Benjamin P Taylor antlerboyPublic work library ↗

A practical guide · Benjamin P Taylor

Choose an intervention

Change what people can actually do together. Try a bounded move, notice its effects, and adjust the rhythm.

Match the move to the pattern

If a needed dynamic is weak, create conditions for more of it. If it’s dominating, reduce what keeps amplifying it and consider a balancing dynamic. Don’t assume that adding another meeting changes the experience: the meeting may reproduce the same separation, passivity, or conformity.

A useful experiment says what will change, who has the authority and resources to support it, what you expect to see, and when you’ll review it. Include the people who will bear the consequences.

Practices to strengthen each dynamic
DynamicPractices to tryWhat to watch
SegmentProtected specialist work; communities of practice; clearer responsibilities; skills development; an expertise directory; purpose-based teams.Specialisation becomes territory. Local targets make helping another team costly.
BlendReciprocal visits; shared meals; demonstrations of a day’s work; learning exchanges; short joint projects; a marketplace of team capabilities.A cross-functional team can import every old dispute. Make the work and dependencies discussable.
EmpowerA real delegated decision with resources; individual coaching; protected experiments; personal feedback; development aligned with useful work.Permission is withdrawn on first use, or individual recognition rewards damage to colleagues.
HarmoniseAgree purpose and priorities; coordinate commitments; connect individual contributions to outcomes; review dependencies; celebrate contributions to the whole.Agreement is performed rather than tested. Mission language becomes a reason to silence dissent.

Direct, vertical, and diagonal moves

Use this arrangement to follow the relationships: segment and blend above; empower and harmonise below. The examples develop the relationships in the source material; they’re hypotheses to test in a particular situation.

Segment ↔ blend

Direct. Teams with rich expertise but little contact may need shared work and learning across boundaries. Where everyone consults everyone and expertise has become thin, protect specialist responsibilities and development.

Empower ↔ harmonise

Direct. If individual action is pulling the work apart, agree shared priorities and commitments. If conformity is suppressing judgement, give people room to disagree, decide, and try.

Segment ↔ empower

Vertical. A team’s unity can suppress its own members’ judgement. Give individuals permission to speak beyond the team line. Conversely, connect isolated individuals through a specialist practice in which they can learn and develop.

Blend ↔ harmonise

Vertical. People can enjoy exchanging ideas without acting together. Connect the exchange to commitments and purpose. Conversely, a grand common purpose may need ordinary encounters where people learn what others actually do.

Segment ↔ harmonise

Diagonal. Separate teams may need a compelling shared undertaking and explicit mutual commitments. A mission dominated by one undifferentiated view may need distinct groups to develop and articulate different capabilities.

Blend ↔ empower

Diagonal. Endless consultation may need a named person authorised to decide. An isolated individual may need relationships across groups to use their discretion intelligently.

The StretchCon deck describes diagonal effects as less strong than the direct balancing relationship. ‘Vertical’ and ‘diagonal’ name positions in this picture, rather than fixed causal laws or a route that must be followed.

Dose, timing, and rebound

Try the smallest change that can genuinely alter the pattern. The source presentation calls one such move ‘homeopathic’: a small, symbolically potent intervention. That’s a metaphor for dose and meaning. A gesture still needs credible permission, time, resources, and consequences behind it.

For a highly conformist meeting, one protected dissenting contribution may change the conversation. For a team whose every decision needs external approval, an invitation to ‘speak up’ is unlikely to help unless approval rules change.

Check for four possible effects:

  • Rebound: pressure towards agreement produces withdrawal or rebellion.
  • Squeeze-out: a new cross-team routine consumes the time needed for specialist work.
  • Displacement: apparent harmony in one meeting pushes disagreement into private conversations.
  • Different doses for different people: what frees one person may expose another to risk or add an obligation.

Keep the review close enough to notice these effects. Intensifying the same treatment because it hasn’t worked can be how the problem grows.

Enantiodromia in practice

Too much of a thing can produce its opposite. An over-unified team can become aggressively separate from everyone else. Unbounded individual freedom can create enough confusion that people demand tight central control. Forced collaboration can make people defend their boundaries harder.

These are possible sequences, not automatic outcomes. Ask what feedback is amplifying the pattern, where people are protecting something worth keeping, and what a change would threaten. Preserve the useful contribution of the dynamic you’re trying to rebalance.

A repeatable experiment

  1. Name the pattern: ‘We’re seeing X, in these events, with these consequences.’
  2. Choose one change: ‘For the next two work cycles, these people will do Y, with this authority and this support.’
  3. Protect its counterpart: ‘While we strengthen Y, we’ll preserve Z.’
  4. Look for effects: ‘We expect to see this change in decisions, handovers, learning, or experience.’
  5. Set a review and a stop condition: ‘We’ll review on this date; we’ll change or stop sooner if this happens.’

For example, two specialist groups might review one real handover together every fortnight. Preserve time for each group’s own craft. Look for fewer unresolved handovers and clearer mutual requests. Stop the routine becoming a second approval gate.

This is an illustrative design. The source notes distinguish illustrations from documented outcomes.