The Truth Need Not Be Plausible

Two senior people shake hands across a boardroom table. One agrees to step aside, the other agrees to make the move worth his while, and nobody writes any of it down, because writing it down would have felt like distrust between people who had built something together. Years later a court is asked to decide what was actually said in that room. A great deal of money turns on the answer.

What was agreed has two accounts, and they do not align. Which one will the court believe, and why? Most of us assume the honest man will sound honest and the liar will give himself away. The courts abandoned that assumption a long time ago. What they put in its place is worth understanding before you shake on it, not after.

The Truth Need Not Be Plausible

How a court decides which version of events to believe

I first read the judgment in Viviers v Kilian for the pleasure of it, the way one reads a strange old judgment on a slow afternoon.  I laughed out loud at the story it tells, and used it sometimes  in small talk.  Aside from the case dealing with the archaic law of adultery and the proverbial wages of sin, the judgment had something to teach me that applies to modern-day commercial transactions.  It is a 1927 decision of the Appellate Division sitting in Bloemfontein. Four (and not five) judges of the highest court in the land convened, in full seriousness, to consider an act of adultery committed through a fence.

The facts, as the court accepted them, are these. On a bright moonlit night in November, a man named Hughes was sitting in the shadow of his house, teasing coir, which is the slow evening work of pulling coconut fibre apart by hand for rope and matting, a job almost nobody has done in a hundred years. He watched the wife of his landlord come out of her kitchen carrying a coffee pot. He watched the appellant, Viviers, waiting for her in the street. The two met at the fence between them, and there, the wires being slack, they had what the report delicately calls connection, standing, with the fence still between them. Hughes heard the wire rattle, went to investigate, and saw Viviers pull himself free and run off down Alexandra Avenue. He asked the wife who the man was. She grunted, and went indoors.

The husband, who had forgiven his wife and gone on living with her, sued the lover (Viviers) for the insult to his dignity. The trial court awarded him fifty pounds. On appeal the figure was solemnly reduced to five pounds. It is easy to read all this and simply laugh at how distant that courtroom feels from ours, and that is fair enough. But the part of the judgment that has outlived its morality is a single piece of reasoning about why the court chose to believe a story that almost nobody would believe. That reasoning is not a museum piece. It is sitting, today, inside commercial disputes worth a great deal more than five pounds.

The real work a court does, when two accounts collide, is not to catch the liar. It is to find which version is true, which is a harder and stranger task. A believable story and a true story are not the same thing, and a court that mistakes the one for the other will get the answer wrong. How our courts tell them apart, and how far they have travelled from trusting the obvious signs, is the most useful thing a senior executive can understand before a dispute of consequence reaches a witness box.

Why the court believed the unbelievable

Counsel for Viviers made the obvious argument, and it was a good one. The story was grossly improbable. A man does not conduct an affair upright against a fence, by moonlight, a few feet from a stranger teasing coir, on the edge of a public street. People do not behave that way, so the witness must be inventing it. The Chief Justice listened, and observed that the argument cut both ways.

Solomon CJ had been on the bench a long time and had the dry economy of a judge who has heard a great deal of invented evidence. He did not find the improbability suspicious. He found it persuasive. The very improbability of the story, he wrote, was “to some extent evidence of its truthfulness”, because “truth, we know is sometimes stranger than fiction”. Had the tale been concocted, it would have come out simpler. It would, in his words, “have been so much easier to have made up a simple story” than to invent this one, with all its needless absurdities intact. A liar tidies up. He removes the detail that serves no purpose, because invention is laborious and he has no reason to spend the effort. The pointless, faintly ridiculous detail is the fingerprint of something that actually happened.

Mark Twain had said much the same a generation earlier, that truth is stranger than fiction because fiction is obliged to keep within the bounds of the possible and truth is under no such obligation. It is a fine instinct, and I have watched it prove right more often than wrong. But it is double edged, and a modern court knows it. Tell a clever liar that improbability reads as honesty and he will simply manufacture the improbability. The fence can be staged. So the principle, sound as far as it goes, cannot be where the enquiry ends. It is where it begins.

How does a court choose between two versions?

Faced today with two irreconcilable accounts, a South African court does not weigh them by feel. It follows a discipline set down by the Supreme Court of Appeal in Stellenbosch Farmers’ Winery Group v Martell et Cie, a case that might have been written for the executive reader. The dispute was whether two parties had agreed, orally and across a boardroom, a sales target for Martell brandy of 5,317,000 litres or 5,366,000 litres.  A formula in their contract meant the narrow gap between those figures was worth a fortune. One side had confirmed its understanding in writing. The other had not.

Nienaber JA opened his judgment with a sentence every businessman should keep above his desk. “Recollection can be fallible”, he wrote. “And in business the failure to confirm an event promptly and on paper can be fatal”. He then set out the method. To resolve the dispute the court makes findings on three things: the credibility of each witness, the reliability of each witness, and the probabilities of each version. Credibility asks whether the witness is trying to tell the truth. Reliability asks whether he is able to, given what he could observe and how memory decays with time and retelling. Probability asks whether the version holds up against everything else that is known. Where credibility and reliability leave the contest level, the probabilities decide it.

This is Solomon CJ’s instinct grown up and given a backbone. The strangeness of Hughes’ story was, in modern terms, a question of probabilities. What the method adds is that it refuses to stop there. It tests the improbable account against the documents and against the conduct of the parties before anyone imagined a court would be involved.

Can you trust a face in the witness box?

The 1927 court leaned hard on something its modern successors have learned to distrust, which is the impression a witness makes in person. The trial judges had watched Hughes, judged him honest, and noted that the husband seemed reluctant to bring so embarrassing an action without good cause. That a witness seemed truthful carried real weight on its own.

The belief that we can read honesty off a person’s face has not survived contact with the evidence. In England, the working rule has a clear form - when you want to know whether a witness is telling the truth, test what he says against the facts that can be proved without him, above all the documents, and against his motives and the overall probabilities. The High Court of Australia went further, noting that research has cast real doubt on the ability of judges, or of anyone, to tell truth from falsehood by watching a person testify. The danger is plain to anyone who has sat through a hostile cross-examination. The accomplished liar is often calm and fluent. The honest witness, frightened and out of his depth, is often hesitant and unconvincing. Manner measures composure, not honesty.

To his credit, Solomon CJ did not rest on Hughes’s manner alone. He tested the story against motive and found none worth the name. Hughes had nothing to gain. He had moved away to Queenstown and was tracked down years afterwards, and the journey the husband made to find him struck the court as the act of a man who believed his own complaint rather than one building a case from nothing. That is the right kind of reasoning, and it endures. The wrong kind, the confident conclusion that a man simply looked truthful, is precisely what a modern court labours to avoid.

What does the record remember that people forget?

If manner is unreliable and memory decays, what is left to trust? Increasingly, the paper.  Courts have looked closely at what is actually known about human memory, and the conclusion they reach unsettles a great many litigants. The courts have held that the legal system has never properly absorbed how unreliable memory is, that the process of litigation itself reshapes a witness’ recollection, and that in a commercial case the judge should place little weight on what people remember of conversations and reason instead from the documents and the probabilities. The contemporaneous record, made when nobody had a courtroom in mind, has become the most trusted witness in the room. 

This is the daylight.  The oral evidence of the moonlit glimpse and the vivid account of which the witness seems so sure, must now answer to the cold and unflattering record. An email sent on the day, or a file note made before the dispute soured, does not grow nervous under cross-examination, and does not improve its story with the passing of time.  Those are the footprints in the sand which show the path walked.

There is a complication coming that our courts have not yet resolved, and it bears directly on this. The whole modern method rests on the assumption that a contemporaneous document is more reliable than memory because it cannot change after the event. Generated and edited communications quietly disturb that assumption. The record is now the thing most easily produced to order, made fluent and plausible, made to feel as though it was written at the time. Our courts follow technology rather than anticipate it, so for some years the weight will fall on the party who can show that a document is genuinely what it claims to be. A record that can be corroborated independently will be worth more than one that merely exists.

How will a court read your boardroom?

Strip away the fence and the moonlight and you are left with something that ought to concern anyone who does business on relationships and trust.  When your dispute arrives, the court will not reward you for having been right. It will decide what happened, and it will do so, in the main, on the record you left at the time rather than on the account you are able to give afterwards, however honest you are and however clearly you tell it.

I have watched serious commercial matters turn not on the law and not on the merits, but on whose version the court was prepared to accept, and the deciding factor was seldom who was telling the truth. It was who could prove it. The party with the contemporaneous note beats the party with the better memory and the warmer manner, time and again. The discipline this asks of you is unglamorous and a little distrustful of your own goodwill. Confirm the important conversation in writing while it is still fresh.  Keep the file note. Treat the handshake as the opening of the record, not the close of it. 

None of this asks you to expect the worst of the people you deal with. It asks only that you understand that the room in which your dispute will be decided is lit very differently from the room in which your deal was done.

Hughes saw enough by moonlight to convince four judges of the Appellate Division. A modern court would admire his improbable honesty, and then ask, quite reasonably, what else there was. A letter, or a note made before anyone had a reason to shade the truth. Moonlight makes for a memorable story. It is a poor light to build a case in.

The trust is personal.  The record is business. 

Written by:

Wessel Badenhorst

July 2026

The_Truth_Need_Not_Be_Plausible.pdf

Download PDF • 2.04MB