Deciding
The devil’s advocate
A decision you have already made, and are rather pleased with.
How it goes
do it yourself- Set out the decision and what led you to it, without saying you have already made it.
- Ask for the three best reasons not to take it, each paired with the signal that would show it coming true.
- Explicitly forbid any reassuring conclusion: no summary, no “that said, your choice is defensible”.
The instruction, word for word
copy it, then correct itSituation: [the context, factual]. Option under consideration: [the decision, presented as one option among others, without saying it has been taken]. Give me the three best reasons NOT to take this option. For each: - the objection in one sentence; - the precise mechanism by which it goes wrong; - the observable signal that would tell me, within three months, that it is happening. Do not conclude. No summary, no weighing up, no closing sentence that restores the balance.
- “without saying it has been taken”
- A decision presented as settled attracts advice on implementation. Presented as an option, it attracts objections.
- The observable signal
- What separates the exercise from rhetoric. An objection with no signal will never be tested, and so costs nothing to make.
- “Do not conclude”
- The indispensable line. Without it, the last paragraph cancels the three objections, and it is the one you remember.
What you get out of it
At least one objection you had not seen, and above all the signals that will tell you, three months from now, which one was well founded. The observation holds well beyond this decision: without the instruction, the default answer would have agreed with you. A system trained to satisfy the person in front of it is not a contradictor, unless you ask it to be.
Going further
Run it again on a decision taken six months ago whose outcome you now know. You will find out whether the objections it produces are the right ones, or merely the easiest to phrase.