Guides
Four ways to justify a decision, how to tell which one is in the room, and where each fails
Four ways to justify a decision, how to spot which one a person is using, where each fails, and a procedure for when two give different answers.
Two people can look at the same decision, reason carefully, and reach opposite conclusions without either of them being careless or dishonest. That is usually not a failure of thinking. It is what happens when they are using different frameworks and neither has said which one.
This page is about those frameworks: what they are, how to tell which one someone is actually applying, where each one breaks, and what to do when two of them give different answers. It is the orientation page for everything else on this site about ethical reasoning.
What to take away
- Most workplace disagreements about right and wrong are disagreements about which question counts, not about the facts. Find the question before arguing the answer.
- Every framework produces a defensible answer and a blind spot. The blind spot is the reason to know more than one.
- The practical output of framework thinking is not a verdict. It is a written record of what you weighed, made while you were still uncertain.
Four ways to justify a decision
Almost every argument you will hear in a meeting is a version of one of four claims. Learning to hear which one is being made is the single most useful skill here, because it tells you what evidence would settle the argument.
Four ethical frameworks compared
Outcome
- Claim
- Best outcome overall
- Question
- Who is better or worse off?
- Evidence
- Effects on named groups
Duty
- Claim
- Not permitted to do this
- Question
- What duty or promise applies?
- Evidence
- The commitment or rule
Character
- Claim
- Not who we are
- Question
- What would a good person do?
- Evidence
- Precedent and standard
Agreement
- Claim
- Nobody could reasonably accept
- Question
- Could you justify it to the cost-bearer?
- Evidence
- Affected party would agree
None of these is the professional or grown-up one. They are four different questions, all legitimate. Trouble starts when one person is answering the first and another is answering the second, and both think the other is being obtuse.
A worked illustration of the difference. Suppose a supplier's contract permits you to extend payment terms unilaterally. The outcome argument asks what the delay does to your cash position and to theirs, and whether the supplier can absorb it. The duty argument asks what you led them to expect when you signed.
The character argument asks whether you would use that clause against a supplier who could fight back. The agreement argument asks whether the supplier would have signed had they known you intended to use it this way. Four different conversations, and the answer changes depending on which one you have.
Working out which framework is in the room
You will rarely hear anyone name a framework. You will hear the reasoning style, and you can identify it in about a minute.
Identify the framework in a minute
- Ask what would change their mind
- Ask who they are counting
- Ask what happens if everybody does it
- Listen for the passive voice
Working out which framework
- Ask what would change their mind.An outcome thinker will name a number or a magnitude. A duty thinker will say nothing about consequences would change it, which tells you the argument is not about size. That answer sounds stubborn and is actually informative.
- Ask who they are counting.Outcome arguments quietly set a boundary around whose welfare is being added up. The boundary is where most of the disagreement hides. Ask it out loud: are we counting the contractor, the neighboring team, the customer's customer?
- Ask what happens if everybody does it.This flushes out the difference between a genuine exception and a rule someone has just decided not to follow this once.
- Listen for the passive voice.When the reasoning shifts to what will be done rather than who will do it, you are usually watching a decision that nobody wants to own. That is a governance signal, not a philosophical one, and it is worth acting on immediately.
Where each one fails
Knowing the failure mode matters more than knowing the name, because the failure is what you will actually meet.
Outcome reasoning fails on distribution and on measurement. It is comfortable with a large benefit spread across many people paid for by a severe cost to a few, which is exactly the arrangement most people find unacceptable when they can see who the few are. It also drifts toward whatever is countable, so the harms that come without a unit of measure quietly drop out of the sum.
Duty reasoning fails when duties collide. It gives you no procedure at all for the case where honoring one promise breaks another, which is the case you are most likely to face. It also invites a game of finding the technically compliant path, where every rule is met and the result is still indefensible.
Character reasoning fails under disagreement. It works well when everyone shares a standard and gives you nothing when they do not, because it has no way to arbitrate between two people who both think they are acting well. It is also unusually easy to fool yourself with, since it runs on self-description.
Agreement reasoning fails on absent parties. It asks what someone could reasonably accept, and quietly assumes they had a real choice. Where there is a large power difference, consent is cheap and the framework mistakes resignation for agreement.
When two frameworks give different answers
This is the situation the reasoning is for. A procedure beats a principle here, because the principle is exactly what is in dispute.
- State both conclusions in one sentence each, without the argument attached. If the two sentences turn out to be compatible, you had a communication problem rather than an ethical one, and you have just saved an afternoon.
- Identify who bears the cost under each option, by name or by group.Write it down. Options whose costs land entirely on people not in the room get chosen far more often than they should.
- Ask whether the disagreement is about facts.A surprising share of these collapse when someone actually checks a number, a contract term, or what was said to whom. Do this before the philosophy.
- Check reversibility.Where one option can be undone and the other cannot, and the facts are still thin, the recoverable one deserves the benefit of the doubt. This is not a tiebreaker for every case, but it settles more than you would expect.
- Decide who decides.Consider whether the choice exceeds your authority, and whether a reasonable observer would think your interests compromise your judgment. A decision that should have gone up and did not is a failure regardless of how it turned out.
- Write the reasoning down before you know the outcome.This is the part everyone skips and the only part that has value later. A record made after the result is known is a justification. A record made under uncertainty is the thing that helps whoever faces the next case.
If a decision touches employment, competition, anti corruption, personal data, product safety, financial reporting or sanctions, the frameworks tell you what is at stake but not what is permitted. Those areas are legally governed, the rules differ by country and by sector, and the reasoning above should route the question to a qualified adviser before anyone decides, not after.
A case run through all four
Hypothetical, and constructed to show where the conflict sits rather than to describe anything real.
Your product has a fault that appears rarely and is unlikely to injure anyone. Disclosing it now costs money and market position. A competitor almost certainly has the same fault and is saying nothing.
A case run through
- Outcome.Depends entirely on the numbers you have not got: how rare is rare, how bad is the harm, how many units are in the field. Notice that this framing rewards you for not gathering the numbers, which is why the framework needs a decision rule attached in advance.
- Duty.What did you tell buyers about this product, in writing and in your marketing? If a safety claim was made, the conflict may already be settled and the argument about magnitude is beside the point.
- Character.Would you want your own family using the affected units without being told? The question is glib on its own and clarifying when it contradicts your other answers.
- Agreement.Would a customer, knowing what you know, accept the decision not to tell them? If the honest answer is that the decision depends on them not knowing, you have your finding.
The competitor's silence appears in none of these, which is the point of doing it this way.
What framework thinking will not do
It will not beat an incentive. If a target can only be met by doing the thing your reasoning prohibits, the reasoning loses and the people who follow it get penalized for it. Fix the target.
It will not substitute for a control. Where one person can authorize, execute and check the same transaction, that is a process defect and no amount of good reasoning by that person repairs it.
It will not tell you what is legal. Those are separate questions that overlap confusingly, and the overlap is where people get hurt.
And it will not make the decision painless. Most of the hard cases are conflicts between real obligations where every option breaks something. What the frameworks buy you is seeing the conflict clearly, in time, with the reasoning on record.
Bottom line
The value of knowing four frameworks is not that you get a better answer. It is that you stop mistaking a difference in question for a difference in integrity, and you notice the option whose cost falls on somebody who is not in the room. The four are set out compactly in the Markkula Center's introduction to thinking ethically.
Common questions
Do I need to pick one framework and stick to it?
Is this not just a way of dressing up the answer I already wanted?
Who should be in the room?
How long should the record be?
When does this need a lawyer rather than a framework?
In this guide
- Four lenses, one case: comparing ethical frameworks examplesEthical frameworks examples chosen for divergence: three hypothetical cases where four ways of reasoning disagree, and the failure mode to watch for.
- Ethical Frameworks Framework: When to Escalate, Who DecidesSet the escalation thresholds of an ethical framework for workplace requests: when a case goes up, who decides at each level, and what to record.
- Reading ethical frameworks metrics without fooling yourselfEthical frameworks metrics read in pairs: why every figure here has two opposite readings, four ways measurement goes wrong, and what to show a board.
- Which ethical frameworks tools actually force the cost onto the page?Six named tools that force decision costs onto the page, what each one catches and misses, and what to ask before buying compliance software.
- Ethical Frameworks Questions for Reviewing a Decision RecordUse ethical frameworks questions to review a written decision record before sign-off: connect consequences, duties, fairness, and dissent to evidence.
- Ethical frameworks in 2027: five shifts changing how decisions get madeEthical frameworks trends 2027: five shifts — configured decisions, vanishing internal documents, speed, supply chains, and process expected.
- Ethical frameworks case study: reasoning built to fit a decision already madeAn invented pricing case run end to end: how the reasoning was built to fit a decision already made, what the delay cost, and the changes that followed.
