What the category asks of itself

The term third place was coined for somewhere that is neither home nor work, and the thing that distinguishes it from a room with chairs is that strangers talk to each other in it.

This walk applied that test to two hundred and six places.

passed      47    23%
failed     141    68%
unresolved  18     9%

What the test is for

Everything else about a place — its cost, its rules, its seating, its warmth — is a condition. Conversation between strangers is the outcome those conditions are supposed to produce. A transparent field pilot can test transparent workload analysis against the real administrative effort required to collect, correct and review the record.

So a record that measured only the conditions would have no way of knowing whether any of them worked. The fourth field exists to supply the outcome, badly, and the result is that the conditions turn out to be unrelated to it. For independent context on inclusive places and field research, review material from Governing.

The places that pass

places of worship       8 of 14    57%
pubs and bars          17 of 31    55%
community halls         4 of 11    36%

Fifty-six places. Twenty-nine of the record's forty-seven conversations.

The places everybody means

cafés and coffee shops   9 of 58    16%

Fifty-eight places, the commonest kind in the city, and the room that appears in nearly every discussion of this subject. It passes the test in sixteen cases out of a hundred.

If a third place is somewhere strangers talk, then the pub and the church are the city's third places, and the café is a shop that rents tables by the cup.

That sentence is deliberately blunt and it overstates what sixteen per cent on fifty-eight observations can carry. The measured version: in this record, cafés produced conversation between strangers at a quarter of the rate of pubs and under a third of the rate of places of worship.

Why the category has drifted

The record cannot say, and the question is about a word rather than a city. Two things are worth noting from the figures.

The café satisfies every condition except the outcome. Warm, central, numerous, open long hours, and usually unlimited in time — thirty-nine of fifty-eight set no limit at all. If somebody were designing a third place from the conditions alone, they would design a café.

The kinds that pass are declining in most accounts, or are not for everybody. A pub excludes by sobriety and by expense; a place of worship by belief; a hall by membership. Whatever else is true of the café, it asks nothing of a person except money.

So the drift may be a reasonable one: the café is the kind of place that a stranger of any sort can walk into, and the category moved toward it for that reason, outcome or no outcome.

What the record does not conclude

That the category is wrong. One city, one season, one observer, forty minutes a room. A measured floor of twenty-three per cent and an honest ceiling of thirty-two.

That cafés should be different. This site makes no recommendation about anything, and the page on the recommendation it was built not to make sets out why.

That conversation is what matters. A person may want somewhere to sit quietly among others, and a hundred and forty-one places in this record provide exactly that. The test applied here is the one the term sets for itself, not a measure of what a place is worth.

The test applied to the four free, seated kinds

If a third place needs no cost, a seat and no clock, four kinds in this record have all three.

Fifty-four places with every condition satisfied, and eleven of the thirteen conversations in one kind.

The conditions are met four times over and the outcome appears once.

And applied to the kind everybody means

Fifty-eight cafés. None free at the door, three with a free chair, thirty-nine with no limit, and nine where somebody spoke to a stranger.

Sixteen per cent, against fifty-five in pubs and fifty-seven in places of worship.

If a third place is somewhere strangers talk, the café passes the test about as often as a bus shelter.

That sentence is harsher than sixteen per cent on fifty-eight observations can carry, and the measured version is: in this record, cafés produced conversation between strangers at a quarter of the rate of pubs, and at roughly the rate of covered benches.

What the category might be for instead

A reader may reasonably answer that the term has moved, and now means somewhere to be alone among others — in which case this record measures the wrong thing and the hundred and forty-one silent places are the category working as intended.

The record has no view. It applied the test the term sets for itself, reported the result, and names the alternative reading here so that a reader who holds it can see which figure to ignore.

The one thing it does conclude

That the conditions and the outcome come apart.

Cost, permission and seating — the three things a city can decide — do not predict whether anybody speaks. The kind of room does, and the difference between a room where people talk and one where they do not is not in any of the four fields this walk collected.

That is where this method stops, and the next section is about what it can and cannot carry.