Personal & Professional Development

Why we can't say sorry

“I’m sorry” has become a word we use to end a moment, not to repair one.

In my latest piece, I explore why sincere apologies feel so hard: the status cost, the power dynamics, and the institutional incentives that teach us never to fully own our mistakes.

Apology isn’t weakness. Done well, it’s a deliberate choice to let the other person’s pain outrank your self‑image, then to change your behaviour.

#leadership #culture #communication #psychology

Axis of Action 8 October 2026
Why we can't say sorry

"I'm sorry" has become a word we use to end a moment, not to repair one. It gets said so often, in so many low-stakes contexts, that by the time it's needed for something real, it arrives already worn thin. The person on the receiving end can usually feel the difference between a word chosen to close a conversation and a word chosen to open one.


I recently came across a framework for apology drawn from Gary Chapman and Jennifer Thomas's work on the languages of apology: name the other person's pain rather than your own feelings, ask for forgiveness rather than simply declaring the apology given, own it without qualifiers, and change the behaviour afterwards. It's a good framework. But what struck me more was what it implicitly reveals about why people avoid apologising in the first place, and it isn't mainly about stubbornness or a lack of remorse. It's structural.


Apology as a status transaction

To apologise sincerely, you have to temporarily lower your position relative to the other person. You're saying, in effect: your account of what happened, and your pain, outranks my self-image right now. That's a real cost, and it isn't distributed evenly.


People with more power in a relationship, whether that's professional standing, seniority, financial leverage, or simply being the one who's less invested, tend to apologise less. They have less need to repair, because the other person often can't afford to walk away. Meanwhile the person with less power over-apologises, sometimes for things that clearly aren't their fault, because smoothing things over has become the price of staying safe.


There's also a widely internalised (and mostly wrong) social script that whoever apologises first has "lost" the exchange. Which is backwards. Every step in a genuine apology, naming the harm in the other person's terms, asking rather than declaring, refusing to explain your way out, requires giving up power rather than managing the situation. That's precisely why it's hard, and precisely why it works when it happens.


The fear underneath the avoidance

Here's the part I think gets missed. Many people don't withhold an apology because they lack remorse. They withhold it because a full apology feels like it opens a door they can't close. If I fully admit "I did this, and here is the harm it caused," the next question is almost always "so what are you going to do about it?" If the honest answer is "I don't know if I can fix this" or "the cost of fixing this is more than I'm prepared to pay," a lot of people short-circuit the whole thing at the apology stage rather than face what comes after.


This isn't only psychological. It's reinforced by systems we operate inside every day:

Legal and institutional caution. "Never apologise, it's an admission of liability" is standard advice in corporate, medical, and insurance contexts. It's worth noting that several jurisdictions eventually passed apology laws specifically because clinicians were avoiding apologising to patients out of fear it would be used against them, even in cases where an early, genuine apology would have prevented litigation altogether by giving the patient what they actually wanted: acknowledgment.


Financial exposure. When reparations are material, whether compensation, restitution, or simply giving something up, there's a direct incentive to minimise the stated harm rather than name it in full. The apology gets negotiated down to match what the apologiser is willing to pay, not what the harm actually cost.


Relational obligation. The same dynamic plays out at a smaller scale in personal relationships. A full accounting of harm can feel like it commits you to years of visible behaviour change, so people quietly shrink the apology to shrink the obligation.


When apology reads as weakness

In a lot of cultural and workplace scripts, apologising gets coded as submission, and submission gets read as an invitation for further challenge. So people build identities around never being wrong, because being wrong once feels like it will be held against them indefinitely. This produces a loop: the less people apologise, the more catastrophic and singular each apology feels, which makes people even less willing to offer one.


It's worth noticing how gendered this script often is, and how selectively it's applied. Men are frequently taught that apology signals weakness, while women are taught to apologise reflexively, for taking up space, for having needs, for things that were never theirs to own. Neither pattern reflects the actual size of the harm caused. Both are inherited habits, rehearsed so early and so often that they can feel like personality rather than conditioning, and both make it harder to tell when an apology is actually warranted versus when it's simply the trained response.


Apology at institutional scale

The same dynamics play out, magnified, when the apologiser is an organisation rather than a person. A board or an executive team facing a failure of governance, a safety incident, a mishandled restructure, faces every one of the pressures above at once, plus a few of its own. Legal counsel will usually advise caution before contrition. Insurers will want the wording controlled. Communications teams will want to manage exposure rather than name harm in the terms the affected people would use.


The result is a familiar genre: the statement that expresses regret without ever using the word "we," the passive voice doing quiet work ("mistakes were made," "the situation was not handled as it should have been"), and the review that finds "systemic issues" rather than naming who was accountable for them. These aren't accidents of drafting. They're the organisational version of explaining instead of owning, and people read them exactly as they'd read a personal non-apology: as an attempt to end discomfort rather than repair harm.


The organisations that do this well tend to share one trait, they treat the apology and the fix as genuinely separate commitments, made in that order, rather than trying to fold acknowledgment and remedy into a single hedge-everything sentence. Naming what went wrong, in plain terms, before the remedy is fully worked out, is uncomfortable precisely because it removes the option of controlling the story. That discomfort is the point.


The automatic exits

When someone is confronted with harm they've caused, a handful of defence mechanisms tend to fire before they've even consciously registered what happened:

  • Deny, attack, reverse the roles. Deny it happened, attack the person raising it, then position yourself as the one being wronged by the accusation.
  • Minimise. "It wasn't that big a deal" quietly relitigates the size of the harm instead of addressing it.
  • Explain instead of own. "I was tired," "I didn't mean it." These aren't necessarily dishonest, but offered in place of accountability, rather than alongside it, they function as an exit rather than an explanation.
  • Deflect to the other person's conduct. "Well, you did this to me too," which converts a request for repair into a new front of conflict.
  • Apologise fast and vague, on purpose. Ending the conversation before the other person can fully articulate the harm. "Sorry you feel that way" reads as an insult because it relocates the problem into the other person's feelings rather than your action.


Most of these aren't conscious manipulation. They're well-rehearsed defences of self-image, fired automatically, before the person has had time to process what actually happened.

Which is what makes a structured approach to apology useful. Naming the other person's pain in their terms makes minimising harder to reach for. Asking for forgiveness rather than declaring it given makes DARVO harder to execute. Ruling out explanations in advance closes the exit before you're standing in front of it.


None of this guarantees the apology will be accepted. It shouldn't. You can offer a genuine apology and still not be forgiven, and that has to be something you just have to accept, because forgiveness was is not something you can control to extract. What is yours to control is whether the apology, when it comes, is actually built to repair something, or just built to make the discomfort stop.