The Question They Have Been Asking Since 1985
Every few years, a research team knocks on doors across the country and asks a question that sounds almost too simple to matter: in the last six months, who did you talk to about the things that really concern you? Not your weather-small-talk neighbor. Not the coworker you nod at in the hallway. The person you would actually call.
In 1985, when the General Social Survey first asked this question of a representative sample of American adults, the average respondent named just under three people. By 2004, that number had dropped by nearly a third, to just over two1.
The finding startled the sociologists who published it, and it startled the discipline that received it. It has been cited, contested, replicated, and re-argued for two decades, and the newer data — collected through the pandemic years and into this decade — has not reversed the trend so much as confirmed it2. Americans are not becoming a nation of hermits. Most people still have jobs, marriages, group chats, church pews, softball leagues. What has thinned is something narrower and more essential: the small circle of people to whom a person can say the true, difficult, unpolished thing, and trust that it will be received rather than repeated.
In the previous installment, we followed a man whose neighbors did not learn he had died until the smell reached the hallway — a man with, in the most literal sense, no one to call3. That story read like an extreme case, an outlier, the kind of thing a reader files under tragic but rare. The data suggests otherwise. It suggests that what happened to him in the extreme is happening to the rest of us in the ordinary — not the absence of a name in a phone’s contact list, but the slow erosion of the shorter list within that list: the two or three people a person could actually call.
The Village That Stopped Needing to Trust Its Neighbors
The pattern is not new, even if the measurement is. A German sociologist named Ferdinand Tönnies spent the late nineteenth century trying to name what he watched happening to the towns around him, as farming villages emptied into industrial cities. He proposed a distinction between two kinds of human bonds: Gemeinschaft, the community — relationships built on kinship, shared land, and inherited obligation — and Gesellschaft, the society — relationships built on contract, convenience, and shared proximity without shared history4.
A village did not require its residents to choose their confidants deliberately. Proximity did the choosing. The blacksmith knew the miller’s grief because he had watched the miller bury two children in the same churchyard where his own were buried. Nobody had to schedule that knowledge. It simply accumulated, year by year, the way dust settles on a windowsill nobody remembers to wipe.
A worker arriving in Pittsburgh, Lowell, or Chicago in 1870 inherited none of that accumulation. He might live beside strangers for a decade and never learn their names — not because he was unfriendly, but because nothing in his day required him to. Boarding houses turned over their tenants every few months. A man could eat breakfast across the table from someone for a season and never learn whether he had a wife back home, a debt, a fear he carried into the mill each morning.
So people built things. Lodges. Mutual-aid societies. Immigrant benevolent associations with secret handshakes and modest dues, invented — quite deliberately — to manufacture what the village had once produced for free5. They were not friendless. They were unsupported, and they knew the difference well enough to organize around it.
What is instructive here is not that isolation is new. It plainly is not. It is that every previous version of it provoked a visible, organized response — an entire generation deciding, on purpose, to rebuild what proximity used to give away for nothing. The current version of the pattern has provoked far less. No comparable infrastructure has risen to replace it. What has grown instead is a set of technologies that simulate contact without asking for the deeper commitment that once came bundled with it.
Why any of this should matter to a reader who has never set foot in a nineteenth-century boarding house is not obvious until the question is put plainly: every structure that once did this work for free has to be rebuilt, on purpose, by someone — and most people alive today have never been told that the rebuilding is theirs to do.
An Older Diagnosis Than Any Survey
Long before anyone thought to measure confidant networks, the opening chapters of Genesis offered a diagnosis that reads, in hindsight, almost clinical in its precision. Of everything catalogued as “good” in the creation account — light, land, sea, the animals each after their kind — one condition alone receives the opposite verdict, spoken before sin, before the fall, before anything had gone wrong.
“Then the Lord God said, ‘It is not good that the man should be alone; I will make him a helper fit for him.'” (Genesis 2:18, ESV)
The text does not treat solitude as a moral failure to be corrected through willpower. It treats it as a structural deficiency in an otherwise complete design — something requiring not exhortation, but a second person. Centuries later, the wisdom literature returns to the same observation from an entirely different angle, weighing the practical mathematics of companionship against the isolated life:
“Two are better than one, because they have a good reward for their toil. For if they fall, one will lift up his fellow. But woe to him who is alone when he falls and has not another to lift him up.” (Ecclesiastes 4:9–10, ESV)
The image is not sentimental. It is closer to a load-bearing calculation — what happens structurally, not just emotionally, when a person falls and there is no second hand to catch the weight. The early church, centuries after that, appears to have taken the observation as an operating principle rather than a proverb. The account in Acts describes a community organized, quite deliberately, around the elimination of exactly the isolation Genesis names as insufficient:
“And they devoted themselves to the apostles’ teaching and the fellowship, to the breaking of bread and the prayers… And all who believed were together and had all things in common.” (Acts 2:42, 44, ESV)
What is notable is the word devoted — not an ambient byproduct of shared belief, but a chosen, sustained practice, the kind of thing that has to be renewed on purpose rather than assumed to happen on its own.
Social science cannot establish theological truth, and this essay does not claim that it does. What it can sometimes do is describe, with numbers, a pattern that biblical writers had already named — long before that pattern became something a survey team could measure.
What the Newer Numbers Add to the Old Finding
The 1985-to-2004 comparison remains the anchor of this research because it was the first to measure the phenomenon with a repeated, national instrument, and because the shift it found was, by the standards of social science, unusually large1. The two decades since have offered little reassurance. Later analyses using the same GSS questions found that core network size never rebounded to its 1985 baseline2.
Everyone knows what this feels like, even without a survey to name it. A full contacts list. A calendar that never quite empties. And still, somewhere underneath all of it, a short list of people who would actually pick up — a list that keeps getting shorter without anyone deciding to shorten it.
Independent surveys built for different purposes converge on the same basic shape. Cigna’s national loneliness index measured 54 percent of adults as lonely in 2018. That spiked to 61 percent during the pandemic’s first year and, tellingly, settled back only to 58 percent once the emergency had fully passed6 — evidence, researchers suggest, that 2020 did not simply cause a disruption. It accelerated something that has proven durable.
AARP’s most recent research, tracking Americans 45 and older, found 40 percent reporting loneliness in 2025, up from 35 percent in both 2010 and 2018 — a rate that held flat for eight years before it started climbing7. One detail inside that report deserves more attention than it has gotten: men in this bracket now report higher loneliness than women, 42 percent to 37 percent, reversing the parity recorded as recently as 2018. For a generation of men who built their identity on self-sufficiency — marriage, career, a full calendar — the crack did not show on the surface. It showed underneath, in a number nobody was tracking until it moved.
What has shrunk is not the wide circle. Most Americans know more people by name than their grandparents did — coworkers scattered across time zones, alumni networks, a phone that never stops buzzing. What has thinned is the inner ring. The people who would hear about a diagnosis before the insurance paperwork gets filed. Someone who would notice — structurally notice, not just wonder — if a person simply stopped answering.
That distinction is what a “friendship recession,” as some researchers now call it, is actually measuring. Men reporting six or more close friends fell from 55 percent in 1990 to 27 percent in 2021. Men reporting no close friends at all rose fivefold, from 3 percent to 15 percent8. The friendships that survive tend to survive on paper. What erodes first is the unglamorous maintenance — the call with no occasion attached, the text that says nothing except thinking of you.
Isolation and loneliness are not the same condition. Isolation is objective — how many people, how often. Loneliness is the gap between the connection a person has and the connection a person wanted. A retired man with one weekly call to his daughter can be isolated, by any structural measure, and entirely at peace. A man surrounded by coworkers and a full contacts list can be thoroughly connected and desperately alone, because none of those names would answer at two in the morning without being asked twice why he was calling. The GSS question was built to measure the second condition, not the first — not how many people surround a person, but how many of them he could actually turn to.
The shorter list within the list. That is what has been quietly shrinking for forty years.
A Question That Was Never Really About Sociology
The General Social Survey did not set out to test a biblical claim. It set out to measure a fairly narrow empirical question about American social networks. But the measurement arrived, from an entirely different direction, at something Genesis had already stated — not as a warning about consequence, but as a description of a structural condition, named before any consequence had occurred.
The modal American in 1985 could name three people. The modal American today, more often than not, names none. Between those two points sits a generation’s worth of church suppers, fraternal lodges, and neighborhood porches, quietly emptied of their original function, replaced by very little that does the same work.
The man in the previous installment had, in the end, no one to call. What this installment suggests is that his condition was not an aberration filed away as tragedy, but the visible edge of a pattern most readers are living somewhere inside of — at some distance from its worst outcome, without ever having measured the distance themselves.
None of the lodges or aid societies came back on their own. Someone, in every generation that lost them, had to decide to rebuild — usually starting with something as small as a call that needed no occasion beyond the calling.
A civilization rarely notices its deepest losses while they are still small enough to measure.
The surveys measure the loss. Ecclesiastes explains why it matters.
The trouble is not falling. Falling happens to everyone, eventually. The trouble is falling with no one there.
1. McPherson, M., Smith-Lovin, L., & Brashears, M. E. (2006). “Social Isolation in America: Changes in Core Discussion Networks over Two Decades.” American Sociological Review, 71(3), 353–375. The study’s “modal zero confidants” finding prompted subsequent methodological debate over survey administration and coding differences between the 1985 and 2004 waves (see note 2); later reanalyses have generally confirmed the underlying decline in network size even where they have questioned its precise magnitude.
2. Reanalysis of GSS network-size data, 1985–2022, found that core discussion networks have not returned to their 1985 baseline despite methodological questions raised about the original 2006 study. (“Transformation of social relationships in COVID-19 America,” arXiv preprint, 2023.)
3. Watchman Insight, “The Man Who Had No One to Call,” watchmaninsight.com/the-man-who-had-no-one-to-call.
4. Tönnies, F. (1887). Gemeinschaft und Gesellschaft.
5. Verified historical literature on nineteenth-century American friendly societies and immigrant mutual-aid associations — bibliography to be finalized. Recommended candidate: David Beito, From Mutual Aid to the Welfare State: Fraternal Societies and Social Services, 1890–1967 (2000).
6. Cigna/Evernorth, “Loneliness in America” (2018, 2020, 2023 national surveys).
7. AARP, “Disconnected: The Escalating Challenge of Loneliness Among Adults 45-Plus” (survey fielded August 2025, updated June 2026).
8. Cox, D. A., “American Men Suffer a Friendship Recession,” Survey Center on American Life / AEI (2021).