Why organizations that chase the absence of threat end up manufacturing more of it — and what to aim at instead.
The modern West has removed more violence from ordinary life than any society in recorded history. Safer cars, safer factories, safer food, safer streets, safer childhoods. By the standard of “how likely is something bad to happen to me today,” we are the most fortunate people who have ever lived. Over the same stretch, anxiety, loneliness, and despair have climbed.
And at the same time, we are told to find safe people, safe places, and psychological safety in our personal and organizational worlds.
If safety — the absence of threat — were the thing that produced peace, we should be the calmest civilization in history. We are conspicuously not. Something in the model is wrong, and I’ve come to think the error is not in our execution but in our target.
It is costing organizations more than they know.
The instrument problem
Start with what kind of thing safety is. It’s an absence. It’s defined by what isn’t happening.
And here is the trap: to pursue an absence, you have to continuously detect a presence. There is no way to know how safe you are without checking for what would make you unsafe. The instrument that reads safe is the threat detector, and the only way to keep the reading current is to keep the detector running.
So an organization that orients itself on psychological safety doesn’t install calm. It installs a permanent scanning apparatus. Then, because scanning finds things, it staffs the apparatus, meets about its findings, and quietly promotes the people who are best at spotting.
You cannot become unafraid by monitoring your fear.
And scanning does not stay pointed where you aimed it. A workforce trained to look for what’s wrong becomes extremely good at finding it — in the process, in the numbers, and eventually in each other. Vigilance is a general-purpose capability. You cannot install it for one target and expect it to stay there.
That’s the irony in one sentence: the pursuit of safety runs the threat system, and a running threat system produces the experience of threat. Not because anyone is doing it badly. Because that’s what the machinery is.
A fair word about psychological safety
I’m not picking a fight with Amy Edmondson. Her actual finding is solid and well-replicated: teams perform better when people don’t fear humiliation for speaking up, or face retribution for trying and failing. That’s real, it matters, and I use it.
But watch what happens when a scanning for lack of safety becomes a program. People are taught to focus on the negative, what can be unsafe. This activates the danger-sensing portions of our brains.
There’s a suggestive line of work here worth sitting with. Research in this area indicates that the leaders who most undermine a team’s psychological well-being are not primarily the ones afraid of failure. They’re the ones wary of forming relationships — of relinquishing control, of being known, of the exposure that connection requires. The deficit is relational, not protective.
Which means the road to the thing Edmondson measured doesn’t run through a safety program. It runs through connection.
Two systems, not one dial
This isn’t a metaphor. It’s motivational architecture, and it’s been mapped for thirty years.
Tory Higgins’s regulatory focus theory (1997) distinguishes two systems that are not the same machine at different volumes:
- The prevention system orients toward non-loss, self-protection, and duty. Its strategy is vigilance.
- The promotion system orients toward nurturance, growth, and gain. Its strategy is eagerness.
Follow-on work on social approach and avoidance motives (Nikitin & Freund, 2008) tracks the consequences: avoidance-motivated people show the vigilant, safety-securing behavior you’d expect, and fare worse on wellbeing and on major life transitions. Approach motivation tracks with flourishing.
Two different postures toward the world, running different chemistry, producing different people. And an organization selects one of them — largely with its vocabulary, and largely without noticing.
Even the field is moving on this. There’s a body of work now arguing explicitly that a genuine sense of safety cannot be built on the absence of threat, and has to be built on connection, meaning, and belonging instead. The researchers are reaching for the same thing from the other direction.
What valuing runs instead
Valuing people is a positive target. You can tell whether you’re doing it without checking for danger — which means the instrument is different.
Scanning still happens. Attention is always scanning something. But an organization oriented on valuing scans for opportunity to give, and finds it, because that’s what scanners do.
The physiology is unusually clean here:
Generosity calms the giver. When people give — and this shows up as far back as Harbaugh’s scanner work in 2007, where participants’ money went to a food bank — the giving engages the same reward circuitry that lights up when something good arrives for you. More recent work fills in the second half: prosocial behavior also recruits caregiving pathways that quiet the body’s stress axis. Cortisol comes down in the person doing the giving.
Warmth is contagious, and so is defensiveness. In a randomized trial, oxytocin increased the moment-to-moment linkage between participants’ autonomic responses during group therapy — nervous systems measurably syncing. This runs both ways. A defensive posture reads as a threat signal to the person across from you and gets mirrored back. Defense begets defense. Warmth begets warmth. You’re always broadcasting one of them.
Connection is what quiets the fear center. The medial prefrontal cortex exerts top-down, GABA-mediated inhibition over the amygdala, and perceived social support engages that circuit. Here’s the detail I find most telling: reduce GABA function in that same circuitry and sociability drops. The brake that quiets fear turns out to be required for connection itself. They’re not two systems. They’re one.
And over a lifetime, this is the variable that matters. The Harvard Study of Adult Development followed men across their entire lives and found relationship quality — not wealth, not security, not the absence of hardship — to be the best predictor of health and longevity.
Language matters because language causes focus
If you take one practical thing from this piece, take this one.
A major 2024 review in Psychological Science in the Public Interest (Flusberg, Holmes, Thibodeau, Nabi & Matlock) establishes framing as an unavoidable feature of communication with measurable effects on how people think, feel, and act. Not a rhetorical nicety. A mechanism.
And it operates below deliberation. Imaging work shows that merely reading emotionally-loaded words engages the amygdala and primes approach or avoidance responses — including defensive reflexes — before you’ve decided anything at all.
Surfers have a saying that where you focus, you will go. Look at the rocks, and you’ll land on the rocks. Look to where you want to go, and you will go there.
Which means the words painted on your wall are doing work whether or not anyone believes in them.
Consider one everyday example. Boundary presupposes something that must be kept out. Say the word and the prevention system is already warm before the conversation starts. Now take the identical practical decision and call it investment — where I place finite energy, what I’m choosing to nurture, what I’m building. Same calendar. Same declined meeting. Entirely different machinery recruited.
Run the audit on your own organization. In your last ten company-wide communications, count the threat words — risk, exposure, mitigate, compliance, guardrails, zero tolerance, watch for, red flags, protect against — against the building words — invest, build, grow, give, serve, develop, contribute.
Whichever list is longer is the system you’ve been running. It’s also, on this evidence, the culture you’ll have in three years.
But what about the wolves?
This is where the argument usually gets waved off, and fairly, because most versions of it are naive. So let me be direct: the wolves are real. Some people will take everything you extend and return nothing. Some will read openness as an opening. Anyone who has run a company for a decade has the scar.
Three things, and none of them requires a scanner.
One: valuing detects predators better than suspicion does — it just detects them differently.
You don’t need a suspicion apparatus. You need only pay attention to what comes back when you give. Reciprocity discloses itself fast; a predator reveals himself by what he does with your generosity, usually inside of a quarter.
And notice what you actually register when it happens. It isn’t threat detected. It’s the absence of fit. The energy doesn’t return. There’s nothing to build with here. So you don’t build here — you don’t hire him, you don’t partner with him, you don’t hand him the account.
That’s not scanning. That’s engagement, and it’s a better instrument for a boring reason: a suspicion apparatus generates false positives about decent people, and a genuine predator who knows he’s being scanned for simply performs. Generosity is a harder test to fake than an interview.
Two: protection can run on valuing.
Sometimes you have to remove someone. End a partnership. Fire a manager who is eating your people alive. Stand between a predator and someone who can’t stand up for themselves.
Nothing here says don’t. What it says is: notice which system you’re running it from.
A man who removes a predator because he values the people in that building is doing something categorically different — in his own body, and in what his organization learns from watching — than a man who removes a predator because he’s afraid. The action can look identical from the parking lot. The fuel is not the same, and the fuel is what determines what happens next.
Fear-fuel has no off switch. Once the threat is gone, the machinery keeps running and finds a new object, which is why you can walk into a company that solved its problem two years ago and still feel the vigilance in the hallway.
Three: when you get hurt, there are positive roads back.
This is the piece most people don’t have, and it’s the actual mechanism by which good people become scanners. Nobody sets out to build a suspicion apparatus. They get hurt, they have no way to metabolize it, and vigilance is what’s left over.
There are three better ways out, and all three are relational rather than defensive:
- Acceptance — stopping the fight against what cannot be undone. The endless war against an unchangeable fact is a stress response with no possible resolution, and it will run for years if you let it.
- Forgiveness — releasing a debt that will not be repaid. Not reconciliation (that takes two, and may never be wise). Not excusing. Just closing an account that has been charging you rent. It shows up in the literature attached to lower stress hormones and lower blood pressure, which is worth knowing: you are not doing it for them.
- Grief — letting the loss land. Loss that is never felt keeps the body braced against something it was never permitted to finish.
Each of these preserves the disposition. Vigilance is what installs itself when none of them is available.
The Lindy test
Nassim Taleb popularized the Lindy effect: for things that don’t perish — ideas, books, institutions — life expectancy increases with age. What has survived two thousand years of continuous use has passed a filter that no management framework has been anywhere near.
So it’s worth noticing which instructions keep showing up.
“Love your enemies.” The approach system, applied to the hardest case that exists. Not tolerate, not manage, not neutralize. Approach.
And from the same tradition, without contradiction: “Do not make friends with a hot-tempered man, and do not associate with one easily angered, or you may learn his ways and entangle yourself.” That is discernment without hatred. You may love him. You are not required to apprentice yourself to him. Both instructions, same book — which is exactly the combination this article has been arguing for.
Then the one that names the problem outright: “I am sending you out as sheep among wolves; so be wise as serpents and innocent as doves.”
Read what’s in that sentence and what isn’t. The wolves are named — no pretending, no Pollyanna. There is no instruction to scan, arm, or fortify. There is an instruction to be shrewd. And there is an instruction to stay harmless. Discernment and non-defensiveness, held together in one breath, which is precisely the posture the research keeps landing on.
And the strategy stated plainly: “Do not be overcome by evil, but overcome evil with good.” That is not a sentiment for a wall hanging. It’s an instruction to deliberately run the approach system at a threat.
These were written by people with no instruments, no imaging, and no way to measure a stress hormone. They were describing what they had lived and telling other people how to get there. Make of the correspondence what you like.
What this looks like on Monday
- Audit the vocabulary. Last ten company communications. Threat words versus building words. Count them.
- Rename one thing. Whatever program you have that’s named for what it prevents, rename it for what it builds — and mean it.
- Retire one word. Replace boundary with investment in a single conversation this week and notice how differently it goes.
- Give one specific thank you. Name what the person did, what it made possible, and what it told you about them. Once a week, one person.
- When someone doesn’t reciprocate, don’t escalate scanning. Just stop investing there. No memo, no watchlist, no story about them. Redirect the energy.
- When you get hurt, metabolize it. Name it, grieve it, release it. The alternative isn’t neutrality — it’s a scanner you’ll be running for the next decade.
The reframe
Safety asks: what could go wrong here? That question always has an answer, and the search for it never terminates.
Valuing asks: who could I build here? That question also always has an answer.
Both questions get answered. Both train the people who spend their days answering them. The only real decision a leader makes is which question his organization will be answering for the next ten years.
You cannot build a culture out of an absence. You can only build one out of what you decide to put there.
Sources referenced
- Higgins, E. T. (1997). Beyond pleasure and pain. American Psychologist, 52(12), 1280–1300.
- Nikitin, J., & Freund, A. M. (2008). The role of social approach and avoidance motives for subjective well-being and the successful transition to adulthood. Applied Psychology, 57(s1), 90–111.
- Flusberg, S. J., Holmes, K. J., Thibodeau, P. H., Nabi, R. L., & Matlock, T. (2024). The psychology of framing: how everyday language shapes the way we think, feel, and act. Psychological Science in the Public Interest, 25(3).
- Herbert, C., et al. (2017). Bodily reactions to emotional words. Frontiers in Psychology, 8, 1277.
- Harbaugh, W. T., Mayr, U., & Burghart, D. R. (2007). Neural responses to taxation and voluntary giving. Science, 316(5831), 1622–1625.
- How helping others helps us: neural mechanisms linking prosocial behavior to psychological and physical wellbeing. (2026). Frontiers in Human Neuroscience.
- Oxytocin increases physiological linkage during group therapy: a randomized clinical trial. (2021). Scientific Reports, 11.
- Stevens, F. L., et al. (2016). Functional and neurochemical interactions within the amygdala–medial prefrontal cortex circuit. Brain Structure and Function, 222.
- Decreasing GABA function within the medial prefrontal cortex or basolateral amygdala decreases sociability. (2016). Behavioural Brain Research.
- Lynch, J. M., et al. (2024). The sense of safety theoretical framework. Frontiers in Psychology, 15, 1441493; and Editorial: From safety to sense of safety (2025), Frontiers in Psychology, 16, 1624676.
- Waldinger, R. J., & Schulz, M. S. (2023). The Good Life. [Harvard Study of Adult Development]
- Worthington, E. L., et al. (2007). Forgiveness, health, and well-being. Journal of Behavioral Medicine, 30(4), 291–302.
- Edmondson, A. C. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350–383.
Scripture quotations: Matthew 5:44; Proverbs 22:24–25; Matthew 10:16; Romans 12:21.