Ray Oldenburg had a term for them. Third places: not home (first place), not work (second place), but the third space where community actually happens. The barbershop. The pub. The diner where the owner knows your name and your order and a reasonable approximation of your personal history. The bowling alley. The park with the regulars who show up at the same time every Tuesday. Oldenburg published his foundational work on third places in 1989. He was already worried. He was writing about something in decline, a form of communal infrastructure that American car culture and suburban sprawl had been quietly dismantling for thirty years. He did not live to see us finish the job.

What the Data Shows About What We Lost

Between 2000 and 2020, the United States lost more than 25% of its bowling alleys, according to the Bowling Proprietors' Association of America. The number of bars and taverns declined by 20% over a similar period. Independent bookstores, community libraries (by hours and staffing, not building count), community centers in lower-income areas, and neighborhood barbershops have all contracted in the same window. This did not happen all at once. It happened in the ordinary way that things end — one closure at a time, each individually explainable by local economics, lease costs, changing demographics, the specific particular reason the owner finally decided not to renew. Each one a small event. The aggregate is a structural transformation. A 2018 study in American Sociological Review found that social network size, measured by the number of people outside immediate family with whom Americans report close relationships, declined significantly between 1985 and 2016. The most dramatic loss was in what the researchers called "confidants" — people you could call in a genuine crisis. In 1985, the median American had three such people. By 2016, the median was one. A substantial number reported zero. The infrastructure for making and maintaining those relationships had been quietly disappearing at the same time.

What Filled the Space

The answer to the question of what replaced third places is already in the title of this piece. The third screen: social media platforms, which presented themselves as community infrastructure and have functioned more precisely as community simulacra. There is something worth being careful about here. Online connection is real connection. This is not technophobia and the data does not support a clean "online bad, offline good" frame. What the data does support is that the type of social interaction provided by platforms optimized for engagement is structurally different from the type of interaction provided by a third place — different enough that it does not appear to produce the same outcomes. A 2023 study in Nature Human Behaviour found that passive social media consumption — scrolling, viewing, reacting — showed negative associations with wellbeing, while active social media use (direct messaging, coordinating actual meetups) showed neutral to slightly positive associations. The platform design optimizes heavily for passive consumption. The feed is not designed to get you to put the phone down and go meet the person.

The Three Replacements

The first replacement for third places was subscription services. Netflix replaced the video rental store where you ran into your neighbor and had a conversation about what was good. Spotify replaced the record store where the employee with the strong opinions changed your taste. Amazon replaced the bookstore where you browsed and noticed what other people were buying and talked to the person in the same aisle. These are better products in almost every measurable sense except the one that is harder to measure: the ambient, unplanned social contact they eliminated. The second replacement was professional socialization. Coworking spaces, work happy hours, networking events. These interact with people, but with specific stakes and social performance expectations that are categorically different from a third place. You cannot be nobody in a networking event. You cannot have no agenda. The third place's defining characteristic was that you could be there with no particular reason beyond being there. The third replacement was the group chat. Nominally social, accessible, asynchronous enough to fit into fragmented schedules. And yet: research on social comparison in group chats, the social labor of keeping up with multiple ongoing threads, the way ambient group chat can create a sense of social contact while actually substituting for it — all point in the same direction. It scratches the social itch without fully addressing the social need.

The Tangent About What This Costs

The word "loneliness" has been used with increasing frequency in public health contexts over the last decade, and it is worth being precise about why. Loneliness is not an emotional complaint. It is a health condition. A meta-analysis from 2015, conducted by Julianne Holt-Lunstad at Brigham Young University and covering over 3 million participants, found that social isolation and loneliness increased mortality risk by approximately 26% — comparable in effect size to smoking 15 cigarettes a day, and exceeding obesity as a mortality risk factor. The Surgeon General's 2023 advisory on loneliness cited this and subsequent research in framing loneliness as an epidemic with measurable public health costs. We built no systemic response to the loss of the spaces where people were not lonely. We gave people phones.

What Cities Are Actually Doing

The most interesting policy experiments are happening at the municipal level, largely because federal attention to this problem remains minimal. Some cities are converting underused commercial spaces into community rooms, not rented, not purchased, just available. Rotterdam's approach to social infrastructure investment has included funding neighborhood centers specifically designed to have no programming — spaces to be in without a reason. In the United States, some local library systems have expanded their "no questions asked" social presence mandate: you can be in the library all day, use nothing, speak to no one, and no one will ask you to justify your presence. This turns out to be significant. Most spaces in American life require either a transaction or a purpose. The third place required neither.

The Second Tangent: The Entrepreneurship of Loneliness

There is a significant and growing market for loneliness. Loneliness has become a product category: companion apps, AI chat, subscription friendship services, paid social experiences. These are often dismissed as pathetic, but the demand they represent is real and the dismissal often carries class implications — "real" social connection is available to people with time and proximity and money, and the alternatives are for people who lack those things. The more interesting question is whether technological companionship — something that is reliably present, genuinely responsive, remembered context, no agenda — begins to recalibrate what people expect from connection. Not replace it. Supplement it, in the way that a third place supplemented home and work rather than replacing either.

What Is Not Solved

The infrastructure problem is not solved by individual behavior change. You cannot walk into a third place that has been converted into a storage unit for Amazon returns. The closures happened; the social infrastructure contracted; the consequences are now distributed across millions of individual lives in the form of smaller social networks, fewer confidants, higher loneliness rates, measurable mortality effects. The political will to treat social infrastructure as infrastructure — the way we treat roads and electricity — has not materialized in any significant way. The conversation is happening at the margins, in urban planning departments and public health offices and sociology conferences. In the meantime, people are finding the third screen and wondering why it does not feel like enough. It was not designed to. It was designed to keep you on it.