The ship's log

A ship at sea keeps a log. Watch by watch, the officer of the deck sets down the course steered, the speed made good, the readings of the glass, the state of the sea and sky, and the events of the passage — a sail sighted, a sounding taken, a man sent aloft, a change of heading ordered and the hour it was ordered. The entries run in sequence, dated and timed, each one written close upon the moment it records. The log is not a memoir composed at the end of the voyage. It is a running instrument, kept in the middle of the thing it describes, by a person whose duty is to keep it faithfully whether the passage ends in a safe harbour or on the rocks.

This is a very old discipline, and the sea did not adopt it for the beauty of tidy records. It adopted it because the sea is where memory is most tempted to lie. After a collision, a grounding, a loss of cargo or of life, everyone aboard has an account, and every account has drifted, quietly and often sincerely, toward the teller's interest. The master remembers the fog thicker than it was; the other vessel remembers herself further off. Human recollection is not a recording; it is a reconstruction, assembled after the fact and shaped, without any conscious dishonesty, by everything the rememberer has since learned about what turned out to matter. The log resists this because it was written before any of it was known.

The record that does not know the ending

That is the whole of its power, and it is worth stating plainly. An entry made at four bells on a Tuesday, recording a heading and a wind, was made by someone who did not yet know that a Tuesday heading would ever be questioned. The keeper had no motive to shade it, because there was, at the moment of writing, nothing to shade it toward. The dispute that would one day turn on that heading did not exist. This is what contemporaneity buys: a record kept in ignorance of the outcome cannot be curated to fit a conclusion it does not yet possess. It is honest not because the keeper is virtuous but because the keeper is blind to the future — and a record made blind to which facts will later be convenient is a record that could not have been bent toward convenience.

Compare it to the account given afterward, however sincere. By the time anyone sits down to reconstruct the passage, the ending is known, and knowledge of the ending reaches backward and rearranges everything before it. Details that turned out to matter grow vivid; details that turned out not to matter fade; the sequence quietly reorders itself into a story that makes the outcome feel inevitable and the teller feel reasonable. None of this requires a liar. It is simply what memory does. And it is precisely why a court, offered both the log and the recollection, has long known which to trust. The log had no stake in the answer when it was written. The recollection has nothing but stake.

A record written before anyone knew it would matter is the most honest kind there is — it had no motive yet.

Written in order, at the time, by a keeper bound to it

The log's credibility does not rest on contemporaneity alone. Three properties travel together, and each does work. It is written in order, so that sequence is preserved and cannot be silently rearranged after the fact — and sequence is itself evidence, because what was known before a decision is a different thing from what was learned after it. It is written at the time, close upon the events, so that the record and the world it describes were never far apart. And it is kept by someone bound to keep it — an officer whose duty, whose training, and whose exposure to consequence all press toward setting down what happened rather than what would later be flattering. A record made in the regular course of a disciplined practice, at or near the moment of the events, by a person under an obligation to be faithful, carries a credibility that no after-the-fact narrative can borrow.

The law noticed this and generalised it far beyond the sea. The broad principle now familiar in evidence — that records made in the ordinary course of business, at or near the time of the events they record, by someone with a duty to record them, deserve a special standing that ordinary out-of-court statements do not — is the ship's log written into doctrine. The reasoning is the same reasoning the sea worked out first. Such a record is trustworthy not because we vouch for the character of whoever kept it, but because of the conditions under which it was made: routinely, contemporaneously, and under obligation, before anyone had a reason to lie. When a record is assembled instead after the dispute has formed, that whole scaffold of trust falls away. It may be accurate. But it has forfeited the one thing that made the log worth more than a memory — it now knows exactly what it needs to say.

A log for the decision

Machine decision systems, for all their speed, mostly produce the opposite of a log. What they emit, when they emit anything, tends to be a summary written after the fact — an account assembled once the output is in and, often, once the stakes are known. Or worse: a record that can be edited after the outcome is understood, its timestamps and contents revisable by whoever later wishes the decision had gone differently or been justified better. A record that can be rewritten once the ending is known has surrendered the very property that made the ship's log evidence. It is a recollection wearing the costume of a log.

The remedy is not more elaborate explanation. It is contemporaneity, restored as an engineering property. A decision deserves a record kept the way the sea keeps a log: written as the decision happens, in sequence, capturing the inputs as they actually stood and the rules as they were actually in force at that instant — and then made tamper-evident, so that the entry cannot be quietly revised once the world learns which detail the dispute will turn on. A contemporaneous, sequential, tamper-resistant decision record carries the credibility a reconstruction never can, and for exactly the reason the log does: it was written before anyone knew what it would have to prove. This is what a Decision Receipt is for — not a fluent account composed afterward, but a contemporaneous entry, made in ignorance of the outcome, that a skeptic can hold to account precisely because it could not have been shaped to the skeptic's dispute.

This is a different claim from the one made for the flight recorder, and the difference is worth keeping straight. The flight recorder is a story about a device built to survive the crash and to feed a no-fault inquiry into what went wrong. The ship's log is a story about the evidentiary power of contemporaneity itself — about why a record made in real time, in order, under obligation, is harder to bend than any account produced once the ending is known. The recorder answers can the record survive? The log answers a prior and subtler question: why should we believe a record at all? The sea's answer, tested over centuries and adopted by the courts, is that we believe the record which could not yet know how the story ends. A machine that decides at the pace of the modern world will produce claims faster than any human can check them. The old discipline tells us what to demand of each one — not a persuasive account after the fact, but a faithful entry made before the fact was anyone's to shape.

— Dispatches · Summit Cognitive

---