Should a personal principle say what you'll do or what you won't do?

Quick answer: default to what you will do, because that is the version you can actually act on in the moment it matters — but don't force every principle into that shape. A few of them are genuinely about a line you don't cross, and rephrasing those as a positive action just makes the line blurrier.

Why the wording changes what happens in the moment

"I don't raise my voice at my kids" and "when I feel my voice starting to rise, I lower it and finish the sentence" are aimed at the same failure, but they hand you different things in the ten seconds when it counts. The first tells you what already went wrong. The second tells you what to do right now, before it does. Under pressure — irritation climbing, a deadline that just moved, someone at the door with an ask you don't want to give — there is rarely spare capacity to invent a good move on the spot. A principle phrased as an action is a move you already decided on, standing ready before you need it. A principle phrased as a prohibition is a line you find out you crossed only after the fact.

Where "I won't" is actually the more honest version

Not every principle is about substitution. Some are about a line, full stop, and forcing those into a positive frame produces something mushier than the original. "I don't lie to make a problem sound smaller than it is" doesn't need a replacement action bolted on — the point is the line itself, and rewriting it as "I speak accurately" softens exactly the part that made it useful, which is that it rules something specific out. Integrity-shaped principles — honesty, fidelity, not taking credit for someone else's work — tend to live here. Their value is that they foreclose an option, and a rewritten action version can end up quietly leaving a door open that the original closed.

The problem "I won't" principles run into at check-in

The real trouble with a pure prohibition shows up later, at review time, not while you're writing it. "I don't gossip" is true almost every day, for the boring reason that most days include no opportunity to gossip in the first place — there's nothing to point to and say that's what living the principle looked like today. A behavior standard needs something to observe, and the absence of a thing is a poor thing to observe: it can't tell honoring the principle apart from the situation never coming up. That's why a check-in built entirely from "don't" statements tends to drift into a rubber stamp, reading as a pass whether or not the principle was actually tested that day.

A test for the wording

Ask, of any draft principle: if I were in the exact moment this is about, what would I actually do differently? If a concrete answer comes immediately — pause, ask a question, put the phone down, say the number out loud — write the principle as that action; it's both more usable in the moment and easier to check honestly later. If the honest answer is closer to "nothing, I would just not do the bad thing," the principle may genuinely belong in "don't" form, and the more useful move is to make the line itself sharper and more specific rather than inventing a substitute action that isn't really what you mean.

Most principles want both

The two framings aren't actually competing for the same sentence. A principle can carry a short "don't" that states the line plainly, and separate behavior standards written as the actions that show whether the line held. "I don't let money decide who I stay in touch with" is the line; "I answer a friend who reaches out within the week, regardless of what I might owe them or they might owe me" is the observable behavior underneath it, and that's what actually goes into a check-in. The prohibition tells you what you believe. The action — and eventually the record — tells you whether you lived it.

This is roughly how CreedOS is built: a creed carries the principle statement in whatever words fit the belief, and separately, two or three behavior standards, always written as something that either happened today or didn't, which is what the daily check-in and the review actually run on. Six template categories give you a starting shape to edit, or AI can turn a rough thought into a first draft you rewrite from there. It's free, with no in-app purchases.