Building a values framework your movement can stand on

1 Evidence does not remove the need for a foundation. It isolates it.

There is a hope in circulation, and I have sympathy with it, that a movement committed to following the evidence has escaped the need for a philosophy. Get the analysis right, put the hard questions to a citizens’ assembly chosen by lot, and the ideological argument starts to look like something the previous generation had to have.

I do not think that survives contact with a real decision. Facts tell you what is the case and what follows from a course of action. They are silent on what you should want. You can know everything there is to know about a river and it will still not tell you whether to dam it, because that turns on what you care about, and caring is not a fact you can look up.

So what a superb analysis actually does is strip out every part of a decision that was a knowledge problem, and leave standing, alone and floodlit, the one part that never was. Evidence does not colonise the values question. It isolates it. Once you have removed the disagreements that were really disagreements about facts, everything remaining is a question about what to prefer when two good things cannot both be had — and that is exactly where evidence has nothing to say.

So: a coherent philosophical foundation is not what a movement builds instead of evidence-led policy. It is what evidence-led policy leaves you holding.

And this is where I would put the opportunity rather than the warning, because the new methods are not a threat to a foundation. They are the first serious chance in a long time to exercise one.

A citizens’ assembly is a room built for precisely the kind of decision that has no factual answer. Chosen by lot, it has no career riding on the outcome. Given time and evidence and each other, its members can hold a genuine collision open long enough to decide it rather than duck it. That is a capability representative politics has been largely unable to offer, because a chamber whose members face re-election cannot easily say out loud that a choice cost something.

And the evidence is usually plural too

There is a further thing here that gets skated over, and it matters more than the values point because it hides better.

“Follow the evidence” assumes the evidence speaks with one voice. On the questions a movement actually has to decide, it frequently does not. Competent scientists disagree. Models built by serious people diverge, and diverge more the further out they run. Disciplines look at the same problem and see different problems — an epidemiologist, an economist and a child psychologist assessing the same school closure are not producing rival answers to one question, they are answering three different questions and only one of them was asked.

So the plural framework is not something you apply to your values and then switch off when the experts arrive. It is the same discipline, applied twice. Averaging competing expert reasoning into a single consensus figure does exactly what averaging values does: it produces a number nobody in the room would defend, and it conceals a choice inside something that looks like a finding. “The science says” can be a fair summary of settled work. It can also be the most effective way yet devised to make a decision without admitting one was made.

Two things follow for a movement that means this seriously.

1. Where experts disagree, publish the disagreement rather than the midpoint. Say which expert thinks what, on what assumptions, and what would have to be true for each to be right. A range with its reasoning attached is more useful to a decision than a point estimate with its reasoning removed, and it is very much harder to weaponise later.
2. Check whether the disagreement is actually about values. A surprising amount of apparent expert conflict is not empirical at all. It is a disagreement about what to optimise, or how much risk is tolerable, or how heavily to weigh people who are not born yet — questions that arrive dressed as technical parameters and are nothing of the sort. A discount rate is a moral position wearing a decimal point. When you find one of these, it moves out of the evidence section and into the values framework, where it can be argued about by the people entitled to argue about it. Learning to spot them is, in my view, one of the more valuable things a movement can teach itself to do.

2 First, work out what you actually mean by a value

Before building anything, run a diagnostic on what you have. The word value quietly does three different jobs, and the muddle between them is what lets a values statement drift into decoration.

• A leaning What you are inclined to favour. This is what sits on almost every values wall you will ever read, and it is also the only one of the three a machine can have.
• A handed-down rule What your people taught you and expect you to pass on. This appears rarely.
• A test of what is rightful What makes a decision legitimate rather than merely popular. This appears almost never, and it is the one that does the work.

The nouns on nearly every published values statement are the first kind. And a leaning forbids nothing. You can lean towards fairness and still do more or less anything.

Which gives you the test to run on every candidate value, before it goes anywhere near a page:

Name one decision this value would have forbidden.

A real value is a thing you can bump into. It rules something out. It costs you something you would otherwise have taken. If there is no situation in which it says no, it was never a value; it was a mood, and putting it on a wall does not change that.

3 Where legitimacy comes from

Two things are easy to confuse and the whole argument turns on keeping them apart.

Capacity is the ability to decide well — to analyse, model, forecast, reason. Legitimacy is the rightful standing to be the one who decides. These come apart completely. Something can run away with capacity and gain no legitimacy at all, because legitimacy is not a skill you can be out-performed at. It is a standing, and it sits with people, with tikanga, and with the room that has to live with the result.

That matters for a values framework because it tells you what your framework is for. It is not a demonstration that your movement reasons well. It is the standard against which your decisions can be judged, held by the people who will live with them.

And here is the part that decides how you write the thing:

A decision earns the word legitimate when it honoured what the group said mattered — which means the group has to have said it, out loud, written down, and settled before the deciding began.

Three consequences follow, and they are the load-bearing rules of the whole exercise.

1. The past supplies content, not authority. Where your values come from matters enormously — the record of what your movement has said and done is the only admissible source of what your values actually are. But being handed down is not what makes a value binding. Your naming it, in advance and out loud, is.
   Why this is not conservatism, and should not be written in its vocabulary

   This is worth being clear about, because it means the exercise is not conservatism and should not be written in its vocabulary. Keeping faith with the past is one good among several, and in a given decision it can lose to doing the new thing. An inherited value your movement has never named has no standing to judge anything. A new one you name and seal has full standing.

2. Draw the values out of your own words. Not from a workshop asked to name its values, which will reach for whatever is hardest to object to, but from what your movement has actually said and done: a decision you are proud of, a time you split down the middle, a line you would not cross whatever it cost. Every value on the finished list should trace back to a sentence one of you actually spoke or wrote. Hold up a mirror. Do not paint a portrait.
3. Name the standard before you decide, then seal it. A standard named after the result is not a standard, it is an excuse with better handwriting. Fixed for this decision, then opened again, out loud, to be rewritten for the next one — because movements grow and what you stand for in five years is yours to change. What you cannot do is change it quietly, halfway through, to make a loss look like a win.

4 Real decisions are plural. Say how you will manage that, not how you will average it.

This is the section most statements do not have, and its absence is the clearest sign that a document was written to be admired rather than used.

Your values will collide. Not occasionally and not because you chose them badly — because the world is like that. Liberty and equality both matter and they pull against each other. Mercy and justice both matter. Feeding a district and leaving the river alone both matter. There is no master value sitting above the others that ranks them once and for all, and no amount of better reasoning dissolves the clash, because the clash was never a mistake in the reasoning.

The temptation at this point is to average. Take two goods that genuinely collide, split the difference, and publish a position that nobody in the room actually holds. It reads as maturity. It is the opposite: it is a movement embarrassed by its own tensions, ironing them flat and calling the result a consensus.

What you publish should be a truthful map of what you hold and exactly where it strains, kept whole.

Not a clean agreed list. A document that smooths the strain into a tidy slogan is lying on your behalf, and a decision measured against a lie only looks legitimate.

Managing a collision, rather than averaging it, means five things in practice.

1. Name the collisions in advance. Your framework should say, in its own text, where your values pull against each other. This is uncomfortable and it is the most credible thing in the document, because everybody knows the tensions are there and only a movement confident in its own position will write them down.
2. Decide case by case, against a standard named first. Not by a standing formula that ranks your values permanently — you have already said no such ranking exists — but by deciding the actual case, with the standard fixed before the deciding began.
3. Record what was set aside. When you choose, something real is given up, and nothing about the procedure repays it. Say what lost. A movement that records the cost of its own decisions is telling the truth about what governing is like, and the alternative — pretending the right choice was free — is the thing that makes political communication unbelievable.
4. Carry dissent up intact. Attributed, verbatim, not summarised into a paragraph that says the group considered various views. A minority position that survives as itself is evidence that the process was real. One that has been blended away is evidence of nothing.
5. Judge in parts and never as one grade. Was the room fairly made up, was the reasoning any good, can the outcome be answered for, were the voices that should have been present but were not at least named. Each of those is a separate judgement, pinned to something in the record. Rolling them into a single mark would be exactly the move the framework exists to prevent: someone other than your members deciding, in one number, whether you were fair.

5 The structure to publish

Now the artefact. What follows is the shape we built and run, with the reason for each part attached, because a movement that copies the shape without the reasons will drop whichever parts are inconvenient — and the inconvenient parts are the ones doing the work.

Two kinds of claim, kept apart. Upstream sit the values: where they come from, and what your movement takes them to mean. Downstream sit the commitments: the specific things you will and will not do. Readers can tell when the two have been merged, even if they could not say what tipped them off, and merging them is how a document ends up unfalsifiable at the top and trivial at the bottom.

A chain, not a list. Values presented as a list invite the reader to shop among them, agreeing with three and passing over the rest. Presented as a chain, where each follows from the one before, they invite the reader to follow the reasoning and identify the exact point at which they get off.

Four moves per value, in the same order every time.

• Inherited Whose idea this is and what they actually said. Not a quotation chosen for its ring, but the substance, stated well enough that somebody who knows the material would recognise it.
• Interpreted What the idea means in the circumstances you are actually working in. This is the only original move in the sequence and it carries the whole document. Restating a thinker is undergraduate work and everybody can see it. What is worth reading is an account of what the idea demanded of you when you had to make a real decision, and what you decided as a result.
• Committed What you therefore owe people, stated so it could be broken, and so an ordinary person could tell whether it had been.
• Enforced What actually makes the commitment true. A rule, a constitutional provision, an independent body, a constraint that would be costly and embarrassing to abandon.

Grade the last move, and show the grade. Every commitment gets marked with what currently backs it: enforced by a named mechanism, partly enforced with the gap described, or nothing yet. Publish the grade beside the commitment rather than burying it.

And attach an intention to every gap. This is not optional and it is where good intentions do damage.

A cost line on each value. What holding this actually costs your movement. Not a caveat — a cost. If you cannot name one, you are probably looking at a leaning.

Presentation: the argument stays visible, the depth opens inside it. Every step of the chain is on the page at once and reads top to bottom as reasoning. The detail — where the value comes from, what you take it to mean, what backs the commitment, what it costs — opens inside each step for a reader who wants it. What you must not do is collapse the steps themselves behind headings, because a wall of closed headings reads as a list of questions and the argument disappears. And every depth must be true on its own: a reader who opens nothing should leave with something accurate they can repeat, and a reader who opens everything should find it deepened rather than corrected.

6 The process to build it

1. Start from your policies, not from a blank page. A room asked to name its values will produce the same short list as everybody else. Instead, take the positions you already hold and ask what somebody would have to believe for that particular set to hang together. Work backwards to the belief, then forwards again to see whether the belief predicts the programme.
2. Expect to find that some positions do not fit. Two or three will sit outside whatever pattern the others make. That is information, not failure. Either the pattern is wrong or those positions arrived for reasons you have not examined, and both are worth knowing before somebody else notices.
3. Find where each value actually comes from. In your movement’s own record first — the decisions, the arguments, the lines nobody would cross. Then, if the idea belongs to a tradition of thought, say which and read enough of it that somebody who knows the material would not wince. Getting this wrong is worse than omitting it: a misattributed thinker is the first thing an informed reader catches, and everything after it is discounted.
4. Write the interpretation yourself. This is the part nobody can do for you and the part most likely to be outsourced. It should contain at least one real decision, made under the pressure the value describes, including the ones that went badly.
5. Test every commitment against what actually enforces it. Go and look. Not what you believe is in place — what is.
   What happened when we ran this on ourselves

   In our case several did not survive the check. Three claims in our own published constitution turned out to assert enforcement the code contradicted, and had to be corrected. A promise about members being able to take their records out held for one half and not the other, and now says so. That was the most useful part of the whole exercise and the part I would least want a movement to skip, because the alternative is discovering it in public.

6. Name the collisions, and say how they are managed. Per the section above. If you cannot yet say, say that, and say who will decide.
7. Then seal it, and put it where it can be checked. Fixed before the next decision, reopened out loud between decisions, never edited quietly mid-argument.

7 What this asks

More than the usual exercise and in a direction most people are not expecting.

It asks you to name a decision each value would have forbidden, and to drop the ones that cannot manage it. It asks you to say where your values came from precisely enough that somebody can check. It asks you to write down where your own values collide, and what you do then. It asks you to say what each commitment costs you rather than only what it entitles you to. And it asks you to publish the commitments nothing yet backs, alongside what you intend to do about each.

All of that is harder than writing that you believe in fairness, and I think the difficulty is the point. A document of this kind is credible in roughly the proportion that it contains things which would have been easier to leave out.