Work & Leadership
Psychological Safety Is Not Psychological Insulation
The all-hands runs forty minutes and nothing is said in it.
Not nothing spoken — there is a deck, and there are updates, and someone from operations walks through a slide with four arrows on it. But the thing everybody in the room knows, the reason the timeline is not going to hold, does not get said. Two people mention it afterward in the parking lot. One of them mentions it to me, some months later, in an office, because by then it has cost him a promotion and he wants to understand what happened.
This is a company with a psychological safety initiative. There are posters. There was a training.
what the term actually meant
Amy Edmondson’s work in the late nineties produced one of the most useful findings in organizational psychology, and one of the most thoroughly misread.
She was studying hospital teams and expected to find that better teams made fewer errors. What she found was that better teams reported more errors. The good units were not the ones where less went wrong. They were the ones where people said what went wrong.
Psychological safety, in her definition, is a shared belief that the team is safe for interpersonal risk-taking — that you can speak up, disagree, admit a mistake, ask an obvious question, and remain respected. It is a claim about what you can say.
Notice what it is not a claim about. It says nothing about whether you will feel comfortable. In fact it describes the opposite: a team where hard things get said is a team where you will regularly be uncomfortable, because being told your design has a flaw is not a pleasant experience and never becomes one.
Somewhere in the last decade the term slipped its moorings. It stopped meaning you can speak up and started meaning you should never have to feel bad here.
That second thing is not safety. It is insulation. And they behave in opposite directions.
the difference, in one meeting
In a psychologically safe team, a junior engineer says the timeline will not hold, the room takes it seriously, someone senior says tell me more, and the plan changes or does not, and the engineer is treated identically afterward.
In a psychologically insulated team, nobody says it — because saying it would make someone uncomfortable, and discomfort has been reclassified as harm. The team protects the feelings in the room at the cost of the information in the room. The timeline still does not hold. It simply does not hold silently, for another five weeks, until it is expensive.
Safety makes hard conversations possible. Insulation makes them unnecessary and therefore impossible. One of them produces teams that learn. The other produces teams that are pleasant right up until the quarter closes.
how a word gets bigger
There is a good name for what happened here, and it comes from Nick Haslam, who called it concept creep: the tendency of psychological concepts to expand over time, stretching outward to include milder and milder phenomena.
Trauma used to mean something quite specific and now covers a difficult conversation. Bullying once required repetition and a power differential. Abuse has widened. And Haslam’s argument is not that the expansion is malicious — it usually comes from genuine moral progress, from taking harm more seriously than we used to. It is that expansion has costs that nobody accounts for.
One of those costs shows up in later work by Haslam and colleagues, who ran a series of experiments on what happens when a concept is broadened. Their finding is uncomfortable and worth sitting with: broadened concepts of harm are taken less seriously. When the same word covers both a serious injury and a mild slight, the word loses its power to mark the serious one.
Which means the person hurt most by the inflation of I don’t feel safe is the employee who is genuinely not safe. She now has to use a phrase that her manager has heard four times this month about a delayed lunch order.
two sentences that need different responses
This is the practical heart of it, and it is a distinction I would put on the wall of every management training I have ever seen.
“I don’t feel safe” and “I didn’t like that” are not the same sentence, and collapsing them is how organizations end up serving neither person.
The first is a claim about risk — of retaliation, of humiliation, of being made to pay for something later. It requires investigation. Somebody with authority needs to find out what happened, and there are usually facts to establish.
The second is a claim about experience. It is legitimate, it deserves a real response, and the response is a conversation, not an investigation. That landed badly. Tell me what happened for you.
Treat the second like the first and you have started a proceeding over a tone of voice, taught the whole team that ordinary friction escalates, and made every manager afraid to give feedback. Treat the first like the second and you have left someone in a genuinely unsafe situation with a suggestion that they talk it out.
The skill is not choosing which one to believe. It is asking one more question before deciding which conversation you are in.
discomfort is not danger
I want to be careful, because this argument gets used badly.
I am not making the case that people are too sensitive. I do not think that, and I have spent too many hours across from people who were harmed at work by things their organizations declined to look at.
I am making a narrower case: that the emotional experience of this is hard and I do not like it and the situational fact of I am at risk here are different categories, and that we have lost the language to hold them apart. Feeling bad is not evidence of being wronged. It is also not evidence of not being wronged. It is information about you, and it needs a second step before it becomes a conclusion about them.
The best-run teams I have seen are not the calm ones. They are the ones where people say difficult things to each other regularly, on the record, without anyone’s standing changing afterward — and where a person who receives hard feedback can be visibly bothered by it for twenty minutes without anyone treating that as a crisis.
That is what sturdy looks like. Not unbothered. Bothered, and still in the room.
a word about what we owe each other
There is a line in Micah that I have thought about in more workplaces than I expected to: act justly, love mercy, walk humbly.
The order matters. Justice first — behavior, the thing that actually lands on another person. Then mercy, the disposition underneath it. Then humility, the willingness to be corrected about both.
Most organizational culture work starts at the third and never gets to the first. It works on how people feel about each other and leaves untouched the question of what is actually being done, and by whom, and with what consequence. Micah looks at conduct first, and I think that is the right order for a company too.
one thing you can do this week
The next time someone brings you something hard, ask one question before you respond to it: are you telling me you are at risk, or telling me something landed badly?
Ask it plainly and without an edge, because it is not a challenge. Most people can answer it immediately, and are relieved to be asked, because the two situations feel completely different from inside and they have been given only one phrase to describe both.
You will find, more often than you expect, that people did not want an investigation. They wanted someone to know.
Dr. Donetta D. Quinones, PhD, LMHC, LPC. There’s No Crying in Baseball: Psychological Safety, Emotional Fragility, and the New Rules of Work makes an argument almost nobody is making — that emotional fragility at work is substantially produced by management, and at least as often by what leaders fail to do as by what they do.