Medieval Iceland had courts, laws and assemblies — and within them, a spoken word carried unusual weight. On what it means to be the kind of person whose word is simply done.
I had this wrong in an earlier draft, and the correction is worth making rather than hiding. Medieval Iceland was not a society without law. It had assemblies, a body of law recited from memory by the lawspeaker, litigation, and functioning courts — quarter courts were operating by around the middle of the tenth century. Oaths worked inside that system rather than instead of one.
What is true is narrower and still substantial: in a society where law was largely oral and enforcement depended heavily on assemblies, witnesses, kin and reputation rather than on a state with police, a spoken oath carried unusual legal and social weight.
To break an oath was not merely to fail at a task. It was to become a person whose word could not be relied on, in a system where reliance on words was doing a great deal of structural work — and where enforcement, when it came, often depended on other people being willing to back you.
What I take from this is not nostalgia. It is a structural observation: the reliability of a word is a property of the person who gave it, and it is built the way other properties are built — by repetition, mostly in places nobody is watching.
The research most often cited here is on implementation intentions — the finding that specifying in advance when, where and how you will do something substantially improves follow-through compared with holding a general intention.4
That effect is real and has been replicated widely. I would note that the effect sizes in the original meta-analytic work look larger than what more recent and better-controlled studies find, which is the ordinary trajectory of a popular psychological finding, and that the mechanism is probably simpler than the framing suggests: specific plans are easier to execute than vague ones.
The commitment-and-consistency literature is shakier ground. The classic demonstrations are old, several are field studies difficult to replicate, and the broader social-influence framework has taken its share of the replication crisis.5 I mention it because it is the standard citation and not because I would rest anything on it.
What holds up without either: a specific commitment made aloud to somebody is kept more often than a general one held privately. That is enough for the practice below.
The interesting oaths are not the public ones. Public commitments have enforcement attached; you keep them partly because breaking them would be noticed.
The ones that build the property are the ones you could break invisibly. Nobody knows whether you did the thing you said you would do alone, and it is precisely there that you find out what kind of word you have.
The literature is good on this. Saga honour is not primarily about reputation management — the texts are full of characters who keep inconvenient oaths when breaking them would have been easy and undetected, and who are treated as admirable for it in a way that has nothing to do with who was watching.
Keeping a word you should break is not integrity. Circumstances change, information arrives, and a person who cannot revise a commitment is not reliable so much as rigid. The saga material is genuinely bad on this — oath-keeping in those texts regularly produces catastrophe that everyone involved can see coming and nobody can stop, because stopping would mean being oath-breakers.
The second: an honour culture organized around the unbreakable word is also a culture in which you cannot back down, apologize, or accept a loss. Those are the same trait viewed from different angles, and the sagas end in a lot of funerals.
The third is smaller and modern. “My word is iron” can become a way of extracting commitments from other people — a standard held over colleagues or family who did not sign up to a warrior ethic. It is a private discipline. Applied outward it is just pressure with a better vocabulary.
The revision I would make to the tradition: a word kept is a strong default, and the ability to say “I committed to this and I was wrong” is not the opposite of reliability. It is what keeps reliability from becoming a trap.
Note — the unbreakable word as a trap
A culture organized around the word that cannot be broken is also one in which nobody can back down or apologize. “I committed to this and I was wrong” is not the opposite of reliability; it is what stops reliability becoming a trap.
The discipline is mostly about giving your word less often and more deliberately.
These essays draw on Old Norse and Old English literature — the Poetic Edda, Snorri Sturluson’s Prose Edda, the Icelandic sagas, and Beowulf — all of which were written down in Christian societies long after the pagan world they describe. Eddic and saga passages follow the translations named below; where a term or picture is contested, the essay says so rather than the list. Modern research is cited by author and publication.
Reliability is a property of a person rather than a set of kept promises, and it is built where nothing is enforcing it — which makes the invisible commitments the ones that count.
But the unbreakable word is a trap, and this literature demonstrates it repeatedly. “I committed to this and I was wrong” is what keeps reliability from ending in a funeral.