The idea that old age itself cools companionship is very old. In the Nicomachean Ethics, in W. D. Ross’s translation, Aristotle writes that “between sour and elderly people friendship arises less readily, inasmuch as they are less good-tempered and enjoy companionship less...” That line belongs to a wider portrait in which age makes friendships cooler and more concerned with usefulness.
Modern research complicates that portrait. But on one point, Aristotle may have been close to the mark.
What actually disappears from your circle?
Research clearly suggests that, on average, our social circles get smaller as we move through adult life. For instance, a 2013 meta-analysis in the journal Psychological Bulletin, pooling 277 studies and 177,635 participants, found that overall networks grow until young adulthood and then shrink steadily. More specifically, friendship networks shrink throughout adulthood, while family networks stay roughly stable.
The obvious explanation is loss: friends die, health narrows what you can do, chances to meet people thin out. But is that actually the whole story?
Laura Carstensen, a Stanford psychologist and founding director of the Stanford Center on Longevity, has spent decades studying whether smaller networks might “reflect preferences for close social partners and less interest in more peripheral partners,” rather than “deaths, disabilities, and a lack of social opportunities,” as she explains in a 2021 review in The Gerontologist.
In a ten-year study of Bay Area adults aged 18 to 94 at the start, published in the International Journal of Behavioral Development in 2014, Tammy English and Carstensen found that networks grew in young adulthood and then declined steadily. However, most of the loss was peripheral partners, a decline that “accelerated at advanced ages,” while close partners stayed relatively stable.
Similarly, in a 1994 study of 156 Berliners aged 70 to 104, Frieder Lang and Carstensen reported that the very old had networks nearly half the size of the old, “but the number of very close relationships does not differentiate age groups.”
But the narrowing does not seem to wait for old age. In a 1992 paper in Psychology and Aging, Carstensen drew on the Child Guidance Study, which interviewed 50 people over 34 years. Contact with acquaintances, and with close friends, fell from early adulthood on, while emotional closeness with relatives and close friends grew. She concluded that “individuals begin narrowing their range of social partners long before old age.”
If you’re 40 or 50, this isn't a pattern waiting for you in old age. On average, it has already begun.
None of this shows that anyone chose anything, however. Loss and choice can produce the same pattern: peripheral people disappear while close people stay.
The most direct evidence I could find on that comes from Frieder Lang, who followed 206 people aged 70 to 103 in Berlin over four years. Writing in the Journal of Social and Personal Relationships in 2000, he separated “non-deliberate endings, such as loss” from “active choice of social partners,” and found that “more than half of discontinued relationships were ended for deliberate reasons.”
That does not mean half of every older person’s friendships are consciously culled, or that all network shrinkage is voluntary. It does show that loss alone cannot explain what Lang observed.
Age and sense of time
Lang's paper names a candidate in its title, which frames those endings as a way of maximizing what relationships give "when feeling near to death." Carstensen's experiments put that idea to a direct test.
A 1990 study in Psychology and Aging by Barbara Fredrickson and Carstensen, they phoned 380 people aged 11 to 92, at random. Each was asked who they would spend “half an hour of free time” with: a member of their immediate family, a recent acquaintance they seemed to have a lot in common with, or the author of a book they’d read.
In the youngest group, ages 11 to 29, 35% picked the family member. In the middle-aged group, 30 to 64, the figure was 64%, and among those aged 65 to 92 it was 65%. The jump came between youth and midlife, not in old age.
Then came a new scenario: “In just a few weeks, you plan to move across the country—by yourself. No members of your family or current social circle will be accompanying you…” With that ending in view, the young group’s choice of the family member jumped from 35% to 80%.
The effect also runs the other way. In a 1999 study by Fung, Carstensen and Lutz, when older adults imagined an “expansive future”, their preference for familiar partners disappeared.
Taken together, the moving-away and extended-life scenarios suggest that social preferences depend partly on how much future time people feel they have.
Was Aristotle wrong about friendship in old age?
Aristotle actually pointed to the same fact of limited time.
In the Rhetoric, in W. Rhys Roberts’s translation, he writes that older men “live by memory rather than by hope; for what is left to them of life is but little as compared with the long past.” His conclusion, however, is almost the opposite of Carstensen’s. They “neither love warmly nor hate bitterly,” he says, and guide their lives “too much by considerations of what is useful.”
The modern research complicates that portrait. The ties that disproportionately fall away with age are peripheral ones. Given a free half hour, older people tend to pick someone they already know over someone new. And in the 34-year study, emotional closeness with the people who stayed grew.
But these studies were not designed to classify relationships according to Aristotle’s categories of utility, pleasure and friendship. They cannot prove that his claim about utility-based friendships in old age was false. What they can show is that his broader portrait of aging as a cooling of close attachment does not sit easily with the observed pattern.
Aristotle was more persuasive on another point: the limits of close friendship.
“One cannot be a friend to many people in the sense of having friendship of the perfect type with them,” he writes in Book VIII. In Book IX he adds that “it is well not to seek to have as many friends as possible, but as many as are enough for the purpose of living together,” and that “we must be content if we find even a few such.”
The choice
Not all shrinking is chosen. Friends die, workplaces disappear, health changes and opportunities narrow. From a distance, the result can look the same: peripheral people disappear while close people stay.
But Lang found deliberate reasons behind more than half of the relationship endings he recorded, and Carstensen’s experiments suggest that our preferences change when the future feels shorter. Choice, in other words, is part of the story.
That, of course, does not mean fewer relationships are automatically better. Isolation carries its own serious costs.
And less-close relationships have value too. In a 2020 study of 802 US adults aged 40 and older, Oliver Huxhold and colleagues found that the number of weaker social ties was more strongly associated with positive changes in emotional well-being over time than the number of close ties. People with more weaker ties tended to show more positive affect and less depressed affect over time. The findings complicate the idea that keeping only our closest relationships is necessarily better.
Perhaps, then, the point is not to shrink our circle for the sake of shrinking it. The harder question is how to give our closest relationships the time they deserve without assuming that our weaker ties are expendable.