Practical guide, IamHITL

When Participation Is Not Safe, What to Do

By Michael Polzin, Regenerative Architect, IamHITL. Published 2026-07-01.

Sometimes a participant goes quiet, or tenses, or steps out. That is a signal. This is what to do with it, without making it a scene, and without pretending to be a clinician.

Every experienced facilitator has watched a room shift under them. A participant folds inward. A voice that was in the discussion three minutes ago is now clipped. None of these are diagnoses. They are cues that this person, right now, cannot participate the way the current exercise is asking. The move is to give them an exit that costs them nothing, hold the room steady, and offer repair on their timeline, not yours.

The boundary, held up front

This is a pedagogical guide, not a clinical one. Facilitators are not clinicians unless separately licensed as such. Nothing here claims to lower distress, address a diagnosis, or change how a nervous system responds over time (Class F: if a person needs clinical care, a well-timed opt-out is not a substitute, and the correct move is to have referral information ready and name that limit out loud). The posture is consistent with public trauma-informed education literature, which treats the approach as an organizational design posture rather than a treatment protocol, notably SAMHSA's 2014 "Concept of Trauma and Guidance for a Trauma-Informed Approach" (Class E).

Read the cues before you name them

The starting move is attention. When something shifts, notice it before you interpret it. A quiet participant may be tired, sick, thinking hard, or signaling that the exercise is asking for something they cannot give right now. You do not need to know which one to do the right thing next. You need to make sure the next step has an opt-out that costs nothing.

Common cues, offered without diagnosis: a sudden change in posture, someone leaving the room during a personal-share exercise, a break in eye contact that stays broken, one-word answers from a person who was verbose earlier. Any single cue can mean nothing. A cluster in the same person over a few minutes deserves a design response, not an interrogation (Class U: this cue list is our practice synthesis, not a tested instrument).

The move: opt-out with dignity

Make the next step opt-out by default, not by request. Restructure the ask so stepping back is not a visible act. Instead of, "let us go around and each share a moment," try, "take three minutes to write privately, then anyone who wants to can share." Instead of, "who wants to add," try, "put it in the chat or hold it for later." The room does not need to know who took the exit. The person who needed it does.

If the exit is visible, taking it becomes a second cost stacked on the first. Some participants will pay it. Others will stay in an exercise that is not working for them and quietly disengage from the whole cohort. The lower the cost of the exit, the more useful the signal you get from the participants who stay. This is the same logic that makes opt-in a base primitive of the room, treated at length in Opt-In at Every Step.

What not to do

Do not single the person out to check on them in front of the group. Do not name the cue publicly. Do not stop the session and turn it into a processing circle. Do not follow them out mid-exercise. Do not backchannel through another participant. Each move converts a private signal into a public event, and the public event usually costs the participant more than the original moment did.

Offer repair on their timeline

After the session, or at the next natural break, a short and specific check-in is enough. "I noticed the exercise this morning had a rough edge. If it landed badly for you, I want to know, and if it did not, no need to reply." That is the whole message. It names that you were paying attention, it does not diagnose, it puts the next move in their hands, and it accepts silence as a valid answer. For the operational shape of that move, see Repair Loops in a Classroom.

This post is a companion to Trauma-Informed Learning, a Non-Clinical Guide for People Who Teach. That pillar covers the four moves that hold the room. This one covers what to do when those moves are not enough for one person in one moment. It is a claim about the facilitator's job, held to a boundary on purpose.

Evidence classes used in this post: C (configuration and integration: the opt-out-by-default move is a design pattern we run inside our own workshop templates), E (public trauma-informed education literature, SAMHSA 2014), F (falsifier: this is pedagogy, not therapy, and if a participant needs clinical care, a graceful exit does not substitute), U (unverified: the cue list is our practice synthesis, not a validated instrument).