The language of managing your own claims
Eight things people say in this kind of discussion. Match each to the move it makes.
Two people doing it properly
Insulation, heat pumps, and what neither of them knows
A: I would put the money into insulation rather than heat pumps, and I want to be clear about what I am and am not claiming. The direction I am confident about. The size I am not.
B: How do you know the direction?
A: Two things. The retrofit trials in the north of England reported bill reductions in the region of a quarter, and those are measured rather than modelled. The heat-pump figures I have seen are almost all modelled — which is not a criticism, it is a different kind of evidence.
B: That is fair, and I would add a scope condition you have not. Those trials were on solid-wall properties, which are the worst-performing stock in the country. In a housing mix like ours the effect would probably be smaller.
A: Accepted. So my claim is narrower than I made it sound: in areas with a high proportion of pre-1919 stock, insulation appears to deliver more per pound over the first five years. And I am taking the five years on trust, incidentally — I have not read the underlying analysis.
B: Then here is where we actually differ. I think the five-year frame is the whole argument, because a heat pump lasts fifteen years and insulation lasts fifty. Neither of us has a number for what that does to the comparison.
A: No. And that is a better disagreement than the one we started with.
Three questions.
1. What does A split at the very start?
2. What does B add that A had not?
3. What is I am taking the five years on trust doing?
The four moves
1 · Split your own confidence before anybody asks. The direction I am confident about; the size I am not. Fact and mechanism, direction and magnitude, now and later — split whichever pair applies.
2 · Name the source and its kind. Measured or modelled. One study or a review. Your reading or your inference.
3 · Accept a scope condition and narrow out loud. Accepted. So my claim is narrower than I made it sound: …
4 · Mark what you are taking on trust. Every argument has an unchecked link. Naming yours is what makes the rest of the chain believable.
Six lines of one argument, out of order. Put them into the order that builds.
- The evidence I am relying on is the 2022 evaluation, which is measured rather than modelled.
- So the honest version is narrower: on this evidence, frequency appears to buy more ridership per pound here.
- I would spend the money on bus frequency rather than on new cycle lanes.
- I am taking on trust that the same elasticity holds at fifteen minutes, which the report does not test.
- I am confident about that for this city and I would not defend it as a general rule.
- It found that a ten-minute frequency raised ridership by roughly a fifth within two years.
Two minutes of assembly
One word per gap. Eight lines you will need in the next twenty-six minutes.
How do you that? I am not challenging it, I want the source.
I am confident about the direction, not the .
Those figures are measured rather than , which is why I am leaning on them.
That is my inference, not the ’s.
I am taking that on — I have not read the underlying analysis.
Accepted; so my claim is than I made it sound.
That was true in 2019; I do not know now.
Cities of this kind to show a rebound in the third year.
Three questions.
1. What do you say before anybody asks?
2. Somebody offers you a scope condition you had missed. What do you do?
3. What does naming your unchecked link do?
The policy discussion
How the twenty-six minutes run. Five alone to prepare, fifteen in fours to argue under the challenge rule, six to write the claim register. One rule holds throughout: anybody may say how do you know that? at any point, and it is answered in one sentence or the claim is withdrawn.
Stage 1 · Alone · 5 minutes · Prepare
One spending decision your city, region or institution could actually make: insulation against heat pumps, buses against cycle lanes, food waste against packaging, trees against flood defences, meters against tariffs.
Four lines. Line two is the one that will be challenged first.
- 1 · The proposal, in one sentence, with the scope attached. In a city with this housing stock…
- 2 · The evidence: what kind, from where, measured or modelled, one study or several.
- 3 · The split: what you are confident about, and what you are not.
- 4 · The link you are taking on trust. Write it down now — it is much harder to admit under pressure.
Stage 2 · Fours · 15 minutes · Argue it
Two proposals per four, argued in turn, with three challengers and a running rule.
Seven minutes each. Ninety seconds of that is the proposer’s opening.
- Proposer (90 sec): the proposal with its scope, the evidence and its kind, the confidence split, and the unchecked link.
- Challengers, in this order: the source (how do you know that?), the magnitude (where does that number come from?), then the scope (does that hold here?).
- The rule: any claim challenged must be sourced in one sentence or withdrawn on the spot. Withdrawing costs nothing; defending something you cannot source costs the rest of the argument.
- Offer scope conditions to the other side. Finding the limit of somebody else’s claim for them is the strongest move available, and it is not a concession.
Stage 3 · Fours · 6 minutes · The claim register
Write down the claims that survived — properly hedged, in full sentences, agreed by all four.
Four claims, three minutes each proposal. Every slot that matters, filled.
- Two claims per proposal, written out so that every load-bearing slot is filled: scope, source, magnitude, and probability where a recommendation follows.
- Every person in the four has to accept the wording. A claim somebody will not sign is not written down.
- Underneath each, one line: what would change our minds about this?
- Finally, one sentence naming the thing nobody in the four could source. That sentence is the most honest output of the whole discussion.
What survived the question
Two questions.
1. How many of your claims did you have to withdraw?
2. What did offering somebody else a scope condition do?
Source it, or withdraw it
Two questions. Twenty seconds.
1. What do you say before anybody challenges you?
2. A claim you cannot source in one sentence should be…
Next. The errors this area produces in both directions — the stack that cancels, the absolute that misleads, and the hedge that has quietly changed the subject.