---
author: "[[Bob Marshall]]"
tags:
  - blogpost
publish: 2026-08-14
sources: https://flowchainsensei.wordpress.com/2026/08/14/evangelism-is-violence/
---
## Evangelism Is Violence

A colleague of mine adopted a note-taking system a couple of years ago. Within about six weeks, he had become insufferable. Not because he was using it — I was pleased for him, honestly — but because every conversation now had a trapdoor in the floor that led directly to atomic notes and bidirectional linking. A stalled project, a book he’d read, his sister’s engagement: all roads led to the system.

That’s the harmless end of a spectrum. The other end is not harmless, and the two ends are made of the same material.

A word on what I mean by [violence](https://flowchainsensei.wordpress.com/2012/12/18/what-is-violence/), because I don’t mean fists. Marshall Rosenberg, who developed Nonviolent Communication, used the word for any way of speaking that produces hurt or harm — judging, blaming, labelling, demanding — on the grounds that these are all ways of exercising _power over_ a person rather than with them. On that definition, most ordinary persuasion is at least a little violent, and evangelism is violent almost by construction. It begins from the premise that there is something wrong with where you currently stand, and it does not consider itself finished until you have moved.

The title is a provocation, and I’ll defend it in a moment. But hold onto the note-taking colleague, because the point is not that he was a bad man. He wasn’t. The point is that the escalation from him to something much worse is short, and the steps are ones almost anyone will take when the stakes feel high enough.

### Why the Urge Is There at All

The cynical read is that evangelism is about power — I want you to think like me because a world of people who think like me is a world I control more of. That’s real, sometimes. But it’s a poor explanation for most of what I actually observe, which is people badly and unstrategically pestering their friends about sourdough.

Three better explanations, and they braid together.

The first is that beliefs are lonely. A conviction held entirely alone has a thin, unstable quality to it. We are not built to verify the world independently — we’re built to check our readings against other people’s, to go with the herd. When nobody else sees what you see, the common inference is often that you’re the one who’s wrong. So you go looking for corroboration. Converting someone isn’t only about them; it’s much more about shoring up the floor _you’re_ standing on. There’s a striking version of this in the research: when people’s confidence in a cherished belief is challenged or undermined, they don’t go quiet. They advocate harder.

The second explanation is more generous: people evangelise because they care about you. If you genuinely believe you’ve found something that reduces suffering — a treatment, a practice, a faith, a way of thinking about money — then keeping quiet is not neutral. It’s a kind of withholding. The awkward zealot at the party is, in his own frame, watching you limp around on a broken leg while he happens to be carrying a splint. His motive may be love.

The third is that most of what we believe isn’t stored as propositions. It’s stored as identity. ‘I think intermittent fasting has good evidence behind it’ is a claim. ‘I’m someone who fasts’ is a self. When you disagree with the first, I can update. When you disagree with the second, I have to defend myself, and defence frequently takes the form of offence.

None of these motives is malicious. That’s exactly the problem. Violence that knows itself is easy to refuse. Violence delivered as a gift is not.

### The Four Pillars: Fear, Obligation, Guilt and Shame

A word on where this framing comes from, since it gets attributed loosely. Fear, obligation and guilt — FOG — is Susan Forward’s, from her work on emotional blackmail; it describes the atmosphere a manipulator generates so that refusing feels impossible. Shame is the fourth I’d add, and the most corrosive. Rosenberg’s own taxonomy is different again: he identified moralistic judgements, comparisons, denial of responsibility and the communication of desires as demands. The four below are a serviceable map of how persuasion turns coercive, not a canonical list.

What matters is the pattern they share. Each one works by raising the cost of your no.

#### Fear

Fear is the oldest. Convert or burn. Buy this or your children’s futures are forfeit. Vote this way or the country enters a tailspin. The mechanism is simple: attach a catastrophe to non-compliance and the argument no longer has to be good, because you are no longer being asked to evaluate it. You’re being asked to escape it.

Fear is also self-legitimising in a way the others aren’t. If the threat is real — and sometimes it is — then frightening you looks like a public service rather than an assault. This is why the fear-based pitch survives so well in movements that are otherwise sincere and even valid. The trouble is that fear narrows attention onto the threat and away from the reasoning, which means it produces compliance and conviction in roughly equal measure, and you can’t tell from the outside which one you’ve got.

#### Obligation

Obligation is the quiet one. It works by converting an offer into a debt.

The move is usually invisible: I have told you the thing, and now you know it, and now your not-acting has become a choice you are answerable to me for. Once you’ve heard the sermon, indifference is no longer available; only rejection is. Notice how much work is being done by that shift. Before the conversation, you owed nothing. After it, you owe an account of yourself.

Every version of ‘now you have no excuse’ is this pillar. So is the friend who tells you the documentary changed his life and then keeps asking, mildly, whether you’ve watched it yet. (Close to a recent scenario for me).

#### Guilt

Guilt is about what you’ve done. You’re eating that. You’re funding that. You’re voting for that. Your choices, tallied up, are hurting someone.

Of the four, guilt is the most defensible, because sometimes the tally is accurate and someone genuinely is being hurt. The evidence on moral emotions is reasonably clear that guilt — a negative feeling about a specific behaviour — tends to prompt repair: apology, amends, change. That’s a real mechanism and not always an abusive one.

The abuse comes when guilt is manufactured at a volume the facts can’t support, or when it’s applied not to prompt repair but to establish leverage. The tell is what happens after you change: if the guilt lifts when the behaviour changes, it was information. If it just relocates to the next thing, it was a tool.

#### Shame

Shame is about what you are. Not you did a harmful thing — you are the kind of person who does harmful things. Ignorant. Selfish. Asleep. Complicit. Lost.

This is the one I’d call violent without hedging, and the research supports the intuition. Shame and guilt look similar and behave completely differently. Guilt-proneness is associated with empathy and constructive repair; shame-proneness is associated with defensiveness, blame-shifting, withdrawal and hostility. Shame doesn’t motivate change, because it doesn’t leave a route to change — you can amend a behaviour but you can’t amend a self, so the only available moves are to hide or to attack.

Which means that shaming someone into your position is, at best, converting them into a person who performs your position while resenting you. It’s the pillar that most reliably produces the opposite of what the evangelist wanted, and it’s the one people reach for when they’re losing.

That progression is worth noticing, actually. The four pillars tend to appear in that order, and the order tracks the weakness of the case. A strong argument doesn’t need any of them. A weak one starts with fear and ends with shame.

### Why It So Rarely Works

Here’s the cruel joke at the centre of it: the harder you push, the less likely it is to land.

People have a reliable allergy to having their agency overridden. Tell someone they must conclude X and you’ve handed them a reason to conclude not-X that has nothing to do with X’s merits — namely that they’d like to remain the author of their own mind. Psychologists call this reactance. Everyone else calls it being a person.

There’s a second failure that’s subtler. The convert and the prospect are standing in different places, and the convert forgets this. When something has genuinely rearranged your life, the evidence for it isn’t primarily argument — it’s the felt texture of your days since. That texture is exactly the part you cannot transmit. So you end up compressing a lived transformation into a debate you can win in eleven minutes, which is a bit like trying to explain a smell. What arrives at the other end is not your experience. It’s just pressure.

And a third, the least flattering. The evangelist frequently understands the thing less well than the feeling of conviction suggests. Ask almost anyone to explain, step by step and mechanistically, how their cause actually produces its effects, and the confident account thins out fast. This isn’t stupidity; it’s a well-documented feature of how explanatory knowledge feels from the inside. But it does mean the pitch is often being delivered by someone who has mistaken a strong impression for a working model. Example: Agile methods in software development.

### Can Evangelism Ever Be Nonviolent?

I think so, but only under a condition most evangelism can’t meet.

The answer isn’t to hold everything loosely and never advocate. That’s its own pathology, a pre-emptive surrender dressed up as humility. Some things are true and load-bearing, and the people who spread them have done enormous good. Vaccination, abolition, handwashing before surgery, the idea of a woman’s right to own property in her own name: all of these needed evangelists, and the evangelists were widely considered tiresome at the time. A rule that forbids advocacy forbids those too.

Rosenberg’s framework is useful here because it locates the violence somewhere specific. His four components are:

1. Observation
2. Feeling
3. Need
4. (Refusable) request

You say what you noticed without editorialising, what it produced in you, what you need, and what you’re asking for. Three of those are just honesty. The whole ethical weight sits in the fourth.

Because a request and a demand are indistinguishable at the moment of asking. The difference only shows up afterwards, in what happens when the answer is no. If the no is accepted — genuinely, with the relationship intact and the subject dropped — it was a request. If the no triggers a follow-up, a coldness, a sigh, a reframing, another article sent at midnight, then it was a demand all along and the request-shaped wrapping was decoration. This is the whole test, and it’s a behavioural one. You cannot pass it by rewording the ask.

So: nonviolent evangelism is possible, and it looks like this. You say what happened to you. You say plainly what you believe and why, including the parts you’re unsure of. You make one clear request. And then you carry the cost of the refusal yourself, rather than distributing it back onto the person who refused. No fear, because you haven’t attached a catastrophe. No obligation, because you haven’t converted your telling into their debt. No guilt beyond whatever the facts already justified. And no shame, ever, because shame is an attack on the person and not on the question.

_Which turns out to be more effective as well as more decent_, though I’d want the ethics to stand on their own. The interventions that measurably shift entrenched attitudes look almost nothing like debate. In one well-known field experiment, canvassers who spent about ten minutes on a doorstep asking open questions, listening to the answers and declining to argue produced reductions in prejudice that were still measurable three months later. _Motivational interviewing_, developed in addiction treatment, rests on the same finding from the clinical side: argue for change and the other person will argue against it, so the practitioner’s job is to elicit their reasons rather than supply their own.

Two honest caveats. The first is that NVC’s own language can be weaponised — ‘I feel hurt when you dismiss my needs’ is a smuggled accusation in a therapeutic accent, and a certain kind of person has learned to run the whole protocol as a pressure technique. The form isn’t the substance. The second is that the demand test is genuinely hard, and I fail it regularly. Most people discover they were making a demand only at the moment they get a no and feel something rise.

### Living With It

I’ve stopped trying to eliminate this impulse in myself, because I don’t think it’s eliminable and I’m not sure I’d want it gone. The urge to share what you’ve seen is downstream of caring about both truth and people. Suppressing it entirely would mean caring less about one or the other.

What seems achievable is noticing. When I feel the pull to convince, I try to ask what I’m actually after. Do I need this person to be better off, or do I seek an ally? Would I be satisfied if they took my idea seriously, considered it properly and still declined? If the honest answer to that last one is no, then what I’m doing isn’t generosity. It’s recruitment, wearing generosity’s clothes.

So the title overstates it, and deliberately. Evangelism isn’t violence by definition; it’s violence by default, because the four pillars are always to hand and they always feel justified when you’re the one who can see the fire. What makes it nonviolent isn’t sincerity or warmth or good intentions, all of which the worst offenders have in abundance. It’s the willingness to lose.

The best evangelists I’ve known were unnervingly relaxed about this. They’d tell you the thing once, clearly, with real warmth, and then they’d let it sit. No follow-up campaign. And it worked far more often than the hard sell ever did — partly because they were pleasant to be around, and partly because their calm was itself a piece of evidence. Someone who doesn’t need you to agree looks a lot like someone who isn’t worried about being wrong.

My colleague, incidentally, still keeps his notes. He barely mentions them now. Last month I asked him to show me how it works, unprompted, which is the only way any of this ever really happens.

### Further Reading

Brehm, J. W. (1966). _A theory of psychological reactance_. Academic Press.

Broockman, D., & Kalla, J. (2016). Durably reducing transphobia: A field experiment on door-to-door canvassing. _Science, 352_(6282), 220–224. [https://doi.org/10.1126/science.aad9713](https://doi.org/10.1126/science.aad9713)

Brown, B. (2006). Shame resilience theory: A grounded theory study on women and shame. _Families in Society, 87_(1), 43–52. [https://doi.org/10.1606/1044-3894.3483](https://doi.org/10.1606/1044-3894.3483)

Brown, B. (2012). _Daring greatly: How the courage to be vulnerable transforms the way we live, love, parent, and lead_. Gotham Books.

Festinger, L., Riecken, H. W., & Schachter, S. (1956). _When prophecy fails: A social and psychological study of a modern group that predicted the destruction of the world_. University of Minnesota Press.

Forward, S., & Frazier, D. (1997). _Emotional blackmail: When the people in your life use fear, obligation, and guilt to manipulate you_. HarperCollins.

Gal, D., & Rucker, D. D. (2010). When in doubt, shout! Paradoxical influences of doubt on proselytizing. _Psychological Science, 21_(11), 1701–1707. [https://doi.org/10.1177/0956797610385953](https://doi.org/10.1177/0956797610385953)

Miller, W. R., & Rollnick, S. (2023). _Motivational interviewing: Helping people change and grow_ (4th ed.). Guilford Press.

Rosenberg, M. B. (2015). _Nonviolent communication: A language of life_(3rd ed.). PuddleDancer Press.

Rozenblit, L., & Keil, F. (2002). The misunderstood limits of folk science: An illusion of explanatory depth. _Cognitive Science, 26_(5), 521–562. [https://doi.org/10.1207/s15516709cog2605_1](https://doi.org/10.1207/s15516709cog2605_1)

Tangney, J. P., & Dearing, R. L. (2002). _Shame and guilt_. Guilford Press.