Get from obstacles to outcomes with Clean Language
Someone comes to us with a problem they're stuck on.
We ask questions. They answer. We ask more questions.
Fifteen minutes later, they're describing the same problem in more detail, and nobody has moved an inch.
It's not because we weren't listening. It's usually because the questions we ask by default, the ones that feel most natural in the moment, are designed for analysis rather than movement.
"Why is this happening?" asks someone to explain the past. "What's stopping you?" asks them to map their obstacles.
"Why haven't you tried that yet?" does something worse: it embeds a mild accusation inside what sounds like a question.
These questions aren't wrong exactly, but they aim at the problem.
And when attention is aimed at the problem, that's where people stay.
What is Clean Language?
Clean Language is a set of questions developed by psychotherapist David Grove in the 1980s and 1990s.
Grove noticed that therapists kept importing their own metaphors into conversations, subtly steering clients toward the therapist's model of the problem rather than the client's own. His solution was a minimal question set that uses the speaker's exact words, adds nothing the speaker didn't say, and directs attention without leading it.
Wendy Sullivan and Judy Rees brought the approach into organisations and wrote the accessible business application in Clean Language: Revealing Metaphors and Opening Minds (2008).
The core insight translates directly to management: the person who holds the problem also holds the map of it. Our job is to help them read the map, not to draw a new one.
When to reach for it
- When someone keeps describing the same problem in different words and nothing shifts
- In 1:1s where the conversation circles but doesn't land anywhere useful
- In retrospectives that drift toward blame instead of forward motion
- When we notice we're about to offer a solution before we've actually understood the problem
- When someone says "I don't know" and we're tempted to fill the silence with our own interpretation
- In stakeholder discovery and scoping conversations, particularly in creative or ambiguous briefs: when we need to understand what someone actually wants rather than what they've asked for, Clean Language surfaces the real requirement without us projecting our own assumptions onto it
Seeing it in action
The same conversation, twice. Someone opens with: "An obstacle I'm currently navigating is my relationship with a peer. We just seem to be on completely different pages."
Without Clean Language
"Why do you think that is?"
The person thinks for a moment. "Probably because we have really different working styles. She's very process-heavy and I'm more fluid."
"Have you tried talking to her directly about it?"
"Yes, a couple of times. It doesn't really go anywhere."
"Why do you think it doesn't go anywhere?"
"I don't know. Maybe she's just not that open to feedback."
"What have you actually said to her when you've had those conversations?"
The person starts reconstructing past conversations. They're now in defence mode, rehearsing their own reasonableness. The obstacle is as large as it was at the start. And now there's a faint implication that they might be doing something wrong.
With Clean Language
"And what would you like to have happen?"
A pause. "I'd like us to be able to work through disagreements without it becoming a standoff."
"And what kind of standoff is that standoff?"
"Like... neither of us backs down. We both think we're right and neither of us moves."
"And is there anything else about neither of you moving?"
Another pause, longer. "I suppose I'm also not moving. I've been assuming it's her problem."
Something has shifted. Not because of anything we added, but because attention moved somewhere new. The person followed their own words into a different part of the map.
The first conversation analysed the past and stalled. The second moved forward without pushing. The only difference was where the questions aimed.
How to use it
The mechanics apply to every question here: use the speaker's exact words in the blanks, every time. Don't paraphrase, summarise, or interpret. Don't add anything that wasn't in what they said.
The "And" at the start of each question isn't a verbal tic. It signals continuation: we're building on what they offered, not changing direction.
Start with three questions. These are enough for most situations and the right place to begin.
"And what would you like to have happen?" is the entry point for almost any stuck conversation. It redirects attention from the obstacle toward a desired outcome. Use it early. Most conversations spend too long in problem description before this question ever gets asked.
"And what kind of ___ is that ?" uses their exact word in both gaps. If they said "standoff," ask about the standoff. This question invites the person to develop their own description without us steering it. It often surfaces something they didn't know they knew.
"And is there anything else about ___?" keeps attention in the same place and opens space for what hasn't been said yet. It's useful when a person is starting to articulate something but hasn't quite arrived at it.
When the conversation needs structure: the 15-minute FOTO.
FOTO (From Obstacles To Outcomes) is a complete conversation architecture built entirely from Clean Language questions. It's designed to move someone from naming an obstacle to a clear outcome and a path through, inside fifteen minutes.
It works in a loop. Name the obstacle, establish the desired outcome, then develop understanding of whatever has been surfaced -- moving backward to find triggers, forward to find consequences, or sideways to find relationships between things the person has already named.
Once the current obstacle has been sufficiently explored, return to the next one and repeat.
The conversation ends when there are no more obstacles, or when something has shifted enough that the person can move forward on their own.
The whole thing runs on one rule: fill every blank with their words, not yours.
Why it works
Questions direct attention before the person answers them. This is the core mechanism, and it's why the question type matters more than we tend to think.
George Lakoff and Mark Johnson's work on conceptual metaphor (Metaphors We Live By, 1980) shows that the language frame a question uses determines the cognitive terrain available for the answer. "Why is this a problem?" activates problem-space thinking. "What would you like to have happen?" activates outcome-space thinking. The person can't easily access both at once. Whichever frame the question sets, that's where they're working from when they answer.
"Why" questions carry an embedded assumption. When we ask "why haven't you tried X?" we're implying that X was available and the person chose not to. The question already contains a verdict. The person's first cognitive move is to defend against that verdict, not to think productively about the situation. This isn't a character flaw. It's a predictable response to a question that feels like a mild prosecution. Clean Language questions don't carry this load. They contain only what the person themselves put into the conversation.
The metaphor the question asks in determines the solution space. Lakoff and Johnson's research demonstrates that people don't just use metaphors to describe experience: they use them to structure it. If someone says "I'm hitting a wall," they're not being colourful. They're telling us something about the shape of the obstacle as they experience it. Asking "and what kind of wall is that wall?" keeps us working inside their structure. Asking "have you tried going around it?" imports our structure. Our metaphor may not match their actual experience at all, and if it doesn't, we're problem-solving in the wrong room.
What to watch for
The urge to paraphrase. When someone says something interesting, the instinct is to reflect it back in our own words: "so what you're saying is..." This feels collaborative. It isn't. The moment we substitute our words for theirs, we've introduced our frame. Use their words. Exactly. Every time.
Filling silence. When a Clean Language question lands well, there's often a pause. The person is actually thinking, sometimes about something they haven't thought about before. The silence means the question is working. Filling it with a follow-up, a clarification, or a reassurance interrupts the process at exactly the moment it's most productive. Wait.
Adding interpretation. "It sounds like you might be feeling..." is not a Clean Language move. Neither is "I wonder if part of what's happening is..." Both of these import our model. The discipline is to stay with what the person has actually said, not what we think might be underneath it. What's underneath it is their business. They'll find it if we give them the space.
Go deeper
- Wendy Sullivan & Judy Rees, Clean Language: Revealing Metaphors and Opening Minds (2008)
- George Lakoff & Mark Johnson, Metaphors We Live By (1980)
- Caitlin Walker, From Contempt to Curiosity (2014)