Conflict Resolution: Argue About Needs, Not Positions

Two people, one window, both right

In 1925 Mary Parker Follett described two people in a library. One wants the window open, the other wants it shut. They argue about the window. There is no answer — the window is either open or it isn't, and any compromise (half open) leaves both worse off than they wanted.

Then someone asks why. One wants fresh air. The other doesn't want a draught on their papers.

They open a window in the next room. Both get everything.

This isn't a story about being clever. It's about the fact that they spent the whole argument on the wrong object. The window was never the disagreement — it was one proposed solution, and they'd mistaken it for the thing itself.

Most of your arguments are windows. And you can tell, because they have a signature: you keep having the same one. Repetition is what a positional argument does — you can't resolve it, so it comes back, slightly angrier each time.

The structure: Position → Need → Options → Test

1. Position — what each of you is demanding

A position is a proposed solution stated as a demand.

"We're going to my parents' for Christmas." "You need to stop being on your phone in the evening." "I should get the bigger room."

Positions have a specific property: they're binary and they compete. One of you wins. Which is why positional arguments feel like arguments — structurally, they are.

Write both positions down. Yours and theirs, in one sentence each, no editorialising. If you can't state theirs without a sneer in it, you haven't understood it yet — you've built a version that's easy to defeat.

2. Need — the why underneath

Now the move the whole lesson turns on. For each position, ask: what does this get you?

Then ask it again of the answer. Usually two or three levels down, you hit something that isn't about the object at all.

POSITION  "We're going to my parents' for Christmas."
  why?    → "I always go."
  why?    → "My mum's on her own since Dad died."
  NEED    → she isn't alone on the one day it would be obvious.

POSITION  "Stop being on your phone in the evening."
  why?    → "You're not really here."
  why?    → "We haven't had a proper conversation in weeks."
  NEED    → some part of the day where we're actually in contact.

Look at what happened to the second one. The position was about a phone. The need is about contact — and there are a dozen ways to meet it, most of which don't involve the phone at all. The positional version had exactly one solution, and it required someone to lose.

Two rules:

• Needs are never about the object. If your stated need still mentions the window, the phone, or the room, go one level deeper.
• Do both columns. Doing only yours makes you a better advocate for your position, which is the opposite of the point.

3. Options — generate before you evaluate

With four needs on the table, generate solutions — and don't judge any of them while generating.

Evaluating as you go kills this. The first sensible-sounding option becomes the new position and you're back where you started, arguing about a different window.

"Five ideas, no vetoes yet. Then we'll go through them."

Aim for five. The first two will be the original positions restated; the useful ones start at three or four, once you've exhausted the obvious.

For the Christmas example: go for two days not five · go this year, hers next · invite her mum to you · go separately · video call on the day and visit the week after. Some are bad. Bad options are productive — they make the constraints visible.

4. Test — check options against the needs list

Now evaluate, and evaluate against the needs, not against who proposed it.

                          her need:      my need:
                          mum not alone  not lose the whole break
  two days not five            ✅              ✅
  alternate years              ❌ (this year)  ✅
  invite her to us             ✅              ✅
  go separately                ✅              ⚠️ costs something else

The winner is whatever meets both columns. Often it's an option neither of you walked in with — which is the entire return on doing it this way, and also why it defuses: nobody conceded, so nobody lost.

When nothing meets both columns, you've learned something real: this is a genuine conflict of needs, not a misunderstanding. That calls for negotiation and trade — SO-11 — rather than more conversation.

Why this works

1. Positions compete; needs often don't. Two positions on one object have one winner by construction. Two needs frequently have several solutions, and the search only starts once they're named.

2. "Why" moves the argument off the object. As long as you're both talking about the window, there's nothing to discover. The information that resolves it is one level down and neither of you has said it out loud.

3. Separating generation from evaluation stops premature lock-in. Judging as you go means the first plausible option becomes a new position — and you're arguing again, about something else.

4. Nobody has to lose. A solution that came from the needs list wasn't anyone's proposal, so accepting it isn't a defeat. That matters more than it sounds: positional wins get re-litigated, because the loser never actually agreed.