Two women in the same friend group stop speaking to each other. Something happened, and now everybody else has to decide what a loyal friend does next.
The 68-year-old refuses to pick. She’s known both of them for thirty years, and thirty years doesn’t get thrown away over one bad summer. She’ll keep calling both.
The 54-year-old also refuses to pick, for a completely different reason. She knows exactly what happened, because one of them told her, and she hasn’t repeated it to a single person and doesn’t plan to.
The 37-year-old picks the one who drove four hours when her mother was in the hospital.
The 25-year-old picks a side, says so out loud, and unfollows the other one that night.
All four are certain they behaved loyally. All four would say the other three got it wrong.
Loyalty gets treated as the one virtue everybody agrees on. It isn’t. Each generation learned what loyalty was protecting against under different conditions, and whatever you’re protecting against is what determines the shape of a betrayal. It’s the same split that opens up when four generations try to agree on what makes someone a good friend, only with a heavier penalty attached. A bad friend has let you down. A disloyal one has turned on you.
Boomers think loyalty means you don’t leave
For a boomer, the test is whether you’re still there. Not whether the friendship is good, not whether it’s balanced, not whether you’d pick the person today. Whether you stayed.
Everything else can be worked around. Leaving is the one move with no defense, and it stays indefensible no matter how well you explain it.
That instinct tracks a measurable gradient.
Moral foundations research treats loyalty as its own moral concern, defined as commitment to one’s group and an aversion to betrayal. A 2026 PLOS One study applying the updated Moral Foundations Questionnaire to US and UK samples found that loyalty and the other binding concerns rose consistently with age, with the expected gap in binding scores between an 18-year-old and an 80-year-old working out to roughly three-quarters of a standard deviation.
The authors are careful about what that can and can’t show. The data is cross-sectional, so it can’t separate getting older from having grown up in a different decade, and the samples came from an online research panel rather than a representative national one. Both explanations land in the same place regardless.
The conditions behind it are easy enough to reconstruct. Boomers came up when the significant relationships were understood as commitments rather than arrangements. One employer for twenty-five years, one marriage, one parish, one street. Staying was the entire proposition, and the people who left were the cautionary story told at the table.
Under those rules, loyalty never needed a quality test. Duration was the quality test.
What it costs them is that a standard built entirely on staying has no exit. A boomer can sit inside a friendship that went bad fifteen years ago and file it under loyalty, because leaving would mean becoming the kind of person who leaves.
It also means any departure by anyone younger registers as betrayal by default. The daughter who ends a friendship over a pattern she could describe in detail gets read as someone who quit, and the detail never comes into it.
For Gen X, loyalty means keeping your mouth shut
Gen X moved the test to information. Tell a Gen Xer something at two in the morning in 1998 and it’s still sitting exactly where you left it.
Discretion is the whole standard. They may not call for eight months, they may not ask how you’re doing, and they will not repeat one word of it to anyone, ever. Betrayal is repetition.
The formation makes sense. They came up in the divorce decade, in houses with nobody home after school, watching institutions they had no reason to trust. Nobody was managing their private business for them, so managing it themselves became the competence. They also built their friendships before anything anyone said was searchable, which made silence a workable defense rather than a quaint one.
The standard is harder to meet than it sounds.
Across nine studies published in the Journal of Personality and Social Psychology, Jessica Salerno and Michael Slepian found that people revealed about 30% of the secrets they had learned about others. When the researchers narrowed it to secrets that someone close had confided in them directly, the figure only fell to 26%.
Moral judgment is what predicted the telling. People passed along the secrets they found objectionable, and the driver was a desire to see the person punished for what they’d done.
Which is worth thinking about, because it means the vault is conditional in a way most people never state out loud. The secret is safe until the listener decides it shouldn’t be.
The flaw in a discretion-based definition is that a friendship can run on it for decades without going anywhere. You can hold the worst fact about someone for twenty years and never once be asked how your own year went. Silence isn’t the same as closeness, and Gen X applies the same silence to itself, which is how a person ends up with four people who would take a secret to the grave and nobody who knows there’s anything to keep.
More Bolde Stories
Millennials think loyalty means showing up when it costs you something
Millennials want evidence. Anyone can say they’re there; the test is who moved something. The flight booked, the shift covered, the four-hour drive, the phone picked up at three in the morning.
Presence that costs nothing barely registers. And the standard runs both directions, which is the part older generations tend to miss. Loyalty is owed to people who return it. In this framework, walking away from a friendship that ran one direction for years is what an accurate reading of the situation looks like. The betrayal already happened, and it happened to them.
They’re the first generation whose friendships came without a location attached. They graduated into a recession, moved for work, moved again for rent, and watched their people end up in five time zones. Nothing local reproduced itself the way it had for their parents. Any friendship they still had at thirty-five, they had kept deliberately, and deliberate effort became the thing they measured.
The generation most willing to close an unbalanced account also reports the emptiest one.
A YouGov survey of 1,254 US adults found that 30% of millennials said they always or often felt lonely, against 20% of Gen X and 15% of boomers, with 22% reporting no friends at all and 30% reporting no best friend.
It’s an online panel poll rather than peer-reviewed work, and the categories behave strangely, since more respondents said they had no acquaintances than said they had no friends. The age gradient has held up in later polling, though.
The trouble with a cost-based standard is that it only reports after the fact. You can’t find out whether someone will rearrange her life for you until the week you need her to, and by then her answer is the entire verdict on a fifteen-year friendship.
It also makes ordinary constraint look like a character reveal. The friend who couldn’t come had a toddler with croup and eleven dollars until Friday, and the ledger doesn’t have a column for that.
For Gen Z, loyalty means taking a side out loud
Gen Z runs the test in public. A loyal friend takes your side, says which side that is where people can see it, and does not maintain a warm relationship with the person who hurt you.
Neutrality is the failure state. Staying friendly with both parties isn’t fairness, it’s a position, and everyone can read it, because the evidence sits on a screen.
They’re the first cohort for whom allegiance leaves a permanent record. Who you follow, who you tag, whose posts you engage with, which group chat you’re in. Affiliation is published by default and archived automatically, so loyalty stopped being something you feel privately and became something you either demonstrate or fail to.
Their willingness to draw those lines does show up in survey data, though it comes tangled up with politics.
In September 2024, the Skeptic Research Center surveyed roughly 3,000 US adults and found that 45% of Gen Z liberals said they would sever ties with a friend over political views they considered inappropriate, against 39% of millennial liberals and 10% or fewer among Gen X and boomer moderates. Age was the strongest single demographic predictor in the whole dataset.
That measures stated willingness rather than anything anybody actually did, and other work points the other way. A 2024 Springtide Research Institute survey of 6,669 people aged 13 to 25 found that 52% disagreed that they would stop speaking to someone who strongly opposed their political views, with just 12% agreeing.
Both findings can be true at once. A person can hold a firm principle about where the lines are and still, in almost every actual case, not reach for it.
The cost of a public standard is that it converts disagreements into loyalty tests. It also eliminates a category that used to be useful, which is the friend who likes you and thinks you’re wrong about something, and says so, and stays.
The other cost is subtler. Loyalty performed for an audience is being graded by people who aren’t in the friendship, and their standards are harsher than either friend’s, because they have nothing at stake.
Every generation is guarding against a different betrayal
Boomers are guarding against abandonment, because their world assumed you’d be there in twenty years and had no plan for what happened if you weren’t. Gen X is guarding against exposure, because private information was the one thing they could still control. Millennials are guarding against being used, because they were the ones doing all the initiating and could feel it. Gen Z is guarding against being left alone in a position, because positions are visible now and standing in one by yourself is exposed in a way it never was before.
Each is a sensible defense against the thing that was genuinely likely to happen to them.
One more number is worth considering here.
The Survey Center on American Life’s 2021 study of American friendship found that only 15% of Americans had actually ended a friendship over politics, with Democrats reporting it at twice the rate of Republicans and liberal women highest at 33%.
Stated willingness runs a long way ahead of behavior. People will tell a pollster they’d cut someone off, and then mostly not do it, which suggests the loyalty rules everyone is so certain about get quietly bent when a real person is standing on the other side of them.
Which brings it back to the friend group that split. Four people watched the same thing happen and reached for four different rulebooks, none of which is written down anywhere, and all of which the other three are somehow expected to know.
