to cutting you, you are going to get cut. The goal is to control where and how badly. The standard advice from people who actually teach knife defense is that you should be looking to control the weapon arm, not the weapon. Grab the wrist or forearm with both hands. Use your body weight. Take them to the ground if you can. And you need other people. One person controlling the arm, another controlling the other arm or the head, someone else on the legs. You're not disarming — you're dogpiling.
The most effective self-defense technique is apparently what happens in a kindergarten soccer game.
It's not elegant but it works. And this is one of those places where the gap between martial arts training and real-world violence becomes apparent. Most martial arts assume a degree of ritualized engagement — even the "practical" ones. Real violence is chaotic, fast, and usually involves multiple people, weapons, and an environment full of obstacles and improvised tools. The Belfast bystanders didn't perform a technique. They overwhelmed through numbers and probably sheer desperate physical effort.
There's also the medical piece. If you're intervening in an attack where someone has already been severely wounded — say, a neck wound from a knife — you're not just dealing with the attacker. You've got a victim who is bleeding out and needs immediate intervention.
This is where the "be a good witness" tier and the physical intervention tier can happen simultaneously with different people. If you're not directly engaging the attacker, you can be applying direct pressure to the wound. For a neck wound, you're pressing directly on the bleeding site with whatever cloth is available. A shirt, a jacket, anything. You're not worried about sterility — you're worried about exsanguination. The UK's St John Ambulance guidance for catastrophic bleeding is straightforward: direct pressure, call nine nine nine, keep pressure until paramedics arrive. If blood soaks through, add more material on top — don't remove the original. You're building a pressure stack.
Tourniquets for limb wounds. That's become much more mainstream knowledge in the last decade.
Largely because of the military experience in Iraq and Afghanistan. The Committee on Tactical Combat Casualty Care found that extremity hemorrhage was the leading cause of preventable death on the battlefield, and widespread tourniquet training brought those numbers down dramatically. Now you see "Stop the Bleed" programs training civilians in exactly these skills. A tourniquet goes high and tight on the limb — as close to the torso as possible, not just above the wound. You tighten until the bleeding stops. You note the time. You do not loosen it. That's a hospital's job.
The fully actualized bystander in a violent attack is simultaneously a first responder, a security guard, and an incident commander. Which is an absurd expectation for a random person on the street, but if you're the one who's able to act, that's the role.
This is why I keep coming back to the pre-loaded scripts. You don't need to be all of those things. You need to be one of them, competently. If you can apply pressure to a wound, you've done something that directly affects survival. If you can call emergency services and give clear information, you've done something that directly affects survival. If you can shout and throw things and disrupt the attacker's focus for even ten seconds, you've done something that directly affects survival. The goal isn't to be the hero. The goal is to be useful.
That's almost a mantra. "The goal isn't to be the hero. The goal is to be useful." I think that reframes the whole conversation.
It takes the ego out of it. And I think ego is actually part of what stops people from acting at the lower tiers. They think, "If I'm not tackling the guy, I'm not really helping." So they don't call nine nine nine either, because that feels insufficient. But calling nine nine nine is the single highest-impact action a bystander can take in most emergencies. It summons people with training, equipment, and legal authority. Everything else is a stopgap until they arrive.
Let's talk about the legal landscape a bit more, because I think people are confused about what they're allowed to do. In the UK, what does the law actually say about intervening to defend someone else?
The UK has something called "common law defense of another," which operates similarly to self-defense. You can use reasonable force to protect another person from an unlawful attack. The key word is "reasonable." It's judged based on what the person honestly believed at the time, even if that belief was mistaken, as long as it was honestly held. And the force used must be proportionate to the threat as they perceived it. There's no duty to retreat in UK law the way there is in some US states — you can stand your ground. But you can't use excessive force. If the attacker is subdued and no longer a threat, continuing to strike them moves from defense to assault.
In practice, "reasonable" and "proportionate" are determined after the fact by people who weren't there and had hours to think about it.
That's the tension. The Crown Prosecution Service guidance says that prosecutors should consider whether the person was acting in the heat of the moment and whether they had an honestly held belief that their actions were necessary. But yes, you're rolling the dice on how your actions will be interpreted. This is where the Belfast case will be interesting to watch. The bystanders subdued a man who was attempting to commit murder. It's hard to imagine a prosecution for their actions. But if someone had used a weapon they were carrying, or if the force had continued after the attacker was clearly incapacitated, the legal analysis changes.
This is the reality that the "good guy with a gun" narrative tends to gloss over. Even if you're legally justified, you're going to be detained, questioned, and possibly charged. Your weapon will be confiscated. You may spend a night or more in custody. Your name may be published. You'll need a solicitor. All of this while processing that you just participated in a violent event.
Which doesn't mean don't intervene. It means go in with clear eyes. The moral calculus and the practical calculus are both part of the decision. And this is where I think the pre-loaded script idea is so powerful — you can include the aftermath in the script. "I will call nine nine nine. I will stay on the line. When police arrive, I will put down anything in my hands and follow instructions exactly." That last part is critical. Police arriving at a chaotic scene don't know who's who. If you're holding a weapon, even an improvised one, you may be perceived as a threat.
Hands up, fingers spread, "I called it in, the attacker is over there.
Follow commands immediately. Don't argue about being one of the good guys — there will be time to sort that out later. The immediate priority for arriving officers is securing the scene and controlling everyone present.
We've covered the tactical hierarchy, the psychology, the legal landscape. What about the aftermath for the intervener? Because I think this is the piece that gets overlooked. You intervene in a violent attack. You see things. You do things.
This is a really important question and it's under-discussed. People who intervene in violent incidents — including off-duty police officers, military personnel, and civilians — can experience acute stress reactions, post-traumatic stress, and a whole range of psychological effects. The UK has something called the National Wellbeing Hub, which provides resources for people who've been involved in traumatic incidents. But most people don't know it exists and wouldn't think to access it.
Because the cultural script is: you're a hero, you should feel good about what you did. And a lot of people don't. They feel horror, or guilt, or they can't stop replaying it.
Or they feel nothing for a while and then it hits them weeks later. The clinical term is "acute stress disorder" if it's within the first month, and it can progress to PTSD if symptoms persist. The recommended approach is psychological first aid — which is not therapy, it's practical support in the immediate aftermath. Making sure the person has somewhere safe to go, someone to be with, basic needs met. Then monitoring for symptoms like intrusive memories, hypervigilance, avoidance, sleep disruption. And seeking professional help if those persist beyond a few weeks.
If you know someone who intervened in something like this, the move is not to call them a hero and expect them to be fine. It's to check in, repeatedly, and normalize the idea that witnessing violence up close does something to a person.
For the intervener themselves: understand that whatever you're feeling is probably normal. There's no correct emotional response to participating in a violent event. Some people feel elation, then guilt about the elation. Some people feel numb. Some people can't stop crying. All of it is within the range of normal acute stress response. The time to worry is when it doesn't fade, when it interferes with your ability to function, when you're avoiding things that remind you of the event to the point where your life constricts around it.
This connects to something broader about how we process these incidents as a society. The video goes viral. People watch it. They feel outrage or fear or disgust. Then the next video comes along and the previous one is forgotten. But for the people who were actually there — the victim, the bystanders who intervened, the witnesses — it doesn't cycle away.
The media coverage rarely follows up on the interveners six months later. We know the broad strokes of what happened in Belfast. We don't know the names of the people who ran toward a beheading. We probably won't know, six months from now, how they're doing. That's a gap in how we think about these events. The story ends when the attacker is in custody and the victim is in the hospital. But the psychological story is just beginning.
Let's pull this back to the practical. If someone listening to this wants to be more prepared — not paranoid, not LARPing as an operator, but more prepared to be useful in an emergency — what's the short list?
First, take a first aid course. Specifically one that covers catastrophic bleeding. Stop the Bleed is available in many countries, including free courses. St John Ambulance and the British Red Cross both offer first aid training in the UK. Knowing how to apply direct pressure, pack a wound, and use a tourniquet is probably the single highest-value skill a civilian can have for violent incidents. Second, do the mental rehearsal. It costs nothing and takes five minutes. Run through the hierarchy we discussed. Decide what your default actions will be. Third, know your local emergency numbers. In the UK it's nine nine nine or one one two. Fourth, understand the legal framework in your jurisdiction. Not so you can game it, but so you know what "reasonable force" means and what to expect afterward.
Fifth, I'd add: make peace with the fact that you might freeze. Preparation reduces the probability but doesn't eliminate it. If you freeze, you're not a moral failure. You're a human with a nervous system that did what nervous systems do under extreme threat.
That's important. The shame spiral after freezing can be as damaging as the trauma itself. People carry that for years — "I didn't do anything, I just stood there." And what the research shows is that freezing is not a choice. It's an autonomic response. The sympathetic nervous system has multiple modes — fight, flight, and freeze are all on the menu, and which one your brain selects in the moment is not under conscious control.
The deer in the headlights isn't making a moral decision.
It's a deeply conserved survival response. In some circumstances, freezing is actually adaptive — predators respond to movement. It's just that in a human context, with a human predator, it's usually not the right response. But your brain doesn't know that. It's running ancient firmware.
That's the human condition in two words.
Running on hardware that was optimized for the savannah, trying to handle situations that involve edged weapons and smartphones and legal liability. It's honestly remarkable that anyone responds effectively at all.
To bring this full circle — the Belfast bystanders. What they did was extraordinary, and the victim is alive because of it. The recommendations for anyone else who might find themselves in something similar: break the bystander effect by giving specific instructions to specific people. Call emergency services immediately. Escalate through the hierarchy — verbal, indirect physical, direct physical — only as far as necessary and as your capacity allows. Use numbers and improvised tools. Know how to stop catastrophic bleeding. Have the pre-loaded scripts. Understand the legal risks. And afterward, take care of your head.
The meta-lesson, which I think is the thing that actually stays with people: being useful is more important than being heroic. You don't need to be the one who tackles the attacker. You can be the one who makes the phone call, or applies pressure to the wound, or shouts from a window, or flags down the police. All of those things save lives. The hero narrative has us fixated on the most dangerous and least likely intervention, and it makes people overlook all the other ways they could help.
The hero narrative is basically the Hollywood-ification of crisis response. The reality is more distributed, more mundane, and in some ways more accessible. Most of the things that save lives in an emergency are not cinematic.
Tourniquets are not cinematic. But they work.
Now: Hilbert's daily fun fact.
Hilbert: In the nineteen forties, a researcher analyzing quipu — the knotted string recording devices of the pre-Columbian Andes — discovered that some of the cotton fibers had been dyed using a mordant derived from aluminum-rich clay deposits found only in a specific river basin in Suriname, suggesting a trade network that spanned the entire South American continent centuries before European contact.
...right.
So if there's one thing to sit with from this episode, it's that question: what kind of person do you want to be in an emergency, and have you done even five minutes of thinking about it? Most people haven't. And that's a fixable problem.
If you want to go deeper on the first aid side, Stop the Bleed has courses in the UK and internationally. It's a few hours. It could save someone's life someday.
Thanks to our producer Hilbert Flumingtop. This has been My Weird Prompts. Find us at myweirdprompts dot com or wherever you get your podcasts.
Until next time.
