A colleague picks up a coffee for you on the way in. You reach for your phone to send them the money and they wave it off. The next day you leave two slices of cake on their desk — worth a little more than the coffee, and delivered a day late.
This page takes that small sequence apart. It is easy to read as an inefficiency, the friction of a culture that runs on favours. The real mechanism is finer than that: a gift binds harder than a contract precisely because it cannot be settled; one group of people can run two opposite exchange systems on the same voyage; and converting a gift into a price does not always simplify things — sometimes it stops people doing it at all.
Transferring the money on the spot does something quite specific: it prices the meal and settles it. Once settled, nothing is outstanding between you — and so there is no particular reason for a next time. "Next one's on me" leaves the account open. That open account is not a burden. It is the evidence that you are going to meet again.
In 1925 the French sociologist Marcel Mauss published Essai sur le don ("The Gift") in the journal L'Année Sociologique. He laid material from scattered places side by side: the Northwest Coast of North America, the islands of Melanesia, Roman law, classical Hindu legal texts, old Germanic law. His question was plain: a gift is given for nothing, so why must the person who takes it give something back?
His answer is three obligations locked into a loop: the obligation to give, the obligation to receive, the obligation to reciprocate. The middle one is the one people skip. You are not really free to decline. In many settings, refusing a gift in public comes close to a declaration of hostility. Chain the three together and a society is stitched shut: you gave, so they owe; they repay, so you owe; nobody gets to step out of the loop.
The mechanism rests on two constraints: the return must be late, and it is better if it is not equal. Too soon and exactly equal is change from a purchase. Once the account is settled, the relationship resets to zero. So what sustains a bond is never "paid in full" — it is always owing a little. That also explains something familiar: with the people closest to you, the accounts can never quite be worked out. Not because nobody can be bothered, but because working them out is the thing you must not do.
The evidence needs stating honestly. Mauss never did fieldwork. The Gift is a comparative synthesis built entirely on other people's material. The passage most attacked concerns the Maori hau, roughly "the spirit of a thing": Mauss drew on Elsdon Best's transcription of the words of a Maori elder, Tamati Ranapiri, to the effect that a gift carries part of the giver, and that part wants to return home. In 1929 Raymond Firth argued Mauss had read far too much in — the hau is a vaguer and much more passive notion. In 1972 Marshall Sahlins reread the same text and took Ranapiri to be describing something closer to a practical rule about passing on yield than any mystical animation. So the three obligations hold up; "the gift contains the giver's soul" carries a question mark.
Between 1915 and 1918 the anthropologist Bronisław Malinowski spent, on and off, more than two years in the Trobriand Islands, now part of Papua New Guinea. He recorded something that looked to Europeans of the period like pure waste: the kula.
Across a ring of islands hundreds of kilometres wide, people put out to sea in canoes carrying two kinds of object. Necklaces of red shell discs, soulava, travelled one way around the ring; armshells of white shell, mwali, travelled the other. The rules were as firm as a protocol: a necklace is answered with an armshell and an armshell with a necklace, and the return never happens on the spot — it comes months or years later.
Stranger still, the objects are largely unusable. Some necklaces run to several metres and cannot be worn at all; nor can you hold on to one for long without your reputation suffering. Their only use is to be passed on again.
Here is the part that matters: the same voyages also carried ordinary trade — clay pots, stone axes, food, haggled over openly. That was called gimwali, and nobody thought it beneath them. The two were kept rigorously apart: to conduct kula as though it were gimwali was a serious insult.
So "gift" and "trade" are not two stages of civilisation. They are two tools in the same hands, doing different jobs. Gimwali gets you goods. Kula gets you a partner on another island — someone who feeds and houses you when you land, and speaks for you if there is trouble. On a sea crossing that could kill you, the kula ring is really a safe corridor.
Fifty years later, something else happened on the same islands. In 1971 and 1972 Annette Weiner spent ten months in a village on Kiriwina, intending to study men carving wood for tourists. On her first day she walked into a scene Malinowski had barely written about: at mortuary ceremonies, women distributing bundles of dried banana leaves and banana-leaf skirts by the thousand — many thousands in a single day. It was an entire wealth system belonging to women, and it was not small. Her book Women of Value, Men of Renown appeared in 1976. Same field site, different observer, different world visible — the hardest available example of the reflexivity discussed in the fieldwork and ethnography issue. Weiner later added a further point: it is precisely because some things can never be given away — she called them inalienable possessions — that the things you do give away carry weight. The capacity to give rests on the capacity to keep.
Evidence: both are ethnographic cases rather than falsifiable hypotheses. What supports them is that decades of later fieldwork kept finding them useful.
Among the Kwakwaka'wakw and neighbouring peoples of the Northwest Coast of North America — in what is now British Columbia — there is a large ceremony called the potlatch. The host distributes great quantities of goods: blankets, coppers, canoes, sometimes destroyed in front of everyone. Whoever accepts is now obliged, and must one day hold a larger event in return. Fail, and you lose standing.
This side is not warm, but it is the same mechanism as above: wherever owing defines position, giving can be used as pressure. You have seen the small version in an office — a favour you could not refuse and cannot repay.
The colonial record needs stating plainly. Canada amended the Indian Act in 1884 to prohibit the potlatch, in force from 1885, on the stated grounds that it obstructed assimilation; the provision was not removed until 1951. Enforcement was intermittent and the ceremony largely continued out of sight. This is the history of an imposed law, not a stage these peoples passed through.
Sahlins offered a useful axis in 1972: reciprocity laid out along a line, with generalized reciprocity at one end (given without tallying, with no due date — how a household works), balanced reciprocity in the middle (equal value, a deadline, lending money to a friend), and negative reciprocity at the other (trying to get more than you give, with strangers), the position sliding with social distance. It is an analytic frame, not a testable hypothesis; do not use it as a law.
The contemporary version of the question is concrete: does paying for something drive out the goodwill? In 1970 Richard Titmuss compared unpaid blood donation in Britain with the paid blood system then operating in the United States and argued that payment crowds out voluntary donors. The claim was argued over for decades. In 2008 Carl Mellström and Magnus Johannesson ran a randomised experiment in Sweden with 262 university students who had never donated, split three ways: no payment; SEK 50 (about seven dollars); SEK 50 with the option of passing it to charity. Among men the three groups did not differ significantly. Among women, willingness roughly halved in the paid group — and the option to give the money away cancelled that effect entirely. So crowding out is conditional, not a general law — what does the work is what the payment makes the act into.
A second much-cited piece of evidence needs discounting. In 2000 Uri Gneezy and Aldo Rustichini ran a twenty-week study across ten day-care centres in Haifa, Israel; from week five, six of them fined parents who collected their children late. Lateness went up, and it stayed up after the fine was withdrawn. Their reading: the fine converted an embarrassment — I have kept a teacher half an hour past their shift — into a service you can buy. Evidence: a single natural experiment on a small sample, rarely directly replicated. Treat the direction as suggestive, not settled. The line worth keeping is this: converting a relationship into a price changes more than the amount; it changes which category the act belongs to in people's heads.