Sanction Rules vs. Signal Rules (Part 2)
Safety + (Must > Should)
In Part 1, I explained an important distinction between two types of rules.
Sanction rules protect what must not be broken—violations damage trust or the system itself, therefore consequences are serious because they defend essential boundaries.
Signal rules also define boundaries, but like the lines of a sports field, they are boundaries that will predictably be stepped over. They intervene and redirect to keep things moving.
Put simply: sanction rules guard integrity—signal rules manage flow—even though both of them are “rules” (i.e. non-optional requirements). Here are the two most useful applications.
1. Ask, “Is This a Safety Issue?”
Sanction rules are best used for issues or concerns that are safety-related or significant and hard-to-reverse. You don’t want certain things to happen, yes, but the key thing is that they would be impossible to repair. A glass heirloom that gets smashed is gone forever. And because some things cannot be made whole, we understandably feel the need to punish violations because the primary goal of sanction rules is disincentivizing certain behavior or outcomes.
So, sanction rules are obviously important, but they are necessarily rigid and heavy-handed. They do not allow discretion or flexibility. If a system has too many of them it quickly chokes under the weight of its own structure. As a result, people will feel frozen and scared to move. These obviously unnecessary constraints inevitably lead to resentment if not actual rebellion (sanction rules really ally emphasize the “force” in “enforcement”).
This means sanction rules as necessary and powerful, but only in small numbers. This is why the question, “Is this a safety issue?” is so important. It gives us clarity on which type of rule or expectation you’re using and therefore what kind of communication and enforcement is appropriate.
Example: Getting My Kid Into the Car
This might be a silly example, but here’s a short story of how my wife and I used the safety question with my 6-year old son. We were trying to get him into the car to go get some professional family photos taken, and he wasn’t being cooperative. Essentially, he didn’t want to go. He wasn’t refusing exactly, but he was having a meltdown which had the same effect.
Anyways, my wonderful wife felt the strong urge to get the show on the road as I was in and out of the garage trying to gather up what we needed. As her stress-level increased, she started down the path of trying to physically force our son into the car—a regrettable but understandable tactic.
But after a few failed attempts (he’s a BIG kid so not easy to wrangle), she paused then took a deep breath. Then she then caught my attention, as I was only partly noticing all of this, and said (with equal parts defeat and humor): “So, I guess this isn’t a safety issue, is it?” “No,” I said. “I guess it’s not.” And with that clarity, he was no longer being pressured to get into the car.
Instead of forcing him, my wife made it clear that he was allowed to not want to go. He was allowed to have some say over when he would be ready to go (if ever). And with that freedom, he happily ran out of the garage and into the front yard.
But he didn’t need to run around for more than maybe 2 minutes before he came back out-of-breath and said, “OK…I’m ready now.” And as we both tried to mask our astonishment, he climbed into the car, settled in, and off we went. (Note: even though I know this is the likely outcome from validating his experience and giving him space, it still feels like magic every time.)
Anyways, the point of asking “Is this a safety issue?” is it preserves the legitimacy of authority and the right to make certain kinds of requirements. As parents, we didn’t feel the need to assert or protect our rights as parents to do what’s best for our child. But the only reason we didn’t feel that need because we had a clear criteria for knowing when we would be justified in using it.
The mistake some would make (and it could have easily been us that day) would be to either continue forcing him into the car, or completely let go of our power and responsibility and become overly permissive—effectively making him parent himself. I probably don’t need to go into why that is bad idea.
So, sanction rules are requirements that are truly non-negotiable, but they must be used with discretion. And the simple version of that discretion is asking, “Is this a safety issue?”
2. Try to Turn “Shoulds” into “Musts.”
The second way to operationalize the distinction between sanction vs signal rules is to recognize that signal rules are still “musts” not “shoulds” (i.e. they are still requirements, not guidelines). However, the way most groups try to operationalize signal rules, likely intuiting they are somehow different than sanction rules, is to try and soften them.
There are a few different ways of trying to soften a rule, but the most common one I see is using the word “should” for what is actually a requirement. For example, a policy like “Everyone should submit their permission slips by Friday,” or “The plane leaves at 10am, so we should get there with enough time to get through security…”
Now, you may be thinking, “Well, using the word ‘must’ for sanction rules makes sense because those are absolute requirements, but doesn’t using ‘must’ for other rules make them seem too constraining or oppressive?”
And my answer is, “Yes,” it often does seem too constraining, but that’s a reason for evolving our relationship to rules, not a justification for leaving things confused.
So, my first point is that while using “must” can seem too rigid, recognize that the rigidity comes from the way they are typically energized and enforced (i.e. treating signal rules as if they were sanction).
The second point is that the word “must” actually gives us a lot of clarity if we can wield it appropriately.
It’s not that “should” is a bad word, after all some normative expectations are most appropriately defined that way, but not when what is actually desired is an agreement or expectation that something happen. The problem with “should” is that it’s too vague. When do you do it? Under what conditions? That’s all left unspecified and up to the individual’s discretion.
That may be fine early on, or if only a vague sense is needed, but if you already know you want the right to expect something and the expectation has been defined, then saying “should” only creates ambiguity.
So, why do we still generally prefer “should” over “must” even when we’re the ones making the rules? Well, I think there are two reasons; 1) there is an inherent administrative cost to managing a “must” and so just saying “should” shifts that burden away from the author to the agent (for more on those nuanced dynamics read the article Strategic Non-Enforcement), and 2) we don’t really know of any other good way to soften the rule.
But there are better ways to “soften” a signal rule without losing its clarity. My recommendation is still to phrase the expectation with a “must” like, “Everyone must submit their permission slips by Friday,” but also…
Clarify exceptions (and/or process to request one). For example, “Let us know if you need an extension,” or, “Late slips will be considered on a case-by-case basis and there’s no guarantee you’ll have a spot.”
Take responsibility for administering the expectation. Afraid people won’t submit the permission slips until right before (or maybe slightly after) the deadline? Send out multiple reminders. Account for human imperfection. Do enough so that any consequences you need to enact can feel justified.
Enforce it neutrally. By definition signal rules will occasionally be violated (though no one should ever intentionally violate them), so enforcement is key. However, shame and blame have no place when it comes to signal rules. You can soften the “must” much more effectively through your stance or energy. For example, think how ridiculous it would be for a referee to shame someone who is playing hard for stepping out of bounds.
Depending on the specific circumstances there are many more ways to make a signal rule feel different than a sanction rule—these are really just some quick tips and examples on this point. Let me know if you want more.
The point is that when it comes to signal rules you still want to use “must,” but unlike sanction rules, you need to consciously prepare for violations. Violations are a key part of their value.
Conclusion
I can’t help but worry that all this content on sanction vs signal comes across as overly academic and conceptual. I’ve tried to make it concrete, but it paradoxically feels like I’ve spent too much and too little time explaining it.
The short version is that we need to embrace using BOTH sanction and signal rules, we just need to operationalize them differently. Particularly when it comes to enforcement.
If we want to be fair to each other, then we need to make sure our intentions and interpretations are aligned. So, if you’re defining rules or expectations in an organization, a team, or even just in your family, it’s worth asking:
Is this a safety issue and therefore a sanction rule?
Or is this a signal rule which means we should actually plan for violations?
When you’re clear on that distinction, rules stop being blunt instruments. They become more specialized tools for more specialized jobs. The rules we make for ourselves can both prevent and redirect—the important thing is that everyone knows the difference.
If this publication has been a source of wisdom for you then please consider helping me sustain it by becoming a monthly or annual contributor.
For just $8.00/month or $80.00/year, you’ll help ensure I can continue dedicating more time to creating thoughtful, quality content. In fact, I’m currently sitting on over 200 draft articles (in various stages of development), which I just haven’t had time to publish. Of course, a third of those articles are probably crap, but that would still mean I’m sitting on tons of content I’d love to share.
As a sponsor, you’ll also be the first to receive future premium content, including a digital copy of my upcoming book, “The Proper Use of Shared Values.”
It’s completely optional and I’ll always provide as much content as I can for free, but your contributions will be critical for helping me spread the word about new ways of collaborating that are both more effective and more meaningful.



Thank you for providing that helpful extra clarity! Two thoughts on this:
(1) The example of the glass heirloom does not seem like a safety issue to me. It is a one-way door for sure. I wonder if an even more precise distinction between sanction rules and signal rules exist. Maybe something along the lines of "is there a low-impact fallback path in case of violations?" - but as concise as "is it a safety issue?" So, while I like the brevity of "is it a safety issue", I'd like to habe something even more precise.
(2) I have often come across sanction rules that did not state the actual sanction. They were more like "Don't do this, or else." It felt weirdly obscure. I never new whether I'd be fired or just "reminded" of the rule, not knowing whether it would maybe still pay off to violate the rule. Is this an antipattern? Should we always state the consequences in case of violations? Is that even desirable in a corporate context?
Thanks Simon. 1) Yeah, I agree about the safety issue question. I mean, I obviously like it, but I think it's more of an exemplar than a definition. "Non-reversibility" is similarly useful, but a lot of insignificant stuff isn't reversible, so my current thinking is to combine them, but yeah, there must be something more requisite, but I don't know what it is yet. It could just be something more like a deal-breaker which more clearly varies depending on the situation. I have an open inquiry around this.
2) Yes! Key point! This is actually a feature of sanction rules that they usually don't state the actual consequence, and it makes sense when you think about it. It is precisely because they are so serious or significant, that the author (i.e. the person whose tension is being addressed via the rule) feels very uncomfortable even imagining that they would be violated. It would be like saying to your partner, "I think we should think through the tangible consequences of infidelity..." You can't even really bring it up without somehow feeling that you're actually and paradoxically creating a permission pathway. But your point about stating the consequences is, in general, one I completely agree with, it's just that when it comes to true deal-breakers, they just don't tend to work that way. Both because of the weirdness I mentioned, but also because you don't often know all of the deal-breakers until they get broken, so the best course of action is to establish a clear inquiry process that clearly includes likely dismissal or dissolution of the relationship.
For lower-level sanction rules—those that are still protective, but not quite deal-breakers, then yes, absolutely the consequences should be clearly stated. And importantly the consequences must be realistic. One of the useful insights that often come from thinking about consequences is realizing that you can't actually enforce something the way you might have assumed you could.
So, yes, the non-statement of consequences is generally a anti-pattern that corresponds to the unconscious trend of authors to deflect responsibility for administrating the rules onto the agents. With the notable exception of deal-breakers.