Conflict Escalation

This is the detailed reference for understanding conflict stages. For practical guidance on handling conflicts, start with Conflict Resolution. This page describes the nine stages of conflict escalation, what interventions work at each stage, and when certain interventions become actively harmful. Based on Friedrich Glasl's model of conflict escalation, adapted for the Noisebridge community.

The Nine Stages at Noisebridge

🟢 Phase 1: Win-Win (Stages 1-3)

Both parties can still walk away satisfied. The conflict is about issues, not people. Self-resolution is possible. Mediation works well here.

For practical guidance on handling Phase 1 conflicts - including scripts and when to handle things privately - see Conflict Resolution#Most Conflicts: Handle Directly.

Stage 1: Tension

What it looks like: Differences of opinion emerge. Someone's frustrated about how a shared resource is being used. There's friction, but people are still talking and assuming good faith.

Examples at Noisebridge:

• "I wish people would clean up the laser cutter after use."
• "The 3D printer queue doesn't seem fair."
• "I don't like how that meeting was run."
• "That person's project is in my way."
• "I feel like my ideas aren't being heard."
• Someone seems cold or distant to you today.
• You disagree about how something should be done.

What works: Direct conversation using Restorative Communication. A quick chat. A DM. Assume good faith - they probably don't know they bothered you. Most conflicts stay here forever and resolve naturally.

What makes it worse: Posting about it publicly. Venting to others instead of talking to the person. Stewing silently until resentment builds.

Stage 2: Debate

What it looks like: Positions harden. People start trying to "win" the argument rather than solve the problem. There's some black-and-white thinking, but it's still about the issue, not the person.

Examples at Noisebridge:

• Back-and-forth in a Discord thread that's getting heated.
• Arguments at Tuesday meetings about policy.
• Two people with different visions for how a project should go.
• Repeated disagreements about the same topic.
• "We keep having this same conversation."

What works:

• Taking a break and coming back to it later.
• Acknowledging the other person's perspective before restating yours.
• Looking for the shared goal underneath the disagreement.
• If you're stuck, asking one neutral person to help facilitate - but keeping it private, not making it a community issue.

What makes it worse: Recruiting allies (see Consensus Spoofing). Taking it to a larger audience. Making it about "winning." Framing it as the other person being unreasonable.

Stage 3: Actions Instead of Words

What it looks like: People stop talking and start acting unilaterally. Communication has broken down. There's distrust, but not yet hostility. People still care about the relationship even if they're frustrated.

Examples at Noisebridge:

• Someone reorganizes a shared space without discussion.
• Passive-aggressive signage appears.
• People start avoiding each other.
• Someone stops responding to messages.
• Unilateral decisions get made to end a stalemate.

What works:

• Recognizing that avoidance is making it worse.
• Reaching out directly: "I feel like we've been avoiding each other. Can we talk?"
• Ask To Disengage if you need space, but communicating that clearly.
• If direct conversation feels impossible, now a mediator might help - but keep it minimal and private.

What makes it worse: Continuing to avoid. Escalating to public forums. Interpreting their avoidance as hostility. Making unilateral decisions that affect them without trying to talk first.

🚨 The Escalation Trap

How Stage 1-3 conflicts become Stage 4+ conflicts:

1. You feel hurt/frustrated (Stage 1)
2. Direct conversation feels too hard/scary
3. You post in Discord or tell friends "for support" (→ you just created an audience)
4. Now there are sides (Stage 4)
5. The other person feels publicly called out
6. They respond defensively or not at all
7. Now it's about reputation, not the original issue (Stage 5)
8. What started as "they didn't clean up after themselves" is now a community conflict

This happens constantly. It's preventable. The prevention is: have the private conversation even though it's uncomfortable.

---

🟡 Phase 2: Win-Lose (Stages 4-6)

⚠️ Once conflict enters Phase 2, mediation becomes problematic. The conflict is no longer about issues - it's about people. Asking parties to "talk it out" may just provide a platform for further harm.

Stage 4: Coalitions

What it looks like: People start recruiting allies. "Sides" form. Rumors spread. The conflict becomes public and political. It's no longer about the original issue - it's about which faction wins.

At Noisebridge: Whisper campaigns. People lobbying others before meetings. "Have you heard what X did?" conversations that aren't aimed at resolution. This is where Consensus Spoofing and Policy Injection often begin.

What works: Structural intervention. Separating parties. Clear boundaries. Mediation is risky - it can be weaponized by the side acting in bad faith (see Policy Injection).

Stage 5: Loss of Face

What it looks like: Personal attacks. Public humiliation attempts. The goal becomes exposing and discrediting the other person. Trust is completely gone.

At Noisebridge: Public callouts. Attempts to get someone banned based on character attacks rather than specific behaviors. Dredging up old incidents.

What works: Mediation is harmful here. It forces the targeted party to repeatedly engage with someone attacking them. Focus on protection: Ask To Disengage, AskToLeave, clear documentation.

Stage 6: Threat Strategies

What it looks like: Ultimatums. "If you don't do X, I'll do Y." Threats of consequences - social, legal, physical. Each threat triggers counter-threats.

At Noisebridge: Threatening to quit, to sue, to "expose" people, to call authorities. Demanding others be banned or else.

What works: Do not mediate. Threats are not a communication style to be worked with - they're a behavior to be stopped. AskToLeave. Documentation. If threats are credible, involve appropriate authorities.

---

🔴 Phase 3: Lose-Lose (Stages 7-9)

🛑 In Phase 3, the only appropriate response is protective action. Mediation is not just ineffective - it's actively harmful. It delays necessary action and exposes community members to ongoing harm.

Stage 7: Limited Destruction

What it looks like: Parties accept damage to themselves if it means damaging the other more. "Winning" means making the other person lose bigger.

At Noisebridge: Sabotaging projects. Deliberately disrupting events. Breaking things to prove a point. Accepting being banned if it means causing maximum chaos first.

What works: Immediate removal. 86. Police if criminal behavior. There is nothing to mediate.

Stage 8: Destroying the Enemy

What it looks like: The goal is complete destruction of the other party - their reputation, their ability to participate, their wellbeing.

At Noisebridge: Sustained harassment campaigns. Attempting to destroy someone's career or relationships outside Noisebridge. Doxxing.

What works: Full community protection mode. 86. Legal action if appropriate. Support for targets. Mediation would be revictimization.

Stage 9: Together Into the Abyss

What it looks like: Self-destruction is acceptable if it takes the enemy down too. No concern for collateral damage.

At Noisebridge: Rare, but: destroying the organization itself to hurt specific people in it. Burning everything down.

What works: External intervention. Legal authorities. Complete separation. The community's survival takes precedence.

When Mediation Becomes Harmful

Mediation is a powerful tool - in the right circumstances. In the wrong circumstances, it causes additional harm:

🎯 The Mediation Trap

The moment you create a "two parties in good faith" container, you've already conceded the premise that both parties are acting in good faith.

Mediation becomes a weapon when one party is manipulating the frame.

A bad-faith actor exploits the mediation structure itself: they get the legitimacy of "we're working this out," the audience of a mediator who must remain neutral, and a process that treats their fabricated rules and manufactured consensus as valid "perspectives." Meanwhile, the target is trapped in a format that assumes their abuser deserves equal consideration.

The process rewards the manipulator for agreeing to participate.

Everyone operating in this bad frame gets harmed:

• The good-faith party - They showed up genuinely trying to resolve something. Instead, they're forced to engage with manipulation while a neutral third party treats it as legitimate. They leave feeling unheard, gaslit, and less trusting of community process.

• The mediator - They're required to act neutral between good faith and bad faith, which is impossible without becoming complicit. They're responsible for communicating some kind of outcome, but there's no good outcome available. If the process goes poorly - and it will - the mediator's own relationships within the community suffer. They may be seen as having "taken sides" or "failed to resolve things." Their credibility and social capital are spent on an unwinnable situation.

• Noisebridge itself - Every failed mediation erodes trust in mediation as a tool. Community members learn that "mediation" means "get trapped in a room with your abuser while someone watches." Future legitimate conflicts become harder to address because people are afraid of the process.

• Future targets - The bad-faith actor learns that agreeing to mediation buys them legitimacy, time, and a platform. They'll use this move again. The community has trained them that manipulation works.

Mediation with a bad-faith actor doesn't fail neutrally - it actively makes things worse for everyone except the manipulator.

❌ Do Not Mediate When:

• The conflict is in Phase 2 or 3 - The parties are no longer trying to solve a problem; one or both are trying to win or destroy
• There's a clear aggressor and target - Mediation treats both parties as moral equals with legitimate grievances; when one party is being harassed, this is false equivalence
• One party is acting in bad faith - Mediation requires good faith from both sides; a bad-faith actor will use mediation as a platform for further manipulation (see Policy Injection, Consensus Spoofing)
• The harmful behavior is ongoing - Mediation delays protective action; the priority is stopping the harm, not facilitating dialogue
• The target has said no - Pressuring someone to "talk it out" with the person who harmed them is revictimization
• Threats have been made - Threats end the possibility of mediation; they require protective response
• There's a significant power imbalance - Mediation can't create safety when one party has power over the other

✅ Mediation Works When:

• Both parties want resolution - Genuine willingness to find a solution using Restorative Communication
• The conflict is about issues, not people - Disagreement about how things should work, not about someone's right to exist in the space
• Good faith exists on both sides - Both parties are being honest and aren't trying to manipulate (no Policy Injection or Consensus Spoofing)
• No ongoing harm - The situation is stable enough to take time for dialogue
• Rough power parity - Neither party is significantly more vulnerable than the other

⚠️ When Assessment Itself Is Compromised

This framework only works if you can accurately assess what stage a conflict is at. Certain manipulation patterns can distort that assessment - making conflicts appear to be at different stages than they actually are, inverting who appears to be the aggressor, or weaponizing the framework itself.

Policy Injection

What it is: Someone states a fabricated rule as if it were established community policy, typically to gain advantage in a dispute. "That's our policy" when no such policy exists.

How it distorts assessment:

• Makes Stage 4-5 behavior look like Stage 1-2 - Coalition-building and public attacks get framed as "just enforcing community norms"
• Inverts aggressor and target - The person being manipulated appears to be "violating policy" while the manipulator appears to be maintaining order
• Bypasses dialogue entirely - "This is already settled" skips the Stage 1-3 conversations where resolution was possible
• Recruits the community as a weapon - "We all agree" language (see Consensus Spoofing) manufactures the appearance of Stage 5 consensus that never actually formed

The danger: Someone using Policy Injection can say "they're at Stage 5, we need to skip mediation" - using this very framework to justify skipping protective dialogue and moving straight to punishment.

Consensus Spoofing

What it is: Claiming community agreement that doesn't exist. "We all decided..." / "Everyone knows..." / "Noisebridge-ers don't do that."

How it distorts assessment:

• Makes manufactured coalitions look organic - What's actually one person recruiting allies (Stage 4) appears to be natural community convergence
• Creates false legitimacy for escalation - "The community has already weighed in" justifies skipping the inquiry phase
• Silences dissent - Anyone questioning the "consensus" appears to be opposing the whole community

Red Flags That Assessment May Be Compromised

• Someone insists a conflict is at a higher stage than the evidence suggests - especially if this justifies skipping dialogue
• One party is being characterized as the aggressor, but they're the one trying to have conversations while the other is avoiding them
• "Community consensus" is cited, but you can't find the meeting notes, wiki page, or people who actually agreed
• The person claiming to enforce norms is the primary beneficiary of those "norms"
• Requests for evidence are met with attacks rather than documentation
• The same person keeps ending up in conflicts where they're "just enforcing policy"

What To Do

• Ask "where is this written?" - Legitimate policies can be pointed to
• Ask "who decided? when?" - Legitimate consensus has a history
• Talk to multiple people independently - Does the claimed consensus actually exist?
• Look at the pattern - Who benefits from this framing? Who keeps ending up as the "policy enforcer"?
• Trust your own assessment - If something feels off about how a conflict is being characterized, investigate before accepting the framing

The framework is only as good as the information going into it. If someone is manipulating the inputs, the outputs will be wrong - and the wrong intervention will be applied.

See Policy Injection, Consensus Spoofing, and Process Weaponization for detailed analysis of these patterns and how to counter them.

What To Do Instead

When Mediation isn't appropriate, other tools exist:

• Ask To Disengage - Immediately separate parties in an escalating situation
• AskToLeave - Remove someone from the space temporarily
• Documentation - Write down what happened, when, with witnesses (see Path_to_86 for how this feeds into community decisions)
• 86 - Permanent ban for those who won't or can't stop harmful behavior
• Support the target - Focus resources on the person being harmed, not on "both sides." Use Restorative Communication with the target, not between target and aggressor.
• Structural changes - Adjust how the space works to prevent future conflicts

See Also

• Conflict Resolution - Main entry point: practical guidance for handling conflicts
• Bravespace - The #bravespace Discord channel for discussing community dynamics
• Mediation - The mediation process (when appropriate)
• Restorative Communication - Communication framework for early-stage conflicts
• Policy Injection - How fabricated rules distort conflict assessment
• Consensus Spoofing - How fabricated agreement manipulates community process
• Process Weaponization - Using legitimate processes as weapons
• Ask To Disengage - De-escalating in the moment
• AskToLeave - Temporary removal from space
• Path_to_86 - How permanent bans happen
• 86 - The ban list

References

• Friedrich Glasl's model of conflict escalation - Wikipedia
• 9 Stages of Conflict Escalation - Die Projektmanager
• Glasl's Nine Stages of Conflict Escalation - Toolshero